Papers
5
Total Citations
29
H-Index
3
About
Huan Yu is a robotics researcher whose work spans simultaneous localization and mapping (SLAM), autonomous navigation, legged locomotion, and motion planning under uncertainty. His most recognized contribution is an online 3D active pose-graph SLAM framework that intelligently selects key poses using graph topology and sub-maps to achieve efficient loop-closure detection in complex three-dimensional environments, earning 17 citations and establishing him as a contributor to the active SLAM community. Complementing this, his work on appearance-based loop closure detection introduces an enhanced bag-of-words approach incorporating inverse depth of feature words, further refining localization reliability in robotic systems. Yu has also made meaningful strides in legged robotics, developing a hierarchical motion planner for the wheel-quadruped platform "BIT-NAZA," enabling robust traversal of unstructured terrain through novel kinematic search strategies. His research extends into probabilistic planning, where he proposed a covariance upper bound technique to accelerate Gaussian belief space planning for underwater robots navigating spatially varying uncertainties. More recently, his work on modular transfer learning addresses disturbance rejection in dynamic environments. Collectively, Yu's research reflects a broad yet cohesive vision of building more capable, adaptive autonomous systems across land and underwater domains.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5