Skip to main content
main-content

Über dieses Buch

This two-volume set LNICST 301 -302 constitutes the post-conference proceedings of the Third EAI International Conference on Advanced Hybrid Information Processing, ADHIP 2019, held in Nanjing, China, in September 2019. The 101 papers presented were selected from 237 submissions and focus on hybrid big data processing. Since information processing has acted as an important research domain in science and technology today, it is now to develop deeper and wider use of hybrid information processing, especially information processing for big data. There are more remaining issues waiting for solving, such as classification and systemization of big data, objective tracking and behavior understanding in big multimedia data, encoding and compression of big data.

Inhaltsverzeichnis

Frontmatter

Modeling Analysis of Network Spatial Sensitive Information Detection Driven by Big Data

The dissemination of sensitive information has become a serious social content. In order to effectively improve the detection accuracy of sensitive information in cyberspace, a sensitive information detection model in cyberspace is established under the drive of big data. By using word segmentation and feature clustering, the text features and image features of current spatial data information are extracted, the dimension of the data is reduced, the document classifier is built, and the obtained feature documents are input into the classifier. Using the open source database of support vector machine (SVM) and LIBSVM, the probability ratio of current information belongs to two categories is judged, and the probability ratio of classification is obtained to realize information detection. The experimental data show that, after the detection model is applied, the accuracy of the text-sensitive information detection in the network space is improved by 35%, the accuracy of the image information detection is improved by 29%, and the detection model has the advantages of obvious advantages and strong feasibility.

Ruijuan Liu, Bin Yang, Shuai Liu

Research on Communication Individual Identification Method Based on PCA-NCA and CV-SVM

In recent years, high-dimensional data has often appeared in the fields of science and industry, such as computer vision, pattern recognition, biological information, and aerospace. Feature dimension reduction and selection are the process of reducing data from high dimensionality to low dimensionality to reveal the nature of the data. In the field of wireless communication, in view of the feature redundancy caused by the high-dimensional features of wireless device startup transient signals, this paper converts the high-dimensional features of signals into low-dimensional features that are conducive to classification through the feature dimensionality reduction and selection method based on PCA-NCA. In addition, this paper also carried out parameter optimization for SVM classifier, and the established CV-SVM classifier improved the classification performance. This paper also carries out simulation devices on the measured start-up signals of ten identical walkie-talkies. When the SNR is greater than 0 dB, the recognition accuracy of the PCA-NCA algorithm is 10% higher than recognition accuracy of the PCA algorithm alone; when the SNR is greater than 10 dB.

Xinghao Guo, Shuai Liu

Termination for Belief Propagation Decoding of Polar Codes in Fading Channels

In additive white Gaussian channel (AWGN), the performance of polar codes under the successive cancellation (SC) decoding is not as good as that of the belief propagation (BP) decoding. However, in a fading channel, the performance of BP decoding is found in our study to be worse than the SC decoding. In this work, we propose a termination criterion for the BP decoding of polar codes to improve the performance and the average number of iterations at the same time. Simulation results show that BP decoding can still achieve a better performance than the SC decoding in fading channels with the proposed termination.

Chen Zhang, Yangzhi Luo, Liping Li

Secrecy Capacity Analysis for Indoor Visible Light Communications with Input-Dependent Gaussian Noise

This paper mainly focus on the performance of secrecy capacity in the physical layer security (PLS) for the eavesdropping channel in visible light communication (VLC) system. In this system, due to the effects of thermal and shoot noises, the main interference of the channel is not only from additive white Gaussian noise (AWGN), but also dependent on the input signal. Considering a practical scenery, based on the input-dependent Gaussian noise, the closed-form expression of the upper and lower bounds of secrecy capacity are derived under the constraints of non-negative and average optical intensity. Specifically, since the entropy of the output signal is always greater than the input signal, on this basis, the derivation of lower bound is using the variational method to obtain a better input distribution. The upper bound is derived by the dual expression of channel capacity. We verified the performance of secrecy capacity through numerical results. The results show that the upper and lower bounds are relatively tight when optical intensity is high, which proves validity of the expression. In the low signal-to-noise ratio (SNR) scheme, the result of bounds with more input-dependent noise is better than less noise. And in the high SNR scheme, the result of bounds with less input-dependent noise outperforms that noise is more.

Bo Huang, Jianxin Dai

The Recursive Spectral Bisection Probability Hypothesis Density Filter

Particle filter (PF) is used for multi-target detection and tracking, especially in the context of variable tracking target numbers, high target mobility, and other complex environments, it is difficult to detect, estimate and track targets in these situations. This paper discusses the probability hypothesis density (PHD) filtering which is widely used in the field of multi-target tracking in recent years. The PHD filter algorithm can estimate the number of targets effectively, however, existing algorithms does not make full use of particle information. This paper proposes a target state extraction method based on the recursive spectral bisection (RSB) node clustering algorithm, which focus on eigenvector centrality, algebraic connectivity, and the Fiedler vector from the established field of spectral graph theory (SGT). The method makes full use of the geometric distance relationship and the weight of particles to construct the particle neighborhood graph, then use the algebraic connectivity and Fiedler vector obtained by the eigenvalue decomposition of the Laplace matrix, finally extracts the target state from each class of particle group. Simulation results demonstrate that the new algorithm provides more accurate state estimations for multi-target detection and tracking.

