基于人工智能模型的自动容错跟踪控制研究

Research on Automatic Fault-Tolerant Tracking Control Based on Artificial Intelligence Model

为了提高多轴运动控制系统的容错性,提出了基于强化学习算法的自动容错跟踪控制方法.该方法适用于工业生产等多代理场景中的故障跟踪控制.设计了新颖的分布式中间估计器,通过修改设计结构来增强估计方法的可行性.设计基于在线强化学习的估计策略来提高估计的性能.实验评估结果表明,提出的方法能实现良好的容错跟踪控制性能.

In order to improve the fault tolerance of multi-axis motion control system,an automatic fault-tolerant tracking control method based on reinforcement learning algorithm is proposed.This method is suitable for fault tracking control in multi-agent scenarios such as industrial production.Firstly,a novel distributed intermediate estimator is designed to enhance the feasibility of the estimation method by modifying the design structure.Then,an estimation strategy based on online reinforcement learning is designed to improve the estimation performance.The experimental evaluation results show that the proposed method can achieve good fault-tolerant tracking control performance.