Skip to main content

Über dieses Buch

This book brings together papers presented at the 2020 International Conference on Communications, Signal Processing, and Systems, which provides a venue to disseminate the latest developments and to discuss the interactions and links between these multidisciplinary fields. Spanning topics ranging from communications, signal processing and systems, this book is aimed at undergraduate and graduate students in Electrical Engineering, Computer Science and Mathematics, researchers and engineers from academia and industry as well as government employees (such as NSF, DOD and DOE).



Optimal Quality of Information Service in an Electronic Commerce Website

Multiple sellers can post their goods information through an online electronic commerce website, e.g., ebay, AliExpress, Amazon. The goods information post by a seller can attract the users who are originally interested in the goods from the other sellers, and thus the information post by a seller may reduce the influence of information from the other sellers, e.g., discourage customers to purchase goods from the other sellers especially when these sellers are competitors. In consideration of two classes of priority for sellers, we develop an admission control algorithm which temporarily removes low-priority sellers when a few high-priority sellers complain about their confusion caused by the others. We proof that the proposed algorithm can achieve the minimal number of sellers who have to be removed, subject to the constraint of an acceptable level of confusion at all of high-priority sellers. Also we investigate the admission control algorithm under a few more general settings, including the multiple-priority setting as well as the setting in which not all of high-priority sellers can attain their acceptable level of confusion even after removing all of low-priority sellers. Finally, under the structure of quite a few typical social networks, we show that the proposed admission control algorithm can efficiently solve the problem of mitigating confusion at high-priority sellers under various structures of social networks.

Weiwei Wu, Di Lin

Weight Forests-Based Learning Algorithm for Small-Scaled Data Processing

This paper proposes an ensemble learning algorithm that assigns a weight to the prediction of trees in each category, and these trees are generated by a specific feature to establish weight forests. The weight of forests is composed of a weight matrix and a weight of full connections, and two types of weight updating methods are employed. On the CICD2017, the generalized weight forest method outperforms the random forests method. The weight forests method can achieve better performance by using incremental learning for weight update. In this paper, we propose an incremental tree algorithm based on weight forests to solve the drift problem, and the occurrence of concept drift can be detected by adjusting the coefficients of incremental trees at a high level of accuracy.

Jiafu Ren, Di Lin, Weiwei Wu

Detecting Anomalous Insiders Using Network Analysis

Collaborative information systems (CISs) can improve efficiency and quality of services, because they enable users to coordinate and collaborate on common works with a large scale. In recent decades, increasing number of hospitals start to adopt CIS to improve efficiency and quality of health care treatment, and CIS is becoming an important platform for hospital employees to treat patients. CIS is rich in resources of relations among users, however, people seldom study networks of users existing in the CIS, especially in health information systems. In this paper, we construct collaborative weighted networks of users from two typical CIS: one from health care respect, Vanderbilt University STARPANEL system; and the other from a more general respect, Wikipedia talk system. We learn characteristics of those two networks, and find both networks follow degree-strength relations, which could be used for detecting anomalous behaviors of users in the CIS, for instance, anomalous behaviors of hospital employees and wiki editors.

Lei Dai, Liwei Zhang, Limin Li, You Chen

Containerization Design for Autonomous and Controllable Cloud Distributed System

With the relative popularity of cloud computing, a large number of services and applications have been deployed in cloud environments. However, the existing cloud technology architecture still has many problems. This paper summarizes the problems existing in the communication module in the development of cloud technology so far, such as the three major problems of complicated cloud’s architecture, homogeneous service, and pseudo-saturation of communication. Based on these problems, we give the corresponding solutions for our system under a distributed architecture container. This solution is implemented under autonomous and controllable conditions and can ensure the robustness and relative security of the system and ensure the smoothness of external communications. Our work has made a certain contribution to the current cloud computing environment and made up for some of the shortcomings of existing cloud environment communication modules.

Xiao Zhang, Yu Tang, Hao Li, Shaotao Liu, Di Lin

Spatially Transformed Text-Based CAPTCHAs

In recent years, deep learning technology has achieved great achievements in the fields of text, image, and speech recognition. As long as there is enough data, deep learning techniques often achieve good results. In the field of text-based CAPTCHA recognition, facing the anti-segmentation technology, using deep learning without using character segmentation techniques to end-to-end recognize CAPTCHAs, can also achieve good accuracy. However, some recent studies have found that deep neural networks are vulnerable to adversarial examples. By adding a very small perceptible disturbance to the input samples, the input of the disturbance will cause the model to output incorrect prediction results with high confidence. In this paper, we propose an adversarial text-based CAPTCHAs based on spatial transformation. And we use four state-of-the-art CNN models to recognize such adversarial CAPTCHAs with and without preprocessing. Experiments show that this type of CAPTCHAs can effectively reduce the recognition rate of the attack models.

Chuanxiang Yan, Yu Tang, Di Lin

Battery Capacity Multi-step Prediction on GRU Attention Network

The prediction of battery capacity plays an important role in estimating battery life information. Many studies use battery capacity as a standard indicator of battery life. But the battery’s capacity will fade irrecoverably due to the electrochemical reactions. The existing lithium-ion battery capacity prediction technologies cannot effectively understand the data relationships between the different charge/discharge cycles. The proposed method uses the attention mechanism to figure out this problem. Also, the proposed method uses the asymmetric loss function that is suitable for the capacity prediction. We use the first 50 $$\%$$ % of the cycle as the training data, then predict the battery capacity after multiple cycles, effectively reducing the dependence of the battery capacity prediction on the experiment data and improving accuracy over traditional recurrent network methods.

Jiazhi Huo, Yu Tang, Di Lin

Heterogeneous Network Selection Algorithm Based on Reinforcement Learning

In the current network environment, multiple wireless access technologies coexist. In order to meet the needs of various services in heterogeneous networks, in order to enable each user to select the most appropriate network to provide services and adapt to the dynamic changes of the network environment, this paper defines the markov decision-making process of network selection based on reinforcement learning, taking the heterogeneous network constructed by PDT and B-trunC as the background A network access control algorithm based on reinforcement learning is proposed, which fully considers the service type of session and the mobility of terminal, and realizes the adaptability of network selection.

Sheng Yu, Shou-Ming Wei, Wei-Xiao Meng, Chen-Guang He

Heterogeneous Wireless Private Network Selection Algorithm Based on Gray Comprehensive Evaluation Value

A variety of wireless access technologies have emerged in public security private network communication systems, and these technologies are heterogeneous in terms of access and business. Under the conditions of such a heterogeneous network, it is an urgent problem for users to achieve the best connection according to their own business needs. Aiming at the network selection of heterogeneous wireless private network, this paper presents a heterogeneous wireless private network selection algorithm based on the gray comprehensive evaluation value. This algorithm calculates the subjective weight of the network attributes by using the analytic hierarchy process, and uses the gray correlation analysis method to obtain the gray correlation matrix that is obtained, and finally, the two are combined to obtain a comprehensive grayscale evaluation value, select the best network to provide users with high-quality services.

Shouming Wei, Shuai Wei, Chenguang He, Bin Wang

A Novel Build-in-Test Method for the Multi-task Radar Warning Receiver Based on a Parallel Radio Frequency Network

A radar warning receiver (RWR) can detect and process radar signals to alarm radar threats in battlefields, and it has become an important sensor in the modern warfare. The multi-task RWR can improve the multi-signal processing capability and warning space, by adding processing hardware. As the amount of hardware increases, the reliability of whole system decreases. Multi-task RWR needs to conduct build-in-test before operation, to check the warning capability and locate the failed hardware. The traditional build-in-test schemes can test each sub-system of a single task independently with an external self-checking signal source. In this paper, we propose a novel build-in-test method for the multi-task RWR based on a parallel radio frequency (RF) network. The RF network makes the hardware of each sub-system as a backup for each other during the build-in-test. By utilizing the redundancy of parallel sub-systems, the joint build-in-test of the multi-task RWR can improve the accuracy without any additional RF signal detector.

Desi Luo, Song Li, Yang Hui, Xu Zhou

Design of FIR Filter Based on Genetic Algorithm

The finite impulse response (FIR) filter has been widely used in wireless communication systems and its design is crucial for modern Space-air-ground-sea integrated communication equipment. Then, genetic algorithm is a simple and effective tool for optimization processing, where it uses fitness functions to guide the search process without the requirement of prior knowledge. So, this paper introduces the genetic algorithm into FIR filter design, where the best frequency-response-error-vector (FREV) is searched, and the filter is produced based on the weighted least squares (WLS) principle. Finally, under the pre-determined filter specifications, we design FIR filters and obtain the satisfactory results.

Yipeng Wang, Yan Ding, Anding Wang, Jingyu Hua, Weidang Lu

Timing Error Detection and Recovery Based on Linear Interpolation

Timing synchronization is a critical and essential component of a satellite communication system. In order to synchronize the received signal with the sampling clock, the synchronization system in the receiving device calculates the sampling clock deviation. The timing error detector (TED) in the Gardner timing algorithm is a module for calculating the clock deviation of the synchronous system, and the recovery performance is related to the S-curve. In this paper, we propose an improved Gardner timing recovery algorithm. This method establishes a TED output timing error lookup table for the S-curve, uses the TED output as an index to determine the S-curve segment, and then uses linear interpolation to recover the timing error estimate. The results show that the estimated performance is still great in the presence of timing error and Rummerler multipath.

Xuminxue Hong, Bo Yang, Jingyu Hua, Anding Wang, Weidang Lu

Robust Interference-plus-Noise Covariance Matrix Reconstruction Algorithm for GNSS Receivers Against Large Gain and Phase Errors

A novel robust interference-plus-noise covariance (INC) matrix reconstruction algorithm is proposed for global navigation satellite system (GNSS) receivers. Instead of using the presumed interference steering vectors (SVs) to reconstruct the INC matrix, the SVs projected onto the interference subspace are utilized so as to mitigate large SV mismatches caused by gain and phase errors. Simulation results indicate that the proposed algorithm performs better than the other INC matrix reconstruction algorithms in output carrier-to-noise (C/N $$_0$$ 0 ) ratio. Moreover, the effectiveness of the proposed algorithm is validated by the GNSS software receiver.

Bo Hou, Haiyang Wang, Zhikan Chen, Zhiliang Fan, Zhicheng Yao

Research on Image Recognition Technology of Transmission Line Icing Thickness Based on LSD Algorithm

Icing on transmission lines has become one of the important factors that endanger the safe and stable operation of transmission lines. The timely identification of ice thickness on transmission lines can effectively prevent the damage caused by ice disasters. In order to improve the measurement accuracy of icing thickness of transmission lines, this paper proposed a linear detection algorithm based on LSD algorithm to detect icing thickness of transmission lines. Firstly, image processing methods such as image pre-processing, morphological processing, and edge detection are used to pre-process the icing image of the transmission line. Then, the edge of the icing image is detected using the LSD algorithm. Finally, the ratio of the image pixels to the actual diameter of the wire is used to find out of ice thickness. The experimental results show that the error of this method is 0.0443, which can be used to measure the thickness of icing.

Shili Liang, Jun Wang, Peipei Chen, Shifeng Yan, Jipeng Huang

A High-Frequency Acceleration Sensor for Monitoring Sloshing Response of Ships

This article introduces a high-frequency acceleration sensor for monitoring sloshing response of ships. The ship is not only threatened by the external environment during the voyage. At the same time, the hull will sway when the frequency of movement in the waves approaches the natural frequency of the liquid in the tank, thereby seriously damaging the hull structure. Therefore, it is necessary to monitor the acceleration outside the hull in real time and respond in time according to the situation. The sensor will move relative when the sensor is installed on the ship and subjected to external acceleration, and the optical fiber connected between the bosses will be stretched or compressed accordingly, both of which change the center wavelength of the fiber Bragg grating (FBG). The magnitude of the acceleration experienced by the vessel will show the change in wavelength on the displacer through the fiber grating. The sensor is designed to face the harsh natural environment of marine ships, which can realize automatic real-time monitoring of acceleration and ensure the safety of the ship during sailing.

Chuanqi Liu, Wei Wang, Libo Qiao, Jingping Yang

Research on cm-Wave and mm-Wave Dual-Frequency Active Composite Detection Guidance Technology

This article proposed a cm-wave and mm-wave dual-frequency active composite detection guidance technology, analyzed the advantages of cm-wave and mm-wave dual-frequency active composite detection guidance. The dual-frequency active composite detection guidance method has the characteristics of long detection distance, strong clutter suppression capability, high target recognition probability, and strong anti-interference capability. On this basis, the detection guidance system scheme and the key technologies were discussed.

Lai-Tian Cao, Chen Yan, Xiao-Min Qiang, Xue-Hui Shao

Research on Modulation Recognition Algorithm Based on Combination of Multiple Higher-Order Cumulant

The algorithm of modulation recognition using a single higher-order cumulant as the characteristic parameter is usually limited, and the recognition performance needs to be improved, for this reason, a recognition method is proposed, which uses the combination of multiple higher-order cumulants to construct the characteristic parameters, so that it contains more signal characteristics, and realizes the recognition of multiple signals of MASK, MPSK, and MFSK. The MATLAB simulation shows that it has a better recognition rate.

Yingnan Lv, Jiaqi Zhen

Indoor Positioning Technology Based on WiFi

In this paper, aiming at the inaccuracy of offline database data acquisition of fingerprint location method in wireless fidelity (WiFi) indoor positioning technology, the traditional mean filtering method is improved, and a median mean filtering method is proposed. Mean filtering, the method first eliminates the maximum and minimum small-probability singular value of RSSI generated by the AP signal source at the same position at different times, and then takes the average value of the RSSI signal strength value after removing the singular value.

Baihui Jiang, Jiaqi Zhen

Power Optimization in DF Two-Way Relaying SWIPT-Based Cognitive Sensor Networks

In this paper, we develop and analyze resource optimization protocols in two-way decode-and-forward relaying SWIPT-based cognitive sensor networks. The current schemes generally ignore the interference caused by secondary users’ transmission to primary users. However, when the channel between them is good, this neglect will have a bad impact on the analysis of the performance of both networks. Therefore, we investigate the maximum transmission rate that can be reached by secondary users under the premise of ensuring the communication quality of the primary users. We derive this scheme by jointly optimizing the transmit power of two secondary users and the power splitting ratio of the relay node, respectively. The tradeoff between the interference threshold and other system parameters is given in the numerical simulations to corroborate the effectiveness of the proposed solutions.

Chenyiming Wen, Yiyang Qiang, Weidang Lu

Compressive Sensing-Based Array Antenna Optimization for Adaptive Beamforming

In the digital beamforming, each antenna element usually corresponds to a front-end chain, which dramatically increases the hardware costs. In this paper, compressive sensing-based array antenna optimization technique is used for adaptive beamforming, which can decrease the hardware complexity while suppressing the interfering signals. Compressive sensing is utilized to reduce the sampling channel number, and the sparse reconstruction based on the convex optimization model is applied to accurately recover the full-array data. Then, the weight vector can be obtained by the recovered data. Simulation results are carried out to demonstrate the effectiveness of the proposed method.

Jian Yang, Jian Lu, Bo Hou, Xinxin Liu

A Fiber Bragg Grating Sensor for Pressure Monitoring of Ship Structure Under Wave Load

Based on the development of high-sensitivity pressure sensor, a kind of FBG sensor for the pressure monitoring of ship structure under wave load is proposed. The pressure sensor is mainly composed of metal circular diaphragm and two FBGs pasted on the sensitized structure, and the difference structure is adopted to greatly improve the measurement sensitivity. Firstly, the structure of the FBG pressure sensor is introduced, and the working principle of the FBG pressure sensor is analyzed theoretically. Then, the pressure sensitivity of the FBG pressure sensor in the range of 0–4.5 MPa is 0.871 pm/KPa through the simulation calculation through the finite element analysis software.

Jingping Yang, Wei Wang, Libo Qiao, ChuanQi Liu

Research on Key Technologies of NoverCart Smart Shopping Cart System

NovelCart is a smart shopping cart system that integrates ultra-wideband indoor positioning technology (UWB), radio frequency identification technology (RFID), associated rule mining algorithms and collaborative filtering algorithms and can be installed directly on traditional shopping carts. Among them, UWB indoor positioning technology is used to provide accurate positioning services with a precision of 30 cm; RFID RF identification technology is used to automatically obtain information on purchased goods; the association rule mining algorithm and collaborative filtering algorithm are used for precision advertising delivery. NovelCart uses a Kivy based graphical interface that allows commercial use of the LGPLv3 protocol to support platforms such as Linux, Windows, Android, and iOS. Through these technologies, NovelCart’s commodity navigation, automatic checkout, accurate promotion information delivery, and automatic scanning of purchased goods are supported.

Chengyao Yang, Gong Chen, Bo Yang, Lu Ba, Jinlong Liu

Human Identification Under Multiple Gait Patterns Based on FMCW Radar and Deep Neural Networks

Human identification has been the crucial and difficult problem of domestic and foreign scholars for a long time. As a novel identification technology, more and more attention is paid to the gait identification, which has proven to be feasible. In this paper, the authors propose a gait identification method based on micro-Doppler signatures obtained by 77 GHz frequency-modulated continuous wave (FMCW) radar. The obtained signal is represented by time-frequency (T-F) spectrum, and then, deep neural network (DNN) is adopted to deal with the spectrums for human identification. It is shown that the method can identify humans under three different gait patterns (walking; jogging; and walking with books) with 95% accuracy for 50 people. In addition, the method can also identify humans even if the subject is walking under other gait patterns that are not included in the training set.

Shiqi Dong, Weijie Xia, Yi Li, Kejia Chen

Design and Application of a High-Speed Demodulator Supporting VCM Mode

With the rapid development of earth observation technology, the resolution of payload is getting higher and higher, and the amount of data transmitted by remote sensing satellite is getting larger and larger. In order to solve the contradiction between the remote sensing data acquired by the satellite and the transmission ability of the satellite-to-ground transmission links. Using new transmission systems, such as high-order modulation of high speed transmission and variable code modulation (VCM), to improve the efficiency of satellite-to-ground transmission links within limited frequency band resources has become the development trend of satellite-to-ground data transmission for new generation remote sensing satellites. In this paper, the high rate digital demodulator supporting VCM mode is designed, and the overall hardware design and software design, as well as the key point VCM design are introduced. Finally, combining with the ground system engineering of GF-7 satellite, the function and performance of the high-speed demodulator supporting VCM are verified, and the engineering application is introduced.

Wang Huai, Li Fan, Han Zhuo

Comparative Analysis of Rain Attenuation Prediction Models for Terrestrial Links in Different Climates

In view of the difference of rain attenuation law between tropical climate and temperate climate, two rain attenuation prediction models based on inverse Gaussian distribution and Gamma distribution are proposed. In this paper, six representative regions of tropical and temperate climate are selected from around the world. The two models are applied to the terrestrial links of these regions and then compared with the ITU-R model based on the measured attenuation of corresponding regions. The simulation results show that the prediction capabilities of the proposed models in the tropical terrestrial links are better than the ITU-R model.

Lijie Wang, Hui Li

Improved Discrete Frequency and Phase Coding Waveform for MIMO Radar

In multiple-input multiple-output (MIMO) radar system, orthogonality transmit waveforms is important. A novel structure of discrete frequency and phase coding waveform is proposed in this paper for MIMO radar system. Non-dominated sorting genetic algorithm is used to optimize the waveform. Simulation results show the performance of waveform proposed in this paper which is better than discrete frequency and phase coding waveform.

Linwei Wang, Bo Li, Changjun Yu

Analysis of Influence of Antenna Azimuth on the Performance in a MIMO System

As a typical technique of the fourth and fifth generation mobile communication, Multiple-Input and Multiple-Output (MIMO) technology can achieve the optimal channel capacity by configuring the antenna azimuth, which can improve the performance of the whole system. In this paper, analysis of MIMO system performance influence is carried out with various environmental factors such as antenna number, Signal-to-Noise Ratio (SNR), Angular Spread (AS), and antenna spacing. The simulation results show that increasing the AS and antenna spacing can improve the channel capacity of the system, but continuously increasing the AS and antenna spacing makes it difficult to implementation on hardware. Therefore, it is possible to increase the number of antennas as much as possible to improve the performance of the MIMO system under a certain AS and antenna spacing.

Ke-Xin Xiao, Hui Li, You Luo, Huan-Yu Li, Yu-Han Wang

Design of a Third-Order Filter with 10 GHz Center Frequency

Filters are extensively applied in the field of communication artificial intelligence. The input–output coupling structure and the stage coupling structure of substrate integrated waveguide (SIW) resonator are introduced in this paper. The transmission characteristics of SIW are studied by means of the electromagnetic field distribution and surface current of rectangular waveguide. Combined with the basic theory and the general design method of coupled resonant filter, as well as the metal via hole design requirements of SIW, a third-order filter with center frequency of 10 GHz, relative bandwidth of about 2.4%, and return loss in the pass band better than −12.5db is designed by using the electromagnetic simulation software HFSS.

Hai Wang, Zhihong Wang, Guiling Sun, Rong Guo, Ming He, Yi Zhang, Shengli Zhang

An Image Acquisition and Processing Technique Based on Machine Vision

This paper takes the two-dimensional code image recognition technology as an example to introduce the image acquisition and processing technology based on machine vision. Our object is to filter noise reduction, edge extraction, and position correction on the randomly collected noise-disturbed images into binary images that can be effectively read. The template for filtering and noise reduction needs to select the appropriate size median filtering rectangular window according to the morphological characteristics of the image and the type of noise. Adaptive threshold method with Otsu algorithm is applied to achieve image binarization, and the image target area is successfully divided. Considering the complex edge details of the two-dimensional code image, this paper selects the Canny operator with the best accuracy and clarity for edge detection and extraction through comparison. Finally, a rotation correction algorithm based on Hough transform is adopted to perform effective positioning correction on the image. The image processing algorithm in this paper is simple and effective. It can be widely used on industrial assembly lines and other fields with a wide range of application prospects.

Huan-Yu Li, Hui Li, Jie Cheng, Yu-Han Wang, Ke-Xin Xiao

Power Allocation Based on Complex Shape Method in NOMA System

Power allocation is critical to non-orthogonal multiple access (NOMA) system. Aiming at the problem that the traditional fixed power allocation (FPA) algorithm and the fractional transmit power allocation (FTPA) algorithm are difficult to achieve optimal power allocation, and based on the theoretical research of the existing complex shape method, this paper proposes power allocation scheme based on complex shape search (CSS) algorithm. The algorithm obtains the initial power value within the constraints, continuously updates the centroid, shrinks the complex shape and finally finds the optimal power value that meets the accuracy requirements to maximize the system throughput. The simulation results show that the performance of the scheme is significantly better than the traditional FPA and FTPA schemes, and it also shows that the performance of the non-orthogonal multiple access system is superior to the orthogonal multiple access (OMA) system.

Jie Cheng, Hui Li, Lijie Wang, Chi Zhang, Huanyu Li

Study on the Feature Extraction of Mine Water Inrush Precursor Based on Wavelet Feature Coding

Coal water inrush acoustic emission (AE) signal is characterized by time-varying, non-stationary, unpredictable, and transient properties. To extract effective features representing coal water inrush information, the AE signal is analyzed by the wavelet characteristic energy spectrum coefficient based on wavelet theory. The feasibility of the wavelet feature coding has confirmed from code scheme’s availability and consistency, and it proves that the coding method can be used as a sign of waveform identification. The inclusion of energy distribution characteristics makes the waveform features more ordered and simplified. While the analysis of the obtained feature encoding in chronological order, it is possible to obtain the state of the time series signals, to lay an important basis for analyzing the evolution of water inrush acoustic emission coal from the time series level, such that a change dynamic characteristic acoustic emission signal becomes possible. And this will lay an important foundation for the time sequence analysis of acoustic emission event’s evolution in mine water inrush.

Ye Zhang, Yang Zhang, Xuguang Jia, Huashuo Li, Shoufeng Tang

Research on Coal and Gas Outburst Prediction Using PSO-FSVM

Coal and gas outburst is a kind of natural disaster in the process of coal mining, which is very destructive. If the outburst can be predicted accurately in time, the corresponding protective measures can be taken before the disaster, and the life safety of the underground workers can be guaranteed to the maximum extent. Because the traditional support vector machine (SVM) has the disadvantages of low noise resistance and easy to be affected by parameters, this paper presents a new prediction method based on particle swarm optimization fuzzy support vector machine (PSO-FSVM). In order to coordinate the global and local optimization ability of PSO, inertia weight and simulated annealing algorithm are introduced to improve the optimization ability and forced local optimal trap probability of PSO. The improved prediction model combined with PSO and FSVM is used to predict coal and gas outburst. The experimental results show that the model has faster training speed and higher classification accuracy than PSO-SVM and FSVM models.

