• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2025, Vol. 61 ›› Issue (2): 346-357.doi: 10.3901/JME.2025.02.346

• 交叉与前沿 • 上一篇    

扫码分享

六自由度液压机械臂运动学标定和逆解研究

梁相龙1, 姚志凯2, 邓文翔1, 姚建勇1   

  1. 1. 南京理工大学机械工程学院 南京 210094;
    2. 南京邮电大学自动化学院 南京 210023
  • 收稿日期:2024-01-07 修回日期:2024-07-13 发布日期:2025-02-26
  • 作者简介:梁相龙,男,1997年出生,博士研究生。主要研究方向为液压机械臂运动控制及机电液伺服系统智能控制。E-mail:xlliang.njust@gmail.com;姚志凯,男,1991年出生,博士,讲师。主要研究方向为机电液伺服系统智能控制和强化学习控制。E-mail:zacyao.cn@gmail.com;姚建勇(通信作者),男,1984年出生,博士,教授,博士研究生导师。主要研究方向为机电液系统伺服控制,动态系统故障检测与容错,半实物动态仿真技术。E-mail:jerryyao.buaa@gmail.com
  • 基金资助:
    国家重点研发计划(2021YFB2011300)和国家自然科学基金(52075262,52275062)资助项目。

Research on Kinematic Calibration and Inverse Solution of 6-DOF Hydraulic Manipulator

LIANG Xianglong1, YAO Zhikai2, DENG Wenxiang1, YAO Jianyong1   

  1. 1. School of Mechanical Engineering, Nanjing University of Science and Technology, Nanjing 210094;
    2. College of Automation, Nanjing University of Post and Telecommunication, Nanjing 210023
  • Received:2024-01-07 Revised:2024-07-13 Published:2025-02-26

摘要: 鉴于运动学标定能够有效地提高机械臂的定位精度且逆运动学求解是控制机械臂正确运动的关键。以六自由度液压机械臂为研究对象,采用M-DH方法和几何映射方法分别实现了液压机械臂Ⅰ型运动学和Ⅱ型运动学误差建模,然后基于列文伯格-马夸尔特(Levenberg-Marquardt, L-M)算法辨识误差参数。另外考虑到机械臂存在几何参数误差,无法应用反变换法求解机械臂的逆运动学,为此提出一种结合解析解法和偏微分理论的逆解计算方法。仿真试验结果表明,经过L-M算法标定后,液压机械臂的末端位置误差从3.032 9 mm下降到0.007 0 mm,效果提升99.77%,液压缸1的位置误差从1.276 2 mm下降到0.000 3 mm,效果提升99.98%,液压缸2的位置误差从1.167 0 mm下降到0.001 2 mm,效果提升99.90%,证明了标定算法的有效性。此外仿真结果还表明所提出的逆解计算方法可有效地获得机械臂逆解。

关键词: 液压机械臂, 运动学标定, Ⅰ型运动学误差建模, Ⅱ型运动学误差建模, 逆运动学

Abstract: In view of the fact that kinematics calibration can effectively improve the positioning accuracy of the robotic manipulator and inverse kinematics solution is the key to controlling the correct motion of the robotic manipulator. Taking the six-degree-of-freedom(6-DOF) hydraulic manipulator as the research object, the M-DH method and the geometric mapping method are utilized to establish the Ⅰ-type and Ⅱ-type kinematic error models of the hydraulic manipulator, and then the kinematic parameter deviations are identified based on the Levenberg-Marquardt(L-M) algorithm. In addition, the inverse transformation method cannot be leveraged to solve the inverse kinematics because of geometric parameter errors in the robotic manipulator. Therefore, an inverse solution calculation method combining analytical solution and partial differential theory is proposed. The simulation results reveal that after calibration by the L-M algorithm, the end position error of the hydraulic manipulator is reduced from 3.032 9 mm to 0.007 0 mm, the effect is improved by 99.77%, the position error of hydraulic cylinder 1 is reduced from 1.276 2 mm to 0.000 3 mm, the effect is increased by 99.98%, and the position error of hydraulic cylinder 2 is reduced from 1.167 0 mm to 0.001 2 mm, the effect is increased by 99.90%, which illustrates the validity of the calibration algorithm. Moreover, the simulation results also reveal that the proposed inverse solution calculation method can effectively obtain the inverse solution of the robotic manipulator.

Key words: hydraulic manipulator, kinematic calibration, Ⅰ-type kinematic error modeling, Ⅱ-type kinematic error modeling, inverse kinematics

中图分类号: