ForeDrive: Foresight-Guided End-to-End Autonomous Driving with a Planning-Relevant Latent World Model

原文：ForeDrive: Foresight-Guided End-to-End Autonomous Driving with a Planning-Relevant Latent World Model

作者：Sinuo Wang, Zichong Gu, Yuhan Huang, Wenxin Wen, Xun Yang, Yiqing Zhang, Xingyu Zhang, Ningyu Che, Jie Ling, Qiankun Yu, Wei Liu, Jing Xu, Xinggang Wang

来源：arXiv cs.CV（计算机视觉）

正文

Computer Science > Computer Vision and Pattern Recognition

arXiv:2609.26299v1 (cs)

[Submitted on 22 Sep 2026]

Title:ForeDrive: Foresight-Guided End-to-End Autonomous Driving with a Planning-Relevant Latent World Model

Authors:Sinuo Wang, Zichong Gu, Yuhan Huang, Wenxin Wen, Xun Yang, Yiqing Zhang, Xingyu Zhang, Ningyu Che, Jie Ling, Qiankun Yu, Wei Liu, Jing Xu, Xinggang Wang

View a PDF of the paper titled ForeDrive: Foresight-Guided End-to-End Autonomous Driving with a Planning-Relevant Latent World Model, by Sinuo Wang and 12 other authors

View PDF

HTML (experimental)

Abstract:Existing latent world models are typically optimized for future predictability, yet the resulting representations are not necessarily useful for planning in autonomous driving. Predictions are commonly used for pretraining or auxiliary supervision rather than as direct conditioning signals for trajectory generation. We propose ForeDrive, which learns a planning-relevant latent representation and couples it asymmetrically to a Diffusion Transformer (DiT) planner. The planner consumes multi-horizon latent future representations learned with a JEPA-style world model; planning gradients update the shared online encoder, while stop-gradient routing trains the latent predictor with forecasting losses only. Because predicted futures have varying reliability across horizons and BEV trajectories are misaligned with image tokens, we use gated visual fusion, future-status injection, and Trajectory-Adaptive Bias (TAB) to inject future latents as guidance without overriding the current observation. Trained with pure imitation learning and using only the current front-view image as visual input at inference, ForeDrive attains 89.9 PDMS on NAVSIM v1 and 90.0 one-stage EPDMS on NAVSIM v2, without reinforcement learning or an external trajectory scorer.

Comments:

9 pages, 4 figures; 8 pages supplementary with 4 figures

Subjects:

Computer Vision and Pattern Recognition (cs.CV)

ACM classes:

I.2.9; I.2.10; I.2.6

Cite as:

arXiv:2609.26299 [cs.CV]

(or

arXiv:2609.26299v1 [cs.CV] for this version)

https://doi.org/10.48550/arXiv.2609.26299

Focus to learn more

arXiv-issued DOI via DataCite (pending registration)

主题

计算机视觉

由「前沿雷达」于 2026-09-22 采集。正文取自原文页面，已保留出处链接。标题与正文版权归原作者所有。