Filtering Design Of Multi-Sensor Multi-Rate Sampling Networked Systems

Multi-sensor network systems are widely used in many fields such as target tracking,space exploration,intelligent transportation systems due to their powerful information collection and collaborative processing capabilities.Among them,networked filtering is an important issue in the analysis and design of sensor networks.Networked filtering mainly focuses on how to model filtering systems in unreliable network environments,and how to handle complex asynchronous information fusion among sensors and their neighboring nodes.Although the research on networked filtering has made some progress,there are still some problems to be solved,such as how to solve hybrid cyber attacks and the networked filtering problem of multi-sensor multi-rate sampling(MSMRS)systems under random topology switching.Based on this,this paper fully considers the networked filter design for MSMRS systems,which includes hybrid cyber attacks,dynamic event-triggered transmission mechanisms,network-induced delays,and time-varying topology switching.The specific research contents are as follows:(1)The design of event-triggered H_∞filtering for MSMRS nonlinear systems under hybrid cyber attacks is studied.The output variables in a system are sampled by multiple sensors with different sampling rates,and the network transmission channel from the sensor to the remote filter is jointly affected by denial-of-service attacks and deception attacks.Considering network resources and sensor energy constraints,a decentralized bandwidth-dependent dynamic event-triggered communication protocol is proposed to reduce data transmission.Considering the impact of different types of attacks on a channel in different instants,a hybrid cyber attacks mode with a compensation mechanism is designed.A time-varying gain filter fusing attack detection and multi-rate sampled triggered data is proposed by introducing the multi-rate sampled data update modes dependent on a discrete-time Markov switch.Based on the Lyapunov-Krasovskii stability method,sufficient conditions that guarantee the mean square asymptotic stability and H_∞performance of the concerned filtering error system are obtained,and a design method of the networked time-varying gain filter is given.Finally,a numerical example is used to demonstrate the feasibility and effectiveness of the proposed scheme.(2)The distributed filtering problem for MSMRS systems with bilateral network-induced delays under random topology switching is studied.Among them,the system output is sampled asynchronously using sensors with different rates,and a set of random variables following the Bernoulli distribution are introduced to characterize the transmission delays among the sensors and the remote filters.Considering that the information interaction among filter nodes with time-varying topology switching may be affected by network uncertainty,a local optimal distributed filter design method based on prediction estimation and measurement innovation is proposed.The filtering method has the advantage of actively compensating for the multi-rate sampling mechanisms and transmission delays.The local minimum upper bound matrix of the estimation error covariance is obtained by designing the gain matrix of the filter.Finally,the superiority of the prediction-based distributed filtering strategy is verified by a simulation example.