Motion Planning of Redundant Robotic Arm Based on Human Arm Motion Principles

How to generate human-like arm extension motion is a challenging problem in the process of human-robot collaboration to understand and judge the robot's motion intention. To address this problem, this paper proposes a multi-task planning algorithm for human-like arm motion based on the energy principle and the joint limit avoidance principle. First, a motion capture device is used to collect human arm motion data to obtain guidelines for human arm motion characteristics. On the basis of analyzing the arm demonstration data, the gravitational potential energy and virtual elastic potential energy models of the arm are established, and the mapping from the virtual variable stiffness of the arm to the virtual variable stiffness of the robot is realized. An end-effector trajectory is generated based on a dynamic movement primitives (DMPs) model and an optimization method, while a spherical linear interpolation (slerp) algorithm is used to generate the attitude sequence. In order to better execute the planned motion, the priority of the motion task is divided, and the design of two priority motion tasks is carried out by combining the gradient projection method. Finally, the humanoid arm motion multi-task planning algorithm proposed in this paper is simulated by simulation experiments, and the experimental results show that the method can generate the corresponding humanoid motions and has better motion accuracy.