Distributed filtering for complex networks under multiple event-triggered transmissions within node-wise communications

This paper focuses on the distributed filtering and fault estimation problems for a class of complex networks, where the communications between filters at different nodes are subject to dynamic event-triggered (DET) transmissions. A filter is constructed at each node by resorting to local measurements and information from neighboring nodes and thus the developed algorithm can be carried out distributedly. Different from the clock-driven signal transmissions in traditional distributed filtering schemes, the transmissions of both state estimates and the upper bounds of filtering error covariances (FECs) between the nodes are monitored by a multiple DET strategy to reduce unnecessary burdens in the links. Under DET transmissions, an upper bound of the FEC is obtained and then minimized via parameterizing the filter recursively. Novel sufficient conditions, which are dependent on locally available information, are provided to guarantee the uniform boundedness of the FEC at each node. The proposed method is used to solve the fault estimation problem in complex networks, where the estimation error is ensured to be exponentially bounded. Some illustrative examples are employed to show the effectiveness of our algorithm.