首页|Distributed UAV swarms for 3D urban area coverage with incomplete information using event-triggered hierarchical reinforcement learning
Distributed UAV swarms for 3D urban area coverage with incomplete information using event-triggered hierarchical reinforcement learning
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
Elsevier
This paper addresses the problem of distributed 3D area coverage with multiple UAVs under incomplete information. It proposes a leader-follower UAV framework that integrates pheromone-based coverage and reinforcement learning dispatching. The 3D space is segmented for dimensionality reduction, and pheromone matrices enable dynamic area coverage. To reduce redundant coverage and achieve multi-layered coordination, the leader UAV uses a virtual decision mechanism and a communication neural network to handle local observations and asynchronous scheduling challenges. Additionally, an event-triggered reinforcement learning is introduced to minimize communication costs and enhance system robustness. Pheromone-based implicit teamwork addresses sparse rewards and improves coverage efficiency. This modular framework enhances general applicability and significantly reduces overall coverage time.
Multi UAVsReinforcement learningPheromone matrix3D area coverageEvent-trigger
Jin Yu、Hui Zhang、Ya Zhang
展开 >
School of Automation, Southeast University, Nanjing 210096, China||Key Laboratory of Measurement and Control of Complex Systems of Engineering, Ministry of Education, Nanjing, China
School of Automation, Southeast University, Nanjing 210096, China||High-Tech Department, Jiangsu Automation Research Institute, Lianyungang, 222061, China
NETL
NSTL
Elsevier
