进化计算在机器人轨迹规划中的应用研究

The Researches of Application Techniques on Robot Trajectory Planning with Evolutionary Computation

本文介绍了采用进化计算思想研制的机器人轨迹规划系统ＲＴＰ－１，提出多层次动态结构化编码方案，利用低中高三层进化算法分别优化距离、路径和关节角度偏差，基于多层综合优化策略解决多目标多约束工程优化问题，建立了次序相关问题求解的通用框架。在进化算法中，利用拉马克效应加快轨迹规划速度，在ＲＭ－５０１机械手上实现的任意空间直线和空间曲线轨迹规划具有良好的鲁棒性，规划轨迹的相邻臂构型间具有良好的柔顺性，规划轨迹的臂构型序列具有良好的平滑性。本文所采用的机器人轨迹规划方法具有通用性，可推广应用于各种动力学系统的研制。

This paper researches into actual robot trajectory planning system (RTP 1) helped with the idea of evolutionary computation,and provides a kind of dynamic coding scheme with multi level stucture,optimizes the distance,path and the displacement of link angles respectively by low,middle,and high level evlutionary algorithms,and solves the problem of engineer optimization with multi goal and multi restrain based on the multi level optimization strategy,and establishes a currency scheme of sequence correlated problem solving.It accelerates the process of robot trajectory planning by take advantage of Lamarck effect.This paper has realized the reasonable trajectory planning with arbitrary space lines or curves on the robot of PM 501.The realized process of trajectory planning has a fine stability,the adjacent arm configurations has a fine flexibility and the sequence of arm configurations has a fine flatness.The method of robot trajectory planning used in this paper can be generalized to apply in different kinds of dynamic systems.