Adaptive output feedback control for cyber-physical systems with unknown nonlinearities and deception attacks

This paper investigates the adaptive output feedback resilient control problem for a class of nonlinear cyber-physical systems (CPSs) subject to deception attacks. Here, the CPSs are characterized by unknown nonlinearities that can vary over time, and the attacks can target both the measurement and control communication channels, as well as disrupt the plant dynamics. Utilizing the only measured yet possibly corrupted system output for feedback, we propose a new adaptive resilient control design scheme by integrating a K-filter method, a barrier function technique, and a dynamic gain-based adaptive backstepping framework. In contrast to the existing methods, our proposed approach neither requires prior knowledge of the structures of the nonlinearities or their bounding functions, nor uses structure approximations to acquire such information. Through a novel analysis strategy, we show that all signals in the closed-loop system remain globally bounded and the system output asymptotically converges to zero. Finally, simulation studies are provided to demonstrate the effectiveness of the proposed approach.