Unsupervised anomaly detection of polluted telemetry data based on a reconstruction model with adaptive weighting

Zhaoping Xu; Zhijun Cheng; Bo Guo

doi:10.1117/12.3025480

28 February 2024 Unsupervised anomaly detection of polluted telemetry data based on a reconstruction model with adaptive weighting

Zhaoping Xu, Zhijun Cheng, Bo Guo

Author Affiliations +

Proceedings Volume 13071, International Conference on Mechatronic Engineering and Artificial Intelligence (MEAI 2023); 130712K (2024) https://doi.org/10.1117/12.3025480
Event: International Conference on Mechatronic Engineering and Artificial Intelligence (MEAI 2023), 2023, Shenyang, China

Abstract

Anomaly detection of telemetry data is a promising way for ensuring the reliability of spacecraft and success of space mission. This study proposes a reconstruction model with adaptive weighting for unsupervised anomaly detection of polluted telemetry data. The reconstruction model integrates the advantages of typical reconstruction models and can generate high-quality results. An adaptive weighting module is designed to against the data pollution issue by assigning different weights to training data samples. Experiments are conducted on two public telemetry datasets, and the results verify the effectiveness of our method.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Zhaoping Xu, Zhijun Cheng, and Bo Guo "Unsupervised anomaly detection of polluted telemetry data based on a reconstruction model with adaptive weighting", Proc. SPIE 13071, International Conference on Mechatronic Engineering and Artificial Intelligence (MEAI 2023), 130712K (28 February 2024); https://doi.org/10.1117/12.3025480

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available