Signal and Information Processing, Networking and Computers

In the massive multiple input multiple output (MIMO) system with transmit antenna selection, considering large-scale fading and power consumption, the trade-off between energy efficiency (EE) and spectral efficiency (SE) is important for green communication and meeting the traffic request. However, the diversities of the solution spaces (decision and objective domain) which the existing method obtained are not enough. In this paper, we adopt the non-dominated sorting genetic algorithm version II (NSGA-II) algorithm to study the EE-SE trade-off, which is summarized as a multi-objective optimization (MOO) problem. Simulation results show that compared with the weighted-sum particle swarm optimization (WS-PSO) algorithm, NSGA-II has obvious advantages on the diversity of solution set in both decision and objective domains.

A secure secret sharing scheme distributes a secret among a group of participants and only when a sufficient number of shares are combined together can the secret be reconstructed. In this paper, we present a secure PHY-based secret sharing (PSS) scheme exploiting radio channel fading coefficients in multiple-input multiple-output (MIMO) systems. $$ (n,n) $$ threshold secret sharing scheme is considered and there is no need for the third party to distribute keys in PSS scheme. A two-step randomness sharing is designed, which reduces the time overhead significantly compared to one-by-one randomness sharing when the number of participants increases. Furthermore, we derive a power allocation scheme under power constraints based on particle swarm optimization (PSO) algorithm. Numerical results demonstrate the efficiency of our power allocation strategies on secret key rates.

In modern wireless systems, channel prediction is an effective way to overcome the feedback delay of channel state information (CSI). When the receiver performs adaptive transmission based on the feedback CSI, the channel prediction algorithm can reduce the system overhead by predicting the future CSI. In this paper, we provide a long short-term memory (LSTM) network for wireless channel prediction. This method can get a smaller prediction error than other intelligence methods. Experiments show that the LSTM model has a lower normalized mean square error (NMSE) and less running time than support vector machine, artificial neural network, and recurrent neural network prediction approaches.

Signal identification is a crucial subject in cognitive radio (CR) systems. In GSM spectrum refarming or spectrum monitoring scenarios, CR is required to identify on-the-air signals like long term evaluation (LTE), global system mobile (GSM). Second-order cyclostationary detection is an identification method robust to noise uncertainty and used widely in spectrum sensing. However, it requires high sampling rate and long processing time. In this paper, we first propose a compressed sensing (CS) based sampling structure to reduce the sampling rate using the second-order cyclostationary features of Time Division-LTE (TD-LTE) and GSM signals. Furthermore, an identification method for TD-LTE and GSM signals based on CS is employed to reduce sensing time. The performance of the method is evaluated by the practical on-the-air-signals measured with a spectrum analyzer. Numerical results show that our method can achieve a high detection probability with a low sampling complexity.

The interference of mobile communication network optimization system is mainly for the interference of intra-system, interference of inter-system and short distance external interference. The interference localization of the mobile communication network, especially the Time Division Duplexing (TDD) system, becomes very complicated. This paper mainly expounds the basis of the generation of atmospheric duct, and how to monitor the interference of atmospheric duct of the Time Division Long Term Evolution (TD-LTE) system. On the basis of this, the paper proposes a centralized scheme to avoid the interference of atmospheric duct, focusing on the optimization measures of avoiding and reducing the interference of atmospheric duct through parameter optimization.

Load balancing and interference management influence the performance of heterogeneous network (HetNet). Cell range expansion (CRE) scheme has been introduced to achieve better load balance by adding a positive bias to the small-cell BSs. Enhanced inter cell interference coordination (eICIC) is recognized as a solution that reduces cross-tier interference by setting the almost blanking subframes (ABS). In this paper, we divide pico area into pico inner part (the normal coverage area of pico) and pico CRE part (bias area assigned to pico). We propose that the macro only transmits data during non-ABS subframes. The pico CRE part transmits data to its users in ABS subframes, and pico inner part transmits over all subframes. The users within different base stations have different interference patterns, and therefore we compute it independently. Then we derive the optimal ABS value by optimizing the system throughput equation. After performing the extensive simulation, we get the proper CRE value. And the proposed ABS scheme has shown remarkable results.

We propose a multi-index evaluation model of quality of experience (QoE) for wireless video, which combines the network quality of service, quality of content, quality of interaction and quality of terminal. In particular, we use the fuzzy analytic hierarchy process (FAHP) to analyze these influencing factors and establish the evaluation system. In addition, we verify the influence of each indicator to QoE for wireless video in network transmission via simulations and subjective evaluation values. The experimental results show that our proposed method is highly correlated with the subjective evaluation, and thus it can more efficiently reveal QoE for the wireless video in the network transmission.