Benlong Zhu, Ye Zhang, Yanjuan Yu, Yang Zhang, Huashuo Li, Yuhang Sun, Shoufeng Tang

Prediction of Coal and Gas Outburst Based on FSVM

Coal and gas outburst are one of the natural disasters in coal mines. It is highly destructive and sudden. It is a complex nonlinear problem that is affected by a combination of factors. Fuzzy support vector machine (FSVM) combines the advantages of fuzzy theory and support vector machine (SVM), has strong recognition ability in the case of small samples, and has better learning ability than traditional SVM. In this paper, the gray correlation analysis (GRA) is used to extract coal and gas outburst indicators, an appropriate fuzzy membership function is introduced, and on this basis, a model of coal and gas outburst prediction based on FSVM is proposed. The comparison of verification and other prediction methods proves that the FSVM model can meet the requirements of coal and gas outburst prediction, and the same set of data is trained using FSVM, PSO-SVM and BP neural network system. Experiments prove that FSVM has better prediction accuracy.

Xuguang Jia, Ye Zhang, Yang Zhang, Yanjuan Yu, Huashuo Li, Yuhang Sun, Shoufeng Tang

Influencing Factors of Gas Emission in Coal Mining Face

Most coal mines in China are distributed below the surface, so the construction of mines plays an important role in coal mining and production. The mining work of coal mines is closely connected with gas. The deeper the mine is, the more gas emission will increase. In the process of coal seam mining, once the gas cannot be discharged in time, gas outburst and gas explosion accidents are prone to occur. Therefore, understanding the laws and influencing factors of gas emission are the premise of gas safety operation and the basis of mine ventilation design. Knowing the distribution rules and influencing factors of gas emission, we can design a more reasonable ventilation system and prevent dangerous gas accidents in mines in advance.

Zhou Zhou, Fan Shi, Yang Zhang, Yanjuan Yu, Shoufeng Tang

Study on Gas Distribution Characteristics and Migration Law Under the Condition of Air Flow Coupling

When the safe concentration of gas in coal face exceeds the upper limit, it will have a great influence on the safe mining work of coal mine and endanger people’s life and property. Therefore, it is of great practical significance to study the gas distribution characteristics and the law of gas migration, which can provide more perfect theoretical guidance for coal mine ventilation management and gas prevention and control. This paper will first study the distribution of wind speed in the mine and the form of gas movement under the condition of air flow coupling, and on the basis of this theory discuss the distribution law of gas concentration on each observation surface, the distribution of gas in the upper corner, and the overall change law of gas under different wind speed.

Yanjuan Yu, Huashuo Li, Yang Zhang, Xuguang Jia, Fan Shi, Yongxing Guan, Shoufeng Tang

Research on Channel Coding of Convolutional Codes Cascading with Turbo Codes

In this paper, we mainly studied the coding and decoding principles of convolutional codes and turbo codes and verified the coding and decoding process and bit error rate performance of the two codes. In different channel conditions (AWGN and Rayleigh fading channels), with different code rates (1/2 code rate and 1/3 code rate), the convolutional codes with different constraint lengths (3 and 7) and turbo codes are simulated and compared. We can know from the simulation results that the performance of the two codes in AWGN channel is better than that in Rayleigh channel. In these two channels, the performance of the two codes with small bit rate is better than that of the ones with large bit rate; the performance of turbo codes improves with the increase of the constraint length. The performance of convolution code decreases with the increase of the constraint length when the SNR is small, but it improves with the increase of the constraint length when the SNR is large.

Chong-Yue Shi, Hui Li, Jie Xu, Qian Li, Hou Wang, Liu-Xun Xue

A Low Pilot-Overhead Preamble for Channel Estimation with IAM Method in FBMC/OQAM Systems

For the channel estimation, a large pilot overhead is required due to the imaginary interference in filter-bank multicarrier employing offset quadrature amplitude modulation (FBMC/OQAM) systems. In this paper, we address the pilot reduction and present a short preamble for the classical interference approximation method (IAM). Compared with the conventional preamble consisting of 3 columns symbols, the pilot overhead is equivalent to 2 column symbols in the proposed preamble. It is proven that there exists no performance loss in the proposed preamble at a significantly reduced pilot overhead. To verify the proposed preamble, numerical simulations are carried out with respect to bit error ratio and mean square error.

Dejin Kong, Qian Wang, Pei Liu, Xinmin Li, Xing Cheng, Yitong Li

Modeling of Maritime Wireless Communication Channel

Influenced by radian of the Earth and the waves, also the shelter of ships and sea waves, wireless channels above the sea have the effects of deep fading and multipath. Microwaves of very high frequency and ultrahigh frequency are analyzed in oceanic environments. We considered the direct path, paths of mirror reflection, and diffuse scattering, and we calculated the power of mirror reflection and diffuse scattering theoretically. Area of effective diffuse scattering and partitioning was introduced to calculate the power of multipath. We built a generalized channel model with different frequency band and communication range.

Yu-Han Wang, Meng Xu, Huan-Yu Li, Ke-Xin Xiao, Hui Li

Research on Message Forwarding Mechanism Based on Bayesian Probability Model in Wireless Multihop Network

Wireless multihop communication network relies on multi-node relays to realize information transmission without the support of infrastructure so as to become an important communication mode in military and civil fields. RREQ flooding in the whole network is always adopted for networking in complex and harsh environment. However, the redundant forwarding and overlapping effects of messages generated by the whole network flooding broadcast will cause the increase of node energy consumption and channel collision probability to decrease of channel utilization. A new message forwarding mechanism is proposed based on Bayesian probability model. The unnecessary forwarding overhead will be reduced by calculating the node density and posteriori probability. Simulation results based on NS2 show that the repeated broadcasting times can be greatly reduced by the Bayesian probability model-based forwarding mechanism proposed in this paper. Compared with similar algorithms, the proposed mechanism can reduce the energy consumption, routing overhead and improve packets success delivery rate effectively on the premise of basically ensuring network throughput.

Yang Yan, Qin Danyang, Guo Xiaomeng, Ma Lin

A S-Max-Log-MPA Multiuser Detection Algorithm Based on Serial in SCMA System

Sparse code multiple access (SCMA) has become one of the most promising key technologies for 5G communication in the future. In the multi-user detection algorithm of SCMA system, the message passing algorithm (MPA) has high complexity because of the exponential (EXP) operation. Because the maximum and approximate algorithms are used in the maximum logarithm message passing algorithm (Max-log-MPA), it will cause some information loss, resulting in poor accuracy of system transmission information. The serial message passing algorithm (S-MPA) combines the user node message update with the resource node message update, which will also cause some information loss, resulting in poor bit error rate of the system. Therefore, a maximum logarithm message passing algorithm based on serial (S-Max-log-MPA) is proposed in the paper, which first converts the exponential operation to the sum of the maximum values and then integrates the user node message update into the resource node message update. In the iterative process, the user node information is updated and then transferred to the next node to update the information of the resource node, which can greatly reduce the information loss and the space occupied by the intermediate variables, and effectively improve the accuracy of the system transmission information. Simulation results show that the bit error ratio (BER) performance of the system is better with the increase of iterations.

Guanghua Zhang, Zonglin Gu, Weidang Lu, Shuai Han

A Method and Realization of Autonomous Mission Management Based on Command Sequence

In order to meet the requirements of more and more remote sensing satellite loads and more and more complex imaging tasks, a method of autonomous mission management based on command sequence is proposed in this paper. The design and encapsulation of the command sequence based on the load operations, reduce the planning of the task details and the amount of the injection. Only the main operation and time of imaging mission can be focused on the ground. This improves the efficiency of mission injection and execution, as well as the autonomous management level of mission. The method is applied to a satellite.

Yiming Liu, Yu Jiang, Junhui Yu, Li Pan, Hongjun Zhang, Zhenhui Dong

Remote Sensing Satellite Autonomous Health Management Design Based on System Working Mode

The work mode of remote sensing satellite in orbit is flexible and diverse, the working state and parameters of the satellite under different working modes are very different, which brings some difficulties to the autonomous health management of remote sensing satellite. In order to associate satellite health management with working mode, a health management method based on finite state machine (FSM) is proposed. Taking the working modes as the state, the switching operation of each mode (such as switching on and off of equipment, etc.) as the state transfer condition, the autonomous health management disposal measures under different states as the execution action, the refined autonomous health management model under different working modes is established. The actual flight verification of a remote sensing satellite shows that the autonomous health management design is reasonable and feasible, and the finite state machine modeling method is universal and reusable.

Li Pan, Fang Ren, Chao Lu, Liuqing Yang, Yiming Liu

An Autonomous Inter-Device Bus Control Transfer Protocol for Time Synchronization 1553B Bus Network

The 1553B bus is widely used in spacecraft avionics system for its high reliability. The single point of failure (SPOF) of bus controller (BC) leads to the bus network paralysis. To improve the service continuity of the bus network, an autonomous inter-device bus control transfer protocol is proposed. The remote terminal (RT) address is used to identify device and initialize it with constant priority and timing parameters. In the situation of start-up and BC failure, the highest priority device autonomously acquires bus control, and the bus communication is self-repaired. After the fault device fixed, the bus control can be transferred under command control. In this paper, the realization scheme, the software implementation and the verification test of proposed protocol is introduced.

Tian Lan, Zhenhui Dong, Hongjun Zhang, Jian Guo

A Hybrid Service Scheduling Strategy of Satellite Data Based on TSN

Using time-aware shaper to ensure the accurate transmission is the key of time sensing network (TSN) scheduling. According to the need of downloading different kinds of massive data from spacecraft to the ground, an efficient scheduling algorithm which is based on time-aware shaper is designed in this paper. By guarantee data arriving exactly and completely, and ensuring the priority of the real-time data transfer, the virtual channel resource utilization can be improved.

Zhaojing CUI, Zhenhui Dong, Hongjun Zhang, Xiongwen HE, Yuling QIU

Campus Bullying Detection Algorithm Based on Surveillance Camera Image

In recent years, the phenomenon of campus violence has gradually come into people’s attention. Detection of campus bullying based on surveillance camera image has become a research hotspot. This paper designs a campus violence detection algorithm with a 3D convolutional neural network. The detection process includes three parts: video image preprocessing, feature extraction, and classification algorithm design. 80% of the samples in the video database are used to train the classification model, 20% of the samples are used as the testing set, and a fivefold cross-validation method is used to evaluate the performance of the classification algorithm. The simulation results show that the average classification accuracy of the proposed algorithm is 95.02%, indicating that the violence detection algorithm has good performance.

Tong Liu, Liang Ye, Tian Han, Tapio Seppänen, Esko Alasaarela

Activity Emersion Algorithm Based on Multi-Sensor Fusion

With the popularity of wearable intelligent devices, the research of activity recognition using wearable sensors has also been developed. This paper proposes an activity emersion algorithm based on multiple movement sensors. The authors gather movement data with wearable sensors and pre-process the data. They extract 57 features in both time-domain and frequency-domain, and select useful features with an SVM-RFE algorithm. SVM is used as the classifier. Recognition accuracy for the waist sensor is 93.4%, and that for the leg sensor is 92.1%. Then, the authors fuse the recognition results and achieve an accuracy of 95.6%, which is better than either single sensor.

Susu Yan, Liang Ye, Tian Han, Tapio Seppänen, Esko Alasaarela

Neural Network for Bullying Emotion Recognition

Bullying is a kind of aggressive behaviour of unjustified actions and speeches which usually occur among students. Bullying behaviours are often followed by verbal disputes such as abuse, crying and other negative voices. Bullying emotion recognition can assist the detection of bullying behaviours. This paper uses neural networks for speech emotion recognition to recognize bullying behaviours. Spectrograms and MFCC features are extracted from speeches, and CNN and RBF neural networks are used for classification. According to experimental results, CNN outperforms RBF and gets an average recognition accuracy of 82.5% for bullying emotions.

Xinran Zhou, Liang Ye, Chenguang He, Tapio Seppänen, Esko Alasaarela

Modulation Recognition Algorithm of Communication Signals Based on Artificial Neural Networks

With the development of wireless communication technologies and computer science, modulation recognition of communication signals has attracted more and more attention. This paper studies a modulation recognition method based on artificial neural networks (ANN). Two kinds of communication channels are tested with the modulation recognition method, namely additive white Gaussian noise (AWGN) channel and Rayleigh channel. Simulations are formed in MATLAB environment. According to the simulation results, the modulation recognition method can achieve an average recognition accuracy of 96% in the AWGN channel and 82% in the Rayleigh channel.

Dongzhu Li, Liang Ye, Xuanli Wu

Deep Learning for Optimization of Intelligent Reflecting Surface Assisted MISO Systems

Intelligent reflecting surface (IRS) has drawn great amount of attention from the researchers recently. It is worth investigating to maximize the spectral efficiency (SE) by jointly optimizing the beamforming at the access point (AP) and the phase shifts of the IRS. Although the traditional iterative algorithm can achieve high SE, it is not suitable for practical implementation due to its high computational complexity. Unsupervised learning could reduce the computational complexity, but as the number of IRS elements increases, the performance of SE becomes unsatisfactory. In this paper, we investigated a new deep learning method to maximize the SE in IRS assisted multiple-input single-output (MISO) communication system. Simulation results show that the performance of SE is better than that of the unsupervised learning method.

Chi Zhang, Xiuming Zhu, Hongjuan Yang, Bo Li

Max-Ratio Secure Link Selection for Buffer-Aided Multiuser Relay Networks

Cooperative communication techniques have proven to be an effective way to achieve physical layer security. In recent years, cooperative networks with buffered relaying have attracted more and more attention. This paper studies the security problems in buffer-aided multiuser relay networks. Based on the proposed maximum ratio secure link selection criterion (Max-ratio-LS for short), we investigate the performances of secrecy rate and secrecy outage probability. The numerical simulation results demonstrate that, for multiuser relay networks, our proposed max-ratio secure link selection criteria outperform previously reported no-buffer-aided cooperative schemes.

Yajun Zhang, Jun Wu, Bing Wang

Reconfigurable Data Acquisition System with High Reliability for Aircraft

Data acquisition is the important measure to acquire aircraft telemetry state and is indispensable components to aerocraft test. A high reliability data acquisition system for aircraft was developed based on master–slave redundant nodes architecture. Dual hot standby redundant architecture was adopted in central node, and control right can be switched rapidly when fault happens, and the reliability and reconfiguration can be achieved. Power modules and signal acquisition modules of acquisition and coding unit were integrated, and then, the complexity and weight of electric cables can be decreased. The experimental results indicate that the acquisition system can meet the demands of isolated acquisition and transmission and fault isolation. The data acquisition system has engineering application value for data acquisition and transmission with high reliability and precision.

Yukun Chen, Lianjun Ou, Gang Rong, Fei Liu

The Algorithm of Beamforming Zero Notch Deepening Based on Delay Processing

In order to solve the problem of the performance degradation of the diagonal loading beamforming algorithm at low signal-to-noise ratio (SNR), a new beamforming method based on the covariance matrix of the received signal delay information is proposed. Firstly, the delay covariance matrix of the received signal is constructed to suppress the influence of the noise, then the delay covariance matrix is transformed to obtain the eigenvalues of the covariance matrix, and finally, the eigenvalues are processed by diagonal loading technology. The simulation results show that the proposed algorithm is productive.

Junqi Gao, Jiaqi Zhen

Mask Detection Algorithm for Public Places Entering Management During COVID-19 Epidemic Situation

COVID-19 now is spreading fast all over the world. Wearing masks has proven to be an effective way to prevent COVID-19 to some extent. This paper studies a mask detection algorithm for public places entering management during COVID-19 epidemic situation. Only people wearing masks are allowed to enter. Cameras are fixed at the entrances of public places and take photos of people who are coming in. Then, a series of pre-processing are performed, including face detection, normalization, etc. Residual network is used as the classifier. Simulation results show that the average recognition accuracy can reach 90%.

Yihan Yun, Liang Ye, Chenguang He

Campus Bullying Detection Algorithm Based on Audio

With the continuous breakthroughs in various technologies, voice recognition has become a research hotspot. It is a method to detect the phenomenon of bullying in time by detecting whether the campus bullying emotion is contained in the voice. This paper builds a convolutional neural network model to recognize speech emotions. Firstly, pre-process the audio data, then extract the MFCC feature parameters from the pre-processed audio data, and finally design a classification algorithm. This paper selects the CASIA database, which has a total of 300 voice audios, including six emotions: angry, scared, happy, neutral, sad, and surprised. Using fivefold cross-validation to test the performance of the model, the accuracy of the classification algorithm is 68.51%. Finally, the classification algorithm is used to perform emotion recognition on a test sample selected from a campus bullying movie section. This section shows “fear” emotion, and the algorithm judges that the audio shows “fear” emotion. The actual scenes are consistent, indicating that the classification algorithm in this paper has certain stability and practicability.

Tong Liu, Liang Ye, Tian Han, Tapio Seppänen, Esko Alasaarela

An End-to-End Multispectral Image Compression Network Based on Weighted Channels

For the spectral correlation of multispectral images, a novel compression framework, named an end-to-end multispectral image compression network based on weighted channels, is proposed. The framework consists of a forward coding network, a quantizer and an inverse decoding network. The multispectral images are fed into the forward coding network, which is composed of the residual block based on weighted channels, to obtain intermediate features based on weighted channels. The intermediate features are quantized and encoded by the quantizer and the entropy encoder to obtain the compressed code stream. The compressed code stream passes through the entropy decoder and the inverse decoding network, which is symmetric to the forward coding network, to reconstruct multispectral images. The results validate that the proposed network shows better performance compared to JPEG2000 and JPEG.

Shunmin Zhao, Fanqiang Kong, Yongbo Zhou, Kedi Hu

Multispectral Image Compression Based on Multiscale Features

We propose a compression framework based on multiscale features using deep learning technique for complex features of multispectral image. The multiscale feature extraction module is taken as the basic block of the framework to constitute the encoder and the decoder. The encoder is used to extract the dominant features of the input multispectral images and the decoder, which is symmetrical to the encoder, is used to reconstruct multispectral images. The balance of the bit rates and the distortion is implemented by the rate-distortion optimizer in loss function. The encoder and the decoder are trained jointly based on minimizing the mean squared error (MSE). Our proposed framework is compared with JPEG2000 in terms of PSNR and the experimental results validate that our proposed framework outperforms JPEG2000.

Shunmin Zhao, Fanqiang Kong, Kedi Hu, Yuxin Meng

Dense Residual Network for Multispectral Image Compression

Considering that multispectral images have a mass of complex features, to extract intact feature data, a compression framework for multispectral images based on dense residual network (DRN) is proposed in this paper. The multispectral images are first fed into the encoder, and residual dense block in it can read all the information learned from the preceding block via a contiguous memory mechanism, then preserves the features adaptively. Then the data is compressed by down-sampling and converted to bit stream by quantization and entropy encoding. With that, global dense features can be extracted completely from the original image. Additionally, rate distortion optimization is used to make the data more compact. Then, we reconstruct the image via entropy decoding, inverse quantization, up-sampling, and deconvolution. The experimental result shows that the proposed method outperforms JPEG2000 at the same bit rate.

Kedi Hu, Fanqiang Kong, Shunmin Zhao, Yuxin Meng

Hyperspectral Unmixing Method Based on the Non-convex Sparse and Spatial Correlation Constraints

Aiming to the row sparse feature and the row sparse feature of the abundance matrix of hyperspectral mixed pixels, a hyperspectral unmixing method based on the non-convex sparse representation and spatial correlation constraints is presented. The non-convex sparse representation and spatial correlation representation models are firstly constructed, which take the non-convex p-norm of the abundance matrix to exploiting row sparse feature and the total variation regularization to exploit the spatial information of the hyperspectral images. Then the non-negativity constraint is taken into the model to exploit the physical property of the hyperspectral images. An iteratively optimization algorithm is developed for the unmixing model by the alternating direction method of multipliers. Experimental results illustrate that the proposed approach can get better unmixing accuracy than the other unmixing methods.

Mengyue Chen, Fanqiang Kong, Shunmin Zhao, Keyao Wen

Deep Denoising Autoencoder Networks for Hyperspectral Unmixing

The autoencoder (AE) is based on reconstruction and unsupervised framework, so its extracted features contain enough components to represent the input signal. Based on this characteristic, AE can be well applied to hyperspectral unmixing. However, due to the low precision of traditional AE and the large influence of noise, in this paper, a deep denoising autoencoder network (DDAE) for hyperspectral unmixing is proposed to improve the accuracy of abundance estimation while realizing the de-noising function. In order to guarantee the abundance to be sum-to-one and nonnegative, the endmembers satisfy the nonnegative, we limit the weight of hidden layer and decoding layer to be nonnegative, add the sum-to-one constraint to the hidden layer, and the L2,1-norm constraint to the objective function as a regular term, which takes advantage of the multiple sparsity between adjacent pixels. Experiments with real data and comparison with other algorithms prove the effectiveness of the DDAE algorithm.

Keyao Wen, Fanqiang Kong, Kedi Hu, Shunmin Zhao

Research on Over-Complete Sparse Dictionary Based on Compressed Sensing Theory

This paper introduces the concept of compression sensing theory, signal sparsity, sparse base and over complete dictionary, and introduces the over complete dictionary of DCT (discrete cosine) and chirplet wavelet in detail. On this basis, with the help of MATLAB simulation tools, the sparse simulation of several commonly used over complete dictionaries is carried out, and the sparse signal is reconstructed by OMP algorithm. The simulation results show that the chirplet wavelet dictionary, and DB wavelet dictionary are better than the DCT.

Zhihong Wang, Hai Wang, Guiling Sun

Spectrum Occupancy Prediction via Bidirectional Long Short-Term Memory Network

In a satellite system, the ability to generate future spectrum occupancy can play an important role in increasing spectrum efficiency, and spectrum prediction is emerging as an efficient approach for increasing spectrum efficiency. In order to predict spectrum occupancy more accurately, we propose a bidirectional long short-term memory network (BiLSTM)-based spectrum prediction (SP) scheme, which can be performed in two stages. Specifically, in the first stage, the historical spectrum data may be pre-processed, and in the second stage, the pre-processed data should be sent to BiLSTM model, which will perform training and generate the optimized hyperparameters firstly. Then, BiLSTM will be activated to perform prediction via the optimized hyperparameters. Performance evaluations show that the BiLSTM-based SP scheme outperforms the LSTM-oriented SP scheme in terms of both accuracy and learning speed.

Lijie Feng, Xiaojin Ding, Gengxin Zhang

Active–Passive Fusion Technology Based on Neural Network-Aided Extended Kalman Filter

The information fusion algorithm based on extended Kalman filter is widely used in the field of information fusion because of its relatively simple algorithm and solid theoretical foundation. This paper uses the advantages of neural networks and their advantages in solving nonlinear and time-varying systems to improve the traditional extended Kalman filtering method, improve the speed and accuracy of filtering, and improve the quality of active and passive data fusion.

Xiaomin Qiang, Zhangchi Song, Laitian Cao, Yan Chen, Kaiwei Chen

Multi-Target Infrared–Visible Image Sequence Registration via Robust Tracking

To solve the problem of multi-target depth difference for infrared and visible image sequence registration, a registration framework based on robust tracking is proposed. Firstly, the curvature scale space corners of targets are extracted, and the descriptors based on the curvature distribution are created to complete feature matching. Since different targets lie on different depth planes, single global transformation matrix is no longer applicable. Robust target tracking algorithm is introduced to dynamically allocate and update the matching reservoir for each target. Finally, the precise registration of multiple targets is realized by computing transformation matrix independently for each target. Experiments on a public dataset of non-planar infrared and visible image sequences show that our framework achieves lower overlapping errors and improves the accuracy of multi-target registration.

Bingqing Zhao, Tingfa Xu, Bo Huang, Yiwen Chen, Tianhao Li

Research on an Improvement of Images Haze Removal Algorithm Based on Dark Channel Prior

In view of the image reduction in haze weather due to aerosol scattering, the reduction of information volume and recognition accuracy of outdoor images, this paper improves the dark channel prior algorithm on the basis of the analysis of the image recovery algorithm, puts forward a solution strategy to obtain the atmospheric light value and the transmission based on the dark channel prior algorithm with different resolutions, and carries on the relevant experimental research, which effectively solves the problems of color distortion and poor real time of image haze removal.

Guonan Jiang, Xin Yin, Menghan Dong

Research on the Human Vehicle Recognition System Based on Deep Learning Fusion Remove Haze Algorithm

The human vehicle identification system is a key module in the field of automatic driving, but the image quality obtained by the system is poor in haze weather, which makes it difficult for the system to identify the target. In view of the above problems, this paper studies the human vehicle recognition system based on deep learning and fusion of remove haze algorithm. This system is based on the dark channel priori with different resolution to research the image of removed haze, in order to effectively detect the categories and behaviors of moving objects in the image sensing area, introduce artificial intelligence into human vehicle recognition, propose a target detection algorithm based on deep learning and Kalman filtering and Hungarian algorithm. This algorithm can ensure the high-speed feedback of the detected object and improve the detection accuracy at the same time, solve the problem that small and medium targets are easy to be missed in the detection process. Finally, the feasibility of the algorithm is verified by experimental research.

Guonan Jiang, Xin Yin, Jingyan Hu

Improved Skeleton Extraction Based on Delaunay Triangulation

The skeleton, which is the center axis of the target shape, is a topological representation of the shape. It has been used in image processing and pattern recognition fields such as target recognition, target matching, text recognition, blood vessel detection, and crack detection. Since Blum first used the grass burning model to extract the skeleton as a shape descriptor, there has been a lot of research on image skeleton extraction algorithms. In view of the problems existing in the existing methods, such as inaccurate skeleton position, discontinuous structure, and sensitivity to noise and small deformation, this paper proposes an improved image skeleton extraction algorithm based on constrained Delaunay triangulation, which effectively improves the performance of the algorithm by means of burr pruning, image pyramid, and other measures. The improved method can meet the requirements of object skeleton extraction in various scenes and has a good effect on noise suppression.

