Hengyu Liu

Xiamen University, University of Alberta

Papers

Total Citations

H-Index

About

Hengyu Liu is a robotics and computer vision researcher whose work focuses on enabling robust autonomous navigation and intelligent infrastructure monitoring. His primary research areas include visual-inertial odometry (VIO), multi-sensor fusion, and deep learning for robotic perception. Liu’s most notable contribution is the development of YO-VIO, a robust multi-sensor semantic fusion localization system designed for dynamic indoor environments. This work addresses a critical limitation of traditional visual SLAM systems—their inability to maintain accurate pose estimation in cluttered or moving scenes—by integrating semantic information from deep learning with inertial and visual data. With 12 citations, YO-VIO has become a reference point for researchers tackling real-world deployment of service robots in hospitals, warehouses, and other unpredictable settings. More recently, Liu has explored the intersection of robotics and structural health monitoring, introducing BoltResvit, an enhanced residual vision transformer for robotic-assisted nondestructive railway bolt looseness detection. This work demonstrates his ability to adapt state-of-the-art transformer architectures for practical, high-stakes industrial applications. Liu’s research is distinguished by its clear focus on bridging the gap between theoretical SLAM advances and deployable, real-world robotic systems.