Ding Wang, Xu Tang, Qun Wan

Research on LAN Network Malicious Code Intrusion Active Defense Technology

Traditional LAN networks had low defense efficiency and poor stability. In order to solve this problem, a new malicious code intrusion active defense technology was studied, and the defense technology structure was designed and the work-flow was studied. The system structure was divided into hardware layer, kernel layer and executive layer. The work-flow was divided into four steps: file judgment, file compression, file processing and file display. The working effect of the technology was verified by comparison with the traditional method. It was known from the experimental results that the studied technology had high defense efficiency and strong stability.

Lei Ma, Ying-jian Kang, Hua Han

Network Information Security Privacy Protection System in Big Data Era

Traditional information security protection is based on an information data set, at least one information can not be distinguished from its own location information. Therefore, this paper studies the network information security privacy protection system in the era of big data. The hardware of network information security privacy protection system is composed of independent monitoring layer, host layer and mixed layer. It disturbs the original data by adding random numbers and exchanging, shields the original data to unauthorized users, and achieves the purpose of privacy protection and recommendation accurate and non-destructive. The system software encrypts information according to the degree of privacy protection set by users, adopts the key management mode, solves the problem of communication security and node key update, and realizes the network information security privacy protection system.

Lei Ma, Ying-jian Kang, Jian-ping Liu

Multi-path Channel Modeling and Analysis of Embedded LTE Wireless Communication Network Under Cloud Computing

The multi-path channel modeling analysis of conventional communication network can analyze the modeling of multi-path channel in communication network. However, for the multi-path channel modeling analysis of embedded LTE wireless communication network under cloud computing, there is a shortage of high analysis error rate. To this end, the multi-path channel modeling analysis of embedded LTE wireless communication network under cloud computing is proposed. The LTE multipath channel modeling and analysis program structure is built, and the single-frequency signal is subjected to time-varying channel technology, and the digital signal is designed by multipath time-varying channel technology to complete the key technology design of multipath channel modeling and analysis. The multipath channel modeling and analysis program is used to determine the multi-path channel modeling Rayleigh distribution and Rice distribution, and the related characteristics analysis to realize multi-path channel modeling analysis of embedded LTE wireless communication network under cloud computing. The experimental data show that the proposed multipath channel modeling analysis is more than the conventional multipath channel modeling analysis, and the analysis error rate is reduced by 14.35%, which is suitable for multi-path channel modeling analysis of embedded LTE wireless communication networks under cloud computing.

Yanhua Qiao, Lin Zhao, Jianna Li

Layered Encryption Method for Monitoring Network User Data for Big Data Analysis

The conventional monitoring network user data layered encryption method had a low security when layered encryption of modern network data. Therefore, a layered encryption method for monitoring network user data for big data analysis was proposed. Big data technology was introduced, and a layered framework of network user data was built to monitor and encrypt network user data. Relying on the determination and layering of different levels of user data, the data layered encryption model was embedded to realize the layering and encryption of monitoring network user data. The test data showed that the proposed layered encryption method for monitoring network user data for big data analysis would improve the security of the data by 46.82%, which was suitable for users of different levels to encrypt their own network data.

Yanhua Qiao, Lin Zhao, Jianna Li

Research on Emergency Communication Command and Scheduling Algorithm Based on Pattern Recognition

In the process of regular pattern recognition, due to the limitation of the algorithm, the emergency communication command and dispatch has a delay and a certain deviation. In order to solve the above problems, an emergency communication command and scheduling algorithm based on pattern recognition is proposed. The delay time constraint is determined by the time interval calculation, and the channel line selection is performed on the basis of the established condition, and the control transmission amount is below the maximum amount supported by the selected channel line, according to the information length, the priority level is divided and the ratio of the command scheduling algorithm is studied. The proposed algorithm is compared with the traditional command and dispatch algorithm, and the balance is more than 10% higher than the traditional algorithm; the scheduling time is also saved by about 30 s, which fully proves the feasibility of the scheduling algorithm.

Jun Li, Jun Xing

Multi-source Heterogeneous Data Acquisition Algorithm Design Different Time Periods

The traditional algorithm was affected by dynamic error and data loss, resulting in low efficiency of collection. In order to solve this problem, a time division collection algorithm based on data format transformation was proposed. According to the data format conversion process multi-source heterogeneous configuration files, and access to the content of the whole configuration file and the GDAL, according to the results of the configuration process design algorithm, under the constraints of the input data for approximate operation, minimize the objective function, through the fixed matrix other factors influence on partial derivatives root, period of time the multi-source heterogeneous data acquisition algorithm design. The experimental results showed that the maximum collection efficiency of the algorithm can reach 90%, which provided an effective solution for scientific researchers to solve the problems caused by differences in data format.

Jun Li, Jun Xing

Research on Data Security Acquisition System Based on Artificial Intelligence

The traditional acquisition system had low collection efficiency and poor acquisition accuracy. In order to solve the above problems, a new data security acquisition system was studied. Artificial intelligence technology was introduced to design the hardware and software parts of the system. The hardware part mainly designed the system A/D acquisition module, serial data acquisition module and parallel port communication module. The software part was divided into three steps: data screening, data analysis and data acquisition. By comparing with the traditional system, the actual working effect of the system was verified. The experimental results showed that the data security acquisition system based on artificial intelligence had higher collection efficiency and better precision. The system was worthy of recommendation.

Yingjian Kang, Lei Ma, Leiguang Liu

Weak Coverage Area Detection Algorithms for Intelligent Networks Based on Large Data

Aiming at the problem of abnormal location often occurring in traditional weak coverage area detection algorithm of intelligent network, a detection algorithm of weak coverage area in intelligent network based on large data is proposed. Firstly, the detection data is collected by data acquisition method based on local characteristics, and then the gray level conversion of these detection data is used to realize the pre-processing of the detection data and the detection after pre-processing. The feature vectors are used to describe the feature points so as to realize the accelerated feature matching of the detected data. Then the region feature detection of the detected data is carried out, and finally the weak coverage area detection algorithm of the intelligent network based on large data is realized. Experiments verify the detection performance of the weak coverage area detection algorithm based on large data in intelligent networks, and draw a conclusion that the detection algorithm based on large data has a much smaller probability of abnormal location than the traditional weak coverage area detection algorithm in intelligent networks.

