自动化学报2025，Vol.51Issue(10)：2293-2301,9.DOI:10.16383/j.aas.c250101
带有输入时滞的非线性系统基于学习的输出反馈控制
Learning-based Output-feedback Control for Nonlinear Systems With Input Time-delay
摘要
Abstract
This paper proposes a new data-driven output-feedback control method to address the direct adaptive optimal control problem for nonlinear systems with input time-delay.The combination of Q-learning with value it-eration(VI)and policy iteration(PI)enables the learning process to be conducted without any knowledge of the system dynamics.Under the condition that the system is uniformly observable,we propose a novel method to recon-struct the state of the system based on output data and input data with time-delay.We then present two iterative methods,VI and PI,to learn the adaptive optimal control policy.Finally,the proposed methods are applied to the classical nonlinear system control——Van der Pol oscillator.The simulation results demonstrate the effectiveness of the proposed methods.关键词
最优控制/输出反馈/时滞/自适应动态规划Key words
Optimal control/output-feedback/time-delay/adaptive dynamic programming引用本文复制引用
刘思彤,高伟男,姜钟平..带有输入时滞的非线性系统基于学习的输出反馈控制[J].自动化学报,2025,51(10):2293-2301,9.基金项目
国家重点研发计划(2024YFA1012702),国家自然科学基金(62373090,62521001),辽宁省兴辽英才计划(XLYC2403177)资助Supported by National Key Research and Development Pro-gram of China(2024YFA1012702),National Natural Science Foundation of China(62373090,62521001),and Liaoning Re-vitalization Talents Program(XLYC2403177) （2024YFA1012702）
