摘要
针对现有基于动态系统稳定估计器(SEDS)的机械臂轨迹规划中无法兼顾运动精度和能量消耗的问题,提出变刚度轨迹规划的方法。首先,采用SEDS对示教轨迹进行拟合,得到从空间任一起点到终点的收敛轨迹。然后,基于人体上肢刚度模型,通过实验得到肌电信号与上肢末端刚度的映射关系,并使用高斯混合模型GMM和高斯混合回归GMR学习示教的刚度。最后,结合学习到的轨迹和刚度进行阻抗控制,然后进行抗干扰和变刚度实验。研究结果表明:与定刚度轨迹规划方法相比,本文所提方法位置误差减小了78.1%,能耗降低16.1%。机械臂能在到达目标点的位置精度与能量消耗之间取得较好的平衡,且当其受到外界干扰后,会根据控制器发出的期望位置调整自身运动,最终仍能收敛到目标点,具有较强的抗干扰能力。
Aiming at the problem that the existing trajectory planning based on the stable estimator of dynamical systems(SEDS)cannot balance motion accuracy and energy consumption,a variable stiffness trajectory planning method was proposed.First,SEDS was used to fit the teaching trajectory to obtain a convergent trajectory from any starting point to the end point in space.Then,based on the human upper limb stiffness model,the mapping relationship between the EMG signal and the upper limb stiffness was obtained through experiments,and the Gaussian mixture models(GMM)and Gaussian mixture regression(GMR)were used to learn the taught stiffness.Finally,impedance control was achieved by combining the learned trajectory and stiffness,and anti-interference and variable stiffness experiments were conducted.The results show that compared with the fixed stiffness trajectory planning method,the proposed method reduces the position error by 78.1%and the energy consumption by 16.1%.The robotic arm can achieve a good balance between the position accuracy of reaching the target point and energy consumption.When the robotic arm is disturbed by the outside,it will adjust its own motion according to the desired coordinate position sent by the controller,and it can still converge to the target point with a strong anti-interference ability.
作者
谢啸
张涵
汤自林
高霄
肖晓晖
XIE Xiao;ZHANG Han;TANG Zilin;GAO Xiao;XIAO Xiaohui(School of Power and Mechanical Engineering,Wuhan University,Wuhan 430072,China)
出处
《中南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2022年第4期1250-1258,共9页
Journal of Central South University:Science and Technology
基金
国家重点研发计划项目(2018YFB2100903)
人因工程国防科技重点实验室开放基金资助项目(6142222180311)。
关键词
示教学习
动态系统
变刚度轨迹规划
learning from demonstration
dynamical system
variable stiffness trajectory planning
