Evaluating Robustness and Adaptability in Learning-Based Mission Planning for Active Debris Removal
Agni Bandyopadhyay, Günther Waxenegger-Wilfing · 2026
Read Evaluating Robustness and Adaptability in Learning-Based Mission Planning for Active Debris Removal by Agni Bandyopadhyay, Günther Waxenegger-Wilfing (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-05091.txt.
Description
A debris removal planner that excels in training conditions may stumble when fuel runs short or time runs out. Agni Bandyopadhyay and Günther Waxenegger-Wilfing compare three planners for multi-debris rendezvous in this 2026 paper, presented at a space robotics conference in Sendai. All three must respect strict limits on fuel and mission duration in low Earth orbit.
The contenders are a learning-based policy trained under fixed conditions, the same policy trained with randomized conditions, and a classic Monte Carlo tree search. Across 300 simulated cases with refueling and realistic transfer dynamics, the fixed-condition learner won when conditions matched its training but degraded sharply when they shifted. The randomized learner adapted better for a modest cost at nominal settings. Tree search handled changes best of all, thanks to online replanning, but took far longer to compute. The authors suggest hybrids.
Catalog record
| CREATOR | Agni Bandyopadhyay, Günther Waxenegger-Wilfing |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 24 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-05091