Intermittent Feedback Optimal Control of Saturated-Input Nonlinear Systems via Adaptive Dynamic Programming
Yuhong Tang,Xiong Yang,Chaoxu Mu,Yongduan Song
DOI: https://doi.org/10.1109/tsmc.2024.3450274
2024-10-19
IEEE Transactions on Systems Man and Cybernetics Systems
Abstract:This article develops an intermittent feedback optimal control scheme for nonlinear systems with asymmetric input saturation using a dynamic event-triggering mechanism. First, an infinite horizon nonquadratic value function with a novel integrand is formulated for the studied system to evaluate the performance, tackle the asymmetric input saturation, and remove certain rigorous assumptions in prior related studies. Second, a critic neural network (CNN) in the adaptive dynamic programming framework is constructed to obtain the optimal event-triggered control (ETC). An improved concurrent learning technique is then developed to update the CNN's weights without requiring the persistence of excitation condition. Compared with the static ETC scheme, the present dynamic ETC strategy consumes fewer computational resources. Third, the uniform ultimate boundedness of the state, the weight estimation error, and the internal dynamic variable are assured, and the Zeno behavior is excluded. Finally, a rotational-translational actuator system is given to validate the developed intermittent feedback optimal control scheme.
automation & control systems,computer science, cybernetics