Uav Path Planning and Obstacle Avoidance Based on Deep Learning and Reinforcement Learning

Abstract

With the rapid development of the Internet of Things and low-altitude economy, unmanned aerial vehicles (Uavs) have been widely used in data collection, environmental monitoring and other fields.Traditional path planning algorithms have defects such as insufficient generalization ability and poor real-time performance in complex and dynamic environments. Deep reinforcement learning technology provides a new solution for autonomous decision-making and collaborative operation of unmanned aerial vehicles (Uavs). In this paper, the UAV path planning technology based on deep learning and reinforcement learning is systematically reviewed. In the aspect of single UAV, the improved Q-Learning algorithm realizes fast local replanning through dynamic exploration factor and artificial potential field reward function. DDPG, DySAC, SMLTO and other algorithms improve the safety of obstacle avoidance and trajectory smoothness in three-dimensional dynamic environment. In terms of multiple Uavs, the MRF-DQN algorithm combines preferential experience replay and maximum reward frequency function to improve the efficiency of collaborative task completion. The DPPDQN algorithm realizes efficient task allocation and path optimization in large-scale regions through double pre-segmentation and CNN feature extraction. The experimental results show that the above algorithms are superior to the traditional methods in terms of collision rate, planning delay, energy efficiency and task completion rate, which provides reliable technical support for multi-UAV cooperative operation.