Sha Hu

Simon Fraser University, Zhejiang University

Papers

Total Citations

165

H-Index

About

Sha Hu is a robotics and artificial intelligence researcher whose work spans robot navigation, motion planning, and morphological design optimization. He is best known for his influential contributions to crowd navigation, particularly his development of a relational graph learning framework that enables robots to navigate safely and efficiently among humans. This approach leverages model-based deep reinforcement learning combined with Graph Convolutional Networks to reason about inter-agent relationships by anticipating future states — a significant advancement in socially aware robot navigation that has garnered over 146 citations, reflecting its strong impact on the field. His earlier work in legged robotics, including the design and control of a hexapod walking robot with 12 degrees of freedom, demonstrates a broad foundation in physical robot systems and motion planning. More recently, Hu has explored the computationally challenging problem of robot morphology optimization, proposing neural fidelity warping techniques to make the design process more resource-efficient. Across his career, Hu's research reflects a consistent drive to make robots more adaptive, intelligent, and capable of operating in complex, real-world environments — contributing meaningfully to both the theoretical and applied dimensions of modern robotics.