Speaker
Description
Ensuring safety in the space environment is critical as human space activities, such as the Artemis program, become increasingly ambitious. In particular, solar energetic particle (SEP) events, triggered by solar flares (SFs) and coronal mass ejections (CMEs), pose significant risks to human health and space systems. To address these risks, Fujitsu Limited and ISEE, Nagoya University, have been collaborating since February 2023 on space weather research focused on predicting particle radiation, primarily for lunar, Mars, and deep space exploration. To further advance this initiative, we launched a collaborative research project with JAXA on February 1, 2025 (#1).
Conventional SEP forecasting methods have primarily relied on proton flux measurements in geostationary orbit. Recognizing the limitations of this approach for lunar and deep-space missions, our research group is working to develop predictive models based on in-situ observations and accurate radiation impact forecasting. This project employs AI to address these challenges.
First, we are developing an explainable AI (XAI) classification model for SEP event prediction using in-situ radiation data from lunar orbit. Second, we are constructing an AI-based regression model to estimate equivalent radiation doses from predicted SEP proton flux/fluence, incorporating the energy spectrum. This will facilitate discussions with JAXA regarding practical considerations for future missions.
(1) We are identifying SFs that cause radiation enhancements in lunar orbit using NASA's LRO/CRaTER data accessed via the CRaTER Web interface (#2). This unique in-situ dataset covers events from July 2009 to the present. We have identified SEP-induced SFs that have not been reported so far, in addition to known events (#3). We are developing a forecasting model for these SFs and associated SEPs using Fujitsu's XAI, "Fujitsu Kozuchi XAI WideLearning" (cf, ESWW2024, CD5.2 Kato et al.). This model provides event probability and prediction reasoning, valuable for operational decision-making. We will discuss our system's potential and advanced pre-flare forecasting models using ISEE NLFFF data.
(2) We are constructing an AI-based regression model to predict SEP proton profiles (flux, fluence, and energy spectrum) using data from SOHO/COSTEP EPHINE and GOES/EPAD, HEAPD, and SGPS. We plan to assess the potential impact of these events on spacecraft and astronauts and will present recent progress.
Finally, we will discuss future prospects for an integrated forecasting system, incorporating JAXA's developing detector. We aim to establish space weather forecasting and real-time dose evaluation using data from instruments like Lunar-RICheS and PS-TEPC (#4), planned for Artemis. We will report on AI collaboration discussions with JAXA's such compact, high-performance radiation measurement and real-time dose assessment for lunar utilization in addition to LRO/CRaTER data.
- https://www.ihub-tansa.jaxa.jp/english/RFP_announcement12_en.html
- https://crater-web.sr.unh.edu/data/craterProducts/events/events.txt
- Rotti et al. 2022, ApJS 262 29
- https://www.nagoya-u.ac.jp/researchinfo/result/2024/11/post-750.html
| Do you plan to attend in-person or online? | In-person |
|---|
