Reliability Analysis And Fault Diagnosis Of Wireless Communication System Using Dynamic Bayesian Network

With the rapid development of wireless communication technology,industrial manufacturing is getting more intelligent and automated.As informatization progresses in each field(military,medical,energy,and transportation),operation form in the proposed field is upgraded with cyberspace to meet the development needs of informatization and intelligence.As a key technology in cyberspace,wireless communication is applied to the industrial upgrading of informatization,therefore,the wireless communication system has been upgraded at the same time.However,the complexity and uncertainty of this system is greatly increased by its intelligent development.Moreover,this system have to face many reliability problems such as frequent failures,performance instabilities and difficulties in pinpointing out weak components.Therefore,considering the actual physical characteristics and unique operation forms of this wireless communication system,reliability analysis and fault diagnosis for this system is the crucial and essential work for military command,medical diagnosis,energy production and transportation during the information age.As the basis of cyberspace,wireless communication system faces many practical problems and technical bottlenecks in system reliability analysis and fault diagnosis,including: The communication scenarios are diverse and it is difficult to build a reliability analysis model for the communication mission under specific scenarios;The equipment has many functions and each functional failure is dynamically related to the failure of corresponding components;The network structure is complex and there is a state combinational explosion in the existing reliability modeling methods for the network;There are many nodes in this system and the traditional method is insufficient in calculation capability to diagnose all the sub-nodes.Therefore,it is difficult to apply the existing analysis theories and diagnosis techniques to the reliability analysis and fault diagnosis of wireless communication systems,and it is urgent to make theoretical innovation and technical breakthrough in this field.In this dissertation,the reliability analysis and fault diagnosis of wireless communication system is carried out considering its different performance indexes and technical requirements,and the research contributions are summarized as follows.(1)A radio communication reliability analysis method based on the propagation loss model is proposed.Considering the influence of environmental factors on the communication reliability under complex scenarios,the propagation loss model is introduced for the reliability analysis of radio communication.Moreover,the existing propagation loss model of very high frequency(VHF)and ultra high frequency(UHF)is introduced,and the uncertainty of communication frequency,equipment parameters,communication distance and other variables have been analyzed.Additionally,the distribution of the communication distance is fitted with the lognormal distribution,and the distribution function of the path loss values under different communication scenarios have been obtained.Finally,the communication reliability of different frequencies under various communication scenario can be calculated through the path loss threshold and the distribution functions.(2)A multi-level system reliability analysis method based on discrete-time Bayesian network(DTBN)is proposed.Considering the failure time of each node,a modeling method of probability table is proposed for the dynamic logic gate in the dynamic Bayesian network(DBN),and the validation of the probability table is implemented by Monte Carlo simulation(MCS).Moreover,considering the engineering background and structural characteristics of the multi-level system,the corresponding DTBN structure of the multi-level system is established.Finally,the DBN for a VHF radio station is built based on its function and structure,and the validation of the modeling techniques is implemented by the quantitatively reliability analysis of this station.(3)Considering the hybrid failure time distributions of different events in wireless communication networks,this dissertation proposes a reliability modeling and analysis method based on hybrid dynamic Bayesian network(HDBN).Moreover,a binary probability table modeling method is proposed to solve the problems of complex modeling and state combinatorial explosion of traditional DBN algorithms,and the validation of the proposed techniques is implemented by comparing with DTBN,MCS and Continuous-Time Bayesian Network(CTBN)algorithms.Additionally,the HDBN modeling method is proposed for various tasks(2-terminal,k-terminal,and k n network reliability)in wireless communication networks.Finally,a reliability analysis case of a military wireless communication network is provided.(4)A hierarchical fault diagnosis method for wireless communication system based on dynamic Bayesian inference is proposed.With the contrastive analysis of the existing DBN inference algorithms,the most suitable algorithm is selected to calculate the posterior probability of multiple sub-events simultaneously.Moreover,considering the dynamic characteristics of failures,a dynamic diagnosis algorithm is proposed to obtain the curve of posterior probability based on the selected inference algorithm.Additionally,the task time is divided into multiple stages according to the diagnostic results,and the most weak node can been find out in each stage.Finally,a hierarchical diagnosis algorithm is proposed which can solve the insufficient computing problems of the DBN algorithm,and the proposed method is implemented to find the weak components of the wireless communication system in different stage.