A Reputation-Based Distributed Controller Reconfiguration Algorithm for Misbehaving Agents in Networked Control Systems

In this paper, we propose a reputation-based algorithm for controller reconfiguration of misbehaving agents in Networked Control Systems (NCSs). The proposed controller is able to reject malicious data injections and maintain the system's desired performance in presence of cyber-attacks. The effectiveness of our proposition is evaluated on a leader-follower vehicular platoon in presence of DoS and Replay attacks. Numerical simulation is provided and shows fast detection and isolation of misbehaving agents and steering them to a predefined safe location. A Consensus among well-behaving agents is achieved via the application of the reputation algorithm.