A Resilient Inverse Optimal Strategy for Industrial Cyber-Physical Systems With Unknown Nonlinearities, Disturbances, and Actuator Attacks

Abstract

This paper proposes a resilient inverse optimal control strategy for industrial cyber-physical systems (ICPSs) subject to unknown nonlinearities, mismatched disturbances, and actuator attacks. To achieve optimality without solving the complex Hamilton-Jacobi-Bellman equation, an inverse optimal control framework is established. Within this framework, fuzzy logic systems and dedicated fuzzy disturbance observers are synergistically employed to online estimate the unknown system dynamics, mismatched disturbances, and the lumped effects of actuator attacks. These real-time estimates are then integrated into a backstepping-based controller design to ensure both stability and performance. The closed-loop system's resilience and optimality are rigorously guaranteed via Lyapunov stability theory. The effectiveness of the proposed method is validated through simulations on an ICPS model and an active vibration isolation platform.