多段连续结构手功能康复外骨骼运动规划与结构优化

Motion Planning and Structure Optimization of a Multi-Segment Mechanism-Based Hand Rehabilitation Exoskeleton

针对现有康复机器人手指运动轨迹仿生度低、轨迹规划策略与正常人手指运动轨迹差异较大的问题,提出了基于Tau-jerk引导策略的手指运动轨迹规划方法。首先建立外骨骼机器人的运动学模型,并用ABAQUS进行运动学仿真,引入修正系数调整运动学模型;然后以jerk值为耦合对象,推导出内部Tau-jerk引导策略,引入最小做功和最小面积确定最优轨迹和最优多段连续结构;最后比较了Tau-jerk规划的运动轨迹与多项式插值方法规划的轨迹同真实人手运动轨迹的偏差。实验结果表明:采用Tau-jerk轨迹规划的外骨骼运动平均关节角度误差小于1.5°,且平均指尖轨迹误差小于1 mm;与3次多项式插值和3、4、5次多项式插值运动规划相比,Tau-jerk轨迹规划方法更符合人手指运动轨迹。

The finger motion trajectory of existing rehabilitation robots has a low bionic degree,and the trajectory planning strategy is mostly far from normal people finger motion trajectory.To solve this difficulty,a method of finger motion trajectory planning following Tau-jerk guiding strategy is proposed.The kinematics model of the exoskeleton robot is established and simulated by ABAQUS,and a modified coefficient is introduced to adjust the model.Then,the internal Tau-jerk guiding strategy is derived by taking jerk value as the coupling object,and the minimum work and minimum area are introduced to search for the optimal trajectory and optimal multi-segment continuous structure.Finally,the deviation between the trajectory of Tau-jerk planning and the other polynomial interpolation methods is compared with that of real human hands.The experimental results show that the average joint angle error of exoskeleton motion using Tau-jerk planning is less than 1.5°,and the average fingertip trajectory error is less than 1 mm.Compared with motion planning using cubic polynomial interpolation and 3,4,5 polynomial interpolation,Tau-jerk trajectory planning model well coincides with human finger motion.