Deployment of the unarmed aerial vehicle (UAV) swarm promises increased efficiency and safety of area search coverage. Multiple UAVs must be capable of autonomous collaboration and area search coverage for this. Therefore, we integrate multi-agent reinforcement learning into the cooperative control method of UAV swarm and propose a decentralized cooperative control for networked multiple UAVs technique based on the extended-proximal policy optimization algorithm (EPPO). The proposed approach not only adopts distributed training for multiple agents, but also allows them to obtain some mutual state information, such as position information and searched sub-areas. So, it can significantly speed up training and increase the effectiveness and safety of task completion in real-world applications. After a simulation, the multi-intelligent UAV can rapidly cover 100% of the mission area and ensure more excellent safety.
A Decentralized Cooperative Coverage Control for Networked Multiple UAVs Based on Deep Reinforcement Learning
13.10.2023
1188505 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Reinforcement Learning Based Coverage Planning for UAVs Fleets
TIBKAT | 2023
|Decentralized optimal control of multiple heterogeneous UAVs in cooperative target tracking
British Library Online Contents | 2010
|