2023 Fiscal Year Research-status Report
Deep Learning for Planetary Rover Localization
Project/Area Number |
20K14706
|
Research Institution | National Astronomical Observatory of Japan |
Principal Investigator |
Wu Benjamin 国立天文台, アルマプロジェクト, 特別客員研究員 (50868718)
|
Project Period (FY) |
2020-04-01 – 2025-03-31
|
Keywords | machine learning / computer vision / deep learning / numerical simulation / orbital dynamics |
Outline of Annual Research Achievements |
We are continuing to apply machine learning to problems in space research and exploration and establish deep learning computational hardware capabilities at NAOJ. In FY2023, scientists from NASA Ames initiated bi-weekly collaboration meetings regarding orbital debris remediation, largely based on techniques from our FY2022 paper "Low-thrust rendezvous trajectory generation for multi-target active space debris removal using the RQ-Law". We also continued to develop the computer vision based methods for lunar rover absolute localization.
|
Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
The original research topic of rover localization progressed slower than planned. However, this is due to several new research directions emerging, which have been very fruitful with multiple publications and collaborations.
|
Strategy for Future Research Activity |
For rover localization, we are exploring the use of vision transformers compared to the previous ResNet baseline. Due high interest in our work on orbital dynamics, we have pursued research collaborations in this direction. We are extending this work by scaling up the number of debris objects that can be targeted for deorbit.
|
Causes of Carryover |
The original plan for the 4th year funding was to support conference and publication expenses and hardware upgrades to the GPU cluster. This did not occur in FY2023, so the remaining budget is planned to be used for these purposes for FY2024.
|