Research on UAV Path Planning Method Based on Reinforcement Learning and Behavioral Tree Modeling

To solve the limitations of traditional path planning algorithms in complex environments, a UAV path planning optimization method based on the reinforcement learning behavior tree was proposed in this paper. It combines the method that the behavior tree generated by reinforcement learning training with the UAV path planning algorithm to improve the ability to make autonomous decisions and deal with emergencies during flight. Firstly, based on the UAV behavior logic, the overall architecture of the behavior tree is designed, and the hierarchical access logic is used for dynamic decision. Then, the Q-learning algorithm of reinforcement learning is used to train the behavior tree. Finally, the behavior tree trained by reinforcement learning is introduced into the UAV path planning, while the A* algorithm is used for auxiliary decision, resulting in the UAV being able to evaluate environmental changes and adjust task priority in real time under complex conditions. Compared with the traditional method, the simulation results show that the proposed method can effectively improve the decision-making accuracy of the UAV in complex environments and reduce the task execution time. In addition, the average search time, average path length, and average search grid number are significantly reduced, and the search return rate reaches 80.49%. In conclusion, the research methods used in this paper can provide a valuable reference for all kinds of agents to improve their autonomous decision-making ability based on reinforcement learning behavior tree.