DriveLaW:Unifying Planning and Video Generation in a Latent Driving World

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Xia, Tianze, Li, Yongkang, Zhou, Lijun, Yao, Jingfeng, Xiong, Kaixin, Sun, Haiyang, Wang, Bing, Ma, Kun, Chen, Guang, Ye, Hangjun, Liu, Wenyu, Wang, Xinggang
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866917414985269248
author Xia, Tianze
Li, Yongkang
Zhou, Lijun
Yao, Jingfeng
Xiong, Kaixin
Sun, Haiyang
Wang, Bing
Ma, Kun
Chen, Guang
Ye, Hangjun
Liu, Wenyu
Wang, Xinggang
author_facet Xia, Tianze
Li, Yongkang
Zhou, Lijun
Yao, Jingfeng
Xiong, Kaixin
Sun, Haiyang
Wang, Bing
Ma, Kun
Chen, Guang
Ye, Hangjun
Liu, Wenyu
Wang, Xinggang
contents World models have become crucial for autonomous driving, as they learn how scenarios evolve over time to address the long-tail challenges of the real world. However, current approaches relegate world models to limited roles: they operate within ostensibly unified architectures that still keep world prediction and motion planning as decoupled processes. To bridge this gap, we propose DriveLaW, a novel paradigm that unifies video generation and motion planning. By directly injecting the latent representation from its video generator into the planner, DriveLaW ensures inherent consistency between high-fidelity future generation and reliable trajectory planning. Specifically, DriveLaW consists of two core components: DriveLaW-Video, our powerful world model that generates high-fidelity forecasting with expressive latent representations, and DriveLaW-Act, a diffusion planner that generates consistent and reliable trajectories from the latent of DriveLaW-Video, with both components optimized by a three-stage progressive training strategy. The power of our unified paradigm is demonstrated by new state-of-the-art results across both tasks. DriveLaW not only advances video prediction significantly, surpassing best-performing work by 33.3% in FID and 1.8% in FVD, but also achieves a new record on the NAVSIM planning benchmark.
format Preprint
id arxiv_https___arxiv_org_abs_2512_23421
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DriveLaW:Unifying Planning and Video Generation in a Latent Driving World
Xia, Tianze
Li, Yongkang
Zhou, Lijun
Yao, Jingfeng
Xiong, Kaixin
Sun, Haiyang
Wang, Bing
Ma, Kun
Chen, Guang
Ye, Hangjun
Liu, Wenyu
Wang, Xinggang
Computer Vision and Pattern Recognition
World models have become crucial for autonomous driving, as they learn how scenarios evolve over time to address the long-tail challenges of the real world. However, current approaches relegate world models to limited roles: they operate within ostensibly unified architectures that still keep world prediction and motion planning as decoupled processes. To bridge this gap, we propose DriveLaW, a novel paradigm that unifies video generation and motion planning. By directly injecting the latent representation from its video generator into the planner, DriveLaW ensures inherent consistency between high-fidelity future generation and reliable trajectory planning. Specifically, DriveLaW consists of two core components: DriveLaW-Video, our powerful world model that generates high-fidelity forecasting with expressive latent representations, and DriveLaW-Act, a diffusion planner that generates consistent and reliable trajectories from the latent of DriveLaW-Video, with both components optimized by a three-stage progressive training strategy. The power of our unified paradigm is demonstrated by new state-of-the-art results across both tasks. DriveLaW not only advances video prediction significantly, surpassing best-performing work by 33.3% in FID and 1.8% in FVD, but also achieves a new record on the NAVSIM planning benchmark.
title DriveLaW:Unifying Planning and Video Generation in a Latent Driving World
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.23421