Ying-jian Kang, Lei Ma, Ge-cui Gong

Research on RF Fingerprinting Extraction of Power Amplifier Based on Multi-domain RF-DNA Fingerprint

The uniqueness of the RF signal is caused by the difference in the hardware structure of the transmitter and the differences between the different devices. Among them, RF power amplifier is one of the key components of RF fingerprinting of wireless transmitter. It is an important breakthrough for RF fingerprint generation mechanism and individual identification. This paper proposes a new identification method of power amplifier based on new intelligent feature set, firstly, processing the received signal. The time domain, frequency domain, time-frequency domain, fractal domain transformation and feature extraction are performed. Secondly, the new intelligent feature set of each power amplifier individual can be characterized, and the RF-DNA fingerprint is visualized. Finally, the support vector machine is used to realize the individual recognition by selecting the optimal RBF kernel function. By simulating and verifying the eight power amplifier signals, a new intelligent feature set can be used to uniquely characterize the power amplifier. Under low SNR, the power amplifier individual can be quickly and effectively identified. The recognition rate of more than 80% can be achieved above the −5 dB signal-to-noise ratio.

Yihan Xiao, Xinyu Li

Fault Feature Analysis of Power Network Based on Big Data

During the operation of the power network, there was a sharp change in current and voltage at the time of failure, which made it difficult for the grid operators to quickly and accurately determine the fault. This paper proposed a big data-based power network fault feature analysis method design. Taking the symmetrical fault component method as the main analysis method, a two-phase short-circuit equivalent model was constructed by accurately analyzing the fault characteristics of the power network, and the fault features were detected and located by the big data network preprocessor. The experimental results shown that the big data power network fault feature analysis method could effectively feedback and locate the fault location and complete the maintenance of the power network in time.

Cai-yun Di

Design of All-Pass Filter System for Power Communication with High Anti-harmonic Interference

In the traditional system, the prediction sensing unit was lacking, which caused the output voltage and current harmonic content to deviate greatly from the actual value. In order to solve this problem, the design of the all-pass filtering system for power communication with high anti-harmonic interference was proposed. According to the hardware structure block diagram of the system, the predictive sensing module was designed to obtain readable and unreadable information. In order to make the system only transmit readable information, the closed switch was designed, and the client module at the end of the hardware was set to display the prediction results and improve the harmonic interference problem; For the above system hardware control module, the software part was designed, and the filter system function was determined according to the software design flow, thereby completing the design of the all-pass filter system for power communication with high anti-harmonic interference. The experimental results showed that the output voltage and current harmonic content of the system were consistent with the actual value, which provided a certain reference for the filter anti-interference.

Caiyun Di, Zhi Zhao

Improved Design of Classification Algorithm in Cloud Computing and Big Data Environment

With the rapid improvement of China’s economic level, science and technology are also progresses, and the scope of the application of science and technology in daily life is becoming more and more extensive, and the large data of cloud computing is also applied to all aspects of our daily life. The classification algorithm is the key to reflect the large data computing ability of the cloud computing. It can further improve the analysis ability of the related data, make the operation of the related data more convenient, more close to the needs of the searcher for information, and avoid a large number of invalid information, because this is very demanding for the classification algorithm. On this basis, we analyzed the operation of the classification algorithm in the cloud computing environment, and used the clustering algorithm to improve the design, improve the efficiency of the related data and improve the accuracy of the data collection.

Yihuo Jiang

Research on Dynamic Access Control Model of Distributed Network Under Big Data Technology

The traditional distributed network dynamic access control management model has the defects of poor control management efficiency and poor expansibility. In order to solve the above problems, the dynamic access control management model of the distributed network is constructed by the large-data technology. According to the requirement of distributed network dynamic access control management model, the mining model is constructed. Based on this model, the direct trust value and indirect trust value are calculated by big data technology, and the final trust value is obtained by combining them. Based on the final trust value obtained, the dynamic access control management process is formulated and executed to realize the control and management of distributed network dynamic access. The simulation results show that compared with the traditional distributed network dynamic access control management model, the distributed network dynamic access control management model greatly improves the efficiency and expansibility of the model. It fully shows that the distributed network dynamic access control management model has better control and management performance.

Yi-huo Jiang

Intelligent Data Acquisition Method for Cross-border E-commerce Guidance and Purchase Considering User Demand

At present, due to the unknown online procurement, and remote distance for the cross-border e-commerce shopping guide, business considerations for users were not so full. Based on this, a cross-border e-commerce shopping guide big data intelligent collection method considering user needs was proposed. Through the mobile port of big data to collect and explore the online social retail collection method of big data, it focused on promoting the way of big data onto intelligent collection, and promoted cross-border e-commerce shopping guide big data to better carry out commodity circulation of user needs. Experiments showed that the big data collection method studied in this paper better combined user needs and merchant profit, and helped to improve user experience.

