基于视觉引导的工业机器人示教编程系统  

Vision guide based teaching programming for industrial robot

作　　者：倪自强[1] 王田苗[1] 刘达[1] 

机构地区：[1]北京航空航天大学机械工程与自动化学院,北京100083

出　　处：《北京航空航天大学学报》2016年第3期562-568,共7页Journal of Beijing University of Aeronautics and Astronautics

基　　金：国家"863"计划(2013AA041201);国家自然科学基金(61175104)~~

摘　　要：针对目前工业机器人所采用的编程方式的局限性,提出了基于视觉引导的工业机器人示教编程方式。首先采用奇异值分解(SVD)法建立双目视觉系统与机器人的坐标映射关系,通过双目视觉系统识别和提取特定的示教工具末端位姿,以得到规划机器人运动所需的位姿数据。然后解析机器人控制器可执行文件的格式,将位姿数据转化为机器人可以执行的文件。最后通过试验验证了视觉引导示教方式的可行性,并对示教跟踪精度进行了分析,结果表明,跟踪误差最大为-1.18 mm,均方根误差为0.47 mm。Most industrial robots used in manufacture are based on teaching programming and offline programming. The shortcomings of these two programming methods limit the further application of industrial robots. A vision guide based programming method was introduced to solve the problem. The singular value decomposition（ SVD） algorithm was used to calculate the registration matrix between computer vision space and robot space. The positions and orientations of robot＇s end-actuator were obtained by measuring the teaching tool,which is the key process to realize vision guide programming. Analytical the format of executable file which running on robot controller,and then transform the positions and orientations＇ data into executable file.An experiment was introduced to verify the feasibility and reliability of the programming method. The results indicate that the maximum error of trajectory tracking is- 1. 18 mm,and the root mean square error is0. 47 mm.

关 键 词：工业机器人 示教编程 视觉 焊接 奇异值分解(SVD) 

分 类 号：V221.3[航空宇航科学与技术—飞行器设计] TB553[理学—物理]