Adaptive cooperative detection method for unmanned planetary vehicles based on deep reinforcement learning

Peng, Yan; Xiao, Huaibao; Guo, Jifeng; Bai, Chengchao; Zheng, Hongxing

doi:10.1109/icus48101.2019.8996016

Cited by 2 publications

(1 citation statement)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Researchers have proposed different biological-based algorithms to tackle path planning for space robots, such as the genetic algorithm [17]. In recent years, with the widespread application of artificial intelligence, many machine learning algorithms and deep learning algorithms have emerged in the research on path planning [18], [19]. Furthermore, some existing works have designed new algorithms to solve the path planning problem for space robots by integrating biological-based algorithms with other algorithms [20].…”

Section: Introductionmentioning

confidence: 99%

Joint Task Allocation and Path Planning for Space Robot

Sun

Liu

2023

IEEE Access

View full text Add to dashboard Cite

Space robots have a broad application prospect in the aerospace industry. It is difficult for space robots to keep running for a long time due to limited fuel. In addition, it is impossible to replenish the fuel for space robots at any time due to the unique working environment. Therefore, a proper path is crucial for the effective operation of the space robot. In this paper, we investigate the allocation of exploration tasks and the path planning of space robots jointly. The goal is to minimize the completion latency of exploration tasks. We propose two algorithms named subbranch insertion task allocation (SI-TA) and parallel search task allocation (PS-TA) to solve the problem. We also customize an algorithm named random path planning task allocation (RTA) as the baseline. At last, we implement extensive experiments to demonstrate that proposed algorithms can obtain lower completion latency than RTA. Compared with RTA, the proposed algorithms SI-TA and PS-TA can reduce completion latency by at least 20% and 40%, respectively. Moreover, both algorithms work more stably than RTA.

show abstract