Jiahua Li

Optimization Design of Cross-Border E-commerce Shopping Guide System Combining Big Data and AI Technology

In the era of Internet economy, cross-border e-commerce shopping guides were conducted under the conditions of virtual network environment. Therefore, the traditional cross-border e-commerce shopping guide system had long been unable to meet the diversified needs of cross-border e-commerce shopping guides. A cross-border e-commerce shopping guide system combining big data and AI technology was proposed and designed. Using big data and AI technology, the hardware and software of the cross-border e-commerce shopping guide system were analyzed respectively, and the optimized design of the cross-border e-commerce shopping guide system was completed. The experimental data showed that the cross-border e-commerce shopping guide system combining big data and AI technology had better performance than the traditional system, and could better meet the technical requirements of cross-border e-commerce shopping guide.

Jiahua Li

Research on Balanced Scheduling Algorithm of Big Data in Network Under Cloud Computing

In order to find the optimal big data balanced scheduling scheme under cloud computing and reduce the completion time of the task, an improved ant colony algorithm based algorithm for large data equalization scheduling under cloud computing was proposed. Firstly, a balanced scheduling algorithm structure was established, then the equilibrium problem to be explored was described, finally, the ant colony algorithm was used to simulate the ant search food process to solve the objective function. And the local and global information deep update methods was introduced to improve, speed up the search speed, and finally the performance test experiments on CloudSim simulation platform was performed. The results show that compared with the discrete particle swarm optimization (DPSO), the algorithm not only greatly reduces the execution time of cloud computing tasks (2.5 s), but also solves the problem of unbalanced data load, and achieves the balanced scheduling of large network data under cloud computing.

Lunqiang Ye

Optimization of Rational Scheduling Method for Cloud Computing Resources Under Abnormal Network

When the traditional heuristic algorithm was used to schedule the cloud computing resources under the abnormal network, there was a problem that the scheduling speed was slow and the effect was poor. Aiming at the above problems, combined with the characteristics of cloud computing and the actual needs of cloud computing resource allocation, based on the advantages of genetic algorithm and ant colony algorithm, a hybrid optimal cloud computing resource scheduling algorithm was designed. The improved algorithm combines the advantages of genetic algorithm and ant colony algorithm, and the genetic algorithm can effectively improve the search efficiency; The ant colony algorithm was used in the later stage of the algorithm to improve the accuracy of the optimal solution and to complete the reasonable scheduling of cloud computing resources under the abnormal network. The results show that the hybrid algorithm was faster than the single genetic algorithm and ant colony algorithm. It only took 10 s, the resource load was more balanced, and the scheduling effect was better.

Lunqiang Ye

Design of Agricultural Product Quality and Safety Big Data Fusion Model Based on Blockchain Technology

In the process of processing agricultural product quality and safety data, the traditional model will have problems such as long delay and redundant storage. Therefore, based on blockchain technology, combined with historical data, real-time data features, external shared data features and previous research results, the data is deeply integrated, the agricultural product quality and safety big data fusion model is designed. Create a big data fusion framework based on blockchain technology to collect and process agricultural product quality and safety data reasonably and efficiently, realizing the quality and Safety of the Integration of Agricultural products big data. The data architecture is proposed in the big data fusion model, and the collection and data storage methods of the quality and safety supervision system are designed to achieve efficient collection and storage of data. The experiment proves that the agricultural product quality and safety big data fusion model has certain advantages over the traditional model.

Kun Wang

Artificial Intelligence Integration Method for Agricultural Product Supply Chain Quality Data Based on Block Chain

Traditional supply chain quality data integration methods costed a lot in integrating product quality, but the integration accuracy was very low and the effect is poor. In order to solved this problem, a supply chain of agricultural products was set up based on the artificial intelligence integration method of block chain using quality data. The framework of agricultural product supply chain was designed. The supply chain included four steps of production, processing, trade and consumption. Based on the frame, the workflow of the supply chain of agricultural products was expounded. The feasibility of the construction of agricultural product supply chain was verified by the experiment. The experimental results showed that the design of intelligent integration method can effectively reduce cost and improve the accuracy of integration.

Kun Wang

Research on Adaptive Scheduling Method of Communication Resource Information in Internet of Things Environment

The continuous expansion of information resources of communication resources under the Internet of things environment had led to information management problems becoming one of the core issues of communication information integration. This paper proposed an adaptive scheduling method for communication resource information in the Internet of things environment. Based on GIS technology, communication resource management was used for database communication resource information sharing. Next the three-level distributed database was used to establish the upload and storage of Internet of things communication resource information with obvious hierarchical relationship. The database synchronization mechanism was used to ensure that each database communication resource configuration was synchronized, and a network channel was established to increase the network load. Finally, based on the browser information, the communication resources were adaptively scheduled. The experimental data showed that compared with the traditional resource information scheduling method, the resource transmission speed of the designed communication resource information resource adaptive scheduling method was increased by 65%, and the information resource transmission matching rate was increased by 27%.

Chao Wang

Research on Dynamic Network Load Evaluation Algorithm Based on Throughput Monitoring

In order to solve the problem of too low total load of individual nodes, a dynamic network load evaluation algorithm based on throughput monitoring was proposed. The node monitoring quantity analysis and evaluation parameter determination were used to complete the monitoring description of the dynamic network throughput rate. On this basis, through the improvement of load effect, evaluation mechanism establishment and correction factor calculation, the new evaluation algorithm was completed. The experimental results showed that after applying the dynamic network load evaluation algorithm based on throughput monitoring, the problem of too low total load of individual nodes was effectively solved.

Chao Wang

Research on the Classification Method of Network Abnormal Data

As people use the network more and more and release more and more personal information to the Internet, it also caused the leakage of personal information. According to the above background, the optimization research on the classification detection method of network anomaly data was proposed. Correlation analysis was carried out for the conventional algorithm, and the related model was constructed. A new algorithm was proposed to detect the network anomaly data to improve the processing ability of the network anomaly data. The experimental data showed that the proposed network anomaly data classification detection optimization algorithm improved the processing range by 31% when processing abnormal data, and the efficiency of processing data was increased by 36%. It proved the effectiveness of the new method and provided a theoretical basis for the processing of future abnormal data.

