基于神经网络的机器人位置逆解

The inverse solution for the position of a robot arm based on neural network

本文运用神经网络求解机器人运动学位置逆解。通过深入分析基于Ｋｏｈｏｎｅｎ网络原理和Ｗｉｄｒｏｗ－Ｈｏｆ误差修正的Ｍ．Ｒ．Ｓ．自组织神经网络及机器人运动学特性，创新了自组织神经网络训练算法。对ＰＵＭＡ５６０机器人的计算机仿真结果表明：本算法在自组织能力和定位控制精度方面有大幅度提高。

Utilizing the neural network this paper solves the inverse solution for the kinematics position of a robot arm.Through a deep analysis on the M.R.S.self organizing neural network based on Kohonen′s principle of network and Widrow Hoff′s revision of errors and the kinematics features of robot,a training algorithm of self organized neural network has been renewed.The result of computer simulation on PUMA 560 robot indicates that this algorithm has made a big ranged improvement on self organizing ability and controlling accuracy of locolization.