Spectrum sensing has attracted much concern of researchers due to its significant contribution to the spectral efficiency. However, the corresponding work mainly focuses on the sensing event of single primary users within a certain band and the investigation of the effect of PU traffic on the sensing performance is considered rarely. In this paper, a spectrum sensing scheme is presented to explore the co-existence of multiple users in the same frequency band based on subspace filtering. To remove uncertain noise as much as possible, subspace filtering is applied to the received signal of a cognitive radio, where the received signal is decomposed into two parts: noise subspace and signal-plus-noise subspace. Then the closed-form solution of the detection and false alarm probabilities with multiple users is given on the basis of the signal-plus-noise subspace in Rayleigh fading channel. Eventually, simulations are made to validate the proposed scheme.

Spectrum sensing is a significant part of technique in a cognitive radio that detecting the presence of primary users in an authorized spectrum. the method that based on the statistical covariance matrix is one of main spectrum sensing techniques, using the difference of statistical covariance between the received signal and noise. In this paper, the new sensing method we proposed is also based on the statistical covariance. The new method compare to some traditional covariance algorithms has decrease the complexity of algorithm, at the same time, ensured the accuracy of detection. We give the statistics of detection, and we also find the threshold of the method when the probability of false alarm is given. The analysis and derivation process of threshold are provided in behind. Using Matlab for simulation to validate the correctness of the method and making the comparison with some typical detection method.

Cognitive radio (CR) is regarded as a perfect technique to cope with the scarcity of spectrum resources. Energy detection is preferred by most of Cognitive Radio researchers, because it is easy to implement and it doesn’t need the prior information about primary user’s (PU) signals. But the performance of energy detection is poor at the low signal-to-noise ratio (SNR) regime. Double threshold spectrum sensing scheme was proposed to increase the reliability of decision. Under the same SNR, if the sensing time is not long enough, a higher noise variance will make the instance energy level falls below the threshold because of the noise uncertainty. In this paper, an enhanced energy detection (EED) scheme combined with double thresholds is proposed, this scheme can reduce the misdetection caused by noise uncertainty. In the proposed method, we compare the average of last M sensing statistic with the preset threshold. It aims at protecting a channel that are underutilizing from being decided to be idle when an immediate signal energy drop happens. The simulation makes a comparison between the proposed method and the traditional method and proves the effectiveness of the new scheme.

In the Cognitive Radio system, cognitive users require perceiving the real-time using of the spectrum accurately. At the same time, the cognitive user stations are usually in severe fading or interference. Energy detection is used widely due to the low computation complexity and an effective method under the high SNR, but it is impressionable by the noise. According to these facts, we propose to use the coefficient of variation (CV) of the sampled signals to amend the judgment result and get the blending final result. After the energy detection, if the test statistic of the signal energy is within a specified range and the signal CV is less than a threshold value, it concludes that the spectrum bands are occupied by another user. The simulation results prove that the rectification method can greatly improve the cognitive user’s accurate detection performance of spectrum usage in real-time while ensuring the computation complexity.

Millimeter wave (mmWave) and cloud radio access network (C-RAN) are two potential candidates for next generation communication. In this paper, we consider user-centric C-RAN in mmWave with the existence of blockages in urban areas. The remote radio heads (RRHs) are deployed according to a Poisson point process in the circular region $$ {\mathcal{D}} $$, of radius R. We employ the stochastic geometry theory to analyze the signal-to-noise ratio (SNR), rate, and outage probability. We emphasize the effect of circular region radius on the performance in this network and evaluate the effect with Monte Carlo simulations. The simulation results show that SNR, rate and outage probability have the same asymptotic trends and have the best performance when replace the circular region $$ {\mathcal{D}} $$ with the line-of-sight (LOS) circular region.

In this paper, the performance of a cloud radio access networks (C-RAN) where remote radio heads (RRHs) operating at millimeter wave (mm-Wave) range are modeled as a homogeneous Poisson point process has been investigated. In this network, due to the high frequency and small wavelength, we use a LOS circle concept from the former study. Considering downlink transmission, we were compared the performance of two different transmission schemes. First is the selection transmission which chooses the best RRH (BR) for transmission. Another scheme is all the RRHs participated (ARP) and transmission the signal together to the typical user. We analysis the outage and capacity of the two schemes and ARP transmission scheme always performed better.

Massive multi-input multi-output (MIMO) improves networks throughput via using plenty of antennas at the base station (BS) to serve multiuser simultaneously. Considering channel hardening, we directly investigate the approximate expressions for network throughput with conjugate beamforming (CBF) and zero-forcing beamforming (ZFBF) schemes and adaptive downtilts of antenna patterns. Building on this, a new adaptive beam switching tilting strategy based on the greedy algorithm which adopts the optimal downtilts to serve different regions are proposed to enhance the performance. We also study intercell interference and propose a new modified cooperative approach to reduce interference. Simulation results demonstrate that the system throughput has been improved compared with the traditional method.