Bozhong Liu

Research on Key Information Retrieval Method of Complex Network Based on Artificial Intelligence

Aiming at the problem of poor retrieval accuracy and slow retrieval speed of information retrieval method based on hyperlink, a key information retrieval method based on artificial intelligence was proposed. The method was mainly divided into three steps, and each step was completed with the help of artificial intelligence. First, file information was preprocessed (information processing and information filtering), then keywords were extracted from information content, and finally semantic similarity calculation and semantic information matching were conducted to complete key information retrieval in complex networks. The results showed that the accuracy of key information retrieval method of complex network based on artificial intelligence was improved by 2.27% and the speed of retrieval was improved by 3.06 s.

Bozhong Liu

Optimized PointNet for 3D Object Classification

Three-dimensional (3D) laser scanning technology is widely used to get the 3D geometric information of the surrounding environment, which leads to a huge increase interest of point cloud. The PointNet based on neural network can directly process point clouds, and it provides a unified frame to handle the task of object classification, part segmentation and semantic segmentation. It is indicated that the PointNet is efficient for target segmentation. However, the number of neural network layers and loss function are not good enough for target classification. In this paper, we optimize the original neural network by deepen the layers of neural network. Simulation result shows that the overall accuracy increases from 89.20% to 89.35%. Meanwhile, the combination of softmax loss with center loss function is adopt to enhance the robustness of classification, and the overall accuracy is up to 89.95%.

Zhuangzhuang Li, Wenmei Li, Haiyan Liu, Yu Wang, Guan Gui

Deep Learning Based Adversarial Images Detection

The threat of attack against deep learning based network is gradually strengthened in computer vision. The adversarial examples or images are produced by applying intentional a slight perturbation, which is not recognized by human, but can confuse the deep learning based classifier. To enhance the robustness of image classifier, we proposed several deep learning based algorithms (i.e., CNN-SVM, CNN-KNN, CNN-RF) to detect adversarial images. To improve the utilization rate of multi-layer features, an ensemble model based on two layer features generated by CNN is applied to detect adversarial examples. The accuracy, detection probability, fake alarm probability and miss probability are applied to evaluate our proposed algorithms. The results show that the ensemble model based on SVM can achieve the best performance (i.e., 94.5%) than other methods for testing remote sensing image dataset.

Haiyan Liu, Wenmei Li, Zhuangzhuang Li, Yu Wang, Guan Gui

Detection for Uplink Massive MIMO System: A Survey

In this paper, we make a compressive survey for the research on detection in uplink Massive multiple input and multiple output (MIMO) system. As one key technology in Massive MIMO system, which is also one primary subject for the fifth generation wireless communications, this research is significant to be developed. As a result of large scaled antennas, the channel gain matrix in Massive MIMO system is asymptotic diagonal orthogonal, and it is an non-deterministic polynomial hard problem to obtain the optimum bits error rate (BER) performance during finite polynomial complexity time. The traditional detection algorithms for MIMO system are not efficient any more due to poor BER performance or high computational complexity. The exiting detection algorithms for Massive MIMO system are able to solve this issue. However, there are still crucial problems for them, including employing the deep learning technology for detection in Massive MIMO system, and not work for the millimeter wave Massive MIMO system in the strong spatial correlation environment even exiting keyhole effect, which is not rich scattering, as well as application in Hetnets wireless communications, and etc. Therefore, the research on detection for uplink Massive MIMO system is still in its early stage, there are lots of significant and urgent issues to overcome in the future.

Lin Li, Weixiao Meng

Big Data-Based User Data Intelligent Encryption Method in Electronic Case System

When the user data of the conventional electronic case system was encrypted, there was a shortage of low analysis accuracy. To this end, an intelligent encryption method for user data of the electronic case system based on big data was proposed. Introducing the big data technology, building a framework for intelligent encryption of user data of electronic case system, and realizing the construction of intelligent encryption of user data of electronic case system; Relying on the determination of the data intelligent encryption algorithm, the electronic case system model was embedded to realize the intelligent encryption of the user data of the electronic case system. The experimental data showed that the proposed big data modeling and analysis method was 61.64% more accurate than the conventional method, which was suitable for intelligent encryption of user data in electronic case system.

Xin Liu

Research on Intelligent Retrieval Technology of User Information in Medical Information System Under the Background of Big Data

Intelligent information technology had made outstanding contributions to strengthening information management in hospitals and improving the level of hospital construction. It had become the key development object in the future of hospital intelligence information data management. As one of the most important data information in the hospital, the management of the user data of the medical information system had an extremely important influence and significance on the decision making of the future management of the hospital, the use of medical data, and the forensics of the judicial materials. The traditional manual operation and manual computer operation of the user data management model of medical information system had a direct impact on the development process of information intelligent management of the future hospital users. Therefore, the intelligent management of the important database of the medical information system user data was of great significance to the exchange and use of data experience within the hospital, and even the world, as well as the intelligent management of the hospital and the scientific and technological development of the medical future. Through the medical record management of the key information of medical system files, this paper effectively removes redundant data and improves data retrieval efficiency; it realizes accurate collection of medical information according to visualization technology. The effectiveness of the proposed method is verified by experiments.

