中国机械工程

• 智能制造 • 上一篇    下一篇

基于BP神经网络的冗余机械臂逆运动学分析

刘世平;曹俊峰;孙涛;胡江波;付艳;张帅;李世其   

  1. 华中科技大学机械科学与工程学院,武汉,430074
  • 出版日期:2019-12-25 发布日期:2019-12-27
  • 基金资助:
    国家重点研发计划资助项目 (2018YFB1306905);
    国家自然科学基金资助项目(71771098,61772481)

Inverse Kinematics Analysis of Redundant Manipulators Based on BP Neural Network

LIU Shiping;CAO Junfeng;SUN Tao;HU Jiangbo;FU Yan;ZHANG Shuai;LI Shiqi   

  1. School of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wuhan, 430074
  • Online:2019-12-25 Published:2019-12-27

摘要: 常见的七自由度冗余机械臂逆运动学解法较为繁杂,且对不同构型机械臂的逆运动学解法通用性较差。为了找到一种通用的七自由度机械臂逆运动学求解方法,建立了神经网络模型,选择了合适的激励函数、隐藏层神经元数量、神经网络层数和学习速率等模型参数。设计轨迹跟踪实验对神经网络模型进行验证,实验数据证明该方法有效且精度较高,是一种可行的冗余机械臂逆运动学求解方法。

关键词: 神经网络算法, 七自由度, 冗余机械臂, 逆向运动学, 精密度

Abstract: General 7 DOF redundant manipulators inverse kinematics solution was complicated, and the inverse kinematics of manipulators with different configurations were less versatile. In order to find a general 7 DOF manipulator inverse kinematics solution, a neural network model was established. Model parameters such as appropriate excitation function, numbers of hidden layer neurons, numbers of neural network layers, and learning rate were selected. Trajectory tracking experiment was designed to verify the neural network model. Experimental data show that the method has efficiency and high precision, which is a feasible method for solving the inverse kinematics of redundant manipulators.

Key words: neural network algorithm;7 degree-of-freedom(DOF); redundant manipulator, inverse kinematics;precision

中图分类号: