paper-with-me

홈 › Papers

SUV: Future Scene Understanding as Video Generation for End-to-End Driving

2026-08-04 · Yibo Yuan, Jiacheng Fu, Jiangtong Zhu, Yi Li, Jianhua Han, Meng Tian, Zhuohan Liu, Zhiwei Xiong, Hang Xu, Jianwu Fang, Jianru Xue arxiv

End-to-end driving requires a coherent understanding of future scenes, yet existing methods model these scenes using task-specific heads and output formats, with limited scalability. Can video generation instead provide a shared predictor? We introduce SUV, a unified end-to-end driving framework that casts future Scene Understanding as Video generation using a pretrained video foundation model. SUV models future appearance, semantics, relative depth, and instance-level dynamics as video streams with a shared video expert, without stream-specific visual prediction heads. Through joint video-action attention, the action expert attends to the latent representations of all future streams and generates the ego trajectory. Experiments show that SUV directly predicts all four future streams, while controlled ablations show that structured future supervision and direct future-stream access yield higher trajectory planning scores. With only a single front camera and no candidate-trajectory selection, SUV outperforms a broad set of recent state-of-the-art methods on both NAVSIM-v2 splits, achieving 91.0 EPDMS on navtest and 36.9 on navhard. On the long-tail WOD-E2E benchmark, SUV achieves a competitive RFS of 7.94.

📄 PDF Abstract BibTeX arXiv:2608.03084

Code (1)

ASH-2046/SUV ★ 6

Tasks

Scene UnderstandingTrajectory PlanningVideo Generation

Similar Papers 제목 키워드 기반

LMGenDrive: Bridging Multimodal Understanding and Generative World Modeling for End-to-End Driving

2026-04-09 · Hao Shao, Letian Wang, Yang Zhou, Yuxuan Hu 외 arxiv

Recent years have seen remarkable progress in autonomous driving, yet generalization to long-tail and open-world scenarios remains a major bottleneck for large-scale deployment. To address this challenge, some works use …

Instruction FollowingAutonomous DrivingVideo GenerationVideo Prediction

HERMES: A Unified Self-Driving World Model for Simultaneous 3D Scene Understanding and Generation

2025-01-24 · Xin Zhou, Dingkang Liang, Sifan Tu, Xiwu Chen 외

Driving World Models (DWMs) have become essential for autonomous driving by enabling future scene prediction. However, existing DWMs are limited to scene generation and fail to incorporate scene understanding, which invo…

Autonomous DrivingLanguage ModelingLanguage ModellingLarge Language Model+3

MoVieDrive: Urban Scene Synthesis with Multi-Modal Multi-View Video Diffusion Transformer

2025-08-20 · Guile Wu, David Huang, Dongfeng Bai, Bingbing Liu arxiv

Urban scene synthesis with video generation models has recently shown great potential for autonomous driving. Existing video generation approaches to autonomous driving primarily focus on RGB video generation and lack th…

Scene UnderstandingAutonomous DrivingVideo Generation

Stag-1: Towards Realistic 4D Driving Simulation with Video Generation Model

2024-12-06 · Lening Wang, Wenzhao Zheng, Dalong Du, Yunpeng Zhang 외

4D driving simulation is essential for developing realistic autonomous driving simulators. Despite advancements in existing methods for generating driving scenes, significant challenges remain in view transformation and …

Autonomous DrivingAutonomous VehiclesScene UnderstandingVideo Generation

UniDrive-WM: Unified Understanding, Planning and Generation World Model for Autonomous Driving

2026-01-07 · Zhexiao Xiong, Xin Ye, Burhan Yaman, Sheng Cheng 외 arxiv

World models have become central to autonomous driving, where accurate scene understanding and future prediction are crucial for safe control. Recent work has explored using vision-language models (VLMs) for planning, ye…

Scene UnderstandingTrajectory PlanningAutonomous DrivingImage Generation