一种冗余机械臂的耦合规划控制算法

Coupling planning control algorithm of redundant manipulator

针对人机交互环境下,存在多次变更机械臂的期望末端位姿和冗余机械臂路径规划时间长、作业卡顿等问题,提出一种兼顾逆运动学、规划和控制的冗余机械臂耦合规划控制算法。基于推导的冗余自由度逆运动学,对其中4个关节的正负取值加入随机因子,扩大解析解的存在空间,作业效率平均提高40.71%;基于动态运动基元,减低作业过程中的卡顿;融合Descartes规划器和改进的逆解器,引入动态运动基元,规划器完成规划任务,执行机构同步接收作业指令。基于机器人操作系统(ROS)建立动态半实物仿真系统,将提出的耦合规划控制算法与常用顺序交替规划控制算法相比较,并应用于人机交互系统进行物理试验。结果表明,提出的耦合控制算法作业时间平均缩短了48.25%,增大了关节执行的最大速度,改善了实体运行流畅性。

Due to the problem of several times changing the desired end pose,long time path planning in redundant manipulator and the operation blocking,a novel coupling path planning algorithm with inverse kinematics,planning and control was proposed in an interactive environment.For the inverse kinematics of redundant freedom,a random factor was added to the positive and negative values of four joints,and the existence space of the analytical solution was expanded,which could improve 40.71%in working efficiency.The operation blocking in the working process was decreased based on Dynamical Movement Primitives(DMP).Through combining Descartes planner with improved inverse solver,the planning tasks was performed and the receiving working instruction was actuated synchronously.The proposed coupling path planning had applied to human-computer interaction system for physical testing,which saved 48.25%in working time comparing with the commonly algorithm of sequential alternating planning control under ROS simulation system.The maximum speed of joint execution was significantly increased,and the physical operation was smoother.