Mobile Robotic Arm for Opening Doors Using Proximal Policy Optimization

Abstract

The traditional robotic arm control method has strong dependence on the application scenario. To improve the reliability of the mobile robotic arm control when the scene is disturbed, this paper proposes a control method based on an improved proximal policy optimization algorithm. This study researches mobile robotic arms for opening doors. At first, the door handle position is obtained through an image-recognition method based on YOLOv5. Second, the simulation platform CoppeliaSim is used to realize the interaction between the robotic arm and the environment. Third, a control strategy based on a reward function is designed to train the robotic arm and applied to the opening-door task in the real environment. In this paper PPO algorithm is used to solve the result. The experimental results show that the proposed method can accelerate the convergence of the training process. Besides, our method can effectively reduce the jitter of the robotic arm and improve the stability of control.