Multi-Information Fusion-Based Fault Diagnosis for Satellite Formation

Considering the incomplete measurement information of a single satellite, a cluster-level fault diagnosis method based on deep reinforcement learning is proposed in this paper. Through the multi-information fusion of satellite formation, the relative position and satellite sensitivity of neighboring...

Full description

Saved in:
Bibliographic Details
Published in2023 42nd Chinese Control Conference (CCC) pp. 5046 - 5051
Main Authors Leng, Jiajun, Zhang, Xiuyun, Zhang, Ruilong, Liu, Da, Zong, Qun
Format Conference Proceeding
LanguageEnglish
Published Technical Committee on Control Theory, Chinese Association of Automation 24.07.2023
Subjects
Decision making
Deep learning
deep reinforcement learning
Fault diagnosis
Heuristic algorithms
multi-agent systems
Reinforcement learning
satellite formation
Satellites
Sensitivity
Online AccessGet full text

Cover

More Information
Summary:Considering the incomplete measurement information of a single satellite, a cluster-level fault diagnosis method based on deep reinforcement learning is proposed in this paper. Through the multi-information fusion of satellite formation, the relative position and satellite sensitivity of neighboring satellites are further considered to realize the in-orbit diagnosis and monitoring of the entire satellite formation. The proposed strategy can realize autonomous decision-making and increase the intelligence and generalization of fault diagnosis. What's more, there is no need to retrain the network when the number of satellites changes. The effectiveness of the proposed method is verified by a simulation example, and the results show that the algorithm has fast convergence speed and high fault recognition accuracy.
ISSN:2161-2927
DOI:10.23919/CCC58697.2023.10240739