Beyond Imitation: Constraint-Aware Trajectory Generation with Flow Matching For End-to-End Autonomous Driving

TL;DR

This paper introduces CATG, a flow matching-based planning framework for autonomous driving that incorporates explicit safety and physical constraints, improving trajectory diversity and adherence to rules.

Contribution

The paper's main novelty is integrating explicit constraints into flow matching for trajectory generation, enabling safety compliance and style control in autonomous driving.

Findings

Achieved 2nd place on NavSim v2 challenge with an EPDMS score of 51.31.

Successfully mitigated mode collapse and improved trajectory diversity.

Enabled explicit safety and kinematic constraints within the generative process.

Abstract

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.