Multi-UAV Path Planning Based on MAA2C
Qiang Yan, Qi Zhao, Ye Zhang · 2024
In order to solve the problem of multi-mission path planning for multiple UAVs in complex scenarios, the APF- MAA2C model is proposed by combining the artificial potential field method with the MAA2C algorithm. The model uses the advantage function as an index to select action values and measure the quality of all action draw values, and adds a repulsive force field in the artificial potential field method generated by other UAVs to the environment as a punishment and reward. The experimental results show that the APF-MAA2C model can better avoid the collision between multiple UAVs, and can successfully complete the task.