Jiayi Wei, Yingguang Hao, Hongyu Wang

An Algorithm of Computing Task Offloading in Vehicular Network Based on Network Slice

In order to solve the diversification requirement of quality of service (QoS) for computing tasks in vehicular network, this study will introduce network slice technology into the RSU uplink and propose an improved genetic algorithm. With the goal of reducing the average delay of computing task offloading in system, optimization function is a nonlinear programming problem, which solved by improving the object of crossover and mutation to the same type genes. Simulation results show that the improved genetic algorithm has good system performance and fast convergence speed.

Peng Lv, Zhao Liu, Yinjiang Long, Peijun Chen, Xiang Wang

A Cluster Routing Algorithm Based on Vehicle Social Information for VANET

In recent years, with the gradual rise and improvement of autonomous driving technology, research on vehicle ad hoc network (VANET) has gradually been paid attention by researchers. In VANET, the vehicle nodes are networked through a routing protocol and communicate with each other with the established routing. However, the high-speed mobility of vehicle nodes will cause rapid changes in the topology of the network, which can increase the information transmission delay. And as the number of vehicles increases, the probability of information transmission collision in the network will also increase. When the delay and collision reach a certain level, it will cause the loss of information. In order to solve these problems, this paper proposes a cluster routing algorithm based on vehicle social information. The communication source node and the destination node will communicate with each other through these cluster heads. This algorithm is superior to the traditional routing algorithm in terms of VANET communication performance.

Chenguang He, Guanqiao Qu, Liang Ye, Shouming Wei

LSTM-Based Channel Tracking of MmWave Massive MIMO Systems for Mobile Internet of Things

We propose a D-step forward prediction-based beam space channel tracking algorithm, which focuses on improving the accuracy of channel tracking in a multi-user massive multi-input multi-output (MIMO) system. Based on the time dependence of long short-term memory (LSTM) algorithm, the prediction results of the previous D-1 time are used as training parameters to track the beam space channel at the current time. It avoids the estimation error expanding with the increase of continuous tracking time in the process of channel tracking. Simulation results show that the proposed algorithm can effectively reduce the symbol error rates in a slowly changing environment.

Haiyan Liu, Zhou Tong, Qian Deng, Yutao Zhu, Tiankui Zhang, Rong Huang, Zhiming Hu

A Polarization Diversity Merging Technique for Low Elevation Frequency Hopping Signals

In the process of wireless signal transmission at low elevation angle, multipath effect is easy to be formed due to the reflection signal of the ground, which can cause signal distortion and interruption. Polarization diversity merging technique is an effective way to combat signal fading, which can improve the quality of wireless communication without increasing transmission power and bandwidth. However, the diversity signals merged with different frequency and phase will cause mutual interference. In the frequency hopping mode, residence time of each frequency hopping point is short and the closed-loop control method is unsuitable to rectify the deviations of each hop. To solve this problem, an open-loop control method is proposed in this paper, which can correct and merge the frequency deviation of polarization diversity signals in real time. It has the advantages of simple structure and fast processing speed, and is very suitable for the reception of low elevation frequency hopping signals.

Gu Jiahui, Wang Bin, Liu Yang, Liu Xin

Systematic Synthesis of Active RC Filters Using NAM Expansion

Active network synthesis is proved an effective method for circuit designer to find new circuits with desired performance. This paper demonstrates method of application of nodal admittance matrix (NAM) expansion to active filter synthesis from the port matrix of voltage-controlled voltage source (VCVS). The Sallen-Key (SK) second-order low-pass filter, Aherberg-Mossberg (AM) second-order low-pass filter, and Deliyannis second-order band-pass filter are synthesized by NAM expansion, and simulation results verify the feasibility of circuit synthesis method.

Lingling Tan, Fei Yang, Junkai Yi

An OFDM Radar Communication Integrated Waveform Based on Chaos

This paper mainly studies the design of radar and communication integrated shared signal based on chaos mapping. Orthogonal frequency division multiplexing (OFDM) technology has been widely used in the field of communication, and phase-coded OFDM radar has received considerable research in recent years. Combining the advantages of these two technologies, an integrated phase-coded OFDM signal based on chaotic sequence is proposed in this paper, where logistic chaos mapping is introduced as the phase coding sequence. The scheme uses logistic chaotic mapping to generate the phase coding sequence of the signal, and then realizes the modulation of the communication data by establishing mapping relationship between the communication data and the coding sequence. In order to improve the communication bit carrying capacity of the integrated signal, each subcarrier of the waveform is modulated with a different phase coding sequence. The analysis and simulation of the ambiguity function of the designed signal demonstrate that the signal not only improves the communication ability of the chaos-based integrated signal but also has better radar performance.

Zhe Li, Weixia Zou

Joint Estimation for Downsampling Structure with Low Complexity

Spectrum sensing and direction of arrival (DOA) estimation have both been thoroughly investigated. Estimating spectrum and DOA is important for many signal processing applications, such as cognitive radio (CR). A challenging scenario, faced by CRs, is that of multiband signals, composed of several narrowband transmissions spread over a wide spectrum and each with an unknown carrier frequency and DOA. The Nyquist rate of such signals is high. To overcome the sampling rate issue, several sub-Nyquist sampling methods, such as multicoset or the modulated wideband converter (MWC), have been proposed. In this paper, we use an ULA-based MWC structure to implement joint carrier frequency and DOA estimation at sub-Nyquist sampling rate. Different from other methods, we reduce the complexity of the hardware with fewer antennas and solve the pairing issues compared to other estimation methods.

Chen Wang, Wei Wang, Wenchao Yang, Lu Ba

One-Bit DOA Estimation Based on Deep Neural Network

This paper established a deep neural network model for DOA estimation of narrowband signals. First, one-bit quantization is considered into implementation for only retaining the symbol information of training data, as it offers low cost and low complexity in actual communication system. Then we investigate the performance of the neural network trained with quantized data and traditional MUSIC algorithm. Finally, simulations are conducted for correctness and validation. The results illustrate that the proposed method can realize meshless DOA estimation and has higher estimation accuracy in the case of low signal-to-noise ratio.

Chen Wang, Suhang Li, Yongkui Ma

A New Altitude Estimation Algorithm for 3D Surveillance Radar

This paper proposes an adaptive altitude estimation (AE) algorithm to improve the 3D surveillance radar’s accuracy. Firstly, the altitude measurement error is derived by the radar’s measurement error matrix theoretically. Then, we design multiple models (MM) for altitude estimation, which both contain maneuvering and constant velocity (CV) models working parallel. The proposed AE algorithm adaptive chooses the optimal result by comprehensively using Kalman filter’s residual and altitude velocity with limited sliding window data. The performance of the proposed AE algorithm is evaluated via simulations of two tracking scenarios. Experiment results show that the proposed AE algorithm may greatly improve accuracy performance of altitude estimation under different scenarios.

Jianguo Yu, Lei Gu, Dan Le, Yao Wei, Qiang Huang

Cause Analysis and Solution of Burst Multicast Data Packet Loss

In some service deployment, in order to realize the purpose of controlling multiple computers by a single computer, it is necessary to use multicast to send control data. However, in the practical application scenarios, if the interval between the two multicast data transmissions is too long, there will be packet loss. In order to solve this problem, this paper discusses the basic principle of multicast, analyzes the technical principle of packet loss in the process of multicast establishment, puts forward the solutions, and tests in the actual network environment to verify the effectiveness of the proposed solutions.

Zongsheng Jia, Jiaqi Zhen

The Prediction Model of High-Frequency Surface Wave Radar Sea Clutter with Improved PSO-RBF Neural Network

The sea clutter of high-frequency surface wave radar (HFSWR) has chaotic characteristics. Using the phase space reconstruction method to extend the one-dimensional sea clutter time series to the multi-dimensional phase space to fully demonstrate the internal dynamics of the sea clutter, and then training the Radial basis function (RBF) neural network to learn the internal dynamics of sea clutter and establishing the prediction model. The initial parameters of the network affect the convergence speed and the accuracy of the network model, so the particle swarm optimization (PSO) algorithm is used to optimize the initial parameters of the RBF neural network. Aiming at the PSO algorithm problems of the slow convergence speed and easily getting into local optimum, this paper proposes an improved PSO algorithm based on stage optimization. The simulation results show that the improved PSO algorithm has higher convergence accuracy; the optimized RBF neural network prediction model has higher stability and accuracy and has a better prediction effect on sea clutter.

Shang Shang, Kangning He, Tong Yang, Ming Liu, Weiyan Li, Guangpu Zhang

CS-Based Modulation Recognition of Sparse Multiband Signals Exploiting Cyclic Spectral Density and MLP

Modulation recognition of sparse multiband signals is a key technology for intelligent signal processing. However, the existing method suffers from high sampling rate and poor anti-noise performance. Therefore, compressed sensing and cyclic spectral density are combined to cope with the shortcomings of existing methods followed by the multi-layer perceptron (MLP) to recognize modulation mode of signal. Some simulations are carried out, and the simulation results show that the proposed algorithm is correct and effective.

Yanping Chen, Song Wang, Yulong Gao, Xu Bai, Lu Ba

Bandwidth Estimation Algorithm Based on Power Spectrum Recovery of Undersampling Signal

With the rapid development of wireless communication and the increase of various types of communication services, cognitive radio based on spectrum sensing is becoming a hot research topic. One of the key tasks of spectrum sensing is signal parameter identification, which is to quickly identify a series of parameters of signal based on signal detection, to provide basis for subsequent spectrum allocation and sharing. Bandwidth is one of the important parameters of communication signal. It is the premise of detecting "spectrum holes" to estimate bandwidth quickly and accurately. Due to the difficulty of Nyquist sampling in wideband spectrum sensing, this paper studies undersampling bandwidth estimation. We introduce the power spectrum estimation algorithm under the assumption of wide stability and put forward the corresponding bandwidth estimation strategy. The limitation of the strategy is analyzed. The performance of the algorithm is simulated under different SNR and compression ratios. Different random undersampling matrix is researched as well. The simulation shows that the algorithm discussed in this paper is feasible and reliable.

Yuntao Gu, Yulong Gao, Si Wang, Baowei Li

An Adaptive Base Station Management Scheme Based on Particle Swarm Optimization

With the rapid development of 5G in recent years, the energy consumption in the information and communication industry is becoming serious day by day. The sleeping strategy of the base station (BS) is to consider the load situation and user distribution of each BS under the heterogeneous cellular network model and close the BS with low load. Meanwhile, some users of the BS with high load are assigned to the BS with low adjacent load, so as to achieve energy consumption balance. The simulation results show that the particle swarm optimization algorithm is superior to traditional distributed algorithm in energy consumption and energy saving efficiency, which can realize green communication, but the time it takes is a little longer.

Wenchao Yang, Xu Bai, Shizeng Guo, Long Wang, Xuerong Luo, Mingjie Ji

An Earthquake Monitoring System of LoRa Dynamic Networking Based on AODV

In recent years, earthquake disasters are increasingly frequent, resulting in great loss to people’s life and property. In order to better evaluate and predict the earthquake, it is of great significance to design a system which can effectively collect the seismic data when an earthquake occurs. LoRa wireless communication has the advantages of low power consumption, long communication distance and strong anti-interference ability. In this paper, LoRa network is used to upload the data collected by the seismic sensor to the LoRa base station. Then, the data is sent to the cloud server through 4G, finally received by the working platform. In view of the node damage during the earthquake, the AODV protocol is used to ensure the transmission of acquisition data in the form of dynamic networking. The result shows that the monitoring system of LoRa network seismic monitoring system based on AODV protocol runs stably and can obtain reliable seismic data.

Long Wang, Wenchao Yang, Xu Bai, Lidong Liu, Xuerong Luo, Mingjie Ji

Track Segments Stitching for Ballistic Group Target

The ballistic group targets are spaced highly close, and they are mutual occlusive. Hence, how to distinguish and precisely track the complex ballistic target under a presented radar working band has become an urgent problem. To solve this problem, we first predict the group center by two-body movement, and then we induce the track segment stitching (TSS) method using track file to construct test hypothesis, then to correlate and smooth the track segments and finally to manage the batch. Simulation results show that the average tracking time is double compared with the traditional algorithm, and the tracking root-mean-square (RMS) errors are significantly decreased. The group targets are consecutively tracking and keep the unique batch, which provides a new precise tracking method for ballistic group target.

Xiaodong Yang, Jianguo Yu, Lei Gu, Qiang Huang

A Physical Security Technology Based upon Multi-weighted Fractional Fourier Transform Over Multiuser Communication System

Inspired by the multiple weighted-type fractional Fourier transform (M-WFRFT) employed over physical layer security of point to point communication systems, we in this paper extend M-WFRFT to the single transmitter to multiple receivers communication system. According to the transform orders relationships between different M-WFRFT, we design the communication system and the frame structure based upon M-WFRFT. It can be demonstrated, from the numerical simulations, that the proposed technology can further improve the physical layer security performance of wireless communication.

Yong Li, Zhiqun Song, Bin Wang

Fast Convergent Algorithm for Hypersonic Target Tracking with High Dynamic Biases

The chirp signal of large time–bandwidth product solves the contradiction between long detection range and fine range resolution, and is widely used in the current radar systems. However, due to the distance-Doppler coupling, high dynamic biases are brought into the radar ranging for the tracking of hypersonic targets, which leads the much longer time to converge for filters. In this paper, we propose a fast convergent algorithm to overcome the problem. The intuition is that the high dynamic bias is caused by the high radial velocity and the long convergence time is introduced by the large estimation error of the radial velocity. The proposed algorithm solves the problem by employing an algorithm with high precision of radial velocity. Experiments show that the proposed algorithm largely improves the convergence time of range, azimuth angle, elevation angle and position.

Dan Le, Jianguo Yu

Optimization of MFCC Algorithm for Embedded Voice System

Feature extraction is the core step to achieve speech recognition and is the key to the correctness of the speech recognition system. Feature extraction is to obtain effective information for speech recognition and remove redundant information. This article briefly describes the MFCC feature extraction process and optimizes the MFCC algorithm by changing the pre-emphasis parameter and the parameter addition method to adapt to the characteristics of short-word embedded speech recognition systems.

Tianlong Shi, Jiaqi Zhen

Energy-Efficient Hybrid Precoding for Adaptive Sub-connected Architecture in MmWave Massive MIMO Systems

An energy-efficient hybrid precoding algorithm with quality of service (QoS) constraints is proposed for adaptive sub-connected architecture in mmWave massive multiple input multiple output (MIMO) systems. The system energy efficiency optimization problem is proposed and then decomposed into an analog domain sub-problem and a digital domain sub-problem. The phase extraction algorithm and iterative optimization algorithm are used to solve the optimization sub-problem in analog domain and digital domain, respectively. Simulation results show that the proposed algorithm is effective, for it could enhance the energy efficiency of networks, and at the same time guarantee the QoS of service.

Li Li, Qian Deng, Weiwei Dong, Yutao Zhu, Tiankui Zhang, Rong Huang, Wei Huang

Sequential Pattern Mining-Based Alarm Correlation Analysis for Telecommunication Networks

Alarm correlation analysis is one of the important contents in fault management of telecommunication networks. There are meaningless redundant alarms and the alarms generated in different network topologies overlapping with each other, which makes it difficult to locate root cause of alarms and analyze faults. In this paper, we propose a alarm correlation analysis approach based on sequential pattern mining. Firstly, an improved K-Modes based on network topology (K-MNT) clustering algorithm is proposed to preprocess alarms. Sequential pattern mining with dynamic sliding time window is applied to obtain the temporal correlation of alarms. Then the root-cause alarms and faults are correlated. Experiments show that the proposed approach can effectively locate the root-cause alarms and increase the proportion of alarms and faults correlation.

Ying Chen, Tiankui Zhang, Rong Huang, Yutao Zhu, Zemin Liu

Multi-Model Ensemble-Based Fault Prediction of Telecommunication Networks

Fault prediction is the critical method to ensure the stability and reliability of the whole communication networks. This paper proposes an ensemble model combining traditional algorithms and deep neural networks with multiple functions of feature combination, text learning, local feature extraction, and sample imbalance processing for fault prediction according to the characteristics of data with attribute relevance, imbalance, and text type. The experiments validate the proposed method can effectively improve fault recognition by 4–9%.

Ying Chen, Tiankui Zhang, Rong Huang, Yutao Zhu, Junhua Hong

Bone Marrow Cell Counting Method Based on Fourier Ptychographic Microscopy and Convolutional Neural Network

In bone marrow examination, the number of bone marrow cells is an essential parameter to judge the degree of myeloproliferative. In this paper, we propose a new bone marrow cell counting method based on Fourier ptychographic microscopy and convolutional neural network. We use Fourier ptychographic microscopy technology to obtain the intensity and phase images of bone marrow cells at first. Then, we combine the intensity and the phase image correspondingly to obtain a dual-channel image. We use the convolutional neural network to extract the characteristics of bone marrow cells in the dual-channel image, which can generate a density map. The number of bone marrow cells is realized by integrating the density map. The experimental results show that both the mean absolute error (MAE = 0.66) and mean square error (MSE = 0.67) of our method are lower than those existing methods.

Xin Wang, Tingfa Xu, Jizhou Zhang, Shushan Wang, Yizhou Zhang, Yiwen Chen, Jinhua Zhang

Identification of Sensitive Regions for Power Equipment Based on Fast R-CNN

Electricity is an indispensable resource in the daily life of people. However, the inspection of power equipment is still in the artificial stage. This way of work is inefficient, consumes a lot of manpower and material resources, and is not accurate enough. In order to realize the intellectualization of patrol inspection system, this paper uses a fast R-CNN technology based on Google LeNet V2, which brings a more efficient and advanced supervisory technology for power equipment. Experimental results of this paper show that the average accuracy rate of the scheme reaches 92.0%, and the average recall rate reaches 79.7%, with good results..

Hanwu Luo, Qirui Wu, Zhonghan Peng, Hailong Zhang, Houming Shen

Dimensionality Reduction Algorithm

Since the information revolution, human society has developed extremely rapidly. As mankind continues to develop computing technologies for processing information data, it is also accompanied by the explosive growth of information data, and various high-latitude data are constantly being produced. A serious problem brought about by this is how to deal with these large amounts of high latitude data, and extract the information that can be used from it. High-dimensional data compression extraction algorithm is an effective way to solve this problem. This paper mainly studies the sparse principal component analysis algorithm (SPCA) and the edge-group sparse principal component analysis algorithm (ESPCA) based on the principal component analysis algorithm (PCA), a high-dimensional data compression algorithm. And this paper focuses on the theory of edge group sparse principal component analysis algorithm, and successfully reproduces the program, and obtains results consistent with the original text on the simulated data.

Wenzhen Li, Qirui Wu, Zhonghan Peng, Kai Chen, Hui Zhang, Houming Shen

An Improved APTEEN Protocol Based on Deep Autoencoder

APTEEN protocol is a popular clustering protocol in wireless sensor networks, which has the problems of too fast energy consumption and large data redundancy. To overcome this problem, a novel improved algorithm of the APTEEN protocol named SAEDF-APTEEN is proposed in this paper. Given that data fusion technology can reduce redundant data transmission in wireless sensor networks, this paper introduces the deep autoencoder into the APTEEN protocol. After training the deep autoencoder model in the base station, the encoder part is deployed in the cluster heads. The cluster heads fuse the data sent by cluster members and transmit the compressed data to the base station. The simulation results show that the SAEDF-APTEEN protocol not only significantly improves the network lifetime and reduces the energy consumption of the whole network, but also effectively reduces the amount of data transmission and enhances the data transmission efficiency.

Yu Song, Shubin Wang, Lixin Jing

APTEEN Protocol Data Fusion Optimization Based on BP Neural Network

In order to reduce the amount of data transmission and reduce the energy consumption of nodes in APTEEN protocol, a data fusion algorithm based on APTEEN protocol and BP neural network is proposed in the paper. The three-layer BP neural network is used to describe the cluster structure. In the process of cluster structure data transmission, BP neural network is used to process the perception data, from which the feature values of the perception data are extracted and forwarded to the sink node. It is proved by the simulation result that this algorithm has better performance in both network life and network consumption than APTEEN protocol. It reduces the amount of communication data, reduces energy consumption, and prolongs the network life.

Lixin Jing, Shubin Wang, Yu Song

Realization of Target Tracking Technology for Generated Infrared Images

Target tracking is to determine the motion parameters of the target in each infrared image in continuous sequence images. In this paper, the target tracking refers to getting the coordinate information in the infrared image. In the infrared image tracking system, there are many methods to determine the target position. In this paper, the traditional infrared tracking methods was applied including weighted centroid tracking, centroid tracking, matching tracking, and Kalman tracking. The tracking performance of the infrared detector system was tested under different signal-to-noise ratios.

Ge Changyun, Zhang Haibei

Review on Wearable Antenna Design

With the rapid popularization of 5th generation mobile networks(5G), wearable devices have also been widely applied in military communications, medical care and entertainment. After the initial development of wearable form, more implantable and attached wearable devices are developed. The progresses of wearable antennas have an important impact on the development of wearable devices. This paper discusses the important performance index and new development of the wearable antenna, introduces the research results of the wearable antenna in recent years, and analyzes the development trend of the wearable antenna.

Licheng Yang, Tianyu Liu, Qiaomei Hao, Xiaonan Zhao, Cheng Wang, Bo Zhang, Yang Li

A Review of Robust Cost Functions for M-Estimation

Robust state estimation plays a key role in mobile robotic navigation, and the M-estimation technique can effectively handle outliers. In this paper, the commonly used robust cost functions for M-estimation are given, and their cost, influence, and weight functions are summarized and compared.

Yue Wang

Research on the Path Planning Algorithm for Emergency Evacuation in the Building Based on Ant Colony Algorithm

In recent years, China has witnessed rapid economic development and rapid population growth. Meanwhile, the population density of cities has continued to increase. Public places in cities are often overcrowded and major accidents occur more frequently. When an emergency occurs, it is of great significance to provide effective evacuation guidance to the on-site people, so that reduce the number of casualties and property losses. This paper first selects the ant colony algorithm (ACO) for research. The ant colony algorithm and the principle of using grid method to generate the simulated map are introduced in detail. According to these principles and standards, the specific environment is modeled, the simulated map is generated, and the ant colony algorithm is used to realize the path planning on the simulated map.

Chenguang He, Suning Liu, Liang Ye, Shouming Wei

Multiple Action Movement Control Scheme for Assistive Robot Based on Binary Motor Imagery EEG

In this paper, a weighted voting system combined with basic signal processing methods is used to classify multi-category motor imagery (MI) scenarios (foot, left-hand, right-hand, tongue) to improve the classification accuracy of MI electroencephalogram (EEG) signal. Meanwhile, a feasible binary coding framework is proposed to control the KUKA robotic arm for grasping to improve online performance of applications on brain–computer interfaces (BCIs). Firstly, two-movement MI with the high classification accuracy is selected from four-action types, i.e., foot as 0, left-hand as 1, and their combination representing the four directions of motion direction of the robotic arm (e.g., 00-front, 01-back, 10-left, 11-right) is generated by two-bit binary coding. Next, the motion of the robotic arm in each direction is achieved by two successive movements of MI. Finally, the accuracy of our integrated classifier reaches 74.6% in four-movement MI data and 92.6% in two-movement MI data. Compared to four-movement MI to control the robotic arm, the binary coding method reduces the time by 6.8% and increases the accuracy more than two times.

Xuefei Zhao, Dong Liu, Shengquan Xie, Quan Liu, Kun Chen, Li Ma, Qingsong Ai

Modeling and Simulation of Ionospheric Surface Scattering Equation for High-Frequency Surface Radar

Ionospheric clutter is the main factor to reduce the detection performance of HFSWR. This paper established the radar range equations of the surface scattering targets. The scattering coefficient of ionospheric RCS is estimated based on the principle of high-frequency coherent scattering for irregularities. Thus, the generalized radar formula based on the physical mechanism of the ionosphere is derived. The simulation shows that the ionospheric echoes under surface scattering close to the vertical direction accord with the pulse-limited model, and the ionospheric echoes close to the oblique direction accord with the beam-limited model, for the case of wide beam of HF surface wave radar.

Yongfeng Yang, Ziqiang Zhang, Xuguang Yang, Jun Du, Yanghong Zhang

Greedy Matching-Based Pilot Allocation in Massive MIMO Systems

With the ability of achieving high data rate transmission and high spectral efficiency, massive multiple-input multiple-output (MIMO) has become a key technology in the fifth-generation wireless network. However, the problem of pilot contamination caused by multiplexing the same pilot between neighboring cell users restricts the performance of massive MIMO systems. To mitigate the impact of pilot contamination, this paper proposes a greedy matching-based pilot allocation (GMPA) algorithm for heterogeneous cellular networks. GMPA transforms the pilot allocation problem into matching optimization problem through matching theory. Then greedy method is used to find sub-optimal pilot allocation solution. Simulation results manifest that the proposed algorithm can effectively improve the performance of system spectrum efficiency and pilot efficiency.

