Reinforcement Learning Path Planning based on Step Batch Q-Learning Algorithm

Abstract

With the rapid development of artificial intelligence, people have put forward higher requirements for robot path planning. As a more commonly used algorithm, reinforcement learning learns from experience by imitating the process of human learning skills and continuously iterates to do. In this paper, an improved Step Batch Q-Learning algorithm is proposed to solve the problem that the traditional Q-Learning algorithm has a slow convergence speed in discrete states. Prediction can detect obstacles and target points earlier, reduce the number of iterations, and significantly speed up the convergence time. Under the same experimental conditions, the method in this paper is compared with the traditional Q-Learning algorithm. The experimental results show that the improved Step Batch Q-Learning algorithm has significantly improved convergence speed and good adaptability.