Sleep is our basic pillar of health, yet many of us consistently fall short of getting enough of it. Sleep health is essential to our well-being, but monitoring it effectively is often difficult. Wearable devices such as smartwatches and fitness trackers have grown in popularity, but not everyone owns one, and many, too, find them uncomfortable to wear while sleeping. This practical issue raises the question of how we can monitor sleep in ways that are less intrusive and more practical for long-term everyday use.
Our research explores unconventional approaches to sleep health sensing that move beyond wearables. The first track is the design of non-wearable systems, such as sensors embedded within the bed or thin overlays that can unobtrusively capture physiological and behavioral signals during sleep. Because sleep health is not a single measure but a multidimensional construct that includes duration, quality, timing, and continuity, the project also examines whether these unconventional sensing modalities can provide data that serve as reliable proxies for these different aspects of sleep. By combining user-centered design with careful validation of accuracy across dimensions of sleep health, this work aims to create sustainable and accessible solutions for long-term sleep monitoring.
GitHub Repository: <to be updated>
Skills and Expertise Involved
- Key areas of work include:
- Sensor Design and Integration: Developing non-wearable, unobtrusive sensing systems for sleep monitoring
- Software and System Development: Implementing data collection, storage, and visualization tools for sleep monitoring
- Human-Centered Design: Ensuring systems are comfortable, practical, and user-friendly for long-term use
- Signal Processing and Data Analysis: Extracting meaningful sleep-related features from raw physiological and behavioral data
- Sleep Science and Health Metrics: Understanding multidimensional sleep health and validating proxy measures
- Core competencies required are:
- Programming in Python
- Data wrangling and cleaning large, noisy datasets
- Statistical modeling and evaluation of analytical methods
We are seeking students who are interested in gaining experience in these areas. Opportunities are on volunteer basis.