• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2024, Vol. 60 ›› Issue (21): 99-111.doi: 10.3901/JME.2024.21.099

• 机器人及机构学 • 上一篇    下一篇

扫码分享

基于带通滤波二阶动量观测器的机器人碰撞检测

谢峰1, 李红杨1,2, 甄圣超3, 魏华伟1   

  1. 1. 安徽大学电气工程及自动化学院 合肥 230601;
    2. 中国电信股份有限公司安徽分公司 合肥 230061;
    3. 合肥工业大学机械工程学院 合肥 230009
  • 收稿日期:2023-11-02 修回日期:2024-05-25 发布日期:2024-12-24
  • 通讯作者: 李红杨,男,1997年出生,硕士研究生。主要研究方向为机器人碰撞检测方法。E-mail:2261949337@qq.com
  • 作者简介:谢峰,男,1963年出生,博士,教授,博士研究生导师。主要研究方向为智能制造,切削加工理论。E-mail:jeexf199@163.com
  • 基金资助:
    国家优秀青年科学基金资助项目(52022077)。

Robot Collision Detection Based on Bandpass Filtered Second-order Momentum Observer

XIE Feng1, LI Hongyang1,2, ZHEN Shengchao3, WEI Huawei1   

  1. 1. School of Electrical Engineering and Automation, Anhui University, Hefei 230601;
    2. China Telecom Co., Ltd. Anhui Branch, Hefei 230061;
    3. School of Mechanical Engineering, Hefei University of Technology, Hefei 230009
  • Received:2023-11-02 Revised:2024-05-25 Published:2024-12-24

摘要: 针对六轴工业机器人在工作过程中可能与周围环境和操作者发生碰撞的问题,提出基于巴特沃斯带通滤波构建二阶动量观测器的机器人碰撞检测方法。通过机器人广义动量与外力矩的解耦性关系,定义二阶动量观测器的观测指标为机器人理论动量与实际动量的偏差值,建立了基于二阶动量观测器的碰撞检测模型,并利用巴特沃斯带通滤波器来优化观测器增益以减小噪声带来的误差,并通过仿真验证了此碰撞观测器的有效性。然后应用基于递减权重的粒子群算法对六轴机器人动力学参数进行整体辨识,采用最小二乘法依次求解出各个关节的最小参数集。最后在ROC6六轴工业机器人平台进行动力学参数辨识和碰撞检测实验,实验结果表明,当机器人在工作中发生碰撞时,所设计的动量观测器可以在0.3 s左右检测到,检测到的碰撞力阈值为21.5 N,可以有效保护操作者和机器人的安全。

关键词: 六轴工业机器人, 碰撞检测, 动量观测器, 参数辨识

Abstract: A robot collision detection method based on Butterworth bandpass filtering to construct a second-order momentum observer is proposed to address the issue of possible collisions between six axis industrial robots and the surrounding environment and humans during operation. By decoupling the generalized momentum and external torque of robots, the observation index of the second-order momentum observer is defined as the deviation value between the theoretical momentum and actual momentum of the robot. A collision detection algorithm based on the second-order momentum observer is derived, and a Butterworth bandpass filter is used to optimize the observer gain to reduce errors caused by noise. The effectiveness of this collision observer is verified through simulation. Then, a particle swarm optimization algorithm based on decreasing weights is applied to identify the overall dynamic parameters of the six-axis robot, and the least squares method is used to sequentially solve the minimum parameter set of each joint. Finally, dynamic parameter identification and collision detection experiments are conducted on the ROC6 six axis industrial robot platform. The experimental results show that when the robot collides during operation, the momentum observer designed can detect it in about 0.3 s, and the detected collision force threshold is 21.5 N, which can effectively protect the safety of humans and robots.

Key words: six axis industrial robot, collision detection, momentum observer, parameter identification

中图分类号: