EUVIMEDEUVIMED entdecken
Crossref

STTRL-DVO: Transformer-Based Reinforcement Learning for Robust Dynamic Target Tracking in Cluttered Environments

Fanghao Wang, Binghong Chen, Youchao Zhang, Xiangyu Guo, Yining Lyu, Chuanjie Liu, Alois Knoll, Di Cui, Huanyu Jiang, Yibin Ying, Mingchuan Zhou

IEEE Transactions on Robotics · 2026 · 42 · 2555-2575

DOI: 10.1109/tro.2026.3701578 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Robotics, 2026. Verfasst von Fanghao Wang und 10 weiteren Autor:innen. Die Arbeit nennt 45 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Key Research and Development Program of Zhejiang Province.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Magnetic microrobots capable of autonomous operation hold significant promise for critical cell and small creature manipulation tasks, including trapping, transportation, sorting, etc. Although conventional microrobot navigation methods have shown notable performance, they often lack adaptability to novel environments. To address these limitations, we propose a learning-based framework for real-world microrobot navigation and dynamic target tracking. Our approach employs spatial-temporal transformer reinforcement learning (STTRL) with a deterministic velocity obstacle (DVO) that processes historical navigation states and virtual light detection and ranging scans to predict optimal actions. The key innovation lies in the model’s ability to extract and utilize contextual information from observation histories, enabling adaptive behavior even in previously unseen environments. Through large-scale model-free reinforcement learning trained in randomized simulation environments, our method achieves remarkable real-world performance with zero-shot transfer capability. Experimental results demonstrate superior navigation agility with an 89.8% success rate in the base environment, representing a 7.4% improvement over state-of-the-art algorithms. Furthermore, the method exhibits robust generalization in diverse unseen environments, validating its adaptability to different environmental characteristics.

Förderung

  • National Natural Science Foundation of China
  • Key Research and Development Program of Zhejiang Province

Zitieren

APA

Wang, F., Chen, B., Zhang, Y., Guo, X., Lyu, Y., Liu, C., Knoll, A., Cui, D., Jiang, H., Ying, Y., & Zhou, M. (2026). STTRL-DVO: Transformer-Based Reinforcement Learning for Robust Dynamic Target Tracking in Cluttered Environments. IEEE Transactions on Robotics, 42, 2555-2575. https://doi.org/10.1109/tro.2026.3701578

Vancouver

Wang F, Chen B, Zhang Y, Guo X, Lyu Y, Liu C, et al. STTRL-DVO: Transformer-Based Reinforcement Learning for Robust Dynamic Target Tracking in Cluttered Environments. IEEE Transactions on Robotics. 2026;42:2555-2575. doi:10.1109/tro.2026.3701578

BibTeX
@article{Wang2026,
  author = {Wang, Fanghao and Chen, Binghong and Zhang, Youchao and Guo, Xiangyu and Lyu, Yining and Liu, Chuanjie and Knoll, Alois and Cui, Di and Jiang, Huanyu and Ying, Yibin and Zhou, Mingchuan},
  title = {STTRL-DVO: Transformer-Based Reinforcement Learning for Robust Dynamic Target Tracking in Cluttered Environments},
  journal = {IEEE Transactions on Robotics},
  year = {2026},
  volume = {42},
  pages = {2555--2575},
  doi = {10.1109/tro.2026.3701578}
}
Kostenlos starten →