.png){kind=icon}
.png){kind=icon}
Return
Multi-UAV Path Planning for Plural Data Collection: Distributed Multiagent Deep Reinforcement Learning Algorithms
.png){kind=icon}
PRE
AI
DOI：10.1109/JIOT.2025.3600404.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Distributed path planning for multiple unmanned aerial vehicles (UAVs) plays a significant role in data collection systems. However, insufficient collaboration among multiple UAVs and inadequate emergency response capability in the case of UAV failure will lead to longer collection time. To address these challenges, we put forward a set of novel distributed path planning schemes in plural data collection scenarios, aimed at minimizing the task completion time and enhancing robustness. Specifically, a distributed framework of multiple UAVs collaboration data collection system (MUC-DCS) is designed, which can effectively avoid repeated data collection and UAV collision through inter-UAV communication. In order to optimize the task completion time in the MUC-DCS, a distributed path planning algorithm based on multiagent deep Q-network (DPP-MDQN) is proposed. In addition, so as to improve the algorithm efficiency, the pheromone is set to represent the state information and reward function. Further, we develop an emergency response strategy and propose a distributed path planning algorithm for emergency response (DPP-ER), enabling response to sudden UAV failure as well as ensuring the robustness and timeliness of MUC-DCS. The simulation results show that DPP-MDQN is superior to existing distributed algorithms and reduces the task completion time, as well as DPP-ER can effectively handle the sudden emergency of UAV failure.
Keywords:
Data collection
distributed multiagent algorithm
multiple unmanned aerial vehicles (UAVs) collaboration
path planning
UAV failure
Journal
IF:
8.9
Papers:
1.4W
Citations:
7.8W