基于强化学习的倒立摆起摆与平衡全过程控制

倒立摆的控制是一种典型的非线性控制问题。本文的目标是在假设不知道任何倒立摆模型的前提下,采用强化学习控制器实现倒立摆的起摆和平衡的全过程控制。为提高学习效率,采用了任务分解的方法,将整个控制任务分解为起摆和平衡两个子任务,对于不同的子任务根据其特点采用不同的强化学习算法。在Matlab/Simulink上进行仿真实验,结果证明,该方法在合理的时间内可以学习到成功的控制方法。

Swing-up and balance control of inverted pendulum based on reinforcement learning

The problem of controlling an inverted pendulum is a typical nonlinear control problem. We assume that no model of the nonlinear system is available. The aim of our work was to swing up and balance an inverted pendulum. For this purpose we use reinforcement learning controller with two sets of rules: first for swinging up the pendulum, and second for balancing the pendulum in upper position. The simulation results (in Matlab/Simulink) demonstrate that our learning system completes successfully in reasonable time exhibiting reasonable performance.;