WinkTPG: An Execution Framework for Multi-Agent Path Finding Using Temporal Reasoning

Highlights: Will present at IROS 2026.

Overview

This work presents WinkTPG (Windowed kinodynamic Temporal Plan Graph Planning), a novel execution framework for Multi-Agent Path Finding (MAPF) that addresses the gap between theoretical MAPF algorithms and real-world deployment requirements. While existing MAPF algorithms rely on simplified kinodynamic models, WinkTPG enables agents to directly follow generated MAPF plans by incorporating temporal reasoning and speed optimization.

Demo Video

Links:

• Paper (IEEE Xplore)
• Preprint (arXiv)
• PDF
• Project Page
• Video

BibTeX

@article{yan2026winktpg,
  title={WinkTPG: An Execution Framework for Multi-Agent Path Finding Using Temporal Reasoning},
  author={Yan, Jingtian and Smith, Stephen F. and Li, Jiaoyang},
  journal={IEEE Transactions on Automation Science and Engineering},
  volume={23},
  pages={9162--9175},
  year={2026},
  doi={10.1109/TASE.2026.3688563}
}

Recommended citation: Jingtian Yan, Stephen F. Smith, and Jiaoyang Li. IEEE Transactions on Automation Science and Engineering. 23, 9162-9175, 2026.
Will present at IROS 2026.
Download Paper | Download Bibtex