Paper and citation

Advancing MAPF Toward the Real World: A Scalable Multi-Agent Realistic Testbed (SMART)

Summary

The SMART testbed provides a realistic environment for evaluating Multi-Agent Path Finding (MAPF) algorithms. It couples physics-engine based simulation to better model kinodynamics and execution uncertainty, uses an Action Dependency Graph for execution monitoring, and focuses on scaling to large numbers of agents.

Access the published paper and preprint:

• IEEE Xplore: https://ieeexplore.ieee.org/document/11495177

• DOI: https://doi.org/10.1109/LRA.2026.3688062

• arXiv preprint: https://arxiv.org/abs/2503.04798

Version information verified from arXiv:

• v1 submitted March 3, 2025

• v2 submitted August 14, 2025

Cite as

Yan, J., Li, Z., Kang, W., Zheng, K., Zhang, Y., Chen, Z., Zhang, Y., Harabor, D., Smith, S. F., & Li, J. (2026). Advancing MAPF Toward the Real World: A Scalable Multi-Agent Realistic Testbed (SMART). IEEE Robotics and Automation Letters, 11(6), 7428–7435. https://doi.org/10.1109/LRA.2026.3688062.