Xin Liu

Research on Embedded Innovation and Entrepreneurship Sharing Platform for College Students Under the Internet of Things

The innovative entrepreneurship project for college students is to strengthen the training of students’ innovative entrepreneurship, enhance their awareness of innovative entrepreneurship, and cultivate innovative entrepreneurship.In order to better realize information sharing, this paper proposes an embedded innovation and entrepreneurship sharing platform for college students under the Internet of Things. With embedded system as the development environment, the design of innovation and entrepreneurship sharing platform is realized through front-end UI interface module, sharing platform module and database module. The experiment shows that the embedded innovation and entrepreneurship sharing platform designed for college students is not only higher than the traditional sharing platform in the amount of information, but also about 20% higher in accuracy than the traditional platform.

Xiao-hui Zhang, Li-wei Jia, Wei Wu

Research on Automatic Estimation Method of College Students’ Employment Rate Based on Internet Big Data Analysis

In order to solve the problem of large error and inaccuracy in employment rate estimation, an automatic employment rate estimation method based on Internet big data analysis is proposed. This method can be divided into four steps: Firstly, the data integration model based on XML middleware is used to select the sample data of employment rate estimation. Secondly, the decision tree C4.5 algorithm is used to classify the attributes of the sample data. Thirdly, the improved KPCA algorithm is used to extract the feature vectors of employment information and calculate the distance between the forecasted samples and all samples. Fourthly, non-linear mapping method is used to transform employment structure data into corner data, and grey theory is used to establish employment rate estimation model. The results show that the average employment rate estimation error of this method is 4.81% lower than that of the statistical method based on support vector machine.

Xiao-hui Zhang, Li-wei Jia, Fa-wei Zhou

Research on Demand Response Model of Electric Power Interruptible Load Based on Big Data Analysis

In order to solve the problem of low precision in the analysis of power interruptible load demand using traditional power interruptible load demand response model, a power interruptible load demand response model based on big data analysis is studied in this paper. Firstly, the demand response is implemented by enabling technology, then the dynamic peak-valley price is determined by the demand response. Finally, the power interruptible load demand response model based on big data analysis is realized by using dynamic peak-valley price. The effectiveness of the demand response model based on big data analysis is verified by experiments.

Chengliang Wang, Hong Sun, Yong-biao Yang

Research on Interruptible Scheduling Algorithm of Central Air Conditioning Load Under Big Data Analysis

The traditional algorithm is a combination of fuzzy dynamic programming and priority-based heuristic rules. The optimization performance of interruptible load scheduling is poor. For this reason, the central air conditioning load interruptible scheduling algorithm is proposed based on big data analysis. The algorithm adopts the characteristics of central air conditioning load management and selects the time scale of central air conditioning load scheduling. By optimizing the flexibility of interruptible scheduling, based on the central air conditioning load interruptible scheduling model, the optimal individual in the last generation population is decoded by binary coding, so as to realize the central air conditioning load interruptible scheduling algorithm. The experiment proves that the central air conditioning load interruptible scheduling algorithm has strong optimization performance.

Cheng-liang Wang, Yong-biao Yang

Simulation Study on Pedestrian Road Planning in Ancient Building Groups Under Cloud Computing Environment

In order to guide the traffic organization in the ancient building group in an orderly manner, the simulation study on the pedestrian road planning of the ancient building group in the cloud computing environment is proposed. Based on the analysis of the characteristics of pedestrian roads, a cloud computing environment road planning simulation model is established. Using the cloud computing environment to determine the maximum number of people on the road, use the simulation software LEGION to set the structure and parameters of the model, and carry out road planning and deduction in the simulator to realize the design of the road planning simulation model. Through the method of experimental argumentation and analysis, the effectiveness of the cloud computing environment road planning simulation model is determined, which can improve the order planning of pedestrians’ road planning.

En-mao Qiao, Da-wei Shang

Design of 3D Reconstruction Model of Complex Surfaces of Ancient Buildings Based on Big Data

Over time, ancient buildings have remained in the natural environment for a long time, and damage often occurs. Therefore, in order to protect the cultural heritage of these materials, repair work is essential. Under this background, a three-dimensional construction method of complex curved surface of ancient buildings based on big data is proposed to complete the three-dimensional modeling of ancient buildings. The model construction is mainly divided into three steps: the first step uses LiDAR to obtain the LiDAR point cloud data and image data of the complex surface of the ancient building; the second step processes the data, including the processing of LiDAR point cloud data and the processing of image data; The method combines LiDAR point cloud data with image data to automatically generate a three-dimensional model of complex curved surface of ancient buildings. The results show that compared with the other two traditional methods, the accuracy of the method is higher, and the model coordinate points are closer to the real coordinates.

Enmao Qiao, Dawei Shang

3D Human Motion Information Extraction Based on Vicon Motion Capture in Internet of Things

In order to solve the problems of unclear contour and poor geometric precision in the process of extracting human body motion information by traditional key point detection method, a 3D human motion information extraction method based on Vicon motion capture was proposed. The motion characteristics of human body were captured by Vicon motion capture technology. According to the capture results, the change characteristics of joint force and angle in the course of human motion were collected and calculated, so that the frequency of human motion wave can be accurately grasped, the distribution of frequency change and the law of change. According to the law of change, the contour features in the process of human motion were perceived and judged, and the geometric accuracy of the moving contour was optimized by using fuzzy algorithm, thus the accurate extraction of three-dimensional human motion information was realized. The simulation results show that this method can effectively extract 3D human motion information, solve the problem of unclear contour in traditional methods, and improve the geometric accuracy of 3D human motion information extraction.

