Motion Planning for Dual-Arm Robot Based on Soft Actor-Critic

In this paper, a motion planning method based on the Soft Actor-Critic (SAC) is designed for a dual-arm robot with two 7-Degree-of-Freedom (7-DOF) arms so that the robot can effectively avoid self-collision and at the same time can avoid the joint limits and singularities of the arm. The left-arm and right-arm of the dual-arm robot each have a neural network to control its position and orientation. Dual-agent training, distributed training structure, and progressive training environment are used to train neural networks. During the training process, the motion of one arm is regarded as the environment of the other arm, and the two agents are trained at the same time. In the input part of the neural network of the proposed method, all parameters come from the angle of each axis and kinematic calculation, no additional sensors are needed, so the method is easier to transplant to different dual-arm robots. With some appropriate neural network inputs and reward functions design, the robot can perform the expected self-collision avoidance and effectively avoid the joint limits and singularities of the arm. Finally, some experiments of the simulation tests in the Gazebo simulator and actual tests in a laboratory-made dual-arm robot are presented to illustrate the proposed SAC-based motion planning method is feasible and practicable in the avoidance of self-collision, joint limits, and singularities.