In this work we consider the user scheduling problem in the massive multiple-input multiple-output (MIMO) wireless multicast system with heterogeneous structures. The dynamic programming (DP) method and Markov decision process (MDP) model is utilized to describe the system behavior. We use asymptotic results for massive MIMO multicast beamforming to estimate the system capacity for the MDP model. The value iteration (VI) method is adopted to solve the MDP problems. The proposed model can enhance the system performance by solving the optimal MDP policy for user scheduling in an off-line manner with maximized average reward. The numerical results show the behavior of the algorithm and evaluate its performance.

The standard bat algorithm is slow convergence and low precision. To overcome this shortcoming, we propose a novel variant of bat optimization algorithm based on centroid strategy. The proposed algorithm has a better global searching capability because the centroid strategy can effectively prevent it falling into a local optimum. Two typical test functions are employed to test its performance. Simulation results show our proposal is both effective and efficient than three other comparison algorithms. Moreover, for high-dimensional function optimization, our proposed algorithm also has excellent approximation performance.

With the development of the electronic system towards high-speed and high-density, the influence of the interconnection topology becomes more and more significant. There are three factors deteriorating the signal: non-ideal effect of the transmission line, heavy load effect and the discontinuous impedance at the link via. These factors frequently cause signal integrity problems, and seriously restrict the realization of the high-speed and multi-load interconnection. In this paper, the limitation of the conventional daisy-chain topology is analyzed, and a novel three-dimensional daisy-chain topology is proposed by considering print circuit board (PCB) as a three-dimensional (3D) structure. The proposed topology can provide an effective method to design the multi-load interconnection of the higher speed and complexity PCB. As is demonstrated in the case study, the proposed topology can effectively reduce the non-ideal effect of multi-load branch lines and vias, which will greatly increase the noise margin of the loads. It can be seen from the results that, the proposed topology makes the received eye height of the load U3 which is the worst affected optimized for 234.8 mv (76.8%).

The development of modern radar signal processing technology put forward higher requirements for processor performance. However, Moore’s law encounters bottlenecks, the computational performance of general-purpose processors is constrained and can not meet application requirements. The high-performance and low-power features of FPGA make them recently become of interest in research as a heterogeneous computing platform together with CPU. Pulse compression algorithm is widely used in the field of radar signal processing, which contains a large number of floating-point computing, the processing effect largely depends on the performance of the processor. Based on Open Computing Language (OpenCL), we first evaluated the Fast Fourier Transform (FFT) of various sample sizes on Arria10 FPGA board and FPGA achieve up to 33.5 times the performance improvement compared to DSP C6678 on processing different sample size of FFT. Then we realize a 4 K × 8 K size pulse compression processing using kernel channel. The results show that the core computation implemented on Arria10 FPGA through OpenCL is approximately 10x faster than DSP C6678 for 4 K × 8 K size pulse compression processing.

With the characteristics of large data volume, high algorithm complexity and large computational complexity, Synthetic Aperture Radar (SAR) technology which makes the signal processing system have to be improved continuously in the aspects of real-time, storage capacity, data throughput and computing capability. As a kind of multi-core architecture, Graphics Processing Unit (GPU) take the advantages of powerful computing capability and efficient storage bandwidth to meet the urgent need in scalability, computing capability and storage bandwidth for large-scale data parallel applications. In this paper, the first thing is to evaluate the FFT performance of the NVIDIA Tesla M6 GPU, which achieves an average 41x speedup ratio compared to TI’s TMS320C6678 DSP. Then, the RD (Range Doppler) algorithm which is the most classical SAR imaging algorithm is implemented on the platform of CPU + GPU using CUDA language, and execution time of the SAR algorithm for 4 K × 8 K point is shortened by 1.18 s and the result shows that GPU achieve 1.9x the performance improvement compared to DSP C6678 on RD-SAR algorithm.

Ship detection in remote sensing imagery has been widely applied in military and citizen applications, such as fishery management, vessel surveillance or marine safety and security. With the development of optical satellite, optical satellite imagery ship detection has caused a lot of attention. In this paper, we propose an offshore ship detection method based on sparse representation. First we employ histogram of oriented gradient (HOG) as the feature descriptor, then the HOG feature are extracted from training dataset. After feature extraction, all of samples are used to adaptively train a dictionary. Next, we encode HOG feature description of patches from test image by the dictionary. Finally, the sparse code and support vector machine (SVM) classification are employed in ship target validation and false alarms elimination. Experiments have shown better detection performance and stronger robustness of our method compared with other methods.

In this paper, we apply the back propagation neural network (BPNN) into the network intrusion detection system (NIDS). To overcome the training speed and local optimality, we propose a new algorithm of simulated annealing back propagation (SABP), incorporating BPNN with simulated annealing algorithm (SAA). The simulations results show that our proposed SABP outperforms the original BPNN in terms of the training speed.