Ze-guo Liu

Human Motion Attitude Tracking Method Based on Vicon Motion Capture Under Big Data

Aiming at the problem that the human body motion posture cannot be correctly and quickly marked in the conventional method, a human body motion attitude tracking method based on Vicon motion capture under big data is proposed and designed. Under the motion capture filtering algorithm, the human body weight measurement function is constructed by the combination of color, edge and motion features, and different images are selected according to the occlusion between limbs to establish a constrained human motion model, and the model is based on Vicon action. The tracking calculation of the capture realizes the tracking process of the human body motion posture. The effectiveness of the method is determined by the method of experimental argumentation analysis. The results show that the method can track the motion posture of the human body quickly and accurately, and the robustness is better. The tracking accuracy is 13.87% higher than the conventional method.

Ze-guo Liu

Simulation of Differential Expression in Root, Stem and Leaf of Ornamental Cunninghamia Lanceolata Under Internet of Things Environment

With the promotion of the use of ornamental Cunninghamia lanceolata (Cunninghamia lanceolata L.), the difference of root, stem and leaf specific expression was simulated under the Internet of things environment. In the process of the research, the model was established. The selective expression of genes was carried out and the score output was calculated. Based on the output probability setting of signal states and differential gene identification, the biological function analysis of peculiar genes in roots, stems and leaves of Cunninghamia lanceolata was realized. Through the experimental analysis, the validity of the simulation is proved effectively.

Hui Liu

Design of Statistical Model for Difference of Leaf Color of Ornamental Begonia Based on Big Data Analysis

Based on big data’s analysis, the statistical model of the difference between different leaf colors of ornamental begonia was designed in order to solve the problem of low statistical accuracy of the difference between different leaf colors of traditional ornamental begonia. Combined with quantitative classification, the variety of ornamental begonia was investigated and analyzed, and the characters of ornamental begonia were analyzed and the coding design was carried out. On this basis, the varieties of ornamental begonia were labeled with different leaf colors. The statistical model design of the difference between different leaf colors of ornamental begonia is realized. The experimental results show that the model design has good accuracy.

Hui Liu

Research on Radiation Damage Characteristics of Optical Fiber Materials Based on Data Mining and Machine Learning

In order to better analyze the damage characteristics of fiber materials under radiation environment, combined with data mining algorithm to calculate the degree of damage of material structure damage. Combine with machine learning method to analyze the calculation results, obtain the damage range of fiber material structure, standardize material damage characteristics and Grade, accurately determine the damage of material structure, and finally improve the radiation damage characteristics of fiber materials. Experiments show that the research on radiation damage characteristics of fiber materials based on data mining and machine learning is accurate and reasonable.

Ang Li, Tian-hui Wang

Analysis of Surface Compressive Strength of Optical Fiber Composites Under Low Velocity Impact Damage Based on Big Data

The traditional method for analyzing the compressive strength of surface under low-speed impact damage of fiber-optic composites uses the aperture equivalent method to calculate the Compressive strength after impact (CAI) value with large error and insufficient accuracy. Aiming at the above problems, a method for calculating the CAI value of the fiber composite under low velocity impact damage by the damage accumulation method is proposed. Firstly, the materials and related equipment used in the experiment were selected, and then the initial damage state was determined by low-speed impact test. The failure state of the fiber composite under different compression loads was analyzed by compression test. Finally, the finite element model was established and the compressive strength was analyzed. The results show that compared with the open equivalent method, the calculation error of CAI value is reduced by 6.48%, the accuracy is improved, and the purpose of accurately analyzing the compressive strength is basically achieved.

Ang Li

Research on Data Mining Algorithm for Regional Photovoltaic Generation

Traditional data mining algorithms have problems such as poor applicability, high false positive rate or high false positive rate, resulting in low security and stability of the power system. For this reason, the regional photovoltaic power generation data mining algorithm is studied. Classification of data sources facilitates correlation calculations, and matrix relationships are used to calculate data associations. Combined with the data relevance, the association rules are output, and the output results inherit the clustering processing and time series distribution of the implicit data, thereby realizing the extraction of hidden data and completing the regional photovoltaic power generation data mining. The experimental results show that the regional PV power generation data mining algorithm has high stability and can effectively solve the system security problem.

Zhen Lei, Yong-biao Yang

Distributed Learning Algorithm for Distributed PV Large-Scale Access to Power Grid Based on Machine Learning

Due to the long prediction time and the large range of data filtering, the traditional algorithm has low system operation efficiency. For this reason, distributed learning based on machine learning is widely used to predict the power grid output. First, establish a grid output prediction model to limit the system’s line loss and transformer losses. Secondly, based on the distributed photovoltaic power generation output prediction model, the vector moment method and the information method are used to narrow the search space. Based on the data concentration and fitness function values, the calculation formula of voltage output prediction of distribution network nodes with distributed photovoltaics is derived to realize the power grid output prediction algorithm. Finally, it is proved by experiments that distributed PV large-scale access to power grid output prediction algorithm can effectively improve system operation efficiency.

Zhen Lei, Yong-biao Yang, Xiao-hui Xu

A Homomorphic Encryption Algorithm for Chaotic Image Coding Data in Cloud Computing

The traditional image data encryption methods tend to ignore the classification of encryption attack types, resulting in poor security and accuracy of encryption results. In order to better guarantee the performance of chaotic image coding, a homomorphic encryption algorithm for chaotic image coding data in cloud computing environment is proposed. First, the homomorphic encryption matrix of coded data is normalized, and the homomorphic encryption parameters are calculated according to the results of the specification. According to the encryption parameters, the image encoding instructions are set, the common encryption attack types are divided, the encryption instructions are selected according to the partition results, and the accurate encryption of chaotic image coding data is finally realized. Finally, the experimental results show that the homomorphic encryption algorithm of chaotic image coding data in cloud computing environment has higher security and accuracy than the traditional encryption algorithm. It is indicated that the proposed encryption algorithm has a certain feasibility.

