Journal article · 2026 · Open access

Adversarial-Training-Based Robust Reinforcement Learning Control for Autonomous Vehicle Trajectory Tracking

Donghyun Lee, Taeyoung Oh, J. H. Kim, Jinwoo Yoo

IEEE Access, 14, 43954–43975 · Published 1 January 2026 · DOI 10.1109/access.2026.3675864

Open full textCite this paper

In brief

“Adversarial-Training-Based Robust Reinforcement Learning Control for Autonomous Vehicle Trajectory Tracking” is a 2026 journal article by Donghyun Lee and 3 co-authors, published in IEEE Access. OpenAlex counts 3 citations as of 22 November 2026, which puts it in the top 4% of works published in 2026 by citations. It is open access under a CC BY licence: the full text is free to read at IEEE Access.

Citations
3
Most cited in
2026 (3)
Field-weighted impact
2.83
References
52
Authors
4

Abstract

Autonomous vehicles must maintain stable and accurate trajectory and speed tracking even in scenarios involving unpredictable disturbances and changing environmental conditions. Reinforcement learning (RL)–based controllers that directly output control inputs often exhibit performance degradation or instability when encountering out-of-distribution scenarios or disturbances. Model-based controllers are sensitive to model mismatch and parameter variations. Notably, even residual RL approaches offer limited robustness if disturbances are insufficiently represented during training. To address these challenges, this paper proposes an integrated tracking control architecture in which a linear quadratic regulator (LQR) serves as the nominal controller, while an RL policy provides residual compensation for steering and longitudinal (acceleration/deceleration) commands. To improve controller robustness, the training process is formulated as a min–max adversarial learning problem: The adversary learns a disturbance-injection policy while the controller minimizes tracking errors and ride-comfort penalties. Evaluation of the proposed method across a range of driving scenarios in the CARLA simulation environment demonstrates that it reduces trajectory errors and constraint violations compared with pure LQR, model predictive control, and an existing RL-based PID controller, both in the presence and absence of disturbances. These results demonstrate the improved tracking stability and robustness of the proposed controller.

Cite this paper

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

APA 7

Lee, D., Oh, T., Kim, J. H., & Yoo, J. (2026). Adversarial-Training-Based Robust Reinforcement Learning Control for Autonomous Vehicle Trajectory Tracking. IEEE Access, 14, 43954–43975. https://doi.org/10.1109/access.2026.3675864

MLA 9

Lee, Donghyun, et al. “Adversarial-Training-Based Robust Reinforcement Learning Control for Autonomous Vehicle Trajectory Tracking.” IEEE Access, vol. 14, 2026, pp. 43954–43975. https://doi.org/10.1109/access.2026.3675864.

Chicago

Lee, Donghyun, Taeyoung Oh, J. H. Kim, and Jinwoo Yoo. 2026. “Adversarial-Training-Based Robust Reinforcement Learning Control for Autonomous Vehicle Trajectory Tracking.” IEEE Access 14: 43954–43975. https://doi.org/10.1109/access.2026.3675864.

Harvard

Lee, D. et al. (2026) ‘Adversarial-Training-Based Robust Reinforcement Learning Control for Autonomous Vehicle Trajectory Tracking’, IEEE Access, 14, pp. 43954–43975. doi:10.1109/access.2026.3675864.

IEEE

D. Lee, T. Oh, J. H. Kim, and J. Yoo, “Adversarial-Training-Based Robust Reinforcement Learning Control for Autonomous Vehicle Trajectory Tracking,” IEEE Access, vol. 14, pp. 43954–43975, 2026, doi: 10.1109/access.2026.3675864.

BibTeX

@article{lee2026adversarial,
  title     = {{Adversarial-Training-Based Robust Reinforcement Learning Control for Autonomous Vehicle Trajectory Tracking}},
  author    = {Lee, Donghyun and Oh, Taeyoung and Kim, J. H. and Yoo, Jinwoo},
  journal   = {IEEE Access},
  year      = {2026},
  volume    = {14},
  pages     = {43954--43975},
  doi       = {10.1109/access.2026.3675864},
  url       = {https://doi.org/10.1109/access.2026.3675864}
}

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

Topic and keywords

Physical Sciences › Computer Science › Artificial Intelligence › Reinforcement Learning in Robotics

  • autonomous vehicles
  • trajectory tracking
  • robust reinforcement learning
  • adversarial training
  • minimax optimization
  • linear quadratic regulator
  • residual compensation
  • disturbance injection

Work with this paper

Most-cited papers in Computer Science

All highly cited Computer Science papers

Questions about this paper

How do I cite this paper in APA style?

Lee, D., Oh, T., Kim, J. H., & Yoo, J. (2026). Adversarial-Training-Based Robust Reinforcement Learning Control for Autonomous Vehicle Trajectory Tracking. IEEE Access, 14, 43954–43975. https://doi.org/10.1109/access.2026.3675864

How many times has this paper been cited?

3 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 IEEE Access: sign in and open the full text from this page.

Who wrote this paper?

Donghyun Lee, Taeyoung Oh, J. H. Kim, Jinwoo Yoo.

What is the DOI of this paper?

10.1109/access.2026.3675864, permanently at https://doi.org/10.1109/access.2026.3675864. 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.