Distributed Filtering For Nonlinear Systems Based On Communication Protocol Under Measurement Fading

Sensor networks are widely used in both civil and military fields due to their advantages of low cost,easy maintenance and high reliability.It should be noted that the transmitted data easily encounters the measurement fading phenomenon due to the disturbance of the instable communication channels,such that the system performance is degraded.On the other hand,if a large amount of sensor nodes transmit data simultaneously,not only will occupy too much network resources but also will reduce the transmission efficiency and filtering performance.Therefore,it is of great theoretical significance and practical application value to consider the distributed filtering problem for nonlinear systems with fading measurements based on commu-nication protocols.This thesis aims to investigate the protocol-based distributed filtering issues for the non-linear systems with fading measurements,and some influence factors in practical engineering applications are taken into account,e.g.,the network-induced phenomenon,the quantization effect,the cyber-attack,the switched topology,the saturation behavior and the nonlinearity per-turbation.The first section mainly introduces the distributed filtering based on the sensor networks,the network-induced phenomenon,the quantization influence,the cyber-attack,the switched topology,the communication protocol,the saturation behavior and the relevant research status.In the second section,the event-based distributed filtering issue is considered for a class of nonlinear systems with fading measurements and quantization effects.First,a measurement model is presented under the influence of the measurement fading,the quantization and the sensor saturation.Then,the event-triggered filter is constructed for each sensor node.Moreover,based on the Lyapunov method and stochastic analysis theory,the mean-square exponential stability and the H_?performance index are analyzed for the filtering error system.Finally,the filter parameter matrix can be calculated by a sequence of linear matrix inequalities.The third section investigates an event-based distributed secure filtering issue for the non-linear systems under the state saturation,the time-varying measurement fading and cyber-attacks.First,The switched measurement fading probability is depicted by a homogeneous Markovian chain,and a Bernoulli random variable is used to describe the cyber attack behav-ior.Second,an event-based distributed secure filter is constructed.Then with the aid of the Lyapunov method,an condition is presented to guarantee the prescribed H_?performance of the filtering error system.Finally,the filter parameter matrices can be determined by solving convex optimization problems.Based on Round-Robin protocol,the distributed dissipative filtering issue is studied in the fourth section for the systems with random nonlinearities,Rice fading measurements and the switching topology.The measurement data is characterized by a Rice fading model,and a se-quence of signals with right continuity is used to depict the topology switching rule.A distribut-ed filter is constructed by Rice fading and Round-Robin protocol.On the basis of Lyapunov method and the average dwell-time technique,a sufficient condition in terms of linear matrix inequalities is presented to ensure the strict dissipative performance for the augmented system.Finally,the distributed dissipative filter can be designed through the matrix transformation.The summary and future research directions are presented in the fifth section.