基于策略迭代的滑动自平衡偏置倒立摆自适应最优控制

Adaptive Optimal Control of Sliding Self⁃Balancing Biased Inverted Pendulum Based on Policy Iteration

针对具有未知参数的滑动自平衡偏置倒立摆系统,提出一种自适应控制方案,能够在线求解系统最优反馈增益,而不需要了解系统内部动态模型。基于策略迭代技术,所提出的算法在策略评估和策略更新步骤之间交替进行,直到控制策略的更新不再改善系统性能为止。因此,该方法提供了一种自适应最优控制算法,能通过收集系统状态数据,求得最优控制策略。通过在滑动自平衡偏置倒立摆系统上仿真验证和进行实物实验证明所提出算法的有效性。

An adaptive control scheme is introduced for a sliding self-balancing biased inverted pendulum system with unknown parameters.This scheme is designed to determine the optimal feedback gain of the system online without requiring a comprehensive understanding of the internal dynamic model.Leveraging the policy iteration technique,the algorithm alternates between policy evaluation and policy update steps until the control policy update no longer enhances system performance.Consequently,this method offers an adaptive optimal control algorithm that can derive the optimal control solution through the collection of system state data.The efficacy of this approach is demonstrated through simulation on a sliding self-balancing biased inverted pendulum system.