Optimization Interval Observer for Nonlinear Networked Systems Subject to Stealthy Attacks

An optimization interval observer is designed for the nonlinear networked systems subjected to stealthy attacks in this paper. The stealthy attacks are difficult to detect and will damage the performance of the system and affect the accuracy of state estimation. In response to this issue, an improved reachable set analysis configuration is built for an interval observer to obtain the boundaries of the stealthy attack signal. Based on the combination of the ellipsoid method, trajectory criterion, and attack signal estimation interval, an optimized interval observer is designed to enhance estimated accuracy. The observer gains are solved in terms of linear matrix inequalities. Furthermore, to compare the estimated accuracy of the traditional interval observer by using the positive system method (PSM) in theoretically, it is shown that the presented method has higher estimation accuracy. Finally, the superiority of the presented method is confirmed with a numerical simulation.