Wen Zhao, He Gao, Yao Ge, Jie Zhang, Sanlei Dang, Tao Lu

Modeling and Simulation of Ionospheric Volume Scattering Equation for High-Frequency Surface Radar

Ionospheric clutter is always the key factor to restrict the detection performance of high-frequency surface wave radar (HFSWR). This paper established the radar range equations of the volume scattering, according to the unique distribution characteristics of ionosphere, and quantified the impact of ionospheric characteristic parameters on radar system. The simulation shows that the ionospheric echoes intensity under volume scattering can increase as the radar beamwidth increases, and the amplitude of irregularities fluctuation is the most important factor affecting the ionosphere intensity.

Yongfeng Yang, Wei Tang, Xuguang Yang, Xueling Wei, Le Yang

Research on the Elite Genetic Particle Filter Algorithm and Application on High-Speed Flying Target Tracking

Resampling is an inevitable process in the standard particle filter, but it also can lead to particles vanish diversity and degenerate the performance. In order to solve this problem, an elite genetic resampling particle filter is proposed in this paper. The global optimization of the genetic algorithm could keep particles move towards real state probability density function. The state estimate is corresponding to the maximum fitness state after several evolution generations. As the maximum fitness of every generation of the algorithm constitutes a non-negative bounded sub-martingale, this algorithm theoretically converges to the optimal global solution with probability 1. The estimate expression of absolute error is also concluded. The simulation demonstrates that this algorithm outperforming the particle filter using genetic operation in resampling could improve the estimation accuracy of high-speed flying targets tracking in the non-Gaussian background.

Lixia Nie, Xuguang Yang, Jinglin He, Yaya Mu, Likang Wang

Fault Estimation and Compensation for Fuzzy Systems with Sensor Faults in Low-Frequency Domain

A new fault estimation and compensation scheme of fuzzy systems with sensor faults is addressed in low-frequency domain. A descriptor observer is proposed to ensure dynamic error’s stability and $$H_\infty $$ H ∞ performance for low-frequency range. The faults estimation is obtained via the observer above. By considering estimation of faults, a $$H_\infty $$ H ∞ output feedback controller is shown such that controlled model with sensor faults considered has certain fault-tolerant function. A simulation proves this results’ effectiveness.

Yu Chen, Xiaodan Zhu, Jianzhong Gu

Anti-jamming Performance Evaluation of GNSS Receivers Based on an Improved Analytic Hierarchy Process

The anti-jamming performance evaluation of global navigation satellite system(GNSS) receivers in complex electromagnetism environment is important. In order to analyze the anti-jamming capability of GNSS receivers in different interference scenarios, an anti-jamming performance evaluation method based on an improved analytic hierarchy process (AHP) is proposed. Firstly, the anti-jamming performance index of GNSS receivers is given by using the constraint of the minimum performance index. Secondly, the improved analytic hierarchy process evaluation model is obtained by using three scales instead of nine scales. Furthermore, the evaluation model is verified by using measured data, which realizes the quantitative anti-jamming performance evaluation of GNSS receivers. The experimental results show that the proposed method can improve the credibility of the anti-jamming performance evaluation results for GNSS receivers.

Yuting Li, Zhicheng Yao, Yanhong Zhang, Jian Yang

Communication Optical Cable Patrol Management System Based on RFID + GIS

With the increasing proportion of optical cable in communication lines, it is very important and urgent to standardize the construction of optical cable and to maintain and manage it after completion. Aiming at the current situation of optical cable line maintenance and management, the paper puts forward the solution of communication optical cable patrol management system based on RFID + GIS, which provides stable and reliable communication guarantee for information construction.

Yang Mei, Jiang Yaolou, Zhou Bo, Qing Chao, Chen Zhen

Lip Language Recognition System Based on AR Glasses

With the rapid development of science and technology, researches on AR, speech recognition, and lip recognition are more mature. Speech recognition is used in many fields; however, the problem exists in which speech recognition is not accurate in noisy environments. To solve the problem, lip recognition can be used as an aid, but receiving real-time image information is difficult. AR is easy to carry and can obtain image information in real time. But few people currently combine lip recognition with AR. This primary goal of this research is the combination of lip recognition and AR glasses. The image information is obtained through the camera module of the AR glasses, next face detection is performed based on OpenCV, and then, the lip language information is compared with the corpus by image processing to form text and finally reflected on the screen of the AR glasses. We have found that on the premise of reasonable recognition distance, using AR glasses as a carrier, the lip-reading text is displayed in a three-dimensional scene, real-time mapping is realized, and people can see the results of lip recognition more intuitively. It is concluded that lip recognition based on AR glasses works well, is robust, and has strong real-time performance, after a large number of experiments.

Zhenzhen Huang, Peidong Zhuang, Xinyu Ren

Device-Free Human Activity Recognition Based on Channel Statement Information

The human activity recognition system has been researched in various technical fields for decades. Nowadays, as WiFi is widely used in homes, using WiFi devices for human activity recognition has become a better choice. In recent years, more and more researchers have begun to use channel state information (CSI) to realize human activity recognition. The CSI features with fine-grained information can show the impact of human activity on the channel. But in the existing CSI-based human activity recognition system, there is an issue. Due to the high proportion of error and noise in the phase information in CSI, the information in CSI is not fully utilized during the processing of CSI data. In this paper, we propose a method for extracting phase information in CSI, so that we can completely extract the effective information in CSI as the input feature of the classifier. Then, we use k-means for feature extraction of main feature data. Finally, we use support vector machines (SVM) to learn features and conduct activity recognition. We evaluated the system performance, and the experimental results show that our system has good performance.

Ruoyu Cao, Xiaolong Yang, Mu Zhou, Liangbo Xie

Wi-Breath: Monitoring Sleep State with Wi-Fi Devices and Estimating Respiratory Rate

As one of the human health detection indicators, breath is becoming the research emphasis. Traditional approaches based on wearable devices or pressure sensor devices are expensive to manufacture, and not suitable for daily use. In this paper, we present Wi-Breath, a monitoring system that extract the disturbing parts from received signals and estimate respiratory rate of the rest. We use wavelet transform to remove the noise of the signal amplitude. Then, by analyzing the time–frequency information of the signal, we use the sliding window method to segment the signal. In addition, we use the Fourier transform (FT) to calculate the respiratory rate of the rest. Experimental results show that Wi-Breath can reliably remove the disturbing part of the signal and estimate the approximate respiratory frequency.

Xin Yu, Xiaolong Yang, Mu Zhou, Yong Wang

Hard Examples Mining for Adversarial Attacks

This paper focuses on adversarial attacks and security on machine learning and deep learning models. We apply different methods of perturbation on pictures in ImageNet and record the success rate of examples which are successfully attacked and wrongly recognized and then conclude a graph to describe the relationship between the intensity of attack and the accuracy of recognition. Then, we figure out the reason of examples hard to be attacked which is determined by models or determined by examples themselves. Besides, we analyze the pictures which are extremely defensive to the attacks and find out some visual characters to support them stay strong.

Jiaxi Yang

Positioning Parameter Estimation Based on Reconstructed Channel Statement Information

Channel state information (CSI) is widely used in the wireless communication systems, but its application has been limited by the huge feedback overhead between the transmitter and the receiver. In this paper, the compressed CSI information is used to estimate positioning parameters, such as arrival of angle (AoA) and time of flight (ToF), and the experimental results show that the estimation accuracy of original CSI can be realized under the condition of reducing the feedback overhead of CSI in Wi-Fi systems.

Xiaolong Wang, Xiaolong Yang, Mu Zhou, Wei He

Three-Dimensional Parameter Estimation Algorithm Based on CSI

Considering that the problem of the estimation accuracy of angle of arrival (AoA) and time of flight (ToF) is limited by the number of antennas and channel bandwidth of commercial Wi-Fi, we propose a unified model, which provides three-dimensional parameter estimation based on AoA, ToF, and Doppler frequency shift (DFS). Our proposed algorithm analyzes the channel state information (CSI) data, constructs a three-dimensional matrix containing AoA, ToF, and DFS information, and then reduces the dimensionality of the constructed three-dimensional matrix. Finally, the spatial spectrum of the signal is estimated after dimensionality reduction. Simulation results show that the parameter estimation accuracy and the signal resolution of the proposed algorithm are better than the existing two-dimensional joint estimation method.

Yuan She, Xiaolong Yang, Mu Zhou, Wei Nie

Edge Cutting Analysis of Image Mapper for Snapshot Spectral Imager

The image mapper used in snapshot image mapping spectrometer (IMS) is a key optical element, which can slice the input image to different parts and reflect each part of the image to different directions. Edge cutting appears between adjacent mirrors when fabricating the image mapper using diamond raster fly cutting technique. The edge cutting on mirrors can introduce the loss of light throughput of the system and make stripe noise in the reconstruction datacube, so that this paper analyzes the effect of the edge cutting and makes optimization for the manufacture of the image mapper.

Xiaoming Ding, Yupeng Li, Xiaocheng Wang, Cheng Wang

Object Detection of Remote Sensing Image Based on Multi-level Domain Adaption

Due to the limited amount of labeled data, remote sensing object detection is faced with great difficulties. The problem of insufficient labeled data usually can be solved by domain adaption. However, current methods mostly focus on feature alignment, without paying attention to the context information or discussing the level of features, which makes it impossible to effectively apply them to remote sensing images. In this paper, we construct our method based on Faster-RCNN model, and design three domain adaptive components for remote sensing object detection at image-level, instance-level, and pixel-level. Image-level alignment enhances global recognition ability by image weight redistribution. Instance-level alignment makes global awareness possible by combining context information. Pixel-level alignment reduces local differences between domains by focusing on small features and enhancing semantic information. Moreover, we collect domain adaption dataset to verify the proposed method, and the experimental results show that our method is superior to other current methods.

Peiran Kang, Xiaorui Ma, Hongyu Wang

Research on Camera Sign-In System Based on SIFT Image Splicing Algorithm

This paper focuses on the image splicing algorithm based on SIFT to help teachers check in more quickly. Multiple images with different angles and overlapping areas monitored by the camera in classroom are selected for image splicing so as to solve the problem of limited view field of the camera lens. At the same time, spatial domain method is used to process the images. The result shows splicing degree is complete that can meet the application requirements of sign-in system with high precision and wide perspective.

Feng Long Yan, Hai Bei Zhang, Changyun, Ge

Research on Performance Optimization Scheme for Web Front-End and Its Practice

This paper focuses on the performance optimization of web front-end, proposing an optimization scheme of the system from the front end. It takes an information management website as an example, proposes a set of systematic, practical and operable website performance optimization solution based on web services. The solution includes loading external CDN resources, optimising component packaging process, lazy loading in route, Qiniu cloud acceleration, file compression, etc. It can improve the usability and performance of web application system, so as to improve the satisfaction of users through the optimization. At the end of the project, the comparative analysis results of website access performance before and after optimization are listed. The result shows that the size of top 5 resources reduced form 20 Mb to 79.7 kb. The image size can be reduce 94.3%, and the loading time can be reduced greatly, which decrease from the initial 25.27 s to the final 0.3 s.

Fenglong Yan, Zhao Xu, Yu Shan Zhong, Zhang HaiBei, Chang Yun Ge

Automatic Counting System of Red Blood Cells Based on Fourier Ptychographic Microscopy

Red blood cell (RBC) counting is of great medical significance in clinical examination. Commonly, the cell counting task is completed by microscopic examination, which requires a high resolution. This paper proposes an automatic counting system of red blood cells based on Fourier ptychographic microscopy (FPM) and estimates the RBC number via a convolutional neural network (CNN). The counting network is based on a regression model, using a VGG-16 network combined with a feature pyramid network (FPN). The experimental results show that the mean absolute percentage error (MAPE) of our counting network can reach 0.86%, which means a high accuracy.

Shushan Wang, Tingfa Xu, Jizhou Zhang, Xin Wang, Yiwen Chen, Jinhua Zhang

Implemention of Speech Recognition on PYNQ Platform

With the development of voice technology, human–computer interaction is widely used in the society. The audio processing in the front end of the Internet of Things needs a faster speed and lower power consumption. In this paper, on the PYNQ development board with more powerful embedded system, the combination of CNN network and FPGA is used to realize the rapid conversion of speech into text.

Wei Sheng, Songyan Liu, Yi Sun, Jie Cheng

Multi-target Tracking Based on YOLOv3 and Kalman Filter

This paper studies multi-target tracking based on YOLOv3 combined with Kalman filter. This method takes the detected target of the current YOLOv3 frame as the detection input of the next frame and iterates the predicted value of Kalman filter and the detection result to modify the tracking trajectory. This method can achieve high real-time performance and improve robustness at the same time.

Xin Yin, Jian Wang, Shidong Song

Load Balance Analysis-Based Retrieval Strategy in a Heterogeneous File System

In order to adapt to heterogeneous storage environment and improve the overall I/O efficiency and system throughput, we address the optimization of replica management and retrieval strategies in a distributed file system. The proposed strategies can enhance the system's adaptability to complex environment, improve the system's rationality for load balance, and reduce the overall storage cost of the system. Specifically, this paper presents a performance metric of measuring the load of nodes. Firstly, we address the concept of comprehensive load, and then propose the method of computing comprehensive load, which is based on the multi-dimension analysis on the system. In addition, we propose the replica management and retrieval strategies, which consider the comprehensive load of nodes in the distributed file system, and optimize the allocation of loads among nodes systematically. Based on the abovementioned strategies, we address the replica placement strategy, the replica management strategy, and the retrieval algorithm in a distributed file system, in consideration of the heterogeneity of nodes in the cluster, the difference between files, and the real-time performance of nodes. All these strategies and algorithms can provide an optimization of replica and retrieval process in a distributed file system.

Di Lin, Weiwei Wu, Yao Qin

Brain-Inspired Maritime Network Framework with USV

Communication has been completely integrated into our life and has laid a foundation for our investigation and exploration in maritime. Intelligent unmanned surface vessels (USVs) are envisioned to perform front services in space-air-ground-sea integrated network. It is obvious that USV improves the safety level of navigation, while faces a series of challenges such as data fusion and signal processing in space-air-ground-sea integrated network. Inspired by the brains’ powerful integration processability of vestibular and visual information, we propose a novel brain-inspired maritime information intelligent integration framework. First, we study signal processing, lidar and vision sensors detection of USV. Then, we propose to solve the problems of chaotic integration and low efficiency by using Bayesian theory. Finally, the simulation results analyze the discrepancy of likelihood function on information integration, and fully verify the superiority of brain-inspired information integration framework is carried by USV.

Xin Sun, Tingting Yang, Kun Shi, Huapeng Cao

Line-of-Sight Rate Estimation Based on Strong Tracking Cubature Kalman Filter for Strapdown Seeker

This paper addresses the inability of the strapdown seeker to indirectly measure line-of-sight (LOS) rate and the strongly nonlinear relative motion equations of missile and target. Strong tracking theory is combined with cubature Kalman filter (CKF), and a novel nonlinear filter for LOS rate estimation is presented; this filter is called the strong tracking CKF (STCKF). On the basis of the relative motion relationship of missile and target, the LOS rate reconstruction model is derived. Then, the STCKF algorithm is proposed by introducing a suboptimal fading factor into the predicted error covariance of the CKF. Finally, the validity and feasibility of the proposed algorithm are verified by simulation. Simulation results show that the proposed STCKF maintains the strong tracking ability for abrupt state changes. A comparison with CKF shows that STCKF can improve the LOS rate estimation precision for a strapdown seeker with better robustness and adaptability.

Kaiwei Chen, Laitian Cao, Xuehui Shao, Xiaomin Qiang

Malicious Behavior Catcher: An Intrusion Detection System Based on VAE

Network intrusion system is an important part of maintaining the security of the network. In this paper, we present a novel VAE-based network intrusion system. VAE is used to get a stable latent representation of the input vector to realize the function of dimension reduction. After that, the lower-dimensional latent representation is sent to the random forest classifier to get the final result. We compare our model with baseline method in KDD Cup’99 and NSL-KDD and the evaluation shows that our model outperforms baseline model in accuracy, recall, F-score and false alarm rate. Besides, our model needs much less consuming time and cost lower computational and storage resources which is more suitable for lower resource scenarios.

Linna Fan, Jiahai Yang, Tianjian Mi

Campus Physical Bullying Detection Based on Sensor Data and Pattern Recognition

Campus bullying is one of the primary problems in education around the world which makes teenagers drop out of school or even suicide. Campus bullying can take various forms, such as physical bullying, verbal bullying, Internet bullying and so on. Physical bullying is considered as the most harmful to teenagers. Therefore, it is necessary and significant to develop anti-bullying methods. In view of the current popularity of smartphones among the student population, this article proposes a scheme for using the smartphone’s built-in acceleration sensor and gyroscope to collect student activity data and using pattern recognition technology to identify students’ status. This paper uses Relief-F algorithm for feature selection and then uses PCA for feature dimensionality reduction and finally extracts three kinds of features from acceleration and gyrodata. The author used the k-NN algorithm as classifiers. In the final test, bullying and non-bullying recognition accuracies of k-NN were 84.13% and 76.92%, respectively. The result indicated that motion recognition based on k-NN can obtain a good classification effect in physical bullying detection.

Bo Long, Jian Liu, Jun Wang, Chenguang He

Speaker Recognition System Using Dynamic Time Warping Matching and Mel-Scale Frequency Cepstral Coefficients

Automatic speaker recognition is to recognize speech given by speaker automatically. The difference between speaker recognition and speech recognition is that it does not focus on the information of text symbols and semantic content with the speech signal included but fixes on the personal features with the speech signal included, furthermore, extracts these personal information features of the speaker to achieve the purpose of identifying speakers. Research on speaker recognition began in 1930s [1]. The early stage research mainly payed attention to the ear of human hearing recognition experiment as well as discovering possibility of listening recognition. Research work has moved from the pure human ear. Later, with improvement of electronic technology and computer technology, there is the possibility of automatically recognizing human voices through machines.

Yang Xue

MAN: A Multidimension Attention-Based Recurrent Neural Network for Stock Price Forecasting

Recently, deep learning has been applied to many fields, including finance. One limitation of existing neural network-based solutions is that their optimization target is the accuracy of the classification or regression task, rather than the ultimate return on investment. Another limitation is that they often treat stocks as independent entities, ignoring the interrelationships among them that are informative for stock forecasting. We put forward a time-feature-stock-wise attention-based RNN to consider the relationship between stocks and distill valuable information from multi-feature temporal data to forecast the stock price. At the same time, we take both the prediction accuracy and the sorting error of the forecast result as the optimization goal to maximize the final yield. To validate our method, we use our model to predict the opening prices of 10, 20, and 40 stocks randomly selected from Chinese stock market and made investment based on the prediction results. The back-test results showed that the return rate of this model was better than that of traditional stock forecasting solutions.

Weipeng Zhang

Driver Multi-function Safety Assistance System

With the increasing number of motor vehicles, the problem of traffic accidents follows. The survey shows that fatigue driving and heart diseases have become important causes of traffic safety accidents. In addition to real-time monitoring of the driver’s physiological health, the design also uses the electrocardiogram and grip strength signals to determine the driver’s fatigue level. Based on the important physiological information contained in the electrocardiogram signal and the basis of the time domain and time–frequency domain feature transformation of the grip strength signal, the fatigue characteristics were analyzed. The design uses STM32 as the control core, the steering wheel cover as the medium, and built-in electrode chips and pressure sensors to collect real-time ECG signals to determine the physiological health of the driver; at the same time, the ECG signal and the grip signal are analyzed at the same time, and the fatigue of the driver is analyzed. Real-time monitoring of the degree ensures the accuracy and stability of the system. In addition, the Bluetooth module is connected to the mobile phone terminal to realize the wireless communication function, and the terminal monitors various indicators of the driver’s body and the degree of fatigue in real time. When the body changes suddenly, the danger level is predicted, and the voice reminds the driver; and when the danger level is high, the satellite accurately locates and transmits the position to the family and the hospital in time, which is convenient for finding and timely rescue. When the driver feels tired or drowsy, the mobile phone terminal reminds or wakes up the driver through voice to realize the human–computer interaction function and prevent the occurrence of traffic accidents.

Wanqi Wang, Peidong Zhuang, Shiwen Zhang

Edge Computing-Enabled Dynamic Multi-objective Optimization of Machining Parameters

Dynamic events such as arrival urgent parts, due date changes, tool wear and so on are inevitable occurrences in machining processes. Optimizing the machining parameters in real time to respond to the dynamic events can significantly improve multiple machining performances. In this paper, an edge computing-enabled dynamic multi-objective optimization approach has been developed to achieve the real-time optimization of machining parameters. In the approach, edge servers are scheduled to provide the optimal computing resources. Based on the edge optimal computing resources, an improved dynamic two-archive evolutionary algorithm is developed to optimize the machining parameters and respond to the dynamic events. The proposed method is compared with edge computing resources’ random selection mechanism, normal dynamic two-archive evolutionary algorithm and NSGAII. The experiment results illustrate the high performance of the proposed method in the dynamic machining process.

Zhibo Sui, Xiaoxia Li, Jianxing Liu, Zhengqi Zeng

Improved Dark Channel Prior Algorithm Based on Wavelet Decomposition for Haze Removal in Dynamic Recognition

Environmental perception and precise positioning of area targets are key technologies in dynamic recognition. However, the perceptual information that is dynamically acquired in some haze weather cannot provide an accurate basis for decision-making and dynamic planning. In addition, under some outdoor condition, such as in the bright light, the quality of the perception information obtained by the system is usually low, and the robustness of the area target is relatively poor. In this paper, the implementation strategy of haze removal for images is comprehensively studied, and an improved dark channel prior algorithm is accordingly proposed by introducing the wavelet decomposition. The related experimental research is further carried out in the MATLAB development environment. As a result, haze can be effectively removed, and the real time of the improved dark channel prior algorithm can be largely enhanced.

Peiyang Song

A Novel Broadband Crossover with High Isolation on Microwave Multilayer PCB

The paper proposed a novel broadband crossover based on microwave multilayer PCB technique with two-layer structure. Low-pass PI filter is introduced to lower coupling capacitance between crossed transmission lines of slight RF performance degradation. The proposed crossover is optimized and simulated by commercial EM software HFSS; among the working band 1–18 GHz, return loss is better than 13 dB, insertion loss is better than 0.5 dB, and isolation between two transmission lines is better than 18 dB. The dimension of the proposed crossover is less than 2 × 2 mm, which is much smaller than traditional crossover on single layer.

Xu Yang, Jiancheng Liu, Meiying Wei, Xiaoming Li, Anping Li, Xiaofei Zhang

A Recognition Algorithm Based on Region Growing

Aiming at the uncertainty of object detection in the field scenic spot cloud data, combine the region growing algorithm and PointNet++. A recognition method based on regional growing is proposed. The method first divides the point cloud data into small regions, and regional growth algorithm was used to cluster the local regions; then the PointNet++ method is used to identify cluster objects. The experiment result shows that this method can achieve the segmentation and recognition of objects in the scenic spot cloud at the same time. The segmentation rate can be achieved 84.3%, and the recognition rate was 90.7%.

Luguang Wang, Yong Zhu, Chuanbo Wang

Point Cloud Simplification Method Based on RANSAC and Geometric Filtering

At this stage, point cloud data is applied to all walks of life. Different needs lead to different processing methods, and the point cloud collected by the Kinect camera contains a little noise, so it is necessary to process it to a certain extent. First, the data is taken by the Kinect camera to obtain the depth image; then the depth image is converted into a point cloud according to the camera parameters and the matrix transformation relationship, and the experimental object is obtained by segmentation using the RANSAC algorithm; finally, it is processed using geometric filtering. The experimental results prove that the method has good effect.

Chuanbo Wang, Yong Zhu, Luguang Wang

Compressive Autoencoders in Wireless Communications

Autoencoder, one of the most promising and successful architectures in deep learning (DL), has been widely used in wireless communications. However, fast-increasing size of neural networks (NNs) in autoencoder leads to high storage requirement and heavy computational overhead, which poses a challenge to practical deployment of autoencoders in real communications systems. In this paper, we investigate two representative NNs compression methods and propose two compressive autoencoder schemes for wireless communications by combining the compression techniques with autoencoder architecture. Our proposed schemes are capable to reduce memory consumption and execution time. Numerical experiments demonstrate that our methods can effectively compress the autoencoder’s size without degrading the model performance or distorting the constellations of transmitted signal.

Peijun Chen, Peng Lv, Hongfu Liu, Bin Li, Chenglin Zhao, Xiang Wang

Activity Segmentation by Using Window Variance Comparison Method

In the activity recognition, it is very important to extract the data when the activity occurs from the whole data. In order to solve this problem, we propose a window variance comparison algorithm. Calculate the window variance of the preprocessed data and compare it with the window variance before or after, to determine the occurrence and continuation of the activity. Experimental results show that this algorithm can accurately and effectively achieve activity segmentation.

