【摘要】 针对神经网络和梯度投影法求解七自由度机械臂逆运动学存在的可靠性差及累积误差等问题,充分考虑机械臂构型特点和自运动特性,提出一种几何法和解析法相结合的冗余机械臂逆运动学新方法并进行仿真验证,通过位置误差和姿态误差来衡量该方法的可行性。实验结果表明:采用该方法对机械臂逆运动学进行求解,机械臂末端位置最大误差不超过0.015mm,平均运行时间为0.455ms,满足实时控制要求,为具有横滚转动副机械臂关节角度求解提供了一种新的思路。更多还原

【Abstract】 Aiming at the problems that solving inverse kinematics for seven degree-of-freedom manipulator has inferior reliability and cumulative error by using neural network and gradient projection method.Fully considered the characteristics of arm configuration and its self-motion,a new algorithm combined with geometry and analysis are proposed for solving redundant manipulator inverse kinematics and do some simulation verification.Position and orientation error weigh up this method feasibility.The simulation result shows that the maximum end error of manipulator position is less than 0.015 mm and mean run time is 0.455 ms by using this way.It is satisfied with online control requirement.It can provide a new idea for solving the manipulator inverse kinematics with rotate joint angle.更多还原