EUVIMEDEUVIMED entdecken
Crossref

Transformer-enhanced deep Q-Learning for adaptive robot path planning in dynamic environments

Harish Sharma, Ritu Tiwari, Shubham Shukla, Sushant Kumar

Cluster Computing · 2026-06 · 29 (5)

DOI: 10.1007/s10586-026-06072-2 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Cluster Computing, online erschienen am 01.06.2026. Verfasst von Harish Sharma und 3 weiteren Autor:innen. Die Arbeit nennt 40 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Efficient navigation in dynamic environments remains a critical challenge for autonomous robots in industrial manufacturing, search-and-rescue operations, and automated warehousing. Traditional path-planning algorithms struggle to adapt to real-time obstacle movements, while conventional reinforcement learning (RL) approaches cannot model long-range spatial dependencies. This paper presents Transformer-Enhanced Deep Q-Learning (Transformer-DQN), a novel framework that integrates transformer architectures with Deep Q-Networks (DQN) to address these limitations. The model dynamically captures obstacle interactions by integrating multi-head self-attention mechanisms and Cartesian positional encoding and optimizes navigation in cluttered environments. The Transformer’s self-attention mechanism to solve specific, well-known limitations of DQN in robotic path planning, namely the capture of long-range spatial dependencies and dynamic obstacle relationships in a non-Markovian setting. Hyperparameter tuning via Optuna ensures a balance between exploration and exploitation, while prioritized experience replay enhances training stability. Experimental results demonstrate significant advancements over baseline methods: 20\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\times$$\end{document}20 Grids: Reduces average pathfinding time by 33.65% (209.5s \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\rightarrow$$\end{document} 139s) and collisions by 37.5% compared to vanilla DQN. 30\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\times$$\end{document}30 Grids: Achieves a 94.6% time reduction (9125s \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\rightarrow$$\end{document} 492.5s) and 33.3% fewer collisions, showcasing superior scalability. Adaptive Performance: Outperforms PPO (70% vs. 85% success rate) and classical planners (RRT*/D*) in dynamic settings, approaching optimal path lengths (28 vs. 25 steps). The Transformer-DQN’s ability to generalize across grid sizes and dynamically replan in real-time positions it as a robust solution for time-sensitive applications. Theoretical analysis confirms convergence guarantees, while empirical validation demonstrates substantial improvements in both time optimization and traversal efficiency and reduced computational overhead. This work provides a foundational step towards bridging the gap between simulated and real-world robotic systems by establishing a robust learning framework in a dynamically controlled environment. Offering a scalable framework for autonomous navigation in unstructured, dynamic environments.

Zitieren

APA

Sharma, H., Tiwari, R., Shukla, S., & Kumar, S. (2026). Transformer-enhanced deep Q-Learning for adaptive robot path planning in dynamic environments. Cluster Computing, 29(5). https://doi.org/10.1007/s10586-026-06072-2

Vancouver

Sharma H, Tiwari R, Shukla S, Kumar S. Transformer-enhanced deep Q-Learning for adaptive robot path planning in dynamic environments. Cluster Computing. 2026;29(5). doi:10.1007/s10586-026-06072-2

BibTeX
@article{Sharma2026,
  author = {Sharma, Harish and Tiwari, Ritu and Shukla, Shubham and Kumar, Sushant},
  title = {Transformer-enhanced deep Q-Learning for adaptive robot path planning in dynamic environments},
  journal = {Cluster Computing},
  year = {2026},
  volume = {29},
  number = {5},
  doi = {10.1007/s10586-026-06072-2}
}
Kostenlos starten →