Xinxing Tang, Xiaolong Yang, Mu Zhou, Lingxia Li

Multi-carrier-Based Positional Modulation Design

With DM technology, the signal transmitted by the transmitter can be received in the desired direction/directions correctly, while in other directions, the eavesdropper cannot obtain the information. However, if eavesdroppers aligned with the desired direction/directions, the DM design will cause a problem as the received signals by eavesdroppers are similar to the one received by the desired user. In this paper, positional modulation (PM) as an extension of directional modulation technology is proposed, which overcomes the shortcoming of DM technology by finding the appropriate weight vector to make a given modulation pattern at certain desired positions but scrambles the patterns at other directions. The advantage of the proposed multi-carrier-based design is that multiple signals over multiple frequencies can be achieved at receivers simultaneously.

Qingnuan Hu, Bo Zhang, Jinjin Zhang, Jin Gao, Yang Li, Xiaonan Zhao, Cuiping Zhang, Cheng Wang

Directional Modulation Design with a Transmission Power Constraint

Directional modulation (DM) has been studied from many different aspects recently. However, the overall transmission power of the antenna array for each symbol is not considered. In many practical scenarios, the power supply to the system is rather limited, and therefore, the transmission power of the antenna array may need to be restricted. In this paper, a fixed power constraint for DM design is proposed, and the non-convex constraint can be modified into a convex form, which can be easily solved by existing convex optimisation toolboxes.

Jinjin Zhang, Bo Zhang, Jin Gao, Qingnuan Hu, Wei Liu, Yang Li, Xiaonan Zhao, Cuiping Zhang, Cheng Wang

Circular Antenna Array-Based Directional Modulation Design

Directional modulation (DM) technology as an extension of beamforming technology has drawn much attention recently. With DM technology, the transmitted information is modulated at the transmitter’s RF end, where the phase information of the transmitted signal is changed by adjusting the phase shifter at the RF end, and then, the desired signal can be formed in the desired direction. In other directions, the received signal not only changes in power, but also changes in phase information, making information difficult to be cracked by eavesdroppers. In this paper, a circular antenna array-based DM design is proposed for the first time, and different circular radius on directional modulation technology is studied.

Jin Gao, Bo Zhang, Qingnuan Hu, Jinjin Zhang, Yang Li, Xiaonan Zhao, Cuiping Zhang, Cheng Wang

An Improved Visual Inertial Odometry in Low-Texture Environment

Recent trends focus on the applications of the autonomous positioning of mobile robots. To solve the limitations of pure vision simultaneous localization and mapping (SLAM) system in unknown environment with few textures, repeated scenes and obstructions, this paper adopts a tightly-coupled visual-inertial odometry (VIO) system as the overall framework. We exploit a method of feature matching by combining both point and line features to improve the system robustly. Compared with point features, line features provide more structured information of the environment and can complete the localization tasks where it is difficult to extract point features. Additionally, the line segments are further filtered and merged by the least squares algorithm for more effective tracking. Based on Plücker coordinates and orthonormal representation, the observation model of line features is given. The experiment results of the Euroc datasets show that our improved system can effectively overcome the low-texture issues and it has higher positioning accuracy. Our method supports both monocular and stereo camera, which makes it possible for more application scenarios.

Zi-Yuan Song, Chen-Yang Yan, Fei Zhou

A Pole Extraction Algorithm Based on Sliding Windowed Matrix Pencil Method

This paper proposes a pole extraction algorithm for robust radar target recognition based on Sliding Windowed Matrix Pencil Method (SW-MPM). The time domain scattering data of the target are windowed and the poles are extracted by Matrix Pencil Method (MPM) with a sliding window. The number of poles parameter is changed to improve the reliability of poles extraction, and the azimuth consistency of poles extracted from multiple azimuths is analyzed. We simulate and verify the effectiveness of the pole extraction method under different directions of the aircraft target.

Tianbao Zhang, Yang Zhou, Hongliang Gao, Zhian Deng, Chunjie Zhang, Jiawei Shi, Jianguo Yu

A Recursive Algorithm for the Energy Minimization of the Data Packets Transmission with a Low Computational Complexity

5G technology provides reliable connectivity to the Internet of things (IoTs) and faster data transfer to users. With the increasing number of the devices connecting to the Internet and the requirements of the transmitting information, the energy used for the transmitting data increases rapidly. For the perspective of environmental protection and sustainable development, it is significant to take strategies to reduce energy consumption during transmission. In this paper, we provide a recursive algorithm to minimize the energy consumption for the data transmission, and the data arrival rates and the circuit power are considered. The algorithm proceeds iteratively and achieves the purpose of energy saving by allocating the start times and the transmission duration of the data packets. Then, we verify the validity of the proposed algorithm by numerical simulation.

Yunhai Du, Chao Meng, Zihan Zhang, Wenjing Ding

Speech Emotion Recognition Algorithm for School Bullying Detection Based on MFCC-PCA-SVM Classification

School bullying is common in school life, which has a negative impact on students’ physical and mental health. Nowadays, the research on school bullying at home and abroad mostly relies on human resources. In this paper, a school bullying detection algorithm based on pattern recognition techniques is presented. The authors firstly collect and pre-process the emotional speech data, and extract the MFCC features. Then they reduce the dimension of features to 6 by the PCA algorithm. They design a two-level SVM model in series with linear kernel for classification. The algorithm proposed in this paper effectively achieves a high recognition performance. The accuracy of the algorithm reaches 86.59% and the F1-Measure reaches 87.36%.

Yuhao Wang, Xinsheng Wang, Chenguang He

3D Anchor Generating Network:3D Object Detection via Learned Anchor

In both 2D and 3D object detection algorithms, the reference boxes obtained by the anchor mechanism lay the foundation for the next detection tasks of the network. Most of the existing anchor mechanisms are designed by hand to generate a dense array of anchors. The size of the anchors obtained in this way is single and the anchors are densely distributed in the image space. It has poor robustness and a large number of redundant anchors. In views of these defects, we propose the 3D anchor generating network. It predicts the 3D anchors by learning the semantic features of the picture, in which the anchors are sparsely distributed around the objects in the image with different sizes in different positions of the image. We applied it to TLNet’s baseline monocular network for 3D object detection on KITTI dataset, the result showed a significant improvement on the performance of 3D object detection algorithms.

Huinan Li, Yongping Xie

Handoff Decision Using Fuzzy Logic in Heterogeneous Wireless Networks

Visible light communication (VLC) systems are becoming a promising means of wireless communication, due to high data rate, low implementation cost, and immunity to radio frequency (RF) interference. However, the VLC systems often suffer from service disruptions because of the limited coverage of light. In this paper, we consider a VLC–WiFi heterogeneous system to take the advantages of both technologies, including increased capacity and ubiquitous coverage. We integrate VLC and WiFi and propose a handoff decision using fuzzy logic in heterogeneous wireless networks. The design of handover decision algorithm is described to dynamically distribute resources. Experiments show the system works well and can avoid disconnection effectively.

Liwei Yang, Qi Zhang

Campus Bullying Detection Based on Speech Emotion Recognition

Campus bullying is now receiving worldwide attention and a great number of measures have been introduced. This article introduces speech emotion recognition into the detection of the bullying incidents on campus. In this paper, the audio data are pre-processed through the frame division and endpoints detection. Based on the victim’s emotional changes during bullying, 36 MFCC features are extracted. This article uses the Wrapper algorithm, combined with KNN classifier for features selection and dimensionality reduction, which constructs the systemic model. The accuracy of model learning performance is about 80%, and it has been proved to be stable, which shows a promise in campus bullying detection.

Chenke Wang, Daning Zhang, Liang Ye

MTFCN: Multi-task Fully Convolutional Network for Cow Face Detection

Automatic cow detection with computer vision is a hitting topic in recent years. Recent studies found that adding facial alignment task can improve the performance of the detection. Instead of using the cascaded method to fine tune the candidate boxes, we propose an end-to-end multi-task fully convolutional network (MTFCN) which outperforms multi-task cascaded convolutional networks (MTCNN) on the collected cow dataset. In addition, focal loss is adopted to focus on hard samples that are hard to be classified by the original model. The addressed network has an average precision (AP) with 91.71%, while the AP of MTCNN is 88.11% on our own cow dataset.

Ziyan Wang, Fuchuan Ni, Na Yao

Boosted Personalized Page Rank Propagation for Graph Neural Prediction

Graph neural networks have recently achieved impressive success on many graph learning problems, including semi-supervised graph node classification. Previous message-passing algorithms are confronted with the dilemma of aggregating more neighbor’s information and avoiding over-smoothed node representations. The recently proposed personalized page rank propagation scheme separates the node prediction and information propagation into independent stages and achieves state-of-the-art performance on public datasets. However, the personalized page rank propagation scheme results in a computationally effective computation graph in model training. In this paper, we propose a boosted personalized page rank propagation scheme to reduce the model complexity and computational cost. The proposed algorithm is based on the principle of compressing multiple iterations in personalized page rank propagation into a one-step operation to simplifying the computation graph. Experiments on three public datasets reveal that the proposed algorithm reduces more than 50% training time, especially when the iteration number increases, without sacrificing the model’s prediction accuracy.

Wenwen Xia, Fucai Luo

An Improvement Single Beacon Positioning Algorithm Using Sparse Extended Information Filter for AUV Localization

This paper presents an algorithm for single beacon positioning that enables autonomous underwater vehicle (AUV) to obtain exact location. A dynamic calculating model is proposed, utilizing the slant ranges measured by single beacon and navigation sensor information equipped with AUV. Furthermore, an algorithm based on sparse extended information filter is proposed to improve the localization accuracy and efficiency. Finally, simulation based on field data is conducted to verify the performance of the proposed algorithm.

Wanlong Zhao, Huifeng Zhao, Jucheng Zhang, Yunfeng Han

An Algorithm Design for Fish Recognition Based on Computer Vision

Fish image recognition has always been a challenging task. In this paper, we propose a recognition algorithm based on computer vision for the four major species, silver carp, bighead carp, herring and grass carp. Firstly, the algorithm uses guided filtering to filter out noise and blur points in the sample image, which can help obtain accurate contour feature values of the fish body. Then the image is enhanced in order to avoid the loss of the contour of the fish body due to the influence of brightness. After the contour of the fish is drawn, the feature points on the contour are extracted according to fishtail angle and aspect ratio, based on which the fish kind could be concluded. The research results show that the recognition algorithm provided in this study can accurately classify and identify four kinds of fish with an accuracy of 96.25%.

Guangyao Chen, Yiran Zhou

An Iterative Multi-channel DDPSK Receiver for Time-Varying Underwater Acoustic Communications

In mobile underwater acoustic (UWA) communications, Doppler spreads have a serious disturbance in communications quality. In this paper, we propose a novel iterative multi-channel double differential phase shift keying (DDPSK) receiver to improve the performance. Meanwhile, we enhance the DDPSK receiver by adopting the multi-symbol estimation to track the channel variation, and an extended trellis diagram is constructed. Furthermore, the multi-channel soft-input soft-output (SISO) based on maximum a posterior (MAP) criterion is proposed. Besides, the effective bit-interleaved coded modulation with iterative decoding (BICM-ID) is applied. Specially, extra gain can be obtained by utilizing spatial diversity with multi-channel receiver. Theoretical simulation shows that the receiver equipped with five arrays can obtain around 5.5 dB gain over the single array.

Zhihui Wu, Chao Gao, Feng Gao, Junhong Peng, Jie Wu

An Improved LC Visual Attention Detection Algorithm

Image visual attention detection is a hot research topic in recent years and an important research direction in the field of computer vision. An excellent image visual attention detection algorithm will greatly improve the accuracy of subsequent image segmentation, target recognition, and image retrieval so that we can have a more comprehensive and accurate understanding of the target contained in the image. This paper in view of LC image attention detection algorithm without considering the spatial information of image pixels places saliency map, this paper proposes an improved LC image attention detection algorithm, the method fully considers the pixel grayscale contrast and the spatial distance of saliency map, and the influence of the mathematical model is established for each pixel point distribution of the significance of exclusive value. Finally, we used MATLAB to program the algorithm and tested the algorithm on MSRA 10K, the test set of public image data. The experimental results showed better performance than LC algorithm, which proved that our idea was effective.

ZhiYang Zhao, BaoJu Zhang, CuiPing Zhang, JiaZu Xie, Man Wang, WenRui Yan

Joint Source-Channel Coding Scheme Based on UEP-Raptor

Raptor code is a solution used for forward error correction at the application layer during multimedia transmission. The amount of multimedia transmission data is huge. In order to transmit high-quality image and video under a fixed bandwidth, data must be compressed before transmission, and compression techniques such as jpg picture compression, MPEG-based video compression, and perceptual compression ROI coding are used. In this way, the compressed data has different reliability levels during transmission. This paper proposes a fountain code with unequal error protection, combined with image compression to construct a joint source-channel coding scheme based on UEP-Raptor code. Finally, the experiment proves that the adoption of this scheme has effectively improved the data transmission quality.

Chang Liu, Wenchao Yang, Dezhi Li, Zhenyong Wang, Zhenbang Wang, Haibo Lv

Energy-Efficient UAV-Based Communication with Trajectory, Power, and Time Slot Allocation Optimization

The unmanned aerial vehicle (UAV) can not only build temporary communication network in the area of lack of ground facilities, but also easily expand the network capacity in hot areas, which has aroused significant research interests in recent years, especially in 5G network. In this paper, our goal is to minimize the total transmission power of UAV for specific tasks and achieve energy-efficient UAV-based communication by optimizing UAV trajectory, power, and time slot allocation. Firstly, a mixed integer non-convex optimization problem is constructed. Secondly, a time slot allocation algorithm based on channel gain and the amount of user task are proposed. Last, the problem is solved by alternating optimization and sequence convex approximation. Compared with the benchmark, the simulation results show that our policy can significantly reduce the total transmission power of UAV under the conditions of satisfying the UAV range, minimum user transmission rate, and user tasks.

Liping Deng, Hong Jiang, Jie Tian, He Xiao

UAV-Assisted Wireless Communication Network Capacity Analysis and Deployment Decision

Deploying the unmanned aerial vehicle (UAV) as an aerial base station to establish or assist wireless communication networks is a promising technology. However, there are also some significant challenges involved in planning, designing, and deploying the UAV base station. In this paper, we aim to find out the optimal location of UAV, by maximizing the capacity of all users. First, we model the rician factor as a function of the elevation angle between the terrestrial user and UAV, and derive the exact ergodic capacity expression, and in addition, give the approximate closed-form expression with arbitrary small error. Subsequently, an iterative algorithm based on particle swarm optimization (PSO) is proposed to solve the optimal deployment of UAV. The effectiveness of the algorithm is demonstrated on the different distribution of the terrestrial users.

Qiuyun Zhang, Hong Jiang, Qiumei Guo, Jie Tian, Fanrong Shi, Mushtaq Muhammad Umer, Xinfan Yin

An Edge Detection Algorithm Based on Fuzzy Adaptive Median Filtering and Bilateral Filtering

As the basic feature of image, edge can be used to identify target, extract feature and provide valuable feature parameters. In the process of image acquisition, transmission and processing, it is inevitable that images will be affected by different degrees of noise. Gaussian noise and salt and pepper noise, as common noises, are often the causes of image blurring and deformation. As the traditional Canny algorithm does not have the disadvantage of removing the salt and pepper noise, this paper adopts the hybrid de-noising method of fuzzy adaptive median filtering and bilateral filtering to remove the salt and pepper noise while achieving the effect of maintaining the edge and smoothing the noise reduction. At the same time, 5 × 5sobel operator and Oust adaptive threshold are selected to better obtain edge information and improve edge connection. The experiment shows that the improved edge detection algorithm is better than the traditional Canny algorithm when adding high-density salt and pepper noise. It can filter out the sundries and identify the subject target.

Man Wang, Bao Ju Zhang, Cui Ping Zhang, Jia Zu Xie, Feng Juan Wang

Fusion Dynamic Time Warping Algorithm for Hand Gesture Recognition in FMCW Radar System

In this paper, we propose a fusion dynamic time warping (FDTW) method for hand gesture recognition using a frequency modulated continuous wave (FMCW) radar. First, we use two-dimensional Fast Fourier Transform (2D-FFT) to estimate the parameters of the radar raw data, and construct the range-time spectrogram (RTM) and Doppler-time spectrogram (DTM). Second, a central time–frequency trajectory is clustered for each hand gesture spectrogram using the k-means algorithm, and then the hand gesture is recognized by the proposed Fusion Dynamic Time Warping (FDTW) algorithm. Experiments with radar data show that the proposed algorithm can significantly improve the recognition accuracy with the time complexity reduced by more than 50%.

Aihu Ren, Yong Wang, Mu Zhou, Xiaolong Yang, Liangbo Xie

Vital Signs Detection Using a FMCW Radar Sensor Based on the Discrete Wavelet Transform

Due to the peculiarity of vital signs, breathing and heartbeat signals are so weak that they are easily submerged in radar noise and clutter. Therefore, it is difficult for the existing technology to directly use the frequency-modulated continuous wave (FMCW) general radar for vital sign recognition. A rigrsure adaptive soft threshold noise reduction based on discrete wavelet transform (RA-DWT) algorithm is proposed to separate and reconstruct the respiratory and heartbeat signals. The algorithm is proved by measured data that it can effectively suppress harmonic and noise interference, and greatly improve the accuracy of respiration and heartbeat detection using the FMCW radar.

Wen Wang, Yong Wang, Xiaobo Yang, Mu Zhou, Liangbo Xie

PHCount: Passive Human Number Counting Using WiFi

Aiming at the problem that the low accuracy and limited detection range in traditional passive human number counting algorithms, this paper proposes a passive human number counting system using WiFi, which is called PHCount. We use the amplitude fluctuation of the channel state information (CSI) to indicate the change of the number of people in the target indoor environment, based on which, combining the regression prediction model and the classification model to estimate the number of people with low training overhead. The experimental results show that the PHCount estimates the accuracy rate of the number of people in the training sample range is higher than 95%. Besides, the prediction accuracy rate of the number of people beyond the training sample range is higher than 87%.

Xi Chen, Zengshan Tian, Mu Zhou, Jianfei Yu, Bin Luo

Gait Cycle Detection Using Commercial WiFi Device

Recently, the research of human gait recognition has attracted wide attention in the field of wireless sense, especially in the fields of public security, clinical medicine, indoor security and so on. Most of the existing gait recognition methods are based on radar detection, cameras, infrared sensors, wearable devices, etc. Camera-based methods are vulnerable to obstacles, light intensity and other factors, with the privacy which is poor. The price of radar equipment is expensive and not universal. The method based on wearable equipment needs active cooperation of users. Wearable equipment will also reduce the convenience and comfort. However, the gait feature extraction method based on WiFi signal channel state information (CSI) makes up the above shortcomings. In this paper, we extract useful gait features from the interference CSI signals and extract human motion speed features by establishing multiple links, realizing human gait cycle detection (WGCD). A large number of experiments show that the WGCD achieves detection accuracy of 93.56%.

Gongzhui Zhang, Zengshan Tian, Mu Zhou, Xi Chen

Improved Spectral Angle Mapper Applications for Mangrove Classification Using SPOT5 Imagery

This paper describes the classification principle of the spectral angle mapper (SAM) and combines pixel purification, training sample set optimization, and watershed image segmentation algorithms to improve SAM. A mangrove area in Guangxi’s coastal zone was used as the study site, and the SPOT5 remote sensing image data were used as the data source. The improved SAM was used to classify the mangrove ecosystem, and classification accuracy was assessed. The results showed that the improved SAM had a good classification effect. It not only took into account the particularity of the complex spectral composition of the land type, but also effectively avoided the fragmentation of the results. The overall accuracy was greater than 90%, and the KAPPA coefficient was 0.8804, which showed that improved SAM has application potential for classification and information extraction of mangrove remote sensing. That provides support for commercialized remote sensing monitoring of mangrove ecosystems.

Xiu Su, Xiang Wang, Derui Song, Jianhua Zhao, Jianchao Fan, Zhengxian Yang

A Novel Method for Direction of Arrival Estimation in Impulsive Noise Environments

To overcome the limitation that the alpha-stable distributed variable possesses infinite second-order moment, a novel fractional lower-order correntropy-analogous statistics (FCAS) is defined. Furthermore, a novel DOA estimation method based on FCAS for impulsive noise is proposed. The simulation results demonstrate that the proposed method can get good estimation results.

Li Li, Derui Song, Xiaofei Shi, DeJun Zou

Sea Utilization of Different Marine Industries in the Bohai Economic Rim

Improving the economic efficiency of marine space resource utilization is an important measure for building maritime power. This study analyzes the sea utilization in different marine industries in the Bohai Economic Rim by constructing a mathematical index, including the diversity of the marine industrial structure, economic efficiency, and an index of water quality. The conclusions drawn are as follows: (1) Fisheries and coastal tourism are the major marine industries in the Bohai Rim region, and their distribution is concentrated. (2) Diversity of the marine industrial structure is generally low in the Bohai Rim, in Liaoning, which has a value of less than 20%. (3) Economic efficiency, which has basically been positively related to the diversification index, is good overall in the Bohai Rim. (4) The overall degree of water quality is better than the Grade III of the water environment quality. The impact of the marine industries on water quality is obviously different. Sea utilization for marine transportation, protection of marine resources, and sewage dumping has a significantly negative impact on marine water quality.

Jianli Zhang, Derui Song, Limei Qu, Hao Zhang, Jingping Xu, Yujuan Ma

Interference Analysis Between Satellite and 5G Network

As the development of 5G technology gradually matures, the integration of satellites and 5G cellular networks has gradually become the focus of industry attention. In particular, the current scarcity of spectrum resources is relatively scarce, and spectrum sharing becomes a key technology in integrated satellite-terrestrial network. Frequency reuse between systems is bound to bring a series of interference problems. This paper combines the ultra-dense networking characteristics of 5G cellular systems and analyzes that the interference of 5G systems to satellites mainly originates from a large number of 5G co-frequency base stations. Through simulation, we have obtained that it is possible to effectively control interference by limiting the number of base stations, but it is not feasible to limit the number of base stations due to the requirements of 5G system performance indicators. In addition, we also found that in the case of high base station transmit power, even if the number of base stations is small, the carrier-to-interference ratio at the satellite receiving end is still negative. Therefore, we can conclude that reducing the transmit power of 5G base stations is a feasible method to reduce satellite and terrestrial 5G systems.

Xinxin Miao, Mingchuan Yang

Analysis of Spatial Correlation for Wideband HAP-MIMO 3-D Channel Model

To solve the problem that 2-D wideband HAP-MIMO model cannot accurately describe the spatial channel characteristics, this paper analyzes the spatial correlation of the channel based on 3-D HAP-MIMO model. A 3-D model introduced an elevation parameter, can more appropriately describe the channel to obtain more accurate spatial correlation characteristics of the channel. Simulation results show that by increasing the base station elevation, and reducing the degree of non-isotropic scattering, and increasing the degree of scattering dispersion, and increasing the maximum horizontal distance between scatter and terrestrial mobile station. All of these cases can reduce correlation and increase channel capacity.

Bingyu Xie, Mingchuan Yang

Channel Characteristics Analysis and Modeling of Deep Space Communication Link

Deep space exploration is one of the three major aerospace activities of mankind in the new century, and deep space exploration is inseparable from the research on technologies of deep space communication. In the future, the goal of human space exploration will be extended to more and farther stars, analyzing the basic characteristics of the deep space communication link channel, and studying the specific communication problems of the nearer stars will be the necessary basis for the research of deep space communication. Focus on the characteristics of the communication channel between Earth–Moon and Earth–Mars. According to the influential parameters of different communication link, the impact of the communication link of Ka frequency band and below is simulated and analyzed to clarify each range of loss and the effect of each parameter, and the relevant channel characteristics of deep space communication link are obtained.

Xin Guan, Mingchuan Yang

Design of Global Coverage Satellite Constellation Based on GEO and IGSO

A hybrid global coverage satellite constellation is designed based on GEO and IGSO satellite considering that GEO satellite can provide favorable coverage performance to low latitude region while IGSO satellite can achieve high elevation angle to high latitude area. The designed satellite constellation can offer double-satellite coverage to the Chinese and nearby area, high elevation angle to high latitude area and seamless coverage to global area. The simulation results show that the Chinese and nearby area can achieve 100% single-satellite coverage and 87% double-satellite coverage. Besides, global area can reach more than 95% single-satellite coverage and North Pole can get communication elevation angle over 30°.

Yingzhe Dou, Xiaofeng Liu

Markov Chain-Based Statistical Model of Three-State SIMO Land Mobile Satellite Channel

In order to analyze the applicability of single-input multiple-out (SIMO) technology in land mobile satellite system, this paper establishes a three-state SIMO-LMS channel statistical model based on Markov chain. Due to the lack of accurate measured data, the establishment of the model in this paper combines the experimental results of the SISO-LMS system with the modeling method of the terrestrial wireless MIMO communication system. According to the established model, the performance of the system such as outage capacity and average bit error rate is simulated. The results show that compared with SISO-LMS channel, the outage capacity of SIMO-LMS channel increases and the average bit error rate of the system decreases. These results provide strong support for the application of SIMO technology in satellite mobile communication system.

