Scholarlink

<p>In this paper, we address a cost function reconstruction problem for optimal tracking of AC/DC converter system. For a given input voltage in the converter systems, the goal is to find state weight matrix in the cost function so that the control input is optimal in perspective of energy consumption. First, by augmenting the system state variables and the desired tracking dynamic variables, the optimal tracking problem with a discount cost function is transformed into an optimal regulator problem. Inverse optimal tracking is, thus, simplified as inverse optimal control of the augmented system. Second, an inverse reinforcement learning (RL) algorithm is proposed to compute state weight matrix in optimal augmented control system. The convergence of the algorithm is rigorously analyzed, and stability of the closed-loop system has been confirmed. Finally, the effectiveness of the proposed algorithm is verified by simulation, demonstrating that output voltage accurately tracks a reference signal in converter system.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Inverse Optimal Tracking Control for AC/DC Converter Based on Inverse Reinforcement Learning

  • 期刊论文

  • 发表日期：

    2026年2月9日

查看全文
  • Jingling Zhao,
  • Dengguo Xu,
  • Xiang Zhang,
  • Xinsuo Li

摘要

In this paper, we address a cost function reconstruction problem for optimal tracking of AC/DC converter system. For a given input voltage in the converter systems, the goal is to find state weight matrix in the cost function so that the control input is optimal in perspective of energy consumption. First, by augmenting the system state variables and the desired tracking dynamic variables, the optimal tracking problem with a discount cost function is transformed into an optimal regulator problem. Inverse optimal tracking is, thus, simplified as inverse optimal control of the augmented system. Second, an inverse reinforcement learning (RL) algorithm is proposed to compute state weight matrix in optimal augmented control system. The convergence of the algorithm is rigorously analyzed, and stability of the closed-loop system has been confirmed. Finally, the effectiveness of the proposed algorithm is verified by simulation, demonstrating that output voltage accurately tracks a reference signal in converter system.