Distributed State Estimation for Nonlinear Uncertain Systems Under Estimate-Based Round-Robin Protocols Subject to DoS Attacks

This paper concerns distributed state estimation for nonlinear uncertain systems under bandwidth constraints and subject to denial-of-service (DoS) attacks. Different from most of the existing researches that assume unbounded sensing ranges and Gaussian white noises, this work focuses on distributed sensor networks with limited sensing ranges and colored measurement noises. A transformed system model is reconstructed to eliminate the correlation between sensor noises. To optimize bandwidth utilization, the round-robin protocols based on state estimates are proposed for each node to orchestrate the transmission sequence of the components according to a fixed circular order. A novel extended distributed state estimator is constructed by fusing partial components of estimation results from neighbors transmitted through DoS attacked channels. To address the computational infeasibility of nonlinear uncertainty, the estimator gain is obtained by minimizing the upper bound of the error covariance matrix. Furthermore, the estimation error is proved to be man-square bounded under stochastic-varying dynamic communication topology. Finally, a simulation example is presented to validate the effectiveness of the proposed algorithm.