A novel deep learning driven robot path planning strategy: Q-learning approach

As the basis of mobile navigation technology, path planning has attracted the attention of the majority of scholars. In this paper, the deep learning framework is integrated into Q-learning, and a Deep Q-Network (DQN) algorithm based on memory optimised mechanism is designed to improve the convergence of DQN, so that the robot can carry out good path planning in complex environment. Experimental results show that the proposed method has a good performance in path planning. Experimental results show that in the case of multi-round training using the algorithm proposed in this paper, the path planning steps of the robot are the shortest and the total training time and single round time of the robot in path planning are also the shortest.