Xinxin Miao, Xiaofeng Liu

Network Coding Over Hybrid Satellite-Terrestrial Two-Way Relay Network

This paper proposes a new communication protocol combining MIMO and network coding called MIMO-NC (MIMO and network coding). In this paper, MIMO-NC is applied in hybrid satellite-terrestrial two-way relay communication system with the two source nodes are out of communication range to each other. At the relay node, STBC code and decode and forward (DF) cooperative mode is used. Application of MIMO and coding schemes can get the spatial and coding gain. The process of MIMO-NC protocol is constituted of source broadcast stage and relay forward stage. The theoretical performance analysis of MIMO-NC is based on a binary symmetric relay channel model to get the maximum capacity. Comparison of MIMO-NC and another two-relay protocols is presents by the results of Monte Carlo simulation. The results show that MIMO-NC has the best BER and capacity performance among those three protocols, especially in low SNR case.

Xin Guan, Xiaofeng Liu

Performance Analysis and FPGA Implementation of Layered Space–Time Code for Mobile Satellite Communication

Layered space–time coding is one of space–time codes with the maximum multiplexing gain. In the paper, we investigate the using of iterative detection in the field of mobile satellite communication systems and analyze the performances of the algorithm in three typical channels. The performances of the simulation results meet the law that the open case is better than the urban case and suburban case. During the process of implementing the receiver, a new method of implementation with FPGA is proposed. The comparison between results by soft simulation and ones by hardware implementation shows the performance of soft simulation is a little better than the one of hardware implementation. At last, we analyze the differences of the situation.

Bingyu Xie, Xiaofeng Liu

Performance Analysis of CRDSA Based on M2M Flow Model

Internet of Things (IoT) is an important application scenario in the fifth generation of mobile communications. Machine to machine (M2M) is the most common form of IoT application at present. With the rapid development of IoT in the world in recent years, M2M systems have been widely used in various fields. M2M system data transmission can be combined with satellites. In this paper, the flow model of M2M is presented, and then the access protocol of contention resolution diversity slot ALOHA (CRDSA) is deeply analyzed in the satellite M2M scenario. According to its transmission mechanism, the throughput boundary is derived, and then the continuous interference cancellation is combined with diversity slot ALOHA, which greatly improves the throughput. The simulation results show that the CRDSA scheme can actually run under a standardized load up to 0.4, which allows the transmission of small and medium-sized burst packets without allocating resources, just like the operation mode of satellite broadband network.

Xiaoling Fu, Yanyong Su

Performance Analysis of Satellite Internet of Things Access Protocol

Internet of things is an important application scenario in the fifth generation of mobile communications, yet it is difficult to place IoT nodes or base stations in a series of areas where people are sparsely populated or difficult to reach. In order to solve this problem, the concept of satellite IoT came into being. However, how to solve the problem of collision and interference between users when a large number of low-cost terminals access the network at the same time is a major challenge. Therefore, it is necessary to design an applicable access protocol for the satellite IoT service. Aiming at the current satellite IoT scenarios, several common random access protocols which are more suitable in this scenario are introduced, and their performance simulation and comparison analysis are carried out. It is concluded that CRDSA and IRSA are more suitable for scenarios with sink nodes, and the pure ALOHA and SA are more suitable for scenario without sink node.

Xiaoling Fu, Yanyong Su

SCMA Multiple Access Technology in Satellite Internet of Things System

The satellite communication technology is suitable as a supplement to the ground communication network. It can be applied in the Internet of things (IoT) system to form a satellite IoT system. However, the satellite IoT system has its inherent disadvantages that need to be solved urgently. Therefore, how to improve the spectrum utilization rate of the satellite IoT system while maintaining good error performance has become a key research issue. In this paper, we propose to apply SCMA technology to the satellite IoT system. SCMA technology is a code domain non-orthogonal multiple access technology. Its basic idea is to map user bits into high-dimensional complex digital codewords according to a predetermined codebook set at the transmitting end, and then all users multiplex their codewords on the same time–frequency resource according to a factor graph (or mapping matrix). When the received signal, channel information and codebooks are known, the receiving end can use the message passing algorithm (MPA) to perform correct decoding. The simulation results show that the error performance of the SCMA system is affected by the number of iterations of the decoding algorithm. When the appropriate number of iterations is selected, the SCMA system can improve the spectrum utilization rate while maintaining good system error performance. This feature makes SCMA technology have important advantages in improving the spectrum utilization rate of satellite communication system and supporting large-scale machine-connected communication.

Ruiran Liu, Yanyong Su

Duplex Mode Switching for Underlay Cognitive Radio Systems Based on Outage Probability

Cognitive radio (CR) and full-duplex (FD) communications, which can improve the spectrum efficiency of mobile communications, have drawn significant research interests. To further improve the spectrum efficiency, some researchers have begun to integrate FD into CR system. In this paper, a duplex mode switching scheme based on outage probability is proposed for the underlay CR system. Firstly, the outage probabilities of the second user for two modes are derived on the promise of satisfying the communication quality of the primary user. Then, by comparing the outage probabilities, we can obtain the switching threshold. Finally, the performance improvement of the proposed mode switching scheme is verified by simulations.

Ranran Zhou, Liang Han, Yupeng Li

A New Security-Oriented Multi-dimensional Assessment Method for Perception Layer of Electric Internet of Things

The electric Internet of things (IoT) is supporting the operation and maintenance of smart grid in much deeper and refined ICT-aided decision-making mode. However, known from the architecture of IoT, its two core functions (i.e., ubiquitous sensing and everything interconnecting) depend greatly on the infrastructural service in terms of secure sensing and reliable networking provided by its perception layer. Hence before reinforcing the core functions of electric IoT, it is more important for the operators of smart grid to know the security situations of its perception layer. Based on our previously proposed security-oriented assessment framework for the perception layer of the electric IoT, this paper designs a comprehensive security-oriented assessment method, which combines with the analytic hierarchy process (AHP) and the fuzzy comprehensive evaluation method. In practice, the proposed method gets the weight of the index by AHP and further reduces the dimension of the index system by fuzzy comprehensive evaluation. Finally, taking the smart substation site as an example, this method is used to assess the security-oriented of its power IoT perception layer and verify the feasibility and effectiveness of the method.

Yuxuan Yang, Jingtang Luo, Chenyang Li, Shujuan Sun, Yuanshuo Zheng, Min Zhang

The Security-Oriented Assessment Framework for Perception Layer of Electric Internet of Things

With the widespread deployment of the electric Internet of things (IoT) in smart grid, it is one of crux tasks for the smart grid operators how to effectively assess and enhance the security protection capability of its perception layer. Moreover, the security of electric IoT is paid more and more attentions by both of academia and industry since it has become one of important components of national cyberspace security. Guided by national security requirements and standard documents related to network and IoT security, this paper proposed a security-oriented assessment framework for perception layer of electric IoT. Based on the method of analytic hierarchy process (AHP), the evaluation index system and the index weights were determined, and then the entire process of the assessment framework was completed. Finally, this paper has taken an intelligent substation scene as an example to demonstrate the process of the assessment framework and verify the feasibility and effectiveness of the index system and the security-oriented assessment framework for perception layer of electric IoT.

Jingtang Luo, Quan Tang, Shujuan Sun, Chenyang Li, Yuanshuo Zheng, Min Zhang

Optimizing Frequency Reuse in Multibeam Satellite Communication Systems

Frequency reuse (FR) is one of the most effective approaches for mitigating co-channel interference and improving the capacity of communication systems. However, there are some difficulties in applying it in multibeam satellite communication systems due to the nonuniform beam layout. In areas where beams are crowded, the interference from the adjacent cells is extremely high. By contrary, in areas with sparse beam layout, the interference from the surrounding cells is relatively low. Therefore, different regions require different frequency reuse factors (FRF), which brings about what the performance of traditional FR suffers seriously. In order to solve this problem, an optimized frequency reuse scheme is put forward.

Weizhong Zhang, Wenchao Yang, Zhenyong Wang, Dezhi Li, Qing Guo

A Critical Links Identification Method for Interdependent Cyber-Physical System of Smart Grid

With the deepening of the interdependent of power and communication network components in smart grid, the critical links of smart grid increase. It is one of crux tasks for the smart grid operators how to accurately and quickly identify the critical links of smart grid under interdependent. Aiming at the traditional method of identifying critical links without considering the electrical interdependent characteristics of network nodes, this paper proposed a power network model which combines electrical interdependent and physical topology connection. Then based on this, a new algorithm of N-K critical links identification strategy based on tabu search is proposed. Simulation results show that the algorithm has high accuracy and fast identification speed, so it is more suitable for the identification of high-order critical links in large smart grid.

Jingtang Luo, Hechun Zhu, Jian Zeng, Ganghua Lin, Jiuling Dong, Min Zhang

An Estimation Algorithm of Power System Frequency in Impulse Noise Environment

With the widespread use of distributed power generation devices and nonlinear power electronic devices in smart grid, the noise in the actual environment of the smart grid shows obvious impulse characteristics. To solve the above problems, this paper proposed a generalized weighted linear prediction algorithm based on the least p norm (lp-GWLP). That is, combining the linear prediction property with the weighted least p norm. Firstly, the whitening transform is used to preprocess the data to remove the correlation in the error. Then the least p norm (lp) is used to calculate the independent error data. Finally, the iterative reweighted least square method is used to achieve the frequency estimation. Simulation results show that the proposed scheme can achieve reliable frequency estimation in the case of harmonic, amplitude oscillation, and voltage interruption in the impulse noise environment.

Shiying Yao, Yiming Chen, Jingtang Luo, Liangtao Duan, Jiuling Dong, Min Zhang

Research on Broadcasting of Beidou Differential Information Based on AIS

BeiDou differential information is an important part of BeiDou satellite navigation system application. Based on AIS, it can broadcast BeiDou differential information. It can provide social users with precision navigation positioning services from the centimeter level to the meter level, and increase the Beidou navigation system, improve the service quality of shipping, and provide technical support for shipping upgrading development.

Xianfeng Zhao, Yanjun Fang, Yang Zhang, Chen Wang

Analysis of Electromagnetic Shielding Model Applied in Wireless Energy Transmission System

Electronic wireless equipment will produce electromagnetic interference, which can affect the surrounding environment and the normal operation of other equipment. This paper builds basic model of wireless energy transmission and electromagnetic shielding model, and analyzes the influence of structure and material of electromagnetic shielding to the wireless energy transmission system based on finite element method. The simulation results indicate that the electromagnetic field attenuates effectively after the addition of shielding body, and the shielding effect of the semi-enclosed structure is slightly better than the flat plate structure under the same power supply and system frequency. With the increase of coil distance, the smaller the receiving current is, the worse the shielding effect is. Ferrite is a magnetic shielding material, while metallic aluminum is an electrical shielding material. Adding ferrite will increase the magnetic induction intensity between coils, while adding aluminum plate will reduce the magnetic induction intensity between coils.

Xiu Zhang, Yang Li, Xin Zhang, Ruiqing Xing

Analysis of the Influence of Subarray Beamforming on Frequency-Hopping Communication System

Frequency-hopping phased array system has the advantages of flexible and variable beam, strong multi-beam capability, anti-interference, and anti-intercept, but its front-end beamforming will consume more hardware and software resources. This paper proposes a simplified subarray-level digital beamforming strategy for resource-limited frequency-hopping phased array system. The influence of the simplified strategy on the performance of frequency-hopping signals is analyzed and simulated, which has certain guiding significance for the actual system design.

Zhengyu Zhang, Qinghua Wang, Yongqing Zou, Xin Wang

EEG Feature Selection Based on SFFS for Audiovisual-Induced Emotion Recognition

Emotion is the mental state of human being accompanied by cognitive process. It plays a very important role in human communication, and emotion recognition has broad application prospects in the fields of human–computer interaction and medical rehabilitation. Electroencephalogram (EEG) signals are not easy to disguise, respond sensitively, and have objective results, which are the current research hot spots. At present, the accuracy of emotion recognition in dimension model needs to be improved, which depends on the features extracted from the EEG signal and the feature vectors fed into the classifier. To solve this problem, we proposed a non-uniform sampling multivariate empirical mode decomposition (MEMD) method to obtain the characteristics of emotional EEG signals and a hybrid sequential floating forward selection (SFFS) method to select less redundancy features. The experimental results were verified on the audiovisual-induced database DEAP for emotion analysis.

Yi Wang, Kun Chen, Yue He, Zhilei Li, Li Ma

Radio Resource Allocation for Centralized DU Architecture of 5G Radio Access Networks

In 5G radio access networks (RAN), centralized unit (CU) and distribution unit (DU), splitting of RAN architecture has been proposed. In CU-DU splitting architecture, the limited power and bandwidth of midhaul link become the bottleneck of the RAN. In this paper, a radio resource allocation algorithm is proposed which aims to minimize the bandwidth and power consumption of the midhaul link for centralized DU architecture. A joint optimization problem of the bandwidth and power consumption of the midhaul link is established, which is solved by multi-agent distributed Q-learning algorithm. Simulation results show that the proposed algorithm can effectively reduce the bandwidth and power consumption of the networks.

Mengbin Gao, Tiankui Zhang, Nan Zhao, Jing Li, Liwei Yang

Low-Energy and Area-Efficient Scheme with Dummy Capacitor Switching for SAR ADCs

In low-power SAR ADCs, the energy consumed by capacitor DAC is the main source of the SAR ADC. In order to reduce the energy consumption, this paper proposes an energy-efficient and high area-efficient DAC switching scheme. By introducing dummy capacitor switching for the last two bits’ generation and a new third voltage reference Vq equal to a quarter of the voltage reference Vref, the proposed DAC switching scheme reduces the switching energy by 97.97%. Moreover, the DAC area of the proposed scheme is only 12.5% of the conventional SAR ADC architecture, which achieves a good area efficiency.

Dengke Yao, Liangbo Xie, Ruoyu Cao, Mu Zhou

SVDec: UAV Detection System Using Software-Defined Radio

In recent years, with the increasing popularity of drones, some non-standard operations have caused many safety problems, and this phenomenon has become increasingly serious. In order to solve these safety problems, the existence of unmanned aerial vehicle detection systems is critical. Since many monitoring systems are costly, low-cost detection systems are particularly important. This paper designs a system that can detect drones by using the signal characteristics of the drones as detection standards. The system first uses software-defined radio (SDR) to monitor and collect wireless signals and then extracts the movement and vibration characteristics of the drone and finally classifies the signal through the joint features of the classifier, which can effectively distinguish the drone signal from other movements. The system has a high accuracy for UAV detection.

Yi Li, Wei Nie, Wei He, Mu Zhou

Recent Progress of ISAR Imaging Algorithms

Inverse synthetic aperture radar (ISAR) has the characteristics of all-day, all-weather and long-range imaging. It realizes the imaging and recognition of noncooperative targets and has a wide range of applications in military and civilian fields. In this paper, the range Doppler (RD) algorithm, range-instantaneous Doppler (RID) algorithm, compressive sensing (CS) theory-based ISAR imaging algorithm, as well as three-dimensional interferometric inverse synthetic aperture radar (3-D InISAR) imaging algorithm are introduced. The prospect of ISAR imaging technology is given.

Yong Wang, Yuhong Shu, Xiaobo Yang, Mu Zhou, Zengshan Tian

Research Status and Development of Battery Management System

A rechargeable battery pack built together with a battery management system (BMS) has been used on a large scale for electric vehicles, micro grids and industrial machinery. As an electronic control system, BMS is able to make sure the battery’s safe operation and monitor battery cell’s states such as current, voltage and temperature. Besides, it can also estimate the battery’s state of charge (SOC) and exchange data with the master controller. This paper critically reviews the present research situation by investigating representative BMS products of various companies and research institutions. Based on topological structure, bus communication, signal acquisition and control system, software and hardware architectures are provided.

Panpan Liu, Changbo Lu, Changfu Wang, Xudong Wang, Wanli Xu, Youjie Zhou, Hua Li

Temperature Data Acquisition and Safety Characteristics Analysis of Lithium-Ion Battery’s Thermal Runaway

The analysis of thermal runaway temperature of lithium-ion batteries hit by gunpowder is of great significance to improve the safety of large energy storage facilities and mobile energy storage square cabin. Based on the self-designed experimental device and data acquisition system, a series of experiments shooting on lithium-ion batteries was carried out. Meanwhile, infrared temperature data is acquired and analyzed. The results show that the surface temperature of all the batteries has the first peak due to the gunshot. And the second peak is formed due to thermal runaway inside the battery. The larger the batteries’ state of charge is, the higher the first peak of batteries’ surface temperature is. Overall, the order of the lithium-ion batteries’ safety from high to low is lithium titanate battery, lithium iron phosphate battery, lithium nickel manganese cobalt battery and Li-sulfur battery.

Wanli Xu, Changfu Wang, Changbo Lu, Dongkai Ma, Xudong Wang, Panpan Liu, Yaohui Wang

Image Acquisition and Analysis of Lithium-Ion Battery’s Thermal Runaway

In order to avoid the lithium-ion battery’s security accidents caused by shooting in large military energy storage bases, it makes sense to monitor the entire failure process of the lithium-ion battery in the shooting test. The battery damage form, bullet hole metal droplets and other image data are obtained using high-speed cameras, digital cameras, infrared thermal imaging cameras, etc. And the mechanism of thermal runaway of battery is analyzed from two aspects of mechanical abuse and thermal abuse. The results show that the lithium-ion battery has six types of damages such as explosion, spitfire, fire, smoke, fume and leakage.

Wanli Xu, Changbo Lu, Changfu Wang, Yao Nie, Xudong Wang, Youjie Zhou, Hua Li

Weak Light Characteristic Acquisition and Analysis of Thin-Film Solar Cells

The battery of individual equipment has always been the component that occupies the largest volume. Therefore, the flexible thin-film solar cell is an ideal energy source for individual equipment. This paper tested volt-ampere characteristics of three kinds of solar cells, that are, respectively, made of Si, copper indium gallium selenide (CIGS) and perovskite. The research investigates the open-circuit voltage, short-circuit current, maximum operating power, and photoelectric conversion efficiency, and the test data are analyzed and discussed. The results show that when the light irradiance ≤ 1000 W/m2, with the increase of light irradiance, the short-circuit current and the maximum operating power increase linearly, and the open-circuit voltage increases more and more slowly in logarithmic relation, the photoelectric conversion efficiency of Si solar cells increases rapidly first and then decreases slowly. The photoelectric conversion efficiency of CIGS thin-film solar cells basically remains while the light irradiance ≥ 500 W/m2.

Wanli Xu, Changfu Wang, Changbo Lu, Hui Sun, Xudong Wang, Yanli Sun, Litong Lv

Research Status and Key Technologies of Long-Distance Laser Energy Transmission System

Due to the small energy storage capacity and short endurance of the unmanned equipment, the existing pluggable wired charging mode has become a bottleneck restricting the energy support of the unmanned equipment. This paper introduces the research status of long-distance laser energy transmission, discusses the optimization theory of laser wireless energy transmission system, the technology of dense spectral combing diode laser and high-efficiency multi-junction laser battery energy conversion, etc. With the further improvement of its power, distance and efficiency, laser wireless energy transmission can reduce the dependence of unmanned equipment on energy storage battery or charging cable, significantly extend the endurance time and mileage of unmanned equipment and improve its battlefield viability and combat effectiveness.

Wanli Xu, Changfu Wang, Changbo Lu, Panpan Liu, Xudong Wang, Mengyi Wang, Lei Xu

Research Status of Unmanned System and Key Technologies of Energy Network

Unmanned system has attracted great attention due to its advantages on completing tactical tasks independently and adapting to complex and harsh environment which can effectively improve combat capability and reduce casualties. The research status of the unmanned system which in the global is introduced in the article firstly. The development trend of the unmanned system has been expounded in the aspects of cluster cooperation, autonomy control, power energy, etc. The key technologies and core research directions of the unmanned system are discussed about the short-distance energy transmission network, long-distance energy transmission network and multi-objective dynamic-adaptive power supply network.

Wanli Xu, Changfu Wang, Changbo Lu, Mengyi Wang, Xudong Wang, Panpan Liu, Weigui Zhou

Analysis of the Effect of Coil Offset on the Efficiency of Magnetic Coupling Wireless Power Transfer System

Magnetic coupling wireless power transfer has the advantages of high transmission efficiency, strong environmental adaptability and low hardware cost, which can provide fast wireless power for unmanned equipment in the battlefield. This paper studies the effect of the coil offset (axial offset, lateral offset and angle offset) on the transmission efficiency of the system. It is simulated by a theoretical model and experimented based on a magnetically coupled wireless power transfer platform. The experimental results show that the transmission efficiency of the system decreases with the increase of the axial and lateral offset of the coil and decreases first, then slows down and finally drops sharply with the increase of the angle offset of the coil.

Wanli Xu, Hang Zhang, Changbo Lu, Shizhan Li, Xudong Wang, Panpan Liu, Yaohui Wang

Data Collection and Performance Analysis of Lithium-Titanate Battery Charging and Discharging at Low Temperature

In a low-temperature environment, the performance of lithium-ion batteries is limited greatly due to the slower chemical reaction rate inside the battery and the hindered migration of lithium-ions. Low-temperature charging and discharging tests were carried out on the lithium-titanate battery by using the test device and data acquisition system. During the tests, some parameters were monitored such as the internal resistance, voltage, and temperature, and the low-temperature performance of the lithium-titanate battery was analyzed. The results show that the internal resistance and the charging voltage continue to increase, while the discharging voltage and capacity continue to decrease with the environmental temperature decrease.

Wanli Xu, Changbo Lu, Changfu Wang, Lijie Zhou, Xudong Wang, Youjie Zhou, Hua Li

Research on Feature Extraction of Piston Knocking Vibration Signal and Analysis of Correlation Degree of Wear

Taking the knocking vibration signal of diesel engine piston as the research object, the knocking behavior of piston is analyzed by dynamic simulation, and the changing rule of knocking time and energy with the clearance of cylinder liner are obtained. On this basis, the knock components of the vibration signals of diesel engine cylinder head under different motorcycle hours are extracted by using the matching tracking algorithm. The evaluation model of diesel engine cylinder wear is constructed with the threshold correlation method by taking knock vibration time and knock energy as the key parameters. It is verified that the model can accurately evaluate the wear condition of diesel engine cylinder and has a high availability.

Xudong Wang, Xiaolei Li, Feng Wang, Wanli Xu, Yanli Sun, Youjie Zhou, Lei Xu

Duplex Mode Selection for Underlay D2D Communications Based on Energy Efficiency Maximization

This paper studies the energy efficiency of half-duplex and full-duplex D2D communication in a wireless cellular network. Firstly, the energy efficiency of half-duplex and full-duplex under the optimal power allocation is obtained when the channel capacity requirements are met. Through comparing the energy efficiency of half-duplex and full-duplex, by adjusting the self- interference coefficient and communication distance, we can make the scheme of communication mode selection. Considering that the factors affecting the energy efficiency may not be unique, we use the method of control variables to simulate energy efficiency. Finally, by comparing the simulation results, the best choice strategy of communication mode is obtained.

Xueqiang Ren, Liang Han, Yupeng Li

The Performance Investigation of Direct Detection Optical OFDM System with Different Modulation Formats

Optical orthogonal frequency division multiplexing (OOFDM) has been shown to be able to resist channel dispersion (CD) and polarization mode dispersion (PMD) and it is considered as a promising long distance optical transmission technology. Direct detection optical OFDM (DDO-OFDM) system is a low-cost system and insensitivity to laser phase noise. In this paper, Optisystem and MATLAB are used to jointly build a DDO-OFDM system, and the transmission performance of CO-OFDM based on M-QAM is studied. Through the simulation, we find that in M-QAM-OOFDM system, with the increase of M, the signal is more sensitive to noise, so in general, we use 16QAM to realize signal transmission.

Guoqing Wang, Yupeng Li, Liang Han, Xiaoming Ding

Novel OFDM-Based Self-interference Channel Estimator for Digital Cancellation in Full-Duplex

In order to operate in full-duplex, the self-interference (SI) caused by the simultaneous transmission and reception should be efficiently mitigated. In this paper, a novel self-interference (SI) channel estimation method for digital SI cancellation is proposed in orthogonal frequency division multiplexing (OFDM) full-duplex system where the SI signal is always strong. Conventional least square (LS) algorithm, which directly removes the cyclic prefix (CP) that contains the useful SI channel information, may degrade the SI channel estimation performance. In contrast to the LS algorithm, the proposed SI channel estimation contains the SI channel information in CP. This indicates that the proposed algorithm can achieve better SI channel estimation. Simulation results also show that the proposed algorithm performs better SI suppression especially when dealing with high self-interference-to-noise-ratio (INR) in full-duplex.

Fei Wu, Liang Han

Multiple Hypothesis Tracking with Mixed Integer Linear Programming

The multiple hypothesis tracking (MHT) algorithm is widely used in many tracking applications. In the track-oriented approach, the reconstruction of the global hypotheses is closely related to the maximum weight independent set problem, which is NP-hard. Its high computational complexity makes it not scale to practical large problems. In this paper, we propose a mixed integer linear programming (MILP)-based approach to enumerate m-best global hypothesis efficiently. Compared to the MHT implementation that only generates the best hypothesis, MILP-based approach exhibits better overall tracking performance in the simulation experiment. By efficiently generating multiple optimal global hypotheses, MILP-MHT is able to prune impossible track hypothesis as early as possible.

