Moving On, Even When You’re Broken: Fail-Active Trajectory Generation via Diffusion Policies Conditioned on Embodiment and Task
Gilberto G. Briscoe-Martinez, Yaashia Gautam, Rahul Shetty, Anuj Pasricha, Marco M. Nicotra, Alessandro Roncone · 2026
Read Moving On, Even When You’re Broken: Fail-Active Trajectory Generation via Diffusion Policies Conditioned on Embodiment and Task by Gilberto G. Briscoe-Martinez, Yaashia Gautam, Rahul Shetty, Anuj Pasricha, Marco M. Nicotra, Alessandro Roncone (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-2602-02895.txt.
Description
Damaged robots that keep working are the vision behind a 2026 paper by Gilberto G. Briscoe-Martinez and 5 co-authors. Most robot failures stop the task and need human rescue. The authors focus on actuation failures, where motors or joints no longer move as commanded, and aim for fail-active operation instead.
Their system, DEFT, generates movement paths with a diffusion policy. It takes the robot's current damaged state and task limits as inputs. Tests used a seven-joint arm in simulation and in the real world. DEFT beat classical planners across thousands of failure cases. It also handled failures never seen in training, and completed drawer and whiteboard tasks where classical methods failed.
Catalog record
| CREATOR | Gilberto G. Briscoe-Martinez, Yaashia Gautam, Rahul Shetty, Anuj Pasricha, Marco M. Nicotra, Alessandro Roncone |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-02895