基于人工神经网络的操作手IKP求解  

Solution to the Inverse Kinematic Problem in Robotics by Neural Networks

作　　者：陈国锋[1] 杨昂岳[1] 傅祥志[1] 廖道训[1] 

机构地区：[1]国防科技大学,华中理工大学

出　　处：《国防科技大学学报》1995年第2期48-52,共5页Journal of National University of Defense Technology

摘　　要：人工神经网络以其信息的并行处理与分布式存储、自组织和自适应等特点，在机器人领域显示了极大的应用潜力。本文探讨人工神经网络在机器人操作手ＩＫＰ求解中的应用，试图求得操作手ＩＫＰ的一组或所有可行解。文章利用ＢＰ网络对操作手ＩＫＰ的求解进行了讨论和分析，并给出了具体的求解实例。Neural Networks have been paid more and more attention in robotics owing toits excellent abilities of parallel distributed processing, self-organizing and self-learning.This paper inquire into and analyzes the potential application of BP networks to solve robot manipulator inverse kinematic problem with multi-solutions. The illustrative example shows that the neural network can be trained to generate a fairly accurate solution without costly derivations and programming of the inverse kinematic algorithm.

关 键 词：机器人 操作手 运动学逆问题 神经网络 

分 类 号：TP241[自动化与计算机技术—检测技术与自动化装置]