Observer Based Event-triggered Fault Compensation Control for Nonlinear Systems via Adaptive Dynamic Programming

This paper develops an observer based event-triggered fault compensation control method for a class of nonlinear continuous-time systems with actuator failures via adaptive dynamic programming (ADP). The proposed control method incorporates an event-triggered optimal control policy and an adaptive observer-based compensator. Due to the excellent estimation performance of neural networks, the actuator failure is precisely estimated by the online learning observer. By combining the event-triggered mechanism with ADP, the eventtriggered optimal control policy is obtained by adopting the critic neural network to solve Hamilton-Jacobi-Bellman equation. The triggering condition is provided along with the stability analysis of the close-loop system. Finally, the simulation of single link robot arm system is presented to confirm the efficiency of the proposed fault compensation control method.