タイトル（掲載誌）JAXA Research and Development Report: Journal of Space Science Informatics Japan: Volume 14
一般注記出版タイプ： VoR
The next-generation quick-look system being developed at ISAS/JAXA enables operators to swiftly check the status of scientific satellites and spacecraft. This quick-look system has been firstly used in SLIM, JAXA’s first and latest moon lander, which landed successfully on the moon in January 2024. While this system was able to show crucial data about health and operations of the spacecraft, it did not incorporate active anomaly detection.
An anomaly could be a telltale sign of imminent failure. Therefore, we tried to implement anomaly detection in this quick-look system to benefit current and future JAXA missions. Two approaches have been tested. The first one uses the time series analysis algorithm SARIMA for data prediction on relatively long time scales. The second one uses a custom method based on statistical techniques to detect anomalies on relatively short time scales. Archival data from the Suzaku satellite, where the satellite was known to have suffered from eventual battery degradation, was used to demonstrate the effectiveness of each approach. The results show that both methods are able to effectively detect the battery anomalies on their respective time scales. Elaboration is given about expanding the method for application to current missions like the XRISM space telescope.
Physical characteristics: Original contains color illustrations
形態: カラー図版あり
レポート番号: JAXA-RR-24-004
連携機関・データベース国立情報学研究所 : 学術機関リポジトリデータベース（IRDB）（機関リポジトリ）
提供元機関・データベース宇宙航空研究開発機構 : 宇宙航空研究開発機構リポジトリ