Seeking Secure Adaptive Distributed Discrete-Time Observer for Networked Agent Systems Under External Cyber Attacks

Abstract

In the era of Industry 5.0, the cooperation among components in Cyber-Physical Systems (CPS) plays a crucial role in achieving digital transformation and intelligent development. Distributed observers can simplify the design complexity of cooperative control schemes in intelligent networked industrial systems. However, the open nature of the information exchange network in these systems makes them vulnerable to cyber attacks. With this in mind, this paper designs a discrete-time secure adaptive distributed observer for networked agent systems under Denial of Service (DoS) attacks. The observer enables agents to monitor reference state information effectively, even when the system matrix is globally unknown and unstable. To enhance security, we propose adaptive secure update rules that respond to compromised communication links, ensuring the ability of the observer to generate reference trajectories during DoS attacks. Furthermore, sufficient algebraic conditions are obtained for the convergence of the proposed observer, employing the analysis approach based on infinite products of sub-stochastic matrices. Finally, some numerical simulation results are provided to verify the effectiveness of the proposed secure adaptive distributed observer.