Journal article · 2026 · Open access

LA-RL: Language Action-guided Reinforcement Learning with Safety Guarantees for Autonomous Highway Driving

Yiming Shu, Wenbin Mai, Jiahui Xu, Jiwei Tang, Ruiyang Gao, Chen Sun

IEEE Transactions on Vehicular Technology, 1–15 · Published 1 January 2026 · DOI 10.1109/tvt.2026.3747776

Open PDFView at the publisherCite this paper

In brief

“LA-RL: Language Action-guided Reinforcement Learning with Safety Guarantees for Autonomous Highway Driving” is a 2026 journal article by Yiming Shu and 5 co-authors, published in IEEE Transactions on Vehicular Technology. OpenAlex counts 1 citations as of 22 November 2026. It is open access: the full text is free to read at arXiv (Cornell University).

Citations
1
Most cited in
2026 (1)
Field-weighted impact
1.09
References
35
Authors
6

Abstract

Autonomous highway driving demands a critical balance between proactive, efficiency-seeking behavior and robust safety guarantees. This paper proposes Language Action-guided Reinforcement Learning (LA-RL) with Safety Guarantees, a novel framework that integrates the semantic reasoning of large language models (LLMs) into the actor-critic architecture with an improved safety layer. Within this framework, task-specific reward shaping harmonizes the dual objectives of maximizing driving efficiency and ensuring safety, guiding decision-making based on both environmental insights and clearly defined goals. To enhance safety, LA-RL incorporates a safety-critical planner that combines model predictive control (MPC) with discrete control barrier functions (DCBFs). This layer constrains the LLM-informed policy to a safe action set under a nominal vehicle kinematic model within a designated state space, employs a slack mechanism that enhances solution feasibility against model mismatches and traffic disturbances, prevents overly conservative behavior and allows for greater policy exploration and behavioral flexibility under dynamic environments without compromising safety. Extensive experiments demonstrate that it significantly outperforms several current state-of-the-art methods, offering a more adaptive, reliable, and robust solution for autonomous highway driving. Compared to existing SOTA, it achieves ap proximately 20% higher success rate than the knowledge graph (KG) based baseline and about 30% higher than the retrieval augmented generation (RAG) based baseline. In low-density environments, LA-RL achieves a 100% success rate. These results confirm its enhanced exploration of the state-action space and its ability to autonomously adopt more efficient, proactive strategies in complex, mixed-traffic highway environments.

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

Shu, Y., Mai, W., Xu, J., Tang, J., Gao, R., & Sun, C. (2026). LA-RL: Language Action-guided Reinforcement Learning with Safety Guarantees for Autonomous Highway Driving. IEEE Transactions on Vehicular Technology, 1–15. https://doi.org/10.1109/tvt.2026.3747776

MLA 9

Shu, Yiming, et al. “LA-RL: Language Action-guided Reinforcement Learning with Safety Guarantees for Autonomous Highway Driving.” IEEE Transactions on Vehicular Technology, 2026, pp. 1–15. https://doi.org/10.1109/tvt.2026.3747776.

Chicago

Shu, Yiming, Wenbin Mai, Jiahui Xu, Jiwei Tang, Ruiyang Gao, and Chen Sun. 2026. “LA-RL: Language Action-guided Reinforcement Learning with Safety Guarantees for Autonomous Highway Driving.” IEEE Transactions on Vehicular Technology: 1–15. https://doi.org/10.1109/tvt.2026.3747776.

Harvard

Shu, Y. et al. (2026) ‘LA-RL: Language Action-guided Reinforcement Learning with Safety Guarantees for Autonomous Highway Driving’, IEEE Transactions on Vehicular Technology, pp. 1–15. doi:10.1109/tvt.2026.3747776.

IEEE

Y. Shu, W. Mai, J. Xu, J. Tang, R. Gao, and C. Sun, “LA-RL: Language Action-guided Reinforcement Learning with Safety Guarantees for Autonomous Highway Driving,” IEEE Transactions on Vehicular Technology, pp. 1–15, 2026, doi: 10.1109/tvt.2026.3747776.

BibTeX

@article{shu2026language,
  title     = {{LA-RL: Language Action-guided Reinforcement Learning with Safety Guarantees for Autonomous Highway Driving}},
  author    = {Shu, Yiming and Mai, Wenbin and Xu, Jiahui and Tang, Jiwei and Gao, Ruiyang and Sun, Chen},
  journal   = {IEEE Transactions on Vehicular Technology},
  year      = {2026},
  pages     = {1--15},
  doi       = {10.1109/tvt.2026.3747776},
  url       = {https://doi.org/10.1109/tvt.2026.3747776}
}

Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)

Topic and keywords

Physical Sciences › Engineering › Automotive Engineering › Autonomous Vehicle Technology and Safety

  • large language models
  • model predictive control
  • reward shaping
  • actor-critic architecture
  • safety guarantees
  • mixed traffic
  • vehicle kinematic model

Work with this paper

Most-cited papers in Engineering

All highly cited Engineering papers

Questions about this paper

How do I cite this paper in APA style?

Shu, Y., Mai, W., Xu, J., Tang, J., Gao, R., & Sun, C. (2026). LA-RL: Language Action-guided Reinforcement Learning with Safety Guarantees for Autonomous Highway Driving. IEEE Transactions on Vehicular Technology, 1–15. https://doi.org/10.1109/tvt.2026.3747776

How many times has this paper been cited?

1 times, according to OpenAlex on 22 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.

Is this paper free to read?

Yes. It is open access at arXiv (Cornell University): sign in and open the PDF from this page.

Who wrote this paper?

Yiming Shu, Wenbin Mai, Jiahui Xu, Jiwei Tang, Ruiyang Gao, Chen Sun.

What is the DOI of this paper?

10.1109/tvt.2026.3747776, permanently at https://doi.org/10.1109/tvt.2026.3747776. Sign in and open it from this page.

Bibliographic data from OpenAlex (CC0), last checked 22 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.