A Simple Approach to Constraint-Aware Imitation Learning with Application to Autonomous Racing
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866918131645022208 |
|---|---|
| author | Cao, Shengfan Joa, Eunhyek Borrelli, Francesco |
| author_facet | Cao, Shengfan Joa, Eunhyek Borrelli, Francesco |
| contents | Guaranteeing constraint satisfaction is challenging in imitation learning (IL), particularly in tasks that require operating near a system's handling limits. Traditional IL methods, such as Behavior Cloning (BC), often struggle to enforce constraints, leading to suboptimal performance in high-precision tasks. In this paper, we present a simple approach to incorporating safety into the IL objective. Through simulations, we empirically validate our approach on an autonomous racing task with both full-state and image feedback, demonstrating improved constraint satisfaction and greater consistency in task performance compared to BC. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_07737 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A Simple Approach to Constraint-Aware Imitation Learning with Application to Autonomous Racing Cao, Shengfan Joa, Eunhyek Borrelli, Francesco Machine Learning Artificial Intelligence Robotics Guaranteeing constraint satisfaction is challenging in imitation learning (IL), particularly in tasks that require operating near a system's handling limits. Traditional IL methods, such as Behavior Cloning (BC), often struggle to enforce constraints, leading to suboptimal performance in high-precision tasks. In this paper, we present a simple approach to incorporating safety into the IL objective. Through simulations, we empirically validate our approach on an autonomous racing task with both full-state and image feedback, demonstrating improved constraint satisfaction and greater consistency in task performance compared to BC. |
| title | A Simple Approach to Constraint-Aware Imitation Learning with Application to Autonomous Racing |
| topic | Machine Learning Artificial Intelligence Robotics |
| url | https://arxiv.org/abs/2503.07737 |