Aiming at the problem of satellite resource scheduling for multi-space targets, drawn on the experience of encoding in the Particle Swarm Optimization (PSO) algorithm, we designed an encoding style to represent the constraint and the solutions to the problem and introduced binary artificial bee colony (BABC) algorithm based on Pareto multi-objective optimization. Compared with the artificial bee colony (ABC) algorithm, the only difference is that BABC used Logistics function mapping the values to the binary. In this paper we made some improvements including population initialization which use the constraint conditions to randomly generate then modify to a feasible solution and candidate solutions generation in a way of crossover used in the Genetic algorithm. In the optimal solution search process, the Pareto optimal solution of the population is recorded, which means a set of differentiated solutions with different advantages on different indexes is obtained. It is convenient to select the corresponding optimal solution according to the user’s preference and the actual situation. The experimental results show that the improved binary artificial bee colony algorithm could solve the satellite resource scheduling problem, which provides a new idea for multi-space target satellite resource scheduling problem.

In this paper, a new coarse-to-fine framework is proposed for automatic liver segmentation on abdominal computed tomography (CT) images. The framework consists of two steps including rough segmentation and refined segmentation. The rough segmentation is implemented based on histogram thresholding and the largest connected component algorithm. Firstly, gray value range of the liver is obtained from image histogram, then the liver area is extracted from the rest of an image according to the largest connected component algorithm. The refined segmentation is performed based on the improved GrowCut (IGC) algorithm, which generates the label seeds automatically. The experimental results show that the proposed framework can efficiently segment the liver on CT images.

Spectrum sensing, a basic functionality in cognitive radio, aims at detecting the presence or absence of primary user (PU). As one of the most popular spectrum sensing methods, Covariance-based sensing works based on the correlation between signal samples. However, its performance sharply declines in low Signal Noise Ratio (SNR) environment. To improve detection performance of covariance-based sensing as far as possible, an improved blind spectrum sensing scheme is proposed in this paper on the basis of QR matrix decomposition and support vector machine (SVM). In the proposed scheme, QR matrix decomposition is applied to the co-variance matrix of received signal firstly, and then the main features are constituted by extracting and arranging orderly the upper triangular elements of R matrix. After that, SVM is used to conduct the obtained features and determine whether PU exists. The proposed algorithm does not need the prior information of PU and noise. Simulation results demonstrate that the proposed method has a better performance than conventional covariance-based methods, especially in low SNR scenarios.

The technology of the Internet develops rapidly recent years, the public tends to share their reviews, opinions and ideas on the Internet. The forms of these subjective texts are free and concise, and they contain a wealth of sentiment information. In this paper, a modified latent Dirichlet allocation (LDA) model and support vector machine (SVM) are used for sentiment analysis of subjective texts. Analysis of sentiment could help producer to enhance the products and guide user make better choices as well. We apply a modified LDA model using term frequency-inverse document frequency (TF-IDF) algorithm to mine potential topics, find the most relevant words of the topic and represent the document. Then we use SVM to categorize the texts into two classes: positive and negative. Experiment results show that the performance of the modified LDA approach is better than the traditional LDA model.

Sentiment classification is intended to classify the sentiment color categories expressed by the text. This paper illustrates the sentiment classification method based on the semi-supervised algorithm that aims to improve performance by using unlabeled data. This paper proposes a novel co-training style semi-supervised learning algorithm in order to improve semi-supervised learning ability. In our algorithm, there are three classifiers trained on the original labeled data, where the text representation for each classifier is unigram, bigram, and word2vec, respectively. And then these classifiers can use unlabeled data to update themselves. In detail, any of two classifiers have the same label, then add the new labeled data to a training set of the third classifier. By combining different types of features, our algorithm can extract text information from multiple views which contribute to sentiment classification. In addition, this algorithm doesn’t require redundant and sufficient perspectives. Experiments show that our algorithm is superior to traditional co-training algorithm and partial semi-supervised learning algorithm.

While a vast amount of applications and services are based on the Global Navigation Satellite System (GNSS), GNSS needs to deal with jamming, and how to carry out spoofing jamming recognition is the key of achieving high accuracy performance. In this paper, we apply machine learning approaches to GNSS spoofing jamming recognition. In particular, first, we investigate the scheme by employing the classical isometric mapping (ISOMAP) and Laplacian Eigen mapping (LE) algorithm to extract intrinsic feature vector for classification recognition from the original high-dimensional data. Next, we compare this scheme to another two feature vector extraction algorithms developed from principal component analysis (PCA), wavelet transform and singular value decomposition (WT-SVD). Finally, we consider the recognition rates of the four algorithms based on the support vector machine (SVM) classifier, and the effectiveness and the robustness of our scheme are verified via simulations.

