Papers
10
Total Citations
176
H-Index
7
About
Ninghang Hu’s research lies at the intersection of assistive robotics, human activity recognition, and human-robot interaction, with a strong focus on developing technologies that support ageing populations. As a key contributor to the ACCOMPANY project, Hu helped design acceptable home companion robots that address the multidimensional challenges of human-system interaction for older adults. His work on activity recognition is particularly innovative: Hu pioneered methods for learning from “soft labels”—incomplete or noisy annotations—allowing robots to recognize human activities even when training data is imperfect. This approach, detailed in his 2016 paper (30 citations), significantly advances the practicality of robot-care systems. Hu also contributed to multi-modal sensing, fusing ceiling-mounted cameras with laser range finders for robust people detection and localization, and developed techniques for posture recognition under heavy self-occlusion using top-view cameras. His research on human intent forecasting, leveraging intrinsic kinematic constraints to predict movement, enhances collaborative human-robot tasks. With over 176 total citations across his most-cited works, Hu’s contributions are foundational to creating robots that can perceive, understand, and anticipate human actions in real-world care environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Learning to Recognize Human Activities Using Soft Labels30 citations · 2016
- 4Learning to Recognize Human Activities from Soft Labeled Data18 citations · 2014
- 5
- 6Posture recognition with a top-view camera12 citations · 2013
- 7Human intent forecasting using intrinsic kinematic constraints8 citations · 2016
- 8A two-layered approach to recognize high-level human activities7 citations · 2014
- 9
- 10