Yunfei Shi

Carnegie Mellon University, Hong Kong Polytechnic University

Papers

Total Citations

442

H-Index

About

Yunfei Shi is a leading researcher in multi-robot systems, with a focus on scalable, decentralized control and learning. Their work tackles fundamental challenges in coordinating large teams of robots, from aerial swarms to warehouse automation. Shi’s most influential contribution is the PRIMAL framework (Pathfinding via Reinforcement and Imitation Multi-Agent Learning), a landmark 2019 paper with over 399 citations that revolutionized multi-agent path finding (MAPF). By combining reinforcement learning with imitation learning, PRIMAL enabled decentralized, scalable coordination that outperformed traditional centralized planners, addressing a critical bottleneck in real-world robot deployments. Shi further advanced the field with work on adaptive informative sampling for heterogeneous multi-robot systems, designing algorithms that allow diverse robot teams—with varying mobility, sensors, and battery life—to efficiently explore hazardous environments. Their research also extends to distributed reinforcement learning for articulated mobile robots, where they developed decentralized control architectures that coordinate spatially distributed robot bodies without a central brain. Through these contributions, Shi has established themselves as a key figure in making multi-robot systems more autonomous, scalable, and practical for real-world applications.