Setup time is one of the most critical factors for transitory initial key based pairwise key establishment protocols in wireless sensor networks. In this paper, we propose TinyPEP, which greatly reduces the setup time of key establishment by removing unnecessary information exchanges and introducing backoff mechanism directly. TinyPEP also provides a compensation scheme for unconnected nodes. Setup time and totally connected probability are theoretically analyzed and experimentally simulated. The results show that, by choosing parameters carefully, the proposed protocol is scalable for different network densities. When the size of backoff window is 8192 slots, the setup time is less than 5.2 s and the totally connected probability is larger than 97% for typical network densities.

By studying the characteristics of WIFI fingerprint signals and combining supervised learning methods in machine learning, an innovative indoor location algorithm based on Naïve Bayes and WIFI fingerprinting is presented. In the experiment, the router is selected as the generator of WIFI signal, and the RSSI fingerprint of the signal is collected to form the fingerprint library. The Naive Bayes models are used to train the data, and the server is used to calculate the position in order to realize the fast positioning of the intelligent terminal. Experiment is designed with an indoor environment including 6 positioning points, scanning interval is set to 5 s, and the learning time is set to 10 min. The experiment result shows that the system and algorithm perform well and the accuracy of positioning is higher than 80%.

Object detection is one of the most effective ways to analyze the remote sensing (RS) images. In this paper, we focus on the prevalent object detection framework based on deep learning technology for RS images which contains three different stages, namely the region proposals generation, feature extraction, and classification. The review provides a clear picture of the challenges and possible development trends in this field. Typical methods under this framework are extensively reviewed and analyzed. Comparisons among traditional methods with deep learning methods are presented, in which supervised and unsupervised methods for RS scene target detection are deeply discussed.

The high-precision remote sensors on satellite provide massive image data which brings new challenges to data processing and interpretation. The existing data processing systems are mostly semi-automatic which have the problems as low efficiency and low interpretation quality. This paper designs and implements a ground test system for processing and interpreting remote sensing images automatically and efficiently. The system introduces a two-step feature parameter correlation calculation algorithm to interpret the images. The authors realize the system based on the structure of universal server + FPGA which capabilities can achieve a processing rate of tens of gigabit per second.

The effect of cloud cover on the quality of remote sensing data becomes an unavoidable problem when dealing with a large amount of remote sensing data obtained from satellite sensors. As an important meteorological element, cloud plays a vital role in all areas of atmospheric science. In this paper, we propose a cloud detection method based on multi-feature hierarchical judgement. First, the gray histogram of the object to be interpreted is extracted and the histogram is intercepted to remove the singular value. Then, five types of feature are employed in feature extraction. After that, the objects to be interpreted is divided into single type and mixed type, and mixed type can be further divided into certain mixed type and uncertain type. Finally, threshold method and support vector machine(SVM) are employed to classify these types. Experiment has shown good performance of the proposed method.

This paper proposes a new method for analyzing dual camera pointing direction difference of remote sensing satellite. This method can calculate relative pointing direction difference without reference to any Ground Control Points (GCPs). This enables the analysis of direction change in long image time. The theorem of the method is demonstrated, including initial calibration for dual-cameras. An experiment is also carried out based on one Chinese remote sensing satellite. Results from different images are presented to demonstrate this method can apply to random orbit. This method is widely applied on evaluation of satellite performance and improvement of image quality.

This article analyzes the feasibility of applying Wi-Fi wireless network into Attitude and Orbit Control System (AOCS) based on the previous research of commercial wireless network. Moreover, this article analyzes the wireless transmission delay of WI-FI. And analyzed results are verified by running a specified experiment. In addition, this article proposes a new scheme of wireless AOCS and develops an AOCS with Wi-Fi network by employing modified spacecraft products with six newly developed modules. Furthermore, the simulation results are illustrated and critically analyzed.

The orbit control of remote sensing satellites is the core part of satellite system application. Its performance directly affects the efficiency of satellite system. Aiming at studying the characteristics of Geosynchronous (GEO) remote sensing satellite imaging, this paper proposes the scheme for mission planning and visualization of remote trajectory remote sensing, which is based on in-depth analysis of mission planning objectives and modeling requirements. It adopts the user oriented control system architecture and branch and bound algorithm to complete the mission planning scheme. By using 2D and 3D visual presentation environment, the proposed scheme can display the result of the mission planning. Finally, an application example is given to evaluate the performance of the system. The evaluation results show that the planning scheme is fast and effective for satellite on orbit applications.

With the development of computer and information technology, simulation technology has been promoted worldwide and is playing a more and more important role in the space mission. In this paper, the main requirements of spacecraft’s observing effectiveness test during the general analysis were first analyzed. And the design of based simulation and test platform based on STK (satellite tool kit) was proposed. By using user-oriented system architecture, building up the high precision coordinate transforming relations and using domain decomposition algorithm, the path planning scheme focused on transformation from geographic coordinate space to the view plane were achieved. And the plan results were further tested using the two and three-dimension visual environment in the typical cases to provide an example for using STK simulation technology in engineering project.

