Skip to content
AI News HubLIVE
Original source2 min read

Long-Horizon Consistent and Interaction-Aware World Models for Multi-Style End-to-End Driving

Summary

Yuxuan Han and five co-authors propose StyleDrive, a world-model reinforcement learning framework for end-to-end autonomous driving. It combines temporal consistency regularization via gated cross-attention, explicit disentanglement of ego-relevant and ego-irrelevant traffic states, and Group Relative Policy Optimization for multi-style driving without retraining. On Bench2Drive it reports a driving score of 88.44 and a success rate of 66.82, plus promising sim-to-real transfer on a real AGV platform.

SourcearXiv RoboticsAuthor: Yuxuan Han, Kunyuan Wu, Liyunong Yang, Zilu Wang, Cansen Jiang, Yi Xiao, Liang Hu
Long-Horizon Consistent and Interaction-Aware World Models for Multi-Style End-to-End Driving
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[Submitted on 3 Sep 2026]

Title:Long-Horizon Consistent and Interaction-Aware World Models for Multi-Style End-to-End Driving

View a PDF of the paper titled Long-Horizon Consistent and Interaction-Aware World Models for Multi-Style End-to-End Driving, by Yuxuan Han and 5 other authors

View PDF HTML (experimental)

Abstract:End-to-end autonomous driving has increasingly adopted world model-based reinforcement learning frameworks to improve learning efficiency through \textit{imagined rollouts}. However, existing world models suffer from three key limitations: temporal inconsistency in long-horizon imagined rollouts, inadequate modeling of ego-environment interactions, and limited adaptability to diverse driving styles. To address these challenges, we propose \textit{StyleDrive}, a world-model-based learning framework that jointly enforces long-horizon consistency, explicitly disentangles interactive traffic states, and supports multi-style policy optimization within a unified learning paradigm. First, we introduce a temporal consistency regularization that integrates historical latent states through gated cross-attention, stabilizing long-horizon imagined rollouts and mitigating error accumulation. Second, we design an explicit state disentanglement module that separates ego-relevant from ego-irrelevant interactive states, enabling more interpretable and efficient decision-making in complex traffic scenarios. Third, we enable multi-style driving behaviors through Group Relative Policy Optimization, which replaces per-step reward optimization with trajectory-wise relative advantages, reducing reward variance and supporting diverse driving styles without retraining. We evaluate StyleDrive on the Bench2Drive closed-loop driving benchmark, achieving a driving score of 88.44 (+17.08 over the previous best world model-based method) and a success rate of 66.82 (+16.58). Furthermore, we deploy StyleDrive on a real automated guided vehicle platform and demonstrate promising sim-to-real transfer capability in dynamic driving scenarios.

Subjects:

Robotics (cs.RO)

Cite as: arXiv:2609.03225 [cs.RO]

(or arXiv:2609.03225v1 [cs.RO] for this version)

https://doi.org/10.48550/arXiv.2609.03225

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Yuxuan Han [view email] [v1] Thu, 3 Sep 2026 00:00:39 UTC (20,019 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled Long-Horizon Consistent and Interaction-Aware World Models for Multi-Style End-to-End Driving, by Yuxuan Han and 5 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.RO

new | recent | 2026-09

Change to browse by:

cs

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

ResearchersAdvanced

Key points

  • Gated cross-attention integrates historical latent states to stabilize long-horizon imagined rollouts and curb error accumulation
  • An explicit state disentanglement module separates ego-relevant from ego-irrelevant interactive states for more interpretable and efficient decisions
  • Group Relative Policy Optimization replaces per-step rewards with trajectory-wise relative advantages, reducing variance and enabling multiple driving styles without retraining
  • Bench2Drive results show 88.44 driving score (+17.08) and 66.82 success rate (+16.58), with sim-to-real deployment on a real AGV

Highlights and analysis are generated automatically and may contain errors. Check the original source.