Dingbao Xie, Zhaolei Liu

An Adaptive Filtering Technique for Hypersonic Targets Based on Acceleration Estimation

Hypersonic targets usually adopt high maneuverability and fast flight mode, which cannot be described by accurate motion model without prior information. In this paper, an adaptive filtering technology based on acceleration estimation is proposed. The CA and singer models are used for parallel filtering and interaction. The adaptive process noise of the filtering model is dynamically estimated to improve the matching degree between the filtering model and the actual target motion. The simulation results show that the proposed adaptive filtering technique has higher tracking accuracy in hypersonic target tracking compared with the traditional IMM algorithm.

Lei Gu, Jianguo Yu, Qiang Huang, Zhenlai Xu

TDOA Estimation of a Single Wi-Fi Access Point Based on CSI

Single access point (Access Point, AP) positioning can significantly improve the availability and signal coverage of Wi-Fi positioning. Time difference of arrival (TDOA) is an available positioning signal for single access point positioning. This paper proposes an arrival (TDOA) estimation algorithm of a single Wi-Fi access point based on channel state information (CSI). The algorithm first constructs a spatial spectrum estimation model through the phase shift introduced by each subcarrier of each antenna of a single access point, then use the smoothing algorithm to get a smooth CSI matrix. TDOA does not directly utilize the signal arrival time, compared with TOA, there is no need to maintain accurate clock synchronization, and the accuracy of delay estimation is relatively improved. Experiments verify that TDOA estimates have reached sub-nanosecond error accuracy.

Ming Zhang, Qingzhong Li, Jianguo Yu, Zhian Deng

Numerical Analysis of the Ill-Posedness of Ground-Based 2D Radar Short-Arc Orbit Determination

2D radar is often used in battlefield surveillance but is unable to obtain the pitch information about the target, resulting in inaccurate target state estimation. Therefore, it is significant to study the 2D radar short-arc orbit determination problem. This paper numerically analyzes the ill-posedness of this problem. Firstly, by examining the Fisher information matrix, the solution of 2D radar short-arc orbit determination can be described as a single-parameter orbit family. Secondly, a feasible method to obtain the starting orbit is given based on the Lambert theory. Lastly, the homotopy method is employed to calculate the solution family of this problem. Using the algorithm proposed in this paper, this ill-posed orbit determination problem is solved and the solution is described as a curve in the state space.

Zhengtao Zhang, Qiang Huang, Jianguo Yu

Simulation Study About the Radar Cross Section of a Typical Targets Based on FEKO

The radar cross section is an important parameter to characterize radar target characteristics. The fluctuation of the data represents the size and shape of the target. Target RCS depends not only on radar frequency and polarization mode but also on the attitude of target relative to radar. Taking a cruise missile as the research object and combining with FEKO software simulation platform. Set the corresponding working frequency, working wavelength, material parameters, geometric parameters, excitation, etc., and then define the plane wave, and then conduct mesh generation to generate the calculation model. This paper established an efficient RCS characteristic calculation based on the high-frequency approximation method. The influence of configuration materials and incident wave parameters on RCS characteristics of cruise missile was studied and some significant results were obtained.

Feng Wang, Pengyuan Liu, Zhonglin Wei

Detection of Malicious Nodes in the uav Network

Network is a structure in which multiple objects communicate with each other through wireless connection. Nowadays, Internet of things, wireless sensor networks, mobile ad hoc networks, uav ad hoc networks, and vehicular ad hoc networks have become common networks. These networks are often used to collect useful data or information to monitor or predict the environment. The attacker invades the nodes in the network to carry out malicious attacks such as drop attack, replay attack, and tamper attack, resulting in the loss of data authenticity and integrity, which leads to serious network security problems. Therefore, the detection of malicious nodes in the network has become a research hotspot. Common detection schemes include: Establishing reputation value mechanism to identify malicious nodes; machine learning algorithm is used to classify nodes in the network; through the mutual cooperation of the nodes in the network, monitoring each other's forwarding behavior and detecting malicious nodes, and so on. This paper summarizes and analyzes the existing methods of detecting malicious nodes in the network.

Jia Chen, Yanzhi Zhu, Shaofan Zhu

A UAV-Swarm Control Platform Architecture Based on Cloud

In the recent 20 years, the research of UAVs has attracted the attention of researchers. At the same time, the drone swarm control system has also developed rapidly. However, the way of controlling a single drone through the drone controller cannot meet the control needs of the drone swarm, which significantly limits the interaction of the drone swarm, the sharing of its resources, and the ability of the drone to perform complex tasks. This paper proposes a centralized cloud platform-based drone cluster control platform architecture for collaborative control of drone swarm, which is combined with workflow technology to allow multiple users to control multiple drones to perform tasks and monitor in parallel via the Internet. The state of the man–machine is not restricted by time and distance and can be connected to the cloud platform anytime and anywhere. Besides, the architecture also provides cloud storage and cloud computing support for drones and improves the support of complex tasks for drones. We conducted experiments through the simulator to verify the effectiveness and feasibility of the architecture.

Li Zeng, Zesheng Zhu, Xuzhou Shi, Yulei Liu

A Survey: Development and Application of Behavior Trees

In recent years, artificial intelligence technology has been greatly developed in the game world. Many games use finite-state machines (FSM) to control the intelligent behavior of non-player-controlled characters (NPC). However, due to the complexity of FSM state transition and low reuse, the technology of behavior tree gradually began to develop and replace FSM to a certain extent. This article introduces research on the basic concepts, development trends, and the latest technologies of the behavior tree. In addition, we explore the application and development of the behavior tree in some fields other than the game industry.

Wang Zijie, Wang Tongyu, Gao Hang

A Survey on Security and Privacy in Spatial Crowdsouring

Spatial crowdsourcing (SC) is a new pattern of crowdsourcing. Spatial crowdsourcing increases the potential for crowds to perform tasks related to real-world scenarios involving physical location, which is not feasible in conventional crowdsourcing methods. SC requires workers to move to specific locations to perform spatial tasks (e.g., perception, activity organization). However, SC still faces many challenges in terms of security and privacy. This paper introduces the basic architecture of spatial crowdsourcing and investigates the research status of this field. Then, the security and privacy issues/challenges in SC task assignment are further elaborated, and potential future research directions are proposed.

Mengting Shi, Mengqi Li, Yuping Zhang

Application of Machine Learning in Space–Air–Ground Integrated Network Data Link

The space–air–ground integrated network is an emerging network architecture integrated by satellite, aerial network, and ground communication, which can provide seamless connection on a global scale. However, the limited energy and spectrum resources cannot meet the growing communication needs, and its high heterogeneity, complex variability affect the reliable and efficient end-to-end transmission of services. In addition, machine learning is widely used. Using machine learning algorithms to solve problems in the space–air–ground integrated network is a new research idea for us. Therefore, this paper first introduces the concept and characteristics of apace–air–ground integrated network, summarizes, and analyzes the application of machine learning algorithms in solving the problems of resource allocation, attack detection, target recognition and location and security authentication of the space–air–ground integrated network, and looks forward to its prospects for development in space–air–ground integrated network.

Shaofan Zhu, Shuning Wang, Jia Chen

Survey on UAV Coverage Path Planning Problem

The advanced and highly mobile design of UAVs has led to a wide range of applications in fields such as battlefield monitoring, intelligent agriculture, photogrammetry, and disaster management, all of which are based on the coverage path planning problem (CPP). In recent years, there have been numerous researches on UAV coverage path planning problem. It is very important to plan an effective flight path for UAV to cover the whole target area completely. The purpose of this paper is to summarize and analysis of the existing literature involving different methods to solve the problem of UAV coverage path planning, we will survey the coverage path planning algorithm is mainly divided into two categories: one is geometric algorithm based on polygon region decomposition, and the second is heuristic algorithm for variant TSP and GTSP problems based on grid division and gives out the summary and analysis method for every type of coverage.

Jiankang Xu, Xuzhou Shi, Zesheng Zhu, Hang Gao

Research on Multi-sensor Fusion Algorithm Based on Statistical Track Correlation

First, a multi-sensor sequential track association algorithm is discussed in this paper, and then modified by means of threshold adaptive technique. Due to the correlation of track estimation error among the sensors, the cross-covariance between the local tracks must be considered in the process of track fusion. However, the calculation of cross-covariance is very complicated, and the amount of calculation is very big as well, especially in the condition of existing false association probability when the tracks are being associated. Considering the above factor, a pseudo-measurements combination filtering algorithm based on statistic track association is proposed, having avoided the complexity of cross-covariance computation for distributed system and data association in dense targets and clutter environment for centralized system. It approaches the optimal solution as well. Simulation results show its effectivity.

Jinliang Dong, Lei Bian, Yumeng Zhang, Huifang Dong

Research on the Improved EMC Design of Vehicle-Borne Rotating Phased Array Radar

In this paper, the main contents of EMC design for complex system are introduced firstly, based on a vehicle-borne rotating phased array radar project. Then, based on the initial test results of the radar RE102 project, the EMC characteristics of the system are analyzed, and the improved EMC design measures of the rotating phased array radar are proposed. Finally, it is verified by test that it has passed the RE102 project testing in the radar system including servo device for the first time.

Gan Wang, Jinliang Dong, Yumeng Zhang, Huifang Dong

RBF Neural Network-Based Temperature Error Compensation for Fiber Optic Gyroscopes

A temperature error compensation scheme for fiber optic gyroscope (FOG) based on radial basis function (RBF) neural network is proposed in this paper. By using the data preprocessing and orthogonal least square (OLS) learning method, the training performance of the network is improved and the over-fitting of the network is prevented. The experimental results illustrate that the proposed method has a 15–40% performance improvement compared with the conventional linear regression model.

Wei Cai, Jianyang Wang, Wenhui Hao, Yanguo Zhou, Yiming Liu

A Multi-Beam Forward Link Precoding Algorithm for Dirty Paper Zero-Forcing

This article considers the severe co-channel interference caused by the efficient use of spectrum by multiple beams combined with full-frequency multiplexing. After establishing a forward link model that considers severe rainfall attenuation in higher frequency bands such as Ka, the classic low-complexity precoding algorithm for zero-forcing is improved, and a regularized zero-forcing precoding algorithm considering the influence of system noise is proposed. Based on the dirty paper coding idea, a low-complexity dirty paper regularization zero-forcing precoding algorithm is proposed, which maximizes the signal-to-interference and noise ratio, thereby increasing throughput.

Yumeng Zhang, Jinliang Dong, Lei Bian, Gan Wang

UAV Detection and Recognition Based on RF Fingerprint

Unmanned aerial vehicle (UAV) has brought many threats to security and privacy. This paper designs a UAV detection and recognition system based on RF fingerprint. The system first utilizes a radio frequency front end to receive wireless signal; and second, determine whether the amplitude of the received signal is greater than a preset threshold; thirdly, determine whether the autocorrelation function of the suspicious signal has periodicity to detect UAV. Finally, the signal features are extracted as the fingerprint of the UAV, and the k-nearest neighbor (kNN) algorithm is utilized to complete the classification and recognition of the UAV.

Zhi Chao Han, Wei Nie, Mu Zhou, Qing Jiang

Soil Humidity Classification Based on Confident Learning via UWB Radar Echoes with Noisy Labels

Aiming at the classification problem of soil UWB radar echo signals with noisy labels, this paper proposes a new method for classifying soils with different VWC based on confidence learning and logistic regression. Ten UWB soil echo signals with different VWC were applied to this method. Simulation experiment results show that this method has a better classification performance than the logistic regression model without confidence learning. This proves that this method can solve the problem of noisy labels in the practical application of UWB to determine soil parameters and has certain practical application prospects.

Chenghao Yang, Jing Liang

Laser Point De-Fuzzy Method Based on Stitched Background Mapping

In the process of laser pointing, there will be laser speckle phenomenon, which affects the image observation effect. In this paper, a laser point de-fuzzy method was proposed, which provided a new idea for reducing the influence of laser speckles. Firstly, the background was stitched by affine transformation of the image, and then, the location information of laser point was acquired. Finally, the new optical point was added to corresponding position in the background by calculation. The video with expected clear laser point and no color difference background is obtained. The experimental results show that the background color difference caused by the optical filter has been eliminated, and the new light spot is clearly mapped in the correct position, which achieves the effect of removing the light spot fuzzy and accurately tracking the laser point.

Wenrui Yan, Baoju Zhang, Chuyi Chen, Jingqi Fei, Cuiping Zhang, Zhiyang Zhao

Z-NetMF: A Biased Embedding Method Based on Matrix Factorization

Network embedding represents the graph in low dimensions, improving the processing of big scale tasks. As node2vec can only be modeled as a tensor, and NetMF cannot be generalized to a biased form directly. In this paper, we draw Z-Laplacian, a framework that describes the dynamic process on the graph, into NetMF, so that NetMF can be extends to a biased form. The biased version of NetMF is named as Z-NetMF. We analyze the biased random walk in the view of graph signal processing and prove the effectiveness of Z-NetMF in the node classification task. The result shows that Z-NetMF outperforms the comparison methods, NetMF and node2vec.

Yuchen Sun, Liangtian Wan, Lu Sun, Xianpeng Wang

Hyperspectral Band Selection Based on Improved K-Means Algorithm

Hyperspectral imaging has emerged as a promising technique due to its use of hundreds of contiguous spectral bands. However, the huge number of bands are normally redundant with high inter-band correlation, and this leads to a need of band selection which has become increasingly important in hyperspectral data exploitation. K-means clustering is applied to band selection, but K-means algorithm is sensitive to the initial clustering center. Random selection of the initial clustering center may select data points that are close to each other, resulting in poor results. In this paper, the maximin distance algorithm is combined with k-means clustering-based band selection. Firstly, the maximin distance algorithm is used to calculate the initial clustering center, and then, k-means is used to cluster the bands. Finally, the band with the largest variance is selected as the output from each cluster. The experimental results of two public hyperspectral datasets show that compared with other methods, the band combinations selected by the proposed method can achieve better classification accuracy.

Yulei Wang, Qingyu Zhu, Yao Shi, Meiping Song, Haoyang Yu, derui Song

Joint Classification of Multispectral Image and SAR Image Based on Deep Feature Fusion

Different imaging sensors which are mounted on thousands of remote sensing platforms collect various information of the land covers. Since more sensors provide more information, the classification of multi-sensor data has potential advantages. Multispectral image collects information from visible spectrum, and SAR image reflects information of microwave band. However, due to the redundant information, multi-sensor data will also bring challenge to traditional classification method. This paper presents a joint classification method which combines the information from images from both multispectral sensor and synthetic-aperture radar (SAR) sensor. The proposed method is based on deep feature fusion, which is a deep network with two feature learning branches for multispectral image and SAR image separately, and then, the two branches are merged together into more fully connected layers to perform feature fusion and optimization; finally, a classification layer is added on the top of the network to predict sample label. The proposed method takes advantage of reciprocal information from different sensors and gives a strategy to utilize multi-sensor information. Experimental results demonstrate that our method is able to give better performance than using any single data.

Beimin Xie, Xinrong Wang, Peiran Kang

Advances of Power Supply Technology for Unmanned Aerial Vehicle

This paper introduces the definition and classification of unmanned aerial vehicle, as well as the functional characteristics and technical status of unmanned aerial vehicle at home and abroad. It focuses on lithium battery, fuel cell, solar cell, and new hybrid power technology, and finally discusses the development direction and trend of power technology.

Wanli Xu, Changbo Lu, Youjie Zhou, Xuhui Wang, Weigui Zhou, Mengyi Wang, Lei Xu

Design of Fuel Cell UAV Power System

The use of fuel cells in long-haul drones is a direction of their development. According to the general requirements of a certain type of UAV, this paper adopts modular design, combined with fuel cell selection and self-design of key components to complete the design of fuel cell drone power supply, and finally proposes the next work direction.

Wanli Xu, Changbo Lu, Youjie Zhou, Xuhui Wang, Hua Li, Lei Xu, Mengjie Hao

Overview on Micro-grid Technology Research

Microgrid is an effective way for connecting distributed generation to the power grid. Microgrid technology, as a key technology for renewable energy generation and distribution, has attracted more and more attention from countries and regions in the context of the environmental problems and energy crisis now. The development and extension of microgrids can facilitate the large-scale intervention of distributed power generation and renewable energy, and promote the transition from traditional power grids to smart networks. This article introduces the microgrid technology in detail in terms of basic concepts, research status, and key technologies. We combine the domestic and foreign microgrid theory and experimental research results, put forward some suggestions and prospects of microgrid research.

Wanli Xu, Changfu Wang, Xuhui Wang, Shushuai Zhou, Youjie Zhou, Hua Li, Weigui Zhou, Lei Xu

Performance Test of Mini Solid Oxide Fuel Cell

Based on the construction of a fuel cell test platform, this paper tests the relationship between the startup time of the micro-100-W solid oxide battery developed in the early stage and the quality of gas consumption, and tests the relationship between the current and voltage changes, the amount of electricity discharged and the gas consumption under the rated power discharge state. The results of the data indicate that the power of the cell using propane fuel exceeds 200 W. Finally, the application prospect of the microtubular solid oxide fuel cell is discussed.

Youjie Zhou, Changfu Wang, Wanli Xu, Shushuai Zhou, Xiangjing Mu, Mengyi Wang, Litong Lv

Development of Hydrogen Fuel Cell Technology and Prospect for Its Military Application

This article introduces the technical level, product advantages, and development status of hydrogen fuel cells. The military requirements of hydrogen fuel cells and several typical application scenarios and usage in the military field are analyzed, and finally, the key contents and trends of the development of this technology are discussed.

Youjie Zhou, Changbo Lu, Jian Cheng, Wanli Xu, Yanli Sun, Hua Li, Lei Xu

A Novel Location-Based and Bandwidth-Aware Routing Algorithm for Wireless Ad Hoc Networks

Routing protocol is crucial to the wireless ad hoc network. In this paper, a novel location-based and bandwidth-aware routing algorithm is proposed to solve the problem that the available bandwidth is not considered in the general location-based routing protocol. This algorithm takes distance and available bandwidth into account when selecting routes; therefore, it can improve the utilization of link bandwidth and further improve the overall throughput, packet delivery rate and other performance of the network.

Chunguang Shi, Wenjin Hao, Quan Yin, Bo Zhou, Kui Du

Real-Time Measurement of Carrier Frequency Based on FFT and ANF

In order to accurately measure the carrier frequency and monitor the change process in real time, a real-time carrier frequency measurement method based on fast Fourier transform (FFT) and adaptive notch filter (ANF) is proposed. First, the short-time input sequence is intercepted, and the frequency of each carrier is quickly detected by FFT as the initial value of the notch parameter of ANF. Then, the input sequence is continuously filtered through ANF to update the notch parameters in real time, thereby achieving accurate measurement and real-time monitoring of the carrier frequency.

Wenzhao Li, Weihua Dai, Shen Zhao, XiWei Guo

Algorithm Research on Improving Halo Defects Based on Guided Filtering

In this paper, the improvement of guided filtering on halo defects in spatial color gamut mapping is proposed. The halo phenomenon is the main defect of the spatial color gamut mapping algorithm. This paper analyzes the principle and theoretical calculation method of guided filtering and proposes a new algorithm to improve the halo defect based on guided filtering. The mapping experiment of the new algorithm is carried out, and compared with the HPMinDE algorithm, the secondary algorithm in the paper. The comprehensive evaluation is carried out from three aspects of image color difference value, structural similarity, and algorithm running time. The results show that the new algorithm proposed in this paper has smaller color difference, larger structural similarity index and is closer to the original image, which verifies the feasibility and advantages of this algorithm. The research results of this paper have positive significance in the field of spatial color gamut mapping, and also have certain application value in the fields of edge detection, tone mapping, non-photorealistic rendering, and so on.

Yan Li, Tiantian Guan, Jingwen He, Zichen Cheng

The Smart Grid Vulnerability Analysis and Chain Faults Suppression Under Interdependent Coupling Relationship

With the widespread application of communication and information technology in the power system, the degree of interdependence between information networks and power networks is becoming deeper and deeper. On the one hand, the interdependent coupling of the smart grid greatly improves and optimizes the operation and control efficiency of the grid. On the other hand, it has somehow made the smart grid more vulnerable and has become one of the biggest contributors to the spread of chain failures. In this paper, we review the research progress on how to effectively control or suppress the impact of dependent coupling on the vulnerability of smart grid. This overview is presented from the perspectives of dependent coupling structure and vulnerability analysis of smart grids, respectively. We synthesize and analyze existing work, distill the issues and challenges still faced in the approach to controlling or suppressing the impact of interdependent coupling on smart grid vulnerability, and discuss future research trends and directions in the field.

Jingtang Luo, Shiying Yao, Jiamin Zhang, Yiming Chen, Min Zhang

A Coastline Detection Algorithm with ACM Driven by Diffusion Coefficient for SAR Imagery

To solve the problem for the synthetic aperture radar (SAR) images with a brighter sea surface area, a diffusion coefficient driven active contour model is proposed for coastline detection. By introducing the diffusion coefficient of conditional information into the regular term of active contour model (ACM) with Gamma distribution, the level set evolution equation is obtained. The algorithm could detect coastlines for SAR images where the sea surface appears bright and the sea surface is uniform.

Xiaofei Shi, Xu Zhang, Derui Song, DeJun Zou

Coastline Detection with a Spatial Second-Order Correlation Statistic for SAR Imagery

Aiming at the problem that the coastline detection is not ideal when the bright sea area appears in the synthetic aperture radar (SAR) imagery of the current coastline detection algorithm, a spatial second-order direction-related statistic is proposed to effectively suppress the highlight feature of the sea surface, and the sea–land junction area is obtained based on this statistic. Then, a traditional edge detection method is used to obtain the coastline. Experiments show that the algorithm can accurately detect the coastline in SAR images with some bright regions on the sea surface.

Xiaofei Shi, Xu Zhang, Derui Song, DeJun Zou

A Survey on the Entity Linking in Knowledge Graph

With the rapid development of information technology, the amount of information is increasing exponentially. All kinds of text data are growing explosively. How to understand the meaning of these data quickly and accurately becomes extremely difficult and challenging. Entity linking is proposed for solving the above problem over all kinds of unstructured data. Entity linking is to link the mentions ( also called entity references) in a given text to the correct Wikipedia page without ambiguity. In this paper, we summarize the methods of entity embedding and the realization of each step of entity link in the application of machine learning.

Jingjing Du, Bo Ning

A Multi-focus Image Fusion Method Based on Cascade CNN Networks

Due to the fact that the optical lenses have limited depth of field, it is difficult to capture an image in which all the objects are in focus. Multi-focus image fusion is a popular technique to solve this problem. In this paper, a method based on cascade CNN networks (Pixel Network and Block Network) is proposed. The Block Network develops the rough fusion mask and the Pixel Network refines edges of the mask. Then, the mask is processed by rolling guidance filter. Finally, the fusion image is generated according to the mask. Experiment shows that our method achieves better result than the methods being compared. All the code of this article can be downloaded on .

Zhang HeXuan, Tong Ying

Research on the Effects of Emotional Intervention in Online Learning

Emotions play a major role in a student’s academic performance. As online learning becomes popular in learning institutions around the world, this paper proposes an emotional intervention model for online learning using machine learning. This is to be achieved by two models, an emotion recognition model that takes student’s video data during online classes and classifies the student’s current emotional state, and an emotional intervention model that takes input data from the recognition model and determines the best emotional intervention strategies and techniques for the student. This paper explains the theory behind the need for such a model and gives a hypothetical use-case scenario of its implementation.

Simbarashe Tembo, Jin Chen

Study on Data Fusion Processing Algorithm of Marine Sensor Based on Information Entropy

Addressing the issue of data processing in the measurement of optical dissolved oxygen sensor, a data fusion processing algorithm based on information entropy was designed and presented. Firstly, the probability distribution of discrete sample data was estimated based on the maximum entropy method. Then, the effective confidence interval was calculated based on the sample measurement uncertainty, which would be used to eliminate the gross error. Lastly, the effective sample data were fused based on information entropy. The calibration experiment of typical marine sensor was taken as an example to verify the effectiveness of the algorithm. Compared with other methods, the result of this algorithm had the absolute error of 0.01 and mean square error of 0.0189. The result indicates that it could overcome the calibration data pollution caused by various subjective factors and improve the stability and reliability of sensor measurement data.

Hao Gao, Lin Cao, Lei Yang

Study on Typical Nonlinear System Control Strategies Based on Energy Conversion

Based on the analysis of the typical nonlinear system, the swing-up control strategy based on energy conversion was proposed for the single inverted pendulum. Through introducing the concepts of “Cart Potential Well,” “Velocity Well,” and “Energy Maintenance,” the pure energy-based control strategy was modified. During the whole swing-up process, it could ensure the pendulum could be swung-up to the upright position quickly and smoothly. In the meantime, the cart displacement could be restricted within limited length in both directions. Then, this optimal controller would be implemented and simulated in MATLAB and tested on the cart-pendulum real-time application. The experimental results indicate that the control theories described in this paper could achieve all the swing-up control requirements for the inverted pendulum and get the desired results successfully.