The electromagnetic shielding effectiveness (SE) is essential to a spacecraft, since the electronic equipment of it may be severely interfered by the electromagnetic interference. SE is determined by both the aperture and the cable penetration except the structure material. Former research is mostly on the impact of the aperture. In this paper, both simulations and lab tests are done to investigate the impact of the cable penetration. The results show that the impact of the cable penetration is obvious and non-neglectable, and it is up to 20 dB for the frequency band lower than 1 GHz. The shielding operation can be deployed to reduce the electromagnetic leakage and then SE can be improved by 10–15 dB for the frequency band of 200 MHz–3 GHz. The results also show that the impact is not related with whether a signal is transferring or not in the cable. If there is an original cable perforation and then an additional cable perforation near it does not impact SE obviously. The impact is dependent on the locations of the built-in radiated sources.

Regarding the topic of LTE uplink interference, this article mainly discussed the impact of LTE network on RSSI, and the impact of RSSI on network quality and user perception. Furthermore, this article classified the suggested RSSI thresholds based on the extent of impact, and utilized these thresholds, in order to guide optimization practice and improve user perception. Finally, after taking a cell of an indoor distribution system on campus with high RSSI level as an example, user perception rate has been found significantly increased after the cell splitting.

Customer churn prediction is important for telecom operators to retain valuable users. Accurate features that can characterize customer behaviors, as well as efficient extraction method are key factors in constructing the customer churn analysis model. In literature, Support Vector Machine (SVM) has shown its applicability to the problem of customer churn analysis. This paper identifies the main features that influence the customer churn model from telecom experts’ viewpoints, and proposes a suitable one based on Support Vector Machine (SVM). An experimental results also is illustrated to verify reasonableness of the proposed models.

Development of LTE wireless network raises new ways for customers to communicate, by offering the possibility to access the network anywhere and anytime. As a result of the investments made over the last few years, this network has achieved widespread popularity. Some specialized infrastructures have been raised to address the special needs of customers. One of this need is to give the opportunity to connect to the network while journeying in a high-speed train. Mobile operators have elaborated special cells to manage high-speed train passengers, taking into account the characteristics of this environment. A key issue is to monitor the effective usage of the cells, by checking that passengers connect to those special cells (and not to the neighbor common cells), and that other users connect to common cells (and not to the special ones). For this purpose, a monitoring system of cells based on data analytics is detailed. This system identifies service performance of each cell, by pointing out common cells where high-speed train passengers are attaching to, and special cells where too many non-passengers are connecting to. The whole system helps mobile operators to elaborate strategies to improve service performance, by determining which cells should be tuned.

Traditional analysis of the Mobile Radio Network Operation Stability (MRNOS) is usually based on several indicators or human experiences, which leads to a great difficulty in carrying out a systematic evaluation. A Comprehensive Evaluation Algorithm (CEA) was proposed in order to solve this issue. However, CEA ignores the efficiency and objectivity in practical implementation, therefore, in this paper, an improved High-efficiency Dynamic Evaluation Method (HDEM) is proposed to solve three main problems of the previous method, including high complexity, non-objectivity and low efficiency.

The PCI mod3 interference in the Long Term Evolution (LTE) system will cause degradation of radio access, handover, and quality of service, which seriously pull down user feeling. We propose a novel scheme utilizing advanced intelligent Genetic Algorithm to mitigate PCI mod3 interference which is based on data resources of Drive Test, handover, and the Measurement Report. The practical network trials demonstrate our scheme dramatically reduced the operational complexity and PCI mod3 interference in LTE system.

Effective base station (BS) evaluation can assist telecom operators to find problematic cells and optimize the system performance. This paper proposes a data mining based joint BS evaluation (JBSE) algorithm in LTE cellular systems. Initially, the JBSE algorithm considers four key factors, including the cell energy consumption, the cell revenue, the cells distribution induced interference, the high BS induced cross-boundary coverage. Then, the expert judgement matrix is employed to rank the level of each factor. Finally, the JBSE algorithm evaluates each cell comprehensively. The JBSE algorithm is used in the LTE systems evaluation of a city in China. It can find problematic cells effectively.

In this paper, we incorporate human visual speed perception model into a NSS based VQA method for video sequences transmitted via wireless network. The human visual speed perception contains two parts: one is motion information, which is calculated by using the prior probability distribution of the relative motion in each frame of the video; the other one is the perception noise, derived from the background motion. The weighting factors are defined as perceptual information that minus perception noise from motion information. We extract both spatial and temporal statistical features in videos (NVS-S and NVS-T), and measure their deviations from pristine statistical features. Consequently, the deviations can be synthesized with perceptual information based weighting coefficients to get the video quality score. The proposed blind VQA model is trained and tested in the LIVE database and EPFL-PoliMI database. The experimental results indicate that our model outperforms other blind VQAs.

