Beyond Imitation: Constraint-Aware Trajectory Generation with Flow Matching For End-to-End Autonomous Driving
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908620022611968 |
|---|---|
| author | Liu, Lin Yu, Guanyi Song, Ziying Li, Junqiao Jia, Caiyan Jia, Feiyang Wu, Peiliang Luo, Yandan |
| author_facet | Liu, Lin Yu, Guanyi Song, Ziying Li, Junqiao Jia, Caiyan Jia, Feiyang Wu, Peiliang Luo, Yandan |
| contents | Planning is a critical component of end-to-end autonomous driving. However, prevailing imitation learning methods often suffer from mode collapse, failing to produce diverse trajectory hypotheses. Meanwhile, existing generative approaches struggle to incorporate crucial safety and physical constraints directly into the generative process, necessitating an additional optimization stage to refine their outputs. To address these limitations, we propose CATG, a novel planning framework that leverages Constrained Flow Matching. Concretely, CATG explicitly models the flow matching process, which inherently mitigates mode collapse and allows for flexible guidance from various conditioning signals. Our primary contribution is the novel imposition of explicit constraints directly within the flow matching process, ensuring that the generated trajectories adhere to vital safety and kinematic rules. Secondly, CATG parameterizes driving aggressiveness as a control signal during generation, enabling precise manipulation of trajectory style. Notably, on the NavSim v2 challenge, CATG achieved 2nd place with an EPDMS score of 51.31 and was honored with the Innovation Award. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_26292 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Beyond Imitation: Constraint-Aware Trajectory Generation with Flow Matching For End-to-End Autonomous Driving Liu, Lin Yu, Guanyi Song, Ziying Li, Junqiao Jia, Caiyan Jia, Feiyang Wu, Peiliang Luo, Yandan Computer Vision and Pattern Recognition Planning is a critical component of end-to-end autonomous driving. However, prevailing imitation learning methods often suffer from mode collapse, failing to produce diverse trajectory hypotheses. Meanwhile, existing generative approaches struggle to incorporate crucial safety and physical constraints directly into the generative process, necessitating an additional optimization stage to refine their outputs. To address these limitations, we propose CATG, a novel planning framework that leverages Constrained Flow Matching. Concretely, CATG explicitly models the flow matching process, which inherently mitigates mode collapse and allows for flexible guidance from various conditioning signals. Our primary contribution is the novel imposition of explicit constraints directly within the flow matching process, ensuring that the generated trajectories adhere to vital safety and kinematic rules. Secondly, CATG parameterizes driving aggressiveness as a control signal during generation, enabling precise manipulation of trajectory style. Notably, on the NavSim v2 challenge, CATG achieved 2nd place with an EPDMS score of 51.31 and was honored with the Innovation Award. |
| title | Beyond Imitation: Constraint-Aware Trajectory Generation with Flow Matching For End-to-End Autonomous Driving |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2510.26292 |