基于神经动力学的冗余机械臂恒定转速比算法

提出一种基于神经动力学的恒定转速比（CRR）算法，用于解决冗余机械臂运动规划末端精度问题.在运动规划中，通过神经动力学求解出雅可比矩阵伪逆，由4阶龙格-库塔法得出实时精准的关节角速度，从而获得各关节平均速度比，并将其比值关系带入CRR算法中求解出最优关节角度.与传统的加权伪逆运动规划不同，CRR算法是在关节角位置层面求解，这也正是提高冗余机械臂末端精度的关键环节.在文末，通过平面3自由度和平面6自由度冗余机械臂模型进行运动规划仿真，仿真结果验证了算法的有效性. 

Constant Rotation Ratio of Redundant Robot Manipulators Based on Neural Dynamics

In this paper, a constant rotation ratio (CRR) algorithm based on neural dynamics was proposed to solve the trajectory accuracy of motion planning for redundant robot manipulators. In the motion planning, the pseudo-inverse of the Jacobian matrix was solved by the neural dynamics network solver. The real-time and accurate joint velocity was acquired by the fourth-order Runge-Kutta methods to solve the average velocity ratio of each joint, and to obtain the optimal joint angle for the CRR. Differing from the traditional motion planning by weighted pseudo-inverse solution, the CRR algorithm was designed to solve at the joint angle level, being the key approach to improve the trajectory accuracy of redundant robot manipulators. In addition, computer-simulation models with three-link and six-link planar robot manipulators were established. Simulation results show the validity of the CRR based on neural dynamics algorithm.