Circadia

Inspiration

Sleep in space is tough. Between microgravity and constantly shifting schedules, it’s hard to know not just how long you slept, but how stable that sleep actually was. Circadia makes that visible. It turns your sleep into a simple, real-time score that updates every 30 seconds, using just heart rate, HRV, and motion. So instead of guessing, astronauts can see when their sleep is fragile, respond in the moment, and track how it improves over time.

What it does

Our tool takes in sleep data in the form of PSGs, which include brain activity waves (EEGs), eye movement, and muscle contractions to understand how a person's restful sleep is affected and when patterns are disrupted.

How we built it

We used a mobile device to host the application locally, which allows for minimal overhead and a faster data analysis. Most importantly, it also helps keep personal information secure. We then fine-tuned a machine learning model inspired by Stanford researchers to take in physiological sleep data and analyze the stability of their sleep patterns using metrics such as micro-arousals and fragmentation.

Challenges we ran into

We had used an NVIDIA Jetson TK1 as our hardware device, but we used an older version, which made it difficult to communicate with modern machine learning libraries like PyTorch. We also ran into trouble establishing a network with its host computer, so we used an ethernet cable to create a more stable offline connection.

Accomplishments that we're proud of

We're proud of creating a very user-friendly frontend which allows for new users to easily navigate through the tool's analysis, helping them better understand the results from their sleep data.

What we learned

We learned how to divide tasks efficiently and maintain consistent communication, so all parts of the project clicked together smoothly despite any unexpected challenges.

What's next for Circadia

We plan to scale Circadia by collaborating with current professors in top universities across the US to explore its use in aiding patients with sleep disorders.