Distributed Fixed-Time Resilient Consensus Control of Nonlinear Uncertain Multiagent Systems Under Sign-Switching Deception Attacks

This article studies distributed resilient consensus for nonlinear multiagent systems (MASs) under sensor and actuator deception attacks. The agents are modeled as high-order strict-feedback dynamics with unknown parameters and external disturbances. Sensor attacks introduce time-varying uncertainties into system dynamics, while actuator attacks may change control magnitudes or even reverse the actual control direction, thereby turning a negative-feedback loop into a destabilizing positive-feedback one. Different from existing results, we consider a more realistic and challenging scenario, where control directions experience an infinite number of switches. To address these challenges, we develop a novel backstepping-based resilient controller. Time-varying uncertainties caused by sensor deception are counteracted by nonlinear damping terms. For actuator malicious sign reversals, each agent generates a direction signal and uses performance-triggered switching to identify the real-time control direction, ensuring the closed-loop system frequently recovers a negative-feedback structure. A two-phase control structure with fixed-time control and barrier Lyapunov functions ensures consensus accuracy. We prove that if the minimum switching frequency of direction signals exceeds twice the maximum switching frequency of actuator attacks, all closed-loop signals are globally bounded and consensus errors converge to an arbitrarily small residual set within a fixed time. Simulations verify the effectiveness.