Hao Gao, Lin Cao, Yulong Cai

Relay Selection based on Multiple-Attribute Decision Making for Underwater Acoustic Cooperative Communication

The common issues in cooperation underwater acoustic (UWA) communication include relay selection and power allocation. In this paper, we will consider primarily the issue of relay selection, combined with the characteristic of underwater acoustic channel-long propagation delay. Consider that using equalization to eliminate inter-symbol interference (ISI), we propose a cooperative transmission scheme. In the scheme, propagation delay and steady-state mean squared error (SMSE) will be regarded as two evaluation indices, using an algorithm one of the multiple-attribute decision making (MADM) to choose the optimal relay node. Lastly, we will analyze the simulation results and compare them with the real data to certify the effectiveness of the algorithm.

Miao Ke, Zhiyong Liu

A Chinese Knowledge Graph for Cardiovascular Disease

Knowledge graph as a graph-based data structure can well represent the relationship of medical data in reality, which can much effectively organize and utilize the medical data. Applying the knowledge graph to the medical field can extract effective knowledge from medical information and fuse it to form structured knowledge. On this basis, it can promote the research and application of intelligent diagnosis and treatment, auxiliary analysis and decision support, so as to promote the development of medical intelligence. However, the research on the construction of medical knowledge graph is still in the initial stage. In this paper, we explore the construction method of the Chinese knowledge graph for cardiovascular disease and construct it.

Xiaonan Li, Kai Zhang, Guanyu Li, Bin Zhu

Study and Validation on a Novel Multi-dimensional Marine Electromagnetic Prospecting Nonlinear Inversion Algorithm

In order to study and verify the practical application effect of multi-dimensional nonlinear conjugate gradient inversion in the exploration of marine natural gas hydrate, a novel multi-dimensional marine electromagnetic prospecting nonlinear inversion algorithm was proposed in this paper. On this basis, the data of marine controlled-sources electromagnetic collected by the Scripps Research Institute of the USA were used to analyze the hydrostatic results of this multi-dimensional nonlinear conjugate gradient inversion algorithm. The results show that the nonlinear conjugate gradient inversion proposed in this paper can clearly reflect the conductivity distribution structure of the area below the bottom of the sea surface 1~2 km of formation. The local inversion results of this method are basically consistent with the seismic reflection, seismic sequence and logging analysis results, and it has a strong resolution to small size anomalies and shallow high-resistivity bodies.

Hao Gao, Lin Cao, Guangyuan Chen

Study on the Noise Analysis of Weak Photoelectric Signal Detection in the Marine Environment

Due to the complexity and variability of the marine environment, the weak useful signals of the measured marine optical instruments often have strong background noise and nonlinear characteristics, which seriously affect the detection accuracy of the weak useful signals. In this paper, the interference of background optical noise in in-situ optical instruments is studied. From the perspective of optical signals received by the instrument detection system, the types of optical signals received are analyzed, and the energy level, spectral characteristics, and spectral characteristics of interference signals and useful signals are analyzed. The noise removal scheme is discussed, which provides the theoretical basis and design basis for the in-situ optical sensor design.

Lin Cao, Hao Gao, Xuejun Xiong

A Brief Survey of Graph Ranking Methods

Graphical model is a general term for a class of technologies that use graphs to represent probability distributions. Graphical representation of data plays an important role in data processing. According to a certain standard, the graph ranking to rank the importance of vertices on graphs is one of the classic problems of graph models, and it has been widely used in various applications, such as recommendation systems and search engine optimization. With the development of graph ranking methods, a variety of variants based on the original algorithm have appeared. This article mainly lists five common graph ranking algorithms and compares their respective characteristics.

Mengmeng Guan, Bo Ning

Power Control for Two-Way DF Relay-Aided Underlaid D2D Communications

Device-to-device (D2D) is used for direct communication between adjacent terminal devices and has been considered as one of the key technologies of 5G. However, when the distance between D2D users is far or the channel condition is poor, the effectiveness and reliability of transmission will be reduced. To solve this problem, a promising way is to use relay-aided D2D communication. In this paper, the minimum rate of a pair of D2D users is maximized by controlling the transmit power of D2D users and relay user while fulfilling the minimum rate requirement of the cellular link. The transmit power of each node is adjusted to minimize the useless power to obtain the optimal transmit power of D2D users and relay. The simulation results show that the minimum rate of a pair of D2D users reaches the maximum under the optimal power allocation scheme.

Ning Liang, Liang Han, Yupeng Li

A Survey on Conversational Question-Answering Systems

With the development of artificial intelligence technology and the increasing demand for rapid access to information, question-answering system has been widely studied. In recent years, due to the great achievements in the single round of Q&A, the research focus has shifted from the previous single round of Q&A to multiple rounds of Q&A. In this paper, we divide the Q&A system into task-oriented Q&A system and non-task-oriented Q&A system, and discuss the application of multi-round Q&A, respectively. Finally, we discuss the possible future direction of multi-round Q&A in the dialogue system.

Deji Zhao, Bo Ning

Stable Time Transfer Over 120 km Optical Fiber with High-Precision Delay Variation Measurement

We report the extension of our frequency dissemination system to the time transfer capability, and a synchronized time signal dissemination system over 120 km fiber is proposed and experimentally demonstrated. The optical fiber propagation delay variation-induced phase fluctuation of RF is identically transferred to an IF by the mean of the DHPT scheme. By using the phase-locked loop (PLL) to compensate the phase fluctuation, a stable frequency signal is obtained, at the same time, a high-precision time delay fluctuation measurement is also achieved by monitoring the phase of IF VCO at the local end. The resolution of fiber transfer delay measurement achieves about 60 fs. After the active compensation of propagation delay variation, the time variation of 120 km fiber link is about ±100 ps during 3600 s.

Xiaocheng Wang, Xiaoming Ding, Yupeng Li, Cheng Wang

Constellation Design and Application of Real-Time Space-Based Information Services Supporting Communication, Navigation and Remote Integration

The existing communication, remote sensing and navigation satellite systems are self-contained resulting in system isolation, information separation, service lag, poor information sharing and interaction. Moreover, the applications of space-based information in communication, remote sensing and navigation are mostly independent of each other. By integrating the communication, navigation, remote sensing and other loads with the platform, we can provide the task-oriented space–time services and construct an all in all system with multi-purpose, multi-satellite networking, multi-network integration and intelligent services. In this paper, we propose a constellation design method, which can support the real-time service of space-based information by integrating the communication, guidance and remote. The model of satellite scale prediction is established based on the optimization algorithm of coverage band. Considering all kinds of load performance, phase relationship, inter-satellite link construction and other constraints, the number of satellites, constellation configuration and deployment scheme are determined.

Wang Liyun, Meng Jing

Path Index-Enhanced Incremental Subgraph Matching Algorithm for Dynamic Graph

A dynamic graph is defined by an initial graph and a graph update stream consisting of edge insertions and deletions. Given a query graph and a dynamic data graph, the problem of incremental subgraph matching is to conduct all subgraph isomorphic matchings for edge insertion. In this paper, an incremental subgraph matching, IncISM, is proposed hierarchical path index-based incremental subgraph matching algorithm for dynamic graph is proposed. Firstly, design a positive-negative-possible path index (PNPC-PI) to index the intermediate results and final results. Then, propose a incremental strategy to maintaining the intermediate results and conducting the subgraph matchings for edge insertion. Compared with SJ-Tree and GraphFlow, experimental results show that IncISM is more efficient.

Yunhao Sun, Guanyu Li, Bo Ning, Bing Han

Obstacle Detection and Recognition Using Stereo Vision and Radar Data Alignment for USV

Environment perception technology is the first step and key technology for autonomous navigation of unmanned surface vehicle (USV) under the condition of over-the-horizon. The detection and location of marine obstacle targets are an important research content of environmental perception technology. This paper proposes a method for aligning the information of the stereo vision sensor and the millimeter-wave radar sensor without united calibration to detect and recognize the marine obstacle targets. This method extracts the position and velocity of the target as a feature vector and designs a matching evaluation function to find the best matching pair from two sensors’ data information. Finally, the article verifies the real-time and accuracy of the method by experiments.

Lin Cao, Hailin Liu, Hao Gao, Hui Li

Research on New Progress and Key Technology of Space TT & C and Data Transmission System

With the continuous development of space system, the tracking ability, data transmission ability, and communication ability of TT & C network are required to be higher. This paper studies the development of foreign space TT & C and data transmission system, analyzes its development process and characteristics, summarizes the future development trend, and combs the key technologies of space TT & C and data transmission system. On this basis, analyzing the needs of China’s future space system to TT & C and data transmission system and presenting some suggestions.

Xiyuan Li, Jing Meng

The Radar Echo Extrapolation Based on ConvLSTM

In order to predict radar echo products, a radar echo extrapolation neural network model based on ConvLSTM is studied in this paper. The structure of this model is long-term memory network, and multi-dimensional data can be used as input after making some changes to its structure. In this paper, the reflectivity data of horizonal polarized SA radar is adopted as the dataset. In order to achieve the ideal prediction effect, the following methods are proved to be effective: (1) in the process of datasets’ preprocessing, use quite or enough datasets, and filter, sort, fit, and filter the datasets. After that, the data is standardized, and the data is scaled proportionally, so that the reflectivity value is centered on 0 and distributed in a small interval. (2) After using convolution layer to extract features from reflectivity data and connecting full connection layer and regression layer behind the hidden layer, more optimized and accurate prediction results can be obtained. (3) Using appropriate activation functions and certain regularization methods, these activation functions and regularization methods can reduce model overfitting and improve model training accuracy. The experimental results show that this paper has a good effect on the prediction of radar echo within 18 min and has a good ability to predict the echo whose reflectivity intensity is less than 30.

Zhaoping Sun, Can Lai, Xia Chen, Haijiang Wang

Cross-Age Face Recognition Using Deep Learning Model Based on Dual Attention Mechanism

Although remarkable progresses have been made in the field of face recognition, the cross-age problem is still a huge challenge. The cross-age problem is mainly reflected in the fact that in addition to the unique identity features of each person, facial features also contain age features changing during aging. To address this problem, we propose a novel cross-age face recognition framework based on dual attention mechanism which combines residual-attention mechanism and self-attention mechanism. The introduction of attention mechanism makes the model focus more on identity features, ignoring the influence of age features. Extensive experiments are conducted on two well-known face aging datasets (MORPH and CACD) to show that the proposed method achieves notable improvement over state-of-the-art algorithms.

Jialve Wang, Shenghong Li, Fucai Luo

Research on Classification of Alzheimer’s Disease Based on Multi-scale Features and Sequence Learning

Due to 2D-CNN cannot effectively use the continuous change information in MRI to classify Alzheimer’s disease (AD), an MDLCSTM-LDenseNet model based on multi-scale features and sequence learning is proposed. On the basis of retaining the advantages of DenseNet, 3D Light-DenseNet with fewer parameters is given as the basic network, and the MDLCSTM module combining dilated convolution and ConvLSTM is embedded in the 3D Light-DenseNet to further extract the slice features and the continuous change information between slice sequences in the global slice range of MRI. Based on the experimental of MRI data in ADNI database with other methods, the classification accuracy of AD and CN is 97.25%, and the classification accuracy of CN and MCI is 92.97%. The results show that the model has a high classification accuracy and reliability.

Sen Han, Lin Wang, Derui Song

A Behavior Predictive Control Mechanism Based on User Behavior Characteristics

The detection of the user behavior characteristics is based on the network analysis technology. Through the analysis of the characteristics of user behavior, the network behavior of users is counted, and the characteristic database based on user behavior is established. The user behavior is compared with the characteristic database to predict whether the user behavior belongs to normal network behavior or abnormal network behavior. This detection method is relatively independent of the operating system and has strong universality. It can reflect the changes of the network environment and find the new abnormal behaviors of users.

Mengkun Li, Yongjian Wang, Caiqiu Zhou

Unmanned Aerial Vehicle (UAV) Networking for Ocean Monitoring: Architectures and Key Technologies

Unmanned aerial vehicles (UAVs) communication networks have great potential in ocean monitoring domains. By virtue of the main features such as mobility, agility, easy deployment and adjustability, UAV communication networks will be widely applied in navigation, ocean ranch management and ocean environment protection. In this paper, we present a brief survey on UAV networking for ocean monitoring in terms of architectures and key technologies including UAV deployment and trajectory optimization.

Bin Lin, Yajing Zhang, Xu Hu, Jianli Duan

MEMS Mirror LIDAR System and Echo Signal Processing

Nowadays, LIDAR is widely used to measure the distance and speed of moving targets, especially in the application of robot and driverless vehicle. In our project, we setup a lightweight LIDAR system combined with the MEMS mirror for 3-D scanning, and OSLRF-1 module for transmitting and receiving the laser impulse. In the echo data processing, we find that the peak detection will greatly be affected by noise; specifically, more jittering appears surrounding the echo peak, which reduces the performance of the peak-peak detection algorithm. In order to solve this problem, we develop a series of methods with the cooperation of Kalman filter and Gaussian fitting. Through this series of methods, we greatly decrease the influence from peak-surrounding noise and achieve a great improvement. The absolute error for 60-inch ranging decreases to 0.197 from 0.820, and the root mean square error decreases to 0.344 from 1.209, which achieve a better performance.

Tao Liu, Dingkang Wang, Ruowei Mei, Xinlin Gou

Research on Indexing and KNN Query of Moving Objects in Road Network Environment

In the field of database research, continuous K-nearest neighbor query processing technology is a problem that many scholars pay attention to. It has many applications in real life, such as traffic dispatch control, digital battlefield, personal location service, and other systems must be related to mobile users. Location or trajectory information is managed. However, in these applications, the spatial data and spatial object access of the continuous K-nearest neighbor query service are restricted to the spatial network, such as the road network. This paper introduces the research status of mobile objects in the road network is index technology and k-nearest neighbor query, analyzes, and summarizes the research results of scholars at home and abroad, and finally analyzes the challenges faced by K-nearest neighbor query of moving objects in the road network.

Wei Jiang, Guanyu Li, Jingmin An, Yunhao Sun, Heng Chen, Xiaonan Li

Design and Implementation of a Power Module with Overcurrent Protection

In order to achieve the local substitution for the special power plug-in board of some equipment imported from abroad, a new power module structure and implementation scheme are designed. We firstly made a functional analysis of original circuit, and then, we discussed the alternative scheme analysis and selected the appropriate scheme and designed of overcurrent protection circuit. After testing and practical application, better current overload protection and the similar voltage current curve are carried out and the module can be used normally in imported equipment, achieving the effect of localization of power supply components.

Peng Chen, Qingchun Wu, Yuejun Shi, Qian Zheng, Sen Yang

The Convolutional Neural Network Used to Classify COVID-19 X-ray Images

The COVID-19 outbreak has a significant impact on the health and well-being of the global population. The first step in the fight against COVID-19 is effective screening of infected patients. A key screening method is chest radiography. In early studies, patients showed abnormalities on chest radiographs, which were characteristic of patients with COVID-19 infection. In this paper, we use the open dataset covidx to train the neural network. covidx is composed of 13,800 chest radiographs with 13725 patients from three open case access data repositories. Novel coronavirus pneumonia was used to identify novel coronavirus pneumonia, common pneumonia and new crown pneumonia images.

Qiu Yiwen, Wang Yan, Song Derui, Lv Ziye, Shi Xiaofei

Speech Emotion Recognition Based on Transfer Learning of Spectrogram

In order to solve the problem of feature extraction and classification model construction of speech emotion recognition, this paper proposes a feature extraction method based on spectrogram, which transfers the time-domain signal of speech to the frequency-domain space to obtain its frequency-domain features. At the same time, the existing convolution network parameters are transferred by using transfer learning as its classifier. Finally, the experimental results show that the accuracy of emotion recognition is better.

Lv Ziye, Wang Yan, Song Derui, Qiu Yiwen, Shi Xiaofei

A Novel Kalman Filter Algorithm Using Stance Detection for an Inertial Navigation System

In this paper, a modified Kalman filter with stance detection is proposed. By utilizing this stance information, the proposed Kalman filter can improve the performance of a personal inertial navigation system (INS) using microelectromechanical system (MEMS). Experiment results have confirmed that the proposed system significantly improves the performance of an INS even under harsh magnetic environment.

Zhijian Shi, Ruochen Feng, Rui Lin, Gareth Peter Lewis

A Multi-Strategy Batch Mode Active Learning Algorithm for Image Classification

Active learning is often used to alleviate the problem of insufficient labeled samples. A batch of samples is selected by batch mode active learning for labeling each iteration, but when a single querying strategy is adopted, the selected samples will exist redundant information, resulting in additional labeling costs. To solve this problem, a multi-strategy batch mode active learning is proposed. Based on uncertainty strategy, clustering and similarity measures are introduced into our algorithm. The convolutional auto-encoder (CAE) and clustering are used to measure the representativeness of unlabeled samples, and then, the distance measure is used to measure the similarity between unlabeled samples and labeled samples. Based on the MNIST, CIFAR-10 and Alzheimer’s disease neuroimaging initiative (ADNI) datasets, the experimental results show that the algorithm can effectively improve the model performance.

Sen Han, Lin Wang, Derui Song

Few Shot Object Detection via Training Image Generation

In recent years, deep neural networks (DNN) have reached state of the art performance on object detection. One of the drawbacks of DNN is that DNN needs large amount of training data, which cost a lot of human labors. When the labeled training data is limited, training results will be poor. In this paper, we propose a framework of object detection via training image generation, which uses a small number of cell pictures as samples. On the basis of the samples, a large number of cell pictures containing labeling information are derived by using GAN network and the method of sticking cells in this paper. Finally, Mask R-CNN was used for training. The method proposed in this paper effectively solves the problem of less sample data and saves a lot of manpower and material resources.

Deyuan Zhang, Yixin Zhang, Junyuan Wang

Research of Simulation Method for UAV Fault Generation Facing Emergency Operation Training

In this paper, for UAV simulation training system facing emergency operations training requirements, UAV fault generation simulation methods are researched. According to the characteristics and laws of UAV system failure, fault generation methods based on accumulation of operating frequency, flight state and flight environment, three modes of UAV fault generating method are proposed. Then, this paper applies typical instance to verify the fault generation methods; the results show that the methods can well represent the UAV fault conditions and can realize UAV’s high realistic simulation training.

Sen Yang, Leiping Xi, Hairui Dong, Xuejiang Dang, Peng Chen

Identification of Precipitation Clouds Based on Faster-RCNN Method

Precipitation clouds are visible aggregates of hydrometeor in the air that floating in the atmosphere after condensation, which can be divided into stratiform cloud and convective cloud. Different precipitation clouds often accompany different precipitation processes. Accurate identification of precipitation clouds is significant for the prediction of severe precipitation processes. Traditional identification methods mostly depend on the differences of radar reflectivity distribution morphology between stratiform and convective precipitation clouds in three-dimensional space. This paper proposes a new method for precipitation clouds identification based on deep learning algorithm. It mainly includes two parts, which are constant altitude plan position indicator data (CAPPI) inversion for radar reflectivity, and the precipitation clouds identification based on Faster-RCNN. The testing result shows that the method proposed in this paper performs better than typical existing algorithms in terms of accuracy rate. Moreover, this method boasts great advantages in running time and adaptive ability.

Yuanbo Ran, Li Tian, Haijiang Wang, Jiang Wu, Tao Xiang

MIF: Toward Semantic-Aware Representation for Video Retrieval

Semantic Concept-Based Video Retrieval (SCBVR) has been widely studied recently, which exploring semantic representations of videos to execute user’s retrieval requests. However, the effectiveness of video retrieval often depends on the accuracy of semantic concepts, but the semantic concept is often imprecise or contrary to the ground-truth. In order to solve the above-mentioned problem, we propose a novel multi-information fusion (MIF) approach, and it is beneficial to improving the performance of video retrieval. Firstly, we infer the most relevant semantic concept corresponding to query keywords by using the inherent association information of videos. Secondly, we fuse the probabilities of the candidate semantic concept by minimizing the potential function. We conduct the extensive experiments on real-world datasets which demonstrate the effectiveness and efficiency of the proposed approach for enhancing the performance of video retrieval.

Bo Lu, Xiaodong Duan

Wireless Channel Models for Maritime Communications

The complexity of the maritime environments brings great challenges to maritime communications. To design a maritime communications system with high speed, high reliability, full coverage, and low cost, it is urgent to establish wireless channel models. In this paper, we describe some factors that affect maritime communications and analyze typical wireless channel models of maritime communications.

Bin Lin, Jiaye Li, ZhenWang, Tiancheng Kang

Architecture Design of 5G and Virtual Reality-Based Distributed Simulated Training Platform for Ship Pilots

In this paper, virtual reality (VR) and fifth-generation cellular network (5G) technology are applied to the remote training fields. We propose a new distributed simulator platform for ship pilots. Compared to traditional training methods, the application of the “5G + VR” technology can simulate the movements of various ship types and various sea conditions in different pilot training locations. Moreover, comprehensive training and assessment sub-systems can be provided for trainees through the intelligent auxiliary decision-making support and expert evaluation functions. The proposed new distributed simulator platform aims at solving the problems due to large-scale pilotage training difficulty, high-risk experiments and lack of high-quality teaching resources.

Bin Lin, Linan Feng, Hongyi Xu, Dewei Wang

Design of Verification System on Robot Trajectory Interpolation Algorithm

To solve the problem of verification of the interpolation technology in robot trajectory, a simulation verification system based on VC++ platform is designed and developed. The operation interface and function module of the system are introduced. The system can complete the input and preprocessing of program files, command interpretation, interpolation algorithm and the output of simulation verification results. The system is used to verify the correctness of three kinds of spline curves (NURBS, CUBIC, AKIMA) interpolation code format and arithmetic of spline curve interpolation, as well as precision of the interpolation. The system can obtain the interpolation results in real time, correctly and intuitively, which meets the needs of practical engineering.

Ti Han

Indoor Motion Status Identification Algorithm Based on Decision Tree Model for FM Signal

In order to improve the accuracy and environmental adaptability of the behavior recognition algorithm in the classroom, this paper proposes an indoor behavior recognition algorithm based on FM signal. This method mainly uses decision tree model for classification calculation. The experimental results show that the accuracy of this method can reach 94.4%, and no additional hardware is required.

Shuai Wang, Ying Jin, Xuemei Wang

Localization Algorithm Based on Fingerprint Model for FM Signal

This paper comprehensively analyzes the fingerprint features such as rssi, snr, stereo separation degree of fm signal for positioning calculation, and finds that outdoor positioning technology based on fm signal has limited practical application value, but it has more advantages than wireless positioning technology such as Wi-Fi in indoor positioning field. This paper proposes a fm fingerprint localization algorithm based on bp neural network. This method effectively reduces the influence of fm signal time migration characteristics on fingerprint localization performance. The experimental results show that the indoor localization accuracy of this method is more than 80% above 3 m, and no additional hardware equipment is required, and the deployment is flexible and convenient.

Xuemei Wang, Yonglin Wu, Wei-cheng Xue

The Design and Development of FM Localization Algorithm Based on KNN Model

Through comparing the radio signals used in indoor positioning at home and abroad, it is found that FM broadcasting has certain advantages in indoor positioning, and through the localization algorithm at home and abroad, the position fingerprint localization algorithm based on FM signal features is proposed, and the comprehensive comparison classification algorithm is used. Finally, it is concluded that the accuracy of FM signal feature localization method can reach 1.5 m in 90% cases.

Xue-mei Wang, Beiqi Song, Wei- cheng Xue

Design and Development of Data Acquisition Module of Intelligent Meter Based on LoRa

This paper will introduce a smart meter based on LoRa technology: a smart meter based on radio frequency communication module. At the same time, this design will use a three-phase four-wire smart meter with low design cost. The meter uses HT6015 as its main chip: It integrates Cortex-M0 processor, clock manager, power manager, and other units; and is mainly characterized by multifunction, high performance, and low-power consumption.

Le Wang, Lei Li

Design and Development of Temperature Processing Module in Intelligent Terminal of Internet of Things

This paper aims at the requirements of the common temperature processing module in the Internet of things application system. The peripheral resources of the micro-controller itself are relatively rich, mainly including the interface application and the use of ordinary I/O. The core of the micro-controller is used as the central control brain. All kinds of peripheral sensors are used to collect information, all kinds of peripheral switch controllers are used to control the switch signal, all kinds of display devices are used to realize the display function, and single-chip microcomputer is used to calculate and process the information to complete the overall function of this design. Specifically, this design uses single-chip microcomputer, temperature sensor, key, buzzer, display screen and other devices based on 51 single-chip microcomputer temperature data collection function design.

Le Wang, Yuxiang Du

Indoor Motion Status Identification Algorithm Based on SVM Model for FM Signal

To improve the accuracy and environmental adaptability of personnel and equipment localization algorithms in the construction and operation of underground pipeline network, this paper proposes a fingerprint localization algorithm based on FM signals. This method mainly uses support vector machine model for classification calculation. The experimental results show that the accuracy of this method can reach 91%, and no additional hardware is needed.

Qun Liu, Yonglin Wu, Wei-cheng Xue


Weitere Informationen