双目视觉伺服的4-DOF机械手臂运动控制
|
摘要
为了解决多自由度机械手臂由于目标放置偏离而引起的抓取任务失败问题,给出了一种基于双目立体视觉的目标识别与定位控制方法。以抓取矩形轮廓的目标物体放置于矩形目标位置为控制任务,基于双目视觉目标图像信息,分析了目标特征参数提取、识别、匹配以及目标空间位姿测量的方法。结合四自由度(4-DOF)机械手臂硬件控制系统,运用改进D-H参数法建立了机械手臂正运动学模型,并通过逆运动学方法将目标位姿转化为机械手臂的控制指令,实现对目标抓取和放置的运动控制。实验结果表明,该视觉伺服的机械手臂运动控制系统能够准确、稳定地实现抓取和放置任务,定位精度高,对工业机器人在分拣、装配中的运动控制研究和应用有重要的参考价值。
In order to solve the problem that the multi-degree-of-freedom robot arm failed in the fetching task due to the deviation of the target placement,a binocular vision approach was proposed for the target recognition and positioning. The control task was to place the target object of grabbing the rectangle contour at the rectangular target position. Based on the image information of binocular vision,the methods of extraction,recognition,matching and measurement of spatial pose of the target were analyzed.Combined with the 4-DOF mechanical arm hardware control system,the forward kinematics model of mechanical arm was established by using improved D-H parameter method. The object pose is transformed into the control instruction of the robot arm through the inverse kinematics method to realize the motion control of grabbing and placing the object. The experimental results show that the motion control system of the robot arm of visual servo can accurately and stably realize the grabbing and placement tasks,and the positioning accuracy is high,which is of great reference value for the study and application of motion control of industrial robots in sorting and assembly.
In order to solve the problem that the multi-degree-of-freedom robot arm failed in the fetching task due to the deviation of the target placement,a binocular vision approach was proposed for the target recognition and positioning. The control task was to place the target object of grabbing the rectangle contour at the rectangular target position. Based on the image information of binocular vision,the methods of extraction,recognition,matching and measurement of spatial pose of the target were analyzed.Combined with the 4-DOF mechanical arm hardware control system,the forward kinematics model of mechanical arm was established by using improved D-H parameter method. The object pose is transformed into the control instruction of the robot arm through the inverse kinematics method to realize the motion control of grabbing and placing the object. The experimental results show that the motion control system of the robot arm of visual servo can accurately and stably realize the grabbing and placement tasks,and the positioning accuracy is high,which is of great reference value for the study and application of motion control of industrial robots in sorting and assembly.
引文
[1]王田苗,陶永.我国工业机器人技术现状与产业化发展战略[J].机械工程学报,2014,50(9):1-13.
[2]汤双清,高虹亮,李力.工业机器人的计算机控制[J].计算机测量与控制,2015(2):37-39.
[3]翟志强,杜岳峰,朱忠祥,等.基于Rank变换的农田场景三维重建方法[J].农业工程学报,2015,31(20):157-164.
[4]随婷婷,王晓峰.一种基于视觉知识加工模型的目标识别方法[J].自动化学报,2016,42(5):760-770.
[5]全燕鸣,黎淑梅,麦青群.基于双目视觉的工件尺寸在机三维测量[J].光学精密工程,2013,21(4):1054-1061.
[6]陈科尹,邹湘军,彭红星,等.基于半实物仿真的采摘机器人视觉定位研究[J].系统仿真学报,2017,29(8):1736-1746.
[7]李海鹏,邢登鹏,张正涛,等.宏微结合的多机械手微装配机器人系统[J].机器人,2015,37(1):35-42.
[8]吴凤和.基于计算机视觉测量技术的图像轮廓提取方法研究[J].计量学报,2007(1):18-22.
[9]项荣,应义斌,蒋焕煜,等.基于双目立体视觉的番茄定位.农业工程学报,2012,28(5):161-167.
[10] ILIUKHIN V N,MITKOVSKII K B,BIZYANOVA D A,et al. The modeling of inverse kinematics for 5 DOF manipulator[J]. Procedia Engineering,2017,176:498-505.
[11]刘俸材,谢明红,王伟.双目视觉的立体标定方法[J].计算机工程与设计,2011,32(4):1508-1512.
[12]高瞻宇,顾营迎,刘宇航,等.采用简化Brown模型及改进BFGS法的相机自标定[J].光学精密工程,2017,25(9):2532-2540.
[13]张新良,冷正明. 3-DOF机械手臂运动定位控制系统设计[J].仪表技术与传感器,2017(8):51-55.
[2]汤双清,高虹亮,李力.工业机器人的计算机控制[J].计算机测量与控制,2015(2):37-39.
[3]翟志强,杜岳峰,朱忠祥,等.基于Rank变换的农田场景三维重建方法[J].农业工程学报,2015,31(20):157-164.
[4]随婷婷,王晓峰.一种基于视觉知识加工模型的目标识别方法[J].自动化学报,2016,42(5):760-770.
[5]全燕鸣,黎淑梅,麦青群.基于双目视觉的工件尺寸在机三维测量[J].光学精密工程,2013,21(4):1054-1061.
[6]陈科尹,邹湘军,彭红星,等.基于半实物仿真的采摘机器人视觉定位研究[J].系统仿真学报,2017,29(8):1736-1746.
[7]李海鹏,邢登鹏,张正涛,等.宏微结合的多机械手微装配机器人系统[J].机器人,2015,37(1):35-42.
[8]吴凤和.基于计算机视觉测量技术的图像轮廓提取方法研究[J].计量学报,2007(1):18-22.
[9]项荣,应义斌,蒋焕煜,等.基于双目立体视觉的番茄定位.农业工程学报,2012,28(5):161-167.
[10] ILIUKHIN V N,MITKOVSKII K B,BIZYANOVA D A,et al. The modeling of inverse kinematics for 5 DOF manipulator[J]. Procedia Engineering,2017,176:498-505.
[11]刘俸材,谢明红,王伟.双目视觉的立体标定方法[J].计算机工程与设计,2011,32(4):1508-1512.
[12]高瞻宇,顾营迎,刘宇航,等.采用简化Brown模型及改进BFGS法的相机自标定[J].光学精密工程,2017,25(9):2532-2540.
[13]张新良,冷正明. 3-DOF机械手臂运动定位控制系统设计[J].仪表技术与传感器,2017(8):51-55.