Research On H-infinity Filtering For Cyber Physical Systems With Incomplete Information Transmission

In recent years,the development of network communication,intelligent control and intelligent sensors technology have made it possible for everything to be interconnected.Under this background,Cyber Physical Systems(CPSs)came into being.However,with the increase of system scale and sensor nodes,incomplete information transmission will inevitably occur when data packets are transmitted through the network.This phenomenon is very likely to make the system error,and in serious cases,it will lead to system fault,resulting in great economic losses.Therefore,it is of great significance to study CPSs with incomplete information transmission.There are three main factors that cause incomplete information transmission in network channels:(1)network inherent factors;(2)communication equipment defects;(3)external network attacks,and which specifically includes network delay,packet loss,communication constraints,network congestion,quantization error,coding noise,Denial of Service(DoS)attacks,eavesdropping attacks,etc.On the other hand,filtering problem for CPSs has been widely concerned by scholars all over the world.The H∞filtering does not require that the statistical characteristics of noise signal and the system model be precisely known.Therefore,H∞ filter has been widely used in automatic control and signal processing fields.Delta operator discretization method is widely used in the study of system control and filtering.Under high-speed sampling,Delta operator model overcomes the ill conditioned problem of traditional shift operator,and has great digital characteristics.Based on the delta operator method,this thesis focuses on the H∞ filtering for CPSs with the above incomplete information transmission phenomenon.The main research results are as follows:(1)High-speed sampling is realized by using Delta operator model.The H∞filtering problem for CPSs with random multi-step delay is discussed.The stochastic multi-step delay model has generality which includes both the occurrence of multi-step delay and the occurrence of one-step delay.Using Lyapunov stability theory,asymptotic stability and H∞ performance are obtained.Simulation results under different delay conditions verify the effectiveness of the filter.(2)Based on the delta operator method and the H∞ filtering problem for CPSs with quantization error and continuous random multiple packet loss is considered.The continuous random multiple packet loss model is general,which contains both single packet loss and multiple packet loss.At the same time,consider that quantization error occurs in the coding and decoding process before network transmission.In the end,two theorems guaranteeing asymptotic stability and H∞ performance are obtained.Simulation results of two packet loss cases show that the filtering method is effective.(3)The H∞ filtering problem for CPSs with random packet loss and DoS attack is studied.Firstly,a time homogeneous Markov chain model is used to describe random packet loss.The DoS attack model limits the attack time,but allows it to appear at any time under the condition of satisfying the upper bound.Then,based on the assumption of asymptotic probability,DoS attack and random packet loss are integrated into a communication failure function and the threshold of communication failure rate is obtained.Finally,Lyapunov function is used to analyze the random stability and specified H∞ performance of the system,and the effectiveness of the filtering method is explained by simulation results.