基于RBF神经网络的冗余度机器人轨迹规划  

The Path Planning of Redundant Robot Based on BF Neural Network

作　　者：李景春[1] 齐瑞贵[2] 张国忠[1] 

机构地区：[1]东北大学,辽宁沈阳110004 [2]沈阳航空工业学院,辽宁沈阳110034

出　　处：《机械与电子》2004年第9期53-55,共3页Machinery & Electronics

摘　　要：利用D-H方法完整地推导出了机器人的运动学方程,并将由此得到的运动学正解作为训练样本,对利用RBF神经网络方法建立的机器人运动学神经网络模型进行训练,实现了机器人从工作空间到关节空间的非线性映射,求解了机器人的逆解,并对3R,4R,5R和6R机器人的运动轨迹规划进行了计算机仿真,验证了该方法的可行性。In this paper, a kinematics equation has been deduced, Based on Denavit-Hartenberg method. The kinematics solution getting from it being training sample, and training using RBF to realize non-liner image from working space to joints space and the inverse kinematics of robot have been solved. the planning of the motion path simulation results for the 3R,4R,5R and 6R robot validate it's effectiveness and feasibility.

关 键 词：冗余机器人 RBF 仿真 

分 类 号：TH112[机械工程—机械设计及理论] TP391[自动化与计算机技术—计算机应用技术]