Self-Learning Optimal Regulation for Discrete-Time Nonlinear Systems Under Event-Driven Formulation

• Published in: IEEE transactions on automatic control Vol. 65; no. 3; pp. 1272 - 1279
• Main Authors: Wang, Ding, Ha, Mingming, Qiao, Junfei
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.03.2020; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Adaptation models; Adaptive algorithms; Adaptive systems; Algorithms; Computer simulation; Control; Discrete time systems; Dynamic programming; Event-driven formulation; iterative adaptive critic; Iterative algorithms; Learning; Mathematical model; neural networks; Nonlinear systems; Optimal control; optimal regulation; self-learning control; Stability analysis
• ISSN: 0018-9286; 1558-2523
• DOI: 10.1109/TAC.2019.2926167

The self-learning optimal regulation for discrete-time nonlinear systems under event-driven formulation is investigated. An event-based adaptive critic algorithm is developed with convergence discussion of the iterative process. The input-to-state stability (ISS) analysis for the present nonlinear plant is established. Then, a suitable triggering condition is proved to ensure the ISS of the controlled system. An iterative dual heuristic dynamic programming (DHP) strategy is adopted to implement the event-driven framework. Simulation examples are carried out to demonstrate the applicability of the constructed method. Compared with the traditional DHP algorithm, the even-based algorithm is able to substantially reduce the updating times of the control input, while still maintaining an impressive performance.