作者：王一, 宋志伟, 王祎泽, 张湧涛

作者单位：华北理工大学电气工程学院, 河北 唐山 063009

关键词：空间网格精度;粒子群算法;机器人标定;神经网络

摘要：

该文提出一种将机器人的位置和姿态拆分开，分别进行标定的机器人位姿标定方法。采用空间精度控制网格标定机器人定位误差，粒子群优化算法（particle swarm optimization，PSO）优化神经网络标定机器人定姿误差。该方法以指数积公式 （product of exponentials，POE）为基础建立机器人正向运动学模型，用映射法建立空间网格，用三坐标测量臂测量机器人位姿，用空间网格精度标定定位误差，用PSO优化的神经网络标定定姿误差。其优点在于既标定机器人工具中心点（TCP）的定位误差，又标定机器人工具坐标系的姿态误差，使得机器人定位、定姿误差都得到补偿。实验结果表明机器人的定位、定姿均方根误差减小接近一个数量级。

Robot calibration method based on spatial mesh and PSO optimal neural network

WANG Yi, SONG Zhiwei, WANG Yize, ZHANG Yongtao

College of Electrical Engineering, North China University of Science and Technology, Tangshan 063009, China

Abstract: A robot's posture calibration method which separate the rotation and location of robot to separate calibrate is introduced. Uses spatial mesh precision to calibrate the robot's position error, and applys PSO neural network to calibrate the robot's rotation error. This paper use POE formula to establish the forward kinematics model of robot. A coordinate measuring machine has been used to measure the robot working space. Apply reflection method to set up spatial mesh. Position error of the robot has been calibrated by spatial mesh precision and the rotation error has been calibrated by neural network. The robot calibrate method in this paper is different from the generals. Advantage of this method is that it not only calibrates the position precision of the robot's TCP but also improves the rotation precision of the robot's TCP coordinate at the same time. This approach can further improve the robot's TCP precision. The results of the experiment illustrate: the precision of robot's position and rotation has been improved approximate a magnitude.

Keywords: spatial mesh precision;PSO;robot calibration;neural network

2016, 42(8): 98-102  收稿日期: 2015-10-14;收到修改稿日期: 2015-11-17

基金项目: 国家自然科学基金项目（E051102）；河北省自然科学基金面上项目（E2013209266）；河北省高等学校科学技术研究项目（QN2013114）

作者简介: 王一(1981-),男,河北唐山市人,副教授,博士,研究方向为光电检测。

参考文献

[1] MENG Y, ZHUANG H Q. Autonomous robot calibration using vision technology[J]. Robotics & Computer-Integrated Manufacturing,2007,23(4):436-46.
[2] HE R B, ZHAO Y J, YANG S N. Kinematic-parameter identification for serial-robot calibration based on POE formula[J]. IEEE Transactions on Robotics,2010,26(3):411-423.
[3] LOU Y J, CHEN T N, WU Y Q, et al. Improved and modified geometric formulation of POE based kinematic calibration of serial robots[C]∥IEEE/RSJ International Conference on Intelligent Robots and Systems. New Jersey: IEEE,2009:5261-5266.
[4] SELING J M. Geometric fundamentals of robotics[M]. 2nd ed. New York: Springer-Verlag,2004:72.
[5] 韩翔宇,都东,陈强,等. 基于运动学分析的工业机器人轨迹精度测量的研究[J]. 机器人,2002,24(1):12.
[6] ROTH Z S, MOORING B W, BAHRAM R. An overview of robot calibration[J]. IEEE Journal of Robotics and Automation,1987,23(5):377-385.
[7] OKAMURA K, PARK F C. Kinematic calibration using the product of exponentials formula[J]. Robotica,1996,14(4):415-421.
[8] PARK F C, OKAMURA K. Kinematic calibration and the product of exponential formula[M]∥Advance in Robot Kinematics and Computational Geometry. Cambridge:MIT Press,1994:119-128.
[9] PADEN B. Kinematics and control robot manipulators[D]. Berkeley:University of California Berkeley,1986.
[10] HUNT K H. Kinematic geometry of mechanisms[M]. New York:Oxford University Press,1978:133-140.
[11] 钱东海,王新峰,赵伟,等. 基于旋量理论和Paden-Kahan子问题的6自由度机器人逆解算法[J]. 机械工程学报,2009,45(90):72-76.
[12] 徐元昌. 工业机器人[M]. 北京:中国轻工业出版社,1999:80-90.
[13] HE R B, ZHAO Y J, YANG S N, et al. Kinematic-parameter identification for serial-robot calibration based on POE formula[J]. IEEE Transactions on Robotics,2010, 26(3):411-23.
[14] 涂娟娟. PSO优化神经网络算法的研究及应用[D]. 镇江:江苏大学,2013.