Resilient Distributed Filtering for Multitarget Systems With Coupled Measurements Under Multichannel Deception Attacks

This article investigates the problem of secure state estimation for multitarget tracking systems based on Kalman consensus filtering. In the existing distributed Kalman filters, the filter gain and consensus structure rely on the independence of tracked targets, which cannot maintain the estimation performance when encountering coupled measurements across multiple targets. Moreover, the existing researches mainly focus on the security in single-channel systems, whereas such efforts fail to consider potential attacks in multichannel scenarios. In this case, by establishing a target-dependent augmented system and a link-unreliable composite directed graph, the coupling features and multichannel attacks are depicted. Then, a modified Kalman consensus filter is proposed by specifically designing consensus structure and gain terms to account for the impacts of coupled measurements and attacks. Furthermore, by scaling the Lyapunov function through the Riccati difference equation and matrix inequalities, sufficient conditions are established to ensure the boundedness of estimation errors. Numerical simulations are conducted to demonstrate the effectiveness of the filter.