中国机械工程 ›› 2015, Vol. 26 ›› Issue (12): 1652-1657.

• 科学基金 • 上一篇    下一篇

一种五轴并联机床的机构参数分步辨识方法

党鹏飞;房立金   

  1. 东北大学,沈阳,110819
  • 出版日期:2015-06-25 发布日期:2015-06-29

A Step Identification Method of Mechanisim Parameters of 5-DOF Parallel Machine Tool

Dang Pengfei;Fang Lijin   

  1. Northeastern University,Shenyang,110819
  • Online:2015-06-25 Published:2015-06-29

摘要:

以五轴并联机床为研究对象,基于量子粒子群优化算法,对少自由度并联机床的机构参数辨识问题进行了研究。根据五轴并联机床的结构特点,对运动末端的测量位姿进行优化选取。将并联机构参数辨识问题转化为非线性系统的最优化问题,利用量子粒子群优化算法的全局搜索能力设计一种分步辨识方法对机构参数进行优化、辨识。仿真结果显示,基于量子粒子群优化算法的分步辨识方法能够比较准确地辨识机构参数的真实值。该分步辨识方法同样适用于其他少自由度并联机器人的机构参数辨识。

关键词: 并联机器人, 位姿误差, 参数辨识, 量子粒子群优化

Abstract:

A step identification method of mechanisim parameters was proposed based on QPSO algorithm to improve the accuracy of 5-DOF parallel machine tool. Firstly, according to structural characteristics of 5-DOF parallel machine tool, the measurement configurations were  selected optimally. Then, the identification problem of mechanisim parameters was regarded as a nonlinear optimization problem, and solved through the two-step identification. The simulation results illustrate that the actual values of geometric errors of parallel robot can be identified accurately through the step identification method based on QPSO. Furthermore, the step identification method of mechanisim parameters is feasible for other limited-DOF parallel robots.

Key words: parallel robot, pose error, parameter identification, quantum-behaved particle swarm optimization(QPSO)

中图分类号: