Towards Robust and Unobtrusive Human-Centric Sensing via Intelligent Hardware

The Hong Kong University of Science and Technology
Department of Computer Science and Engineering


PhD Thesis Defence


Title: "Towards Robust and Unobtrusive Human-Centric Sensing via Intelligent
Hardware"

By

Mr. Chenggao LI


Abstract:

Human-centric sensing serves as the bridge between physical reality and digital
intelligence, significantly transforming lifestyles across smart-space
interaction, personalized healthcare, and body motion understanding. However,
the realization of ubiquitous and trustworthy human-centric sensing is hindered
by four critical challenges: (i) the vulnerability of environment-anchored
sensing to dynamic environmental interference, (ii) the underexplored security
risks inherent in the wireless physical layer, (iii) the trade-off between
sensing fidelity and physical unobtrusiveness in human-coupled sensing, and
(iv) the domain shifts across users, sensing configurations, and deployment
conditions. This thesis investigates these challenges through the synergy of
intelligent hardware and domain-aware algorithms. Specifically, we leverage
Reconfigurable Intelligent Surfaces (RIS) to enhance the robustness and
security of environment-anchored wireless sensing, while developing novel
human-coupled sensing interfaces and domain-adaptive algorithms for
unobtrusive blood pressure monitoring and 4D snowboarder pose reconstruction.
Collectively, these contributions enable a robust and unobtrusive sensing
ecosystem.


Date:                   Monday, 29 June 2026

Time:                   10:00am - 12:00noon

Venue:                  Room 3494
                        Lifts 25/26

Chairman:               

Committee Members:      Prof. Qian ZHANG (Supervisor)
                        Prof. Song GUO
                        Dr. Qijia SHAO
                        Prof. Dan WANG (ENVR)
                        Dr. Zhenjiang LI (CityU)