Beyond Imitation: Learning Safe End-to-End Autonomous Driving from Hard Negatives
About
Existing imitation learning methods for end-to-end autonomous driving predominantly learn from successful demonstrations by minimizing geometric deviations from expert trajectories. This paradigm implicitly assumes that spatial proximity implies behavioral safety, leading to a critical objective mismatch: trajectories with nearly identical imitation losses may exhibit drastically different safety outcomes, where one remains recoverable while the other results in collision. To address this limitation, we propose BeyondDrive, a failure-aware imitation learning framework that jointly learns from successful and failed driving behaviors. First, we introduce a flow matching-based negative trajectory generator that synthesizes safety-critical yet expert-proximate trajectories, enabling explicit modeling of safety asymmetry. Second, we develop a diversity-aware sampling strategy that mitigates mode collapse and improves coverage of diverse failure modes during negative trajectory generation. Third, we propose a Repulsive Distance Loss that simultaneously attracts predictions toward expert demonstrations while repelling them from hard negative trajectories, thereby establishing discriminative safety boundaries in trajectory space. Applied to the uni-modal baseline Latent TransFuser, BeyondDrive achieves 89.7 PDMS on the NAVSIMv1 closed-loop benchmark, outperforming prior state-of-the-art methods. Moreover, BeyondDrive generalizes effectively across different autonomous driving architectures, including multi-modal planners, and further demonstrates strong zero-shot transferability on the HUGSIM benchmark.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Closed-loop Planning | NAVSIM v2 (test) | NC98.4 | 24 | |
| Autonomous Driving Planning | NAVSIM v1 (navtest) | NC98.7 | 24 | |
| Autonomous Driving | HUGSIM Easy | Reachability Coverage (RC)76.8 | 12 | |
| Autonomous Driving | HUGSIM Extreme | Reachability Coverage (RC)29.6 | 12 | |
| Autonomous Driving | HUGSIM Overall | Reachability Coverage (RC)46.2 | 12 | |
| Autonomous Driving | HUGSIM Medium | Reachability Coverage (RC)43 | 12 | |
| Autonomous Driving | HUGSIM Hard | Reachability Coverage (RC)35.5 | 12 |