Distributed resilient filtering for nonlinear complex networks with quantification of sensor

This paper studies the distributed resilient filtering problem for nonlinear complex networks with quantification of sensor. A sensor scheduling scheme integrated with the quantization technique is proposed for the cost reduction in communication over the whole network, in which each sensor is capable of sending its measurement to the remote node by an alternative multi-channel communication. Furthermore, a resilient filter design is carried out to relieve the adverse effect caused by the stochastic gain perturbation occurring in filter implementation. In the case of considering multi-channel sensor scheduling and gain disturbance simultaneously, a distributed resilient filter is designed by exploiting the variance-constrained approach, in which the expected filter gain is derived by minimizing the generated upper bound of the error covariance. A sufficient condition is provided to ensure that the upper bound of the estimation error covariance in each node is bounded. An illustrative example involving multi-target tracking is exploited to demonstrate the practical performance of the designed distributed filtering algorithm.