Inappropriate parameters and antenna problems will result in abnormal coverage performance of LTE networks. In order to deal with the problems mentioned above and improve the coverage performance, this paper proposes a downlink coverage Self-optimizing algorithm (DCSA) on the strength of big data analytics. The proposed algorithm obtains and analyzes the data, which records the performance situation of existing wireless networks to locate the cells with abnormal coverage performance. Then a Self-optimizing method is proposed to improve the coverage by adjusting the parameters. In the last part, the analysis results will display that the coverage optimization algorithm is high-efficiency and low-cost for telecom operators.

Wireless network planning is a key component for TD-LTE system deployment. In this paper, we investigate the wireless network planning method for a railway TD-LTE system. This evaluation is based on an actual TD-LTE system deployment for Shuohuang railway and includes frequency planning, network capacity prediction, wireless network coverage planning and cell planning. The corresponding Reference Signal Received Power (RSRP), Signal to Interference plus Noise Ratio (SINR) and throughput of the Railway TD-LTE System are further evaluated by Atoll network planning software, to validate the rationality for wireless network planning in such scenarios.

With the popularity of 4G mobile networks, mobile video service becomes the key service in the 4G era. In order to improve users’ awareness and increase the network optimization efficiency, it is important to establish a scientific and accurate model to evaluate the video service from the user’s perception. In this paper, we focus on the video streaming traffic and propose a modelling approach to evaluate the video service performance. The essential characteristics of video traffic are taken into account. Based on the hierarchical clustering and the Pearson correlation coefficient method, key factors of video service perception are determined. Furthermore, the threshold values of key factors are obtained through extensive user surveys and simulation tests. The results of the application in the realistic network demonstrate the effectiveness of the proposed model. In addition, results show the proposed model enables the telecom operator to evaluate the video service quality of each user or user group, which helps improve the network optimization efficiency.

Drive Test (DT) is a very important approach to measure the quality of the wireless networks. But with the rapid development of wireless networks, the network structure comes to more and more complicated, such as massive network, different systems and multiple manufacturers. The complexity of network leads to a mass expenditure in operation and maintenance which including DT. In this paper, we study a new method, called Virtual Drive Test (VDT) by using Measurement Report (MR) and Call Detail Record (CDR), to measure the quality of the wireless networks in more efficient and less costly ways.

With the popularization of cars in China, traditional auto insurance market experiences fierce competition. Auto insurance company concerns with insurance product innovation, thus improving the insurance service level. For the purpose of auto insurance innovation, this paper proposes a novel driving behavior model to evaluate the driving risk. Through logistic regression algorithm, we analyze the correlation between the driving score and the accident. Then, we discuss the reliability of this model. Furthermore, we employ this driving score to the auto insurance pricing model, in order to improve the risk identification of the pricing model. Both the novel driving behavior model and the auto insurance pricing model can achieve the effective risk segmentation and precise pricing, as well as assisting the auto insurance company to improve the market competition capability.

Entering the 21st century, data are considered as nationally strategic resources for business innovations. Currently, it is a global trend to utilize big data to promote economic development, improve social governance, and enhance governmental services and regulatory capabilities. However, the low-quality data are seriously hindering applications for data analysis and decision support. Testing is a necessary part of quality assurance. This paper presents a theoretical survey on compliance testing for data quality assurance. Firstly, basic concepts of data quality and compliance testing are studied. Secondly, data quality testing models are proposed, involving testing element, testing object and testing process. Finally, the application model of documents data quality testing is demonstrated through requirement analysis of documents.

Mobile cellular networks are experiencing fast development recently. The electromagnetic wave is the key factor to impact the cell coverage and the performance of mobile cellular networks. In this paper, we research the electromagnetic wave under the haze and the rainstorm. Initially, we study the basic theory of electromagnetic wave and research its propagation characteristic. Then, we analyze the theoretical impact of electromagnetic wave under the haze and the rainstorm. In order to verify the theoretical analysis, we employ the telecom big data in a city of China to analyze the cell coverage and the service access performance. Results show the haze has little impact on the cell coverage and the access performance, whilst the rainstorm degrades the cell coverage.

In the age of 4G, mobile video has become the key development strategy of Internet companies and telecom operators. Because the user perception of video service has a direct impact on the reputation of the operators’ networks, the evaluation of video traffic quality becomes one of the most important tasks for operators. Video is different from other types of service and it needs to inspect every packet to analyze video fluency which is a challenge for both the analytic algorithm and the server. This paper proposes a pre evaluation algorithm to evaluate video fluency which can dramatically reduce the inspection overhead as well as the server pressure.

With the development of e-commerce technology, a growing number of people prefer to purchase clothes on the e-commerce websites. Therefore, an effective recommendation system is necessary for customers. User-based Collaborative Filtering (UCF) algorithm is widely utilized to predict the preferences of customers. However, UCF algorithm employs the sparse matrix and the recommendation has low precision. In this paper, an improved recommendation algorithm named Advanced User-based Collaborative Filtering (AUCF) algorithm is proposed and implemented in the clothing recommendation system. The proposed AUCF algorithm introduces user-item linked list, which can overcome the problem of large time complexity. Considering the impact of different popularity of items, AUCF algorithm is capable of publishing the negative influence of popular items, which can increase the recommendation coverage. Experiment results show the AUCF algorithm significantly increases the recommendation coverage and precision.

