林安迪, 干旻峰, 葛涵, 唐宇存, 徐海东, 匡绍龙, 黄立新, 孙立宁. 基于模糊模型参考学习控制的手术机器人人机交互[J]. 机器人, 2019, 41(4): 543-550. DOI: 10.13973/j.cnki.robot.180495
引用本文: 林安迪, 干旻峰, 葛涵, 唐宇存, 徐海东, 匡绍龙, 黄立新, 孙立宁. 基于模糊模型参考学习控制的手术机器人人机交互[J]. 机器人, 2019, 41(4): 543-550. DOI: 10.13973/j.cnki.robot.180495
LIN Andi, GAN Minfeng, GE Han, TANG Yucun, XU Haidong, KUANG Shaolong, HUANG Lixin, SUN Lining. Human-Robot Interaction for Surgical Robot Based on Fuzzy ModelReference Learning Control[J]. ROBOT, 2019, 41(4): 543-550. DOI: 10.13973/j.cnki.robot.180495
Citation: LIN Andi, GAN Minfeng, GE Han, TANG Yucun, XU Haidong, KUANG Shaolong, HUANG Lixin, SUN Lining. Human-Robot Interaction for Surgical Robot Based on Fuzzy ModelReference Learning Control[J]. ROBOT, 2019, 41(4): 543-550. DOI: 10.13973/j.cnki.robot.180495

基于模糊模型参考学习控制的手术机器人人机交互

Human-Robot Interaction for Surgical Robot Based on Fuzzy ModelReference Learning Control

  • 摘要: 为了解决机器人辅助手术环境下人机交互运动过程中的不稳定性问题以及医生个人因素难以建模的问题,提出了一种基于模糊模型参考学习的变导纳人机合作控制方法.首先将人体手臂自然运动的特征作为模糊学习控制的参考模型,通过离线学习机构训练出模糊导纳控制器的变阻尼系数调整参数规则.再以医生对机器人的拖拽力以及机器人速度作为输入、机器人期望的速度作为输出,构建基于变阻尼参数调整的变导纳控制方法.离线训练实验结果表明,该方法经过10次离线训练,可以达到柔顺性要求,人机合作速度最大误差低于17 mm/s,且人机合作轨迹最大误差低于15 mm.相比单纯基于固定导纳参数的模糊控制,该方法具有更好的跟踪速度与精度.

     

    Abstract: In order to solve the instability problem in human-robot interaction in the robot-assisted surgery environment and the difficulty in modeling the personal factors of the doctor, a variable admittance human-robot cooperative control method is proposed based on fuzzy model reference learning. Firstly, the natural movement characteristics of human arm are taken as the reference model for fuzzy learning control, and the variable damping coefficient adjustment parameter rules are trained for the fuzzy admittance controller by an offline learning mechanism. Then a variable admittance control method is constructed based on variable damping parameter adjustment, taking the doctor's dragging force on the robot and the speed of the robot as the input, to output the desired speed of the robot. The offline training experimental results show that after offline training 10 times, the maximum error of the human-machine cooperation velocity is less than 17 mm/s, and the maximum error of the human-machine cooperation trajectory is less than 15 mm. Therefore, the proposed method has both a better tracking speed and a better precision than fuzzy control based on fixed admittance parameters.

     

/

返回文章
返回