基于模糊RBF神经网络的冗余机械手运动学逆解  

A new inverse kinematics of redundant manipulator based on fuzzy RBF neural network

作　　者：罗小平[1] 韦巍[1] 

机构地区：[1]浙江大学电机工程系,浙江杭州310027

出　　处：《电光与控制》2002年第3期49-52,共4页Electronics Optics & Control

基　　金：国家自然科学基金资助项目 (6 990 4 0 0 9);教育部高等学校骨干教师资助计划

摘　　要：利用模糊 RBF神经网络对冗余机械手的运动学逆解函数逼近进行研究。为解决冗余度引起的多解问题 ,文中提出了关节最佳柔顺性附加准则 ,并将多目标多参数优化问题转化为单目标单参数优化问题 ,使网络训练数据能顺利得到。仿真所得结果令人满意 。In this paper, the function approximation of inverse kinematics for a redundant manipulator planning is investigated using the fuzzy RBF neural network. To solve the multi result problem caused by redundancy, the optimization of the multi object and multi parameter is transformed into that of single object and single parameter based on the joints' best compliance rules, so as to get the network training data successfully. The simulation result is satisfactory and shows that this method is effective.

关 键 词：模糊RBF神经网络 运动学逆解 冗余度机械手 关节柔顺 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]