Coverage path planning for multi UAV collaborative environment under communication constraints

The aim of multi-UAV collaborative coverage is to efficiently allocate tasks to multiple UAVs, achieving rapid and effective full coverage of a given area. However, in real-world applications, the distance between UAVs often exceeds the communication range, leading to communication disruptions and challenges in UAV collaboration and information exchange. Therefore, a multi-UAV coverage path planning (CPP) method based on Deep Q Networks (DQN) is proposed. The path quality is evaluated by two indexes, communication disruption rate and maximum communication disruption time, and the autonomous path decision-making for UAV teams is realized by constructing reward functions related to these indexes. Simulation experiments demonstrate that the proposed method can be consistent with the traditional optimization algorithms on the shortest path. Moreover, under a balanced path condition, the communication interruption rate can be reduced by 80% with a 20% increase in path length. Additionally, under full communication path conditions, the communication with connected network throughout the entire process can be achieved by 100%. Therefore, the proposed method can generate efficient paths covering all environmental nodes according to different communication environments. © 2024 Editorial Department of Journal of Chinese Inertial Technology. All rights reserved.