Bin-bin Jiang

Quantitative Evaluation Model for Information Security Risk of Wireless Communication Networks Under Big Data

Quantitative evaluation of information security risk in wireless communication network can effectively guarantee the security of communication network. In order to solve the problem that the traditional network security evaluation method is not effective, a quantitative risk assessment model of wireless communication network information security under big data is constructed. Using the wireless composition and working principle, the risk assessment system of wireless communication is built, and the index weight is determined. On this basis, the network information function interface is deployed, and the initial probability is calculated, and the quantitative risk assessment model of wireless communication network information security under big data is constructed. The experimental results show that under the condition of increasing the frequency of network attack, the security potential value of the model is always at a higher level, which indicates that the model has better performance and is helpful to detect the security of the system. It is convenient to provide accurate safety protection measures in time to resist safety risks.

Bin-bin Jiang

Intelligent Recognition Method of Short Wave Communication Transmission Signal Based on the Blind Separation Algorithm

The traditional signal recognition method can not be quickly and efficiently identified by noise interference. In order to avoid the drawbacks of traditional methods, an intelligent identification method for short-wave communication transmission signals based on blind separation algorithm is proposed. According to the mathematical model, all the transmission signals in short-wave communication are modally decomposed, and the signal can be decomposed into functions of several different feature scales, and the time and frequency are extracted as the physical quantities of the signal characteristics. The blind separation algorithm is used for signal preprocessing. The short-time energy, short-term average amplitude and short-time zero-crossing rate are used as the starting point of the recognized speech signal. Under the fixed background noise, the normal signal and the noise signal are identified. It can be seen from the experimental results that the method has short recognition time and fast rate, which lays a foundation for short-wave communication transmission.

Yan-song Hu

Research on Active Push Method of Multi-source Patrol Information in Large-Capacity Communication Network

The traditional large-capacity communication network multi-source patrol information active push method has the defect of poor push effect. For this reason, the active push method of multi-source patrol information in large-capacity communication network is proposed. The differential filtering method is used to preprocess the collected multi-source patrol information, based on the processed multi-source patrol information obtained above. The multi-source patrol information is grouped and clustered to obtain a multi-source patrol information feature set, and the obtained multi-source patrol information feature set is collaboratively filtered to obtain a user-neighbor neighbor multi-source patrol information set. The active push algorithm is used to actively push the nearest neighbor multi-source patrol information, which realizes the active push of multi-source patrol information in the large-capacity communication network. Through experiments, the proposed multi-source patrol information active push method push response time of the large-capacity communication network is 4.1 S less than the traditional method. The proposed multi-source patrol information active push method for large-capacity communication network has better push effect.

Yan-song Hu

Research on Multi-master-slave Hybrid Coordinated Control Method in VANET Environment

To improve the quality and output efficiency of multi-master-slave hybrid operation, a coordinated control technology of multi-master-slave hybrid operation based on closed-loop error regulation in VANET environment is proposed. The constraint parameter model of multi-master-slave hybrid coordinated control is constructed, and the DC voltage stabilizing capacitance, DC voltage outer loop gain, interharmonic oscillation and electromagnetic loss are taken as the constraint parameters. An input-output closed-loop control system for multi-master-slave hybrid operation of power grid is constructed, and the internal loop feedback adjustment method is used to compensate the inter-harmonic output error. The steady state regulation method of open circuit voltage and short circuit current is used to realize the adaptive identification of control constraint parameters, and the coordinated control of multi-master-slave hybrid operation of power grid is realized. The simulation results show that this method can effectively realize the multi-master-slave coordinated control in VANET environment, increase the output efficiency of the power grid, and improve the output quality and gain of the power grid control.

Chao Song

Design of Power Intelligent Control DCS Module Based on Improved PID

Distributed control system (DCS) is the core of power system control. A power intelligent control DCS module based on improved PID is studied to realize the output gain control of power electrical system and improve the control efficiency of power electrical system. Combined with integrated DSP information processing chip, a design of power intelligent control DCS module based on output power amplification and regulation is proposed. The overall model of DCS power control system is designed, and the DCS power control frequency doubling gain amplifier is constructed. The signal anti-interference design adopts cascade filter and output power amplification adjustment method to obtain the reset circuit of DCS controller. The output power amplification and adjustment algorithm are designed to equalize the gain distribution to improve the power control performance of DCS. The test results show that the output gain of intelligent power control is large, the adaptive performance is good, and the output stability is strong.

Chao Song

Identification of Wireless User Perception Based on Unsupervised Machine Learning

Wireless user perception (WiUP) plays an important role in designing next-generation wireless communications systems. Users are very sensitive with the quality of WiUP. However, the bad quality of WiUP cannot be identified with traditional methods. In this paper, we propose an intelligent identification method using unsupervised machine learning. More precisely, we create an algorithm model based on historical data to realize feature extraction and clustering. The most similar cluster to those cells with bad WiUP is identified according to Euclidean distance. The experiment is conducted on the basis of a large amount of historical data. With several contrast experiments, Simulation results show that the method proposed achieves the accuracy of identification of bad WiUP over 93%. The study manifests that unsupervised machine learning is effective in identifying bad WiUP in wireless networks.

Kaixuan Zhang, Guanghui Fan, Jun Zeng, Guan Gui

Backmatter

Weitere Informationen

Premium Partner

    Bildnachweise