Collaborative Coverage Path Planning of UAV Cluster based on Deep Reinforcement Learning

With the continuous application of unmanned aerial vehicles (UAV) in the field of national defense and civil use, the UAV cluster system in which multiple UAVs cooperate to perform tasks has become a key research in many countries. This paper focuses on the problem of multi-UAV's Coverage Path Planning (mCPP) to exploit all points of interests within an area based on Reinforcement Learning (RL), where each UAV starts with a random position and carries a camera during the mission. There have been a number of optimal algorithms proposed for the coverage path planning of single UAV, however, it is under-explored for multiple UAVs. As such, we leverage Deep reinforcement learning with Double Q-learning Networks (DDQN) to learn a global optimal control policy for a team of UAVs under certain power constraints to cooperate effectively to explore a wider area. Regarding the task area as a 2D plane, we divide it into a collection of uniform grid cells, which represent a section of the environment. The camera field of view of each UAV covers a cell area underneath the UAV. Simulation results demonstrate that wherever the start positions are, the UAV cluster can fully cover the whole task area under the energy constraints and achieve autonomous collaboration. The proposed method has great potential in applying to dynamic environment.