Thesis Defense: Hao He, Contactless Sleep and Physiological Monitoring Using Artificial Intelligence and Radio Waves

Friday, May 9
10:00 am - 11:00 am

Kiva (32-G449)

Doctoral Thesis Title:  Contactless Sleep and Physiological Monitoring Using Artificial Intelligence and Radio Waves
Presenter: Hao He
Presenter’s Affiliation (CSAIL, RLE, LIDS, MTL, etc.): CSAIL
Thesis Supervisor(s): Prof. Dina Katabi

Date: May 9, 2025
Time: 10:00am – 12:00 pm
Location if in person: Kiva (32-G449)

I will also provide a Zoom link for the defense.

Abstract: 

——

In-home monitoring of sleep and physiological signals is crucial for enabling early disease detection, personalized healthcare, remote patient monitoring, and long-term health assessment. This thesis advances this vision by leveraging recent developments in artificial intelligence (AI) and radio-frequency (RF) sensing techniques to create remote, contactless, and non-intrusive sleep and physiological monitoring systems.

We make four fundamental contributions:
First, we introduce a system capable of extracting breathing signals from radio waves reflected off human bodies, even in scenarios where multiple individuals share the same bed.
Second, we present an advanced AI model for sleep monitoring that generates sleep hypnograms and detects respiratory events without requiring wearable sensors.
Third, inspired by large language models, we demonstrate that physiological signals can be interpreted and reasoned similarly to language, facilitating translation between different physiological signals and effective representation learning for downstream clinical tasks.
Finally, we develop AI models that infer key biomarkers, such as blood oxygen saturation ($SpO_2$) and inflammation (C-reactive protein levels), in a fully passive and non-intrusive manner.
We implement and evaluate our AI-driven systems on real-world data, showcasing their significant potential to transform healthcare through continuous, contactless sleep and physiological monitoring.

Details

  • Date: Friday, May 9
  • Time: 10:00 am - 11:00 am
  • Category:
  • Location: Kiva (32-G449)