TY - JOUR
T1 - A Multi-robot Lunar Area Coverage Method Based on Deep Reinforcement Learning
AU - Guo, Yufei
AU - Zheng, Zixuan
AU - Liang, Qiming
AU - Yuan, Jianping
N1 - Publisher Copyright:
Copyright © 2023 by the International Astronautical Federation (IAF). All rights reserved.
PY - 2023
Y1 - 2023
N2 - When exploring celestial bodies like the Moon, lunar surface coverage becomes a critical task, necessitating multiple robots to cover as much area as possible on the lunar surface. However, the intricate lunar terrain and environmental uncertainties render traditional area coverage methods ineffective. Therefore, this paper proposes a multi-robot lunar surface area coverage approach based on deep reinforcement learning (DRL). The approach consists of two phases: the training phase and the execution phase. During the training phase, robots learn a strategy to maximize lunar surface area coverage, employing the multi-agent deep deterministic policy gradient (MADDPG) algorithm for deep reinforcement learning. In the execution phase, robots execute movements based on their current state and learned strategy to achieve lunar surface coverage. Furthermore, a series of simulation experiments were conducted, with coverage rate during the execution phase serving as the evaluation metric. Experimental results demonstrate that our approach significantly enhances lunar surface area coverage compared to traditional methods. In the future, we will further optimize the algorithm to accomplish more efficient and intelligent lunar exploration missions.
AB - When exploring celestial bodies like the Moon, lunar surface coverage becomes a critical task, necessitating multiple robots to cover as much area as possible on the lunar surface. However, the intricate lunar terrain and environmental uncertainties render traditional area coverage methods ineffective. Therefore, this paper proposes a multi-robot lunar surface area coverage approach based on deep reinforcement learning (DRL). The approach consists of two phases: the training phase and the execution phase. During the training phase, robots learn a strategy to maximize lunar surface area coverage, employing the multi-agent deep deterministic policy gradient (MADDPG) algorithm for deep reinforcement learning. In the execution phase, robots execute movements based on their current state and learned strategy to achieve lunar surface coverage. Furthermore, a series of simulation experiments were conducted, with coverage rate during the execution phase serving as the evaluation metric. Experimental results demonstrate that our approach significantly enhances lunar surface area coverage compared to traditional methods. In the future, we will further optimize the algorithm to accomplish more efficient and intelligent lunar exploration missions.
KW - area coverage
KW - deep reinforcement learning
KW - Multi-robot
UR - http://www.scopus.com/inward/record.url?scp=85187988695&partnerID=8YFLogxK
M3 - 会议文章
AN - SCOPUS:85187988695
SN - 0074-1795
VL - 2023-October
JO - Proceedings of the International Astronautical Congress, IAC
JF - Proceedings of the International Astronautical Congress, IAC
T2 - 74th International Astronautical Congress, IAC 2023
Y2 - 2 October 2023 through 6 October 2023
ER -