摘要
针对传统梯度投影法存在关节位移误差积累的不足,在逆运动学算法中加入关节位移校正模块:用末端操作器目标位姿矩阵与中间位姿矩阵的差值生成迭代过程中需要的末端操作器参考速度,从而保证末端操作器位姿向期望值收敛,进而得到满足用户设定精度的关节位移和关节速度。仿真结果表明该方法具有很快的收敛速度,适于冗余度机器人的实时轨迹控制。
Aiming at the shortcomings of the traditional gradient projection method in accumulating the error of joint displacement, a joint displacement correction module is implemented in the inverse kinematics algorithm： the desired end-effectors velocity is generated by D-value between current position-orientation matrix （POM） and the expeeted POM, so that the current POM astringes to the expected POM. Therefore the desired accurate joint displacement can be obtained. Quick convergence of this method is found by the simulation result. It can be applied to the real-time trace tracking control.
出处
《机械科学与技术》
CSCD
北大核心
2009年第5期618-621,共4页
Mechanical Science and Technology for Aerospace Engineering
关键词
冗余度机器人
梯度投影法
逆运动学
redundant robot
gradient projection method
inverse kinematics
