一种机器人轨迹跟踪的迭代学习控制方法

A SORT OF ITERATIVE LEARNING CONTROL ALGORITHM FOR TRACKING OF ROBOT TRAJECTORY

针对机器人轨迹跟踪问题 ,提出了一种带遗忘因子的迭代学习控制算法。给出了学习算法收敛的充分条件 ,该算法在不改变学习控制器结构的前提下 ,对要求跟踪的新的期望轨迹 ,利用系统的历史控制经验 ,合适地选择了初始控制输入 ,使系统的输出能尽快地收敛于新的期望轨迹 ,从而达到了改善系统跟踪性能的目的。

An iterative learning control with forgetting factors for the tracking of robot trajectories was proposed, and conditions were presented to guarantee convergence of the iterative algorithm. The convergence of error can be improved without modifying the structure of the controller by utilizing previous experience in tracking different desired trajectories to select an initial control input for a new desired trajectory tracking. The output of the system can track the desired trajectory better. A simulation shows the fast convergence of the proposed algorithm.