Distributed H_? Filtering For Networked Systems With Multi-rate Sampling And Event-triggered Mechanism

With the wide application of wireless sensor network(WSN)in smart power grid and intelligent transportation system,the based on WSN distributed filtering problem has attracted wide attention in the field of theory and practice.However,in the distributed fil-tering systems,there are many sensors jointly monitoring the target output and sharing data among sensor nodes with different functions(for example,clustering protocol is used to par-tition WSN so that cluster head can receive the sampling information of intra-cluster nodes as estimator and share the estimation information through inter-cluster topology).Therefore how to deal with the complex coupling relationship of multiple execution intervals caused by multi-sensor sampled-data sharing and to design the effective structure of distributed estimation systems are two important problems in the design of distributed filter.For multi-sensor networked systems with adaptive event-triggered,multi-rate sampling and topology switching,the following distributed filtering design problems are studied:(1)The distributed H_? filtering problem for nonlinear systems with distributed adap-tive event-triggered transmission mechanism is studied.The transmission of sampled state among filters based on distributed adaptive event-triggered transmission mechanism depend-s not only on the error between current sampled-data and its latest released data,but also on the error of the latest release sampled-data of itself and the neighbor nodes.Compared with the event-triggered mechanism with the fixed parameters,the adaptive parameters which de-pend on the network condition and the dynamic performance of the system can be changed dynamically,and the communication resources can be saved by adjusting the threshold pa-rameters appropriately.Based on lyapunov-krasovskii functional and distributed adaptive event-triggered transmission mechanism,the sufficient conditions of distributed H? con-sensus filtering performance and the design method of filter parameters are given,respec-tively.(2)The distributed H_? filtering problem for discrete-time systems with topology switch-ing under adaptive event-triggered transmission mechanism is studied.Firstly,considering the uncertainty of the actual network communication,the Markov chain is used to model the topology mode switching process between filters.Secondly,a new distributed adaptive event-triggered transmission mechanism is proposed to reduce computational burden and save limited communication resources.Based on the adaptive event-triggered mechanism,the adaptive parameters depend on the network state among filters and the system dynamic performance.Finally,it is proved by the Lyapunov-Krasovskii stability theory that the de-signed filter can ensure the stability of the filtering error system and the sufficient conditions of the filter parameter design method are given.(3)This paper addresses the distributed H? filtering problem for a class of continu-ous systems over the bilateral sensor networks.Taking into account the multi-rate sampling scheme of multiple physical sensors,a group of matching mechanisms are firstly designed to synchronize the sampled-data successfully transmitted by each transmission channel and the corresponding sampled-data of each local sub-filter,and the framework of the bilateral networked distributed filtering system is constructed based on the multi-rate sampled-data.The considered bilateral networks include two types of data transmission channels,namely,from the system to filters and from one filter to another filter.By integrating the characteris-tics of multi-rate sampling,multi-channel-induced delays and packet losses in the first type of networks,and the topology switching in the second type of networks,the multi-sensor networked distributed consensus filtering system is modeled as a switched delay system.By using a Lyapunov-Krasovskii functional,the design method of the the distributed H? filter parameters are given.Finally,the research results of this paper are summarized and the directions of future research are given.