Traditional methods are generally based on single data source or several simple indicators to achieve coverage efficiency analysis, which makes them neither effectively reflect the full reality of network nor provide sufficient restrictions. In response to this problem, an OSS data based LTE wireless coverage efficiency analysis method is proposed. This method is capable of reflecting network status from the granularity of cell-level indicators and the dimension of multi-source data. Using this method, the improvement in the effectiveness of network state reflection and the enhancement of ability to coverage efficiency analysis have been demonstrated by practical application.

This article presents a new parameter optimization algorithm based on radio coverage grids clustering analysis. First, the coverage zone is divided into some grids based on drive test data and measure reports including GSM/WCDMS/LTE RF signal performance index. Next, using system integration clustering algorithm, all grids are clustered in different groups based on the degrees of RF signal deviation. Finally, we select the optimized parameter configurations based on GSM/WCDMA/LTE signal strength and quality in each group. The performance of the inter-RAT cell reselection and handover are improved.

With the gradual improvement of mobile network technology and the vigorous development of mobile Internet business, data traffic revenue has become telecom operators’ main driver of profit growth. In this paper, a novel value-added service strategy is proposed for 3G mobile communication, in which the priority quality of service (QoS) will be provided in the case of network resource constraints if the user order this service. The main objective of this service is to improve user perception in network resource limitation. The experimental results reveal that the proposed scheme is able to achieve good performance in addressing the user satisfaction problem.

To deal with the ever-growing demand for data service, the small cell network (SCN), a new structure network has been proposed. Although through spatial reuse SCNs may greatly increase the networks’ throughput, it may also lead to severe intra-layer interference and excessive handover (HO) overheads. Besides, the large-amount small cell scenery can also lead to severe network energy consumption. In the article, a dynamic power control scheme is proposed for small cells to handle the interference and HO problems. It can not only guarantee the cell edge users’ perception but also reduce the network energy consumption. The set of experiments in the big data analysis shows the scheme can bring a substantial growth in cell edge users’ throughput as well as decrease system outage probability.

In the traditional marketing activities, the key to success is to identify the potential needs of target users. An excellent recommendation system can help improve the effectiveness and efficiency of the precision marketing. In this paper, an Apriori algorithm based recommendation scheme is proposed, and it’s then applied to the mobile APP marketing field. The scheme is realized through two processes, the frequent item sets generation process and the strong association rule generation process. Experimental results have shown that the scheme can achieve high advertising arrival rate, as well as superior exposure and click conversion rate. The effectiveness of the mobile APP product promotion has been improved dramatically.

It is a convention to verify the validity of a universal propagation model before its local application with the aid of mass local experimental data and make the universal propagation model further localized. This paper proposes detailed comparisons between theoretical predictions of field strength based on ITU-R P.1546 with the data collected from the experiments conducted on the South China Sea, which proved that the recommendation ITU-R P.1546 is valid in China tropical maritime region. Three optimization methods are also proposed for tuning the parameters from ITU-R P.1546 recommendation to improve accuracy of the propagation prediction in China tropical maritime region. At the end, future researches on tropical maritime propagation prediction are suggested.

The real estate industry is a hot topic and the factors of a house which affect the investment benefit is worth of research. This paper designs a novel machine learning assisted real estate industry investment guidance (MLRIG) architecture and a machine learning algorithm, aiming at researching the factors and their weight respectively of a house which have influence on its investment value. The MLRIG architecture is composed of 4 stages: Data collection, Data discretization, Data Mining Process and Factors weight output; the proposed machine learning algorithm, called QSFL-LR (Quantum-inspired Shuffled Frog Leaping Logistic Regression), combines Quantum-inspired Shuffled Frog algorithm with Logistic Regression to select the factors of a house which affect the investment value before data training, then output the weight of the factors respectively. Experiment shows the proposed QSFL-LR algorithm has better performance in accuracy and precision compared with traditional Logistic Regression, proving the superiority of QSFL-LR. The experiment also shows MLRIG architecture can guide both business companies and individuals to reduce investment risk in real estate industry.

The evaluation of region development of network has been a widespread concern in recent years. However, network development of a region, due to its specific characteristics and application scenario, should have a tailor-made evaluation system. In this study, taking into account various factors in multiple fields, a multiple-index evaluation system is established. Then, a principal component analysis-based K-means clustering approach is proposed to address the analyzing problem with an acceptable complexity. A simulation experiment is implemented to verify the algorithm. The results can be used to compare the different areas telecommunication networks, and provide rational and effective suggestions for network planning and construction.