基于双目视觉动态跟踪的机器人标定  

Robot calibration based on binocular vision dynamic tracking

作　　者：应再恩 李正洋 平雪良 蒋毅 刘开明 

机构地区：[1]江南大学机械工程学院,江苏无锡214122

出　　处：《计算机应用研究》2014年第5期1424-1427,共4页Application Research of Computers

基　　金：江苏省科技支撑计划资助项目(BE2011098)

摘　　要：采用双目视觉动态跟踪技术对自主研发的工业机器人进行运动学标定。区别于以往机器人运动学标定中复杂模型计算,在此利用双目视觉动态跟踪系统的静态测量和动态跟踪等优势特性来跟踪测量机器人的连杆参数误差,结合机器人控制系统开放性特点,运用提出的动态标定原理对机器人实施连杆参数测量、辨识、修正及补偿。实验表明,通过参数反馈补偿,自主研发的机器人的定位误差明显降低,且该方法易于实现,为机器人精度研究提供了可靠依据。This paper used binocular vision dynamic tracking technology in the kinematic calibration of independent research and development industrial robots. The advantages of static measuring and dynamic tracking characteristics of binocular vision dynamic tracking system, which was different from previous complex models, it used to track and measure robot link parame- ters errors. It proposed a dynamic calibration method, which combined with open characteristics of the robot control system, to implement the robot link parameters measurement, identification, correction and compensation. The experiments show that the positioning error of the independent research and development robot is significantly reduced, by using parameters feedback compensation. Meanwhile, the method is easy to implement, and also provides a reliable basis for the research of robot accura-cy.

关 键 词：工业机器人 标定 定位误差 运动学模型 双目视觉 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]