This work proposes FPD-Drive, an end-to-end planner that distills privileged future knowledge into a visual driving policy to improve both multi-trajectory supervision and safety selection, and introduces Future-Privileged Multi-Target Distillation to transfer diverse, high-quality trajectory targets from a teacher with access to recorded future observations to the student.
Abstract
End-to-end autonomous driving models are typically trained on real-world driving logs, which record only the trajectory actually executed in each scene. This provides limited supervision for learning alternative valid behaviors. Multi-target supervision can mitigate this limitation, but broader trajectory coverage makes safe candidate selection more challenging. We propose FPD-Drive, an end-to-end planner that distills privileged future knowledge into a visual driving policy to improve both multi-trajectory supervision and safety selection. Specifically, we introduce Future-Privileged Multi-Target Distillation to transfer diverse, high-quality trajectory targets from a teacher with access to recorded future observations to the student. These targets extend supervision beyond the single recorded human trajectory. For candidate selection, we further introduce Safety-Aware Score Correction to correct ranking errors between nearby safe and unsafe candidates and raise underestimated safety scores of safe candidates. At deployment, the student requires neither future observations nor teacher inference. Our base model achieves 94.9 PDMS on NAVSIM v1 navtest and 55.9 EPDMS on NAVSIM v2 navhard. The scaled FPD-Drive-Pro improves these scores to 95.5 PDMS and 58.1 EPDMS, respectively, achieving SOTA performance among the compared methods on both benchmarks. Zero-shot closed-loop evaluation on HUGSIM further yields an HD-Score of 45.1 and RC of 55.3.
This work proposes RFWorld, a Rollout-Free World model for end-to-end trajectory evaluation, which constructs a shared scene memory and shapes it into a future-readable representation via lightweight training-time temporal readout with semantic supervision, keeping training overhead low.
Closed-Loop Aware Planning for Autonomous Driving (CLAP-AD), a plug-in planning head that integrates with existing end-to-end driving models and is the first to enable explicit test-time control over their speed selection without speed-conditioned training, is introduced.
Emad Aghajanzadeh, Thomas Schwarzfischer, Marc Kaufeld et al.· 0 citations
Vision–Language–Action (VLA) models have shown strong potential for end-to-end autonomous driving, yet their post-training commonly relies on expensive simulator interaction or global policy updates. For an already competent pretrained policy, targeted correction is substantially more economical than repeated simulator...
Yun-Han Xu, Ao Xu· Italian National Conference...· 0 citations
World-action models (WAMs) jointly model future traffic states and driving actions for end-to-end autonomous driving in intelligent transportation systems. However, their action learning remains dominated by expert-trajectory imitation, leaving unsafe alternatives close to expert trajectories insufficiently constrained...
Zi-Hao Zhang, Xin-Yu Zhang, Bing Zeng et al.· 0 citations
End-to-end (E2E) autonomous driving integrates perception, prediction, and planning, yet often lacks explicit supervision for high-level driving semantics. We propose Task-Decomposed Language Supervision (TaLS), a training-only framework that organizes DriveLM question–answer annotations into four sequential stages: Ob...
Jiwon Lee, Seok-Cheol Kee· Italian National Conference...· 0 citations
This work proposes FeasibleFlow, a one-step end-to-end generative framework that jointly transports a configuration-space feasibility field and multimodal ego trajectories and introduces the Anchor-relative ranker (ARR) and Pareto-ReinFlow to balance safety and progress in candidate selection and generation.