SleepGuardian

Chest Patch
Content of the Chest Patch
SleepGuardian App
PNEUMA v0.1.0
Logo
Illustration of Alert Usage

Inspiration: SleepGuardian is personal.

My baby cousin Zeinab was diagnosed with CCHS ,Congenital Central Hypoventilation Syndrome. I remember visiting her house and seeing the large monitoring machines in her room while she slept. I remember my aunt constantly pausing movies, stepping away from conversations, and checking on her throughout the night.

Sleep wasn’t peaceful, it was something to monitor.

CCHS affects roughly 1 in 200,000 births, only about 5,000 people worldwide live with it. But for those families, the burden is constant. The fear of silent CO₂ buildup. The fear of missed breathing signals. The exhaustion.

SleepGuardian was built to give families like mine safer nights.

What it does:SleepGuardian is a wearable pediatric chest patch that:

• Continuously monitors transcutaneous CO₂ (TcCO₂) • Detects hyperventilation and hypoventilation • Tracks respiratory impedance (breathing movement) • Uses AI to detect apnea-related patterns • Sends real-time alerts to a caregiver app

Instead of relying on bulky wired machines, SleepGuardian provides lightweight, overnight respiratory monitoring, directly synced to a mobile app.

It turns silent risk into visible, actionable data.

How we built it:We combined design, software, datasets, and AI modeling.

Design & Branding • Built product visuals and slides using Canva • Designed the SleepGuardian brand identity and UI system

App Development • Built a working iOS prototype using Swift • Created live dashboards for CO₂ trends and breathing rate • Integrated alert logic into the app interface

AI Model Development • Used Python to build and train a machine learning model • Leveraged the: • Apnea ECG Database • UCD Sleep Apnea Database • Borealis CapnoBase datasets (borealisdata-sp2-nlb8it)

We used frequency-domain features, respiratory signal patterns, and ECG-derived markers to train a detection model capable of identifying apnea-related respiratory events.

AI & Research Assistance • Used ChatGPT and Claude to: • Refine ML architecture • Structure signal processing pipelines • Improve model training workflow • Validate physiological reasoning

The result: a working app prototype integrated with a trained AI detection model.

Challenges we ran into:

1.Medical Accuracy Ensuring that our physiology assumptions were correct, especially around TcCO₂ measurement and impedance-based respiration. 2.Signal Complexity Respiratory signals are noisy. Real-world breathing data is not clean, especially in children. 3.Data Translation Most apnea datasets are adult-based. We had to carefully adapt logic to pediatric contexts. 4.Hardware Feasibility Designing a wearable that could safely measure CO₂ overnight without overheating or bulky wires was technically challenging. 5.Emotional Weight This wasn’t just a project. It was personal. That made getting it right even more important.

Accomplishments that we're proud of:

• We built a functional iOS app • We trained a working AI detection model • We created realistic wearable hardware architecture • We connected physiological datasets to real application logic • We transformed a personal story into an engineering solution

Most importantly, we proved that intelligent, lightweight overnight monitoring is possible.

What we learned:We learned that living with CCHS is not just a diagnosis, it’s a lifestyle.

It affects sleep, family time, mental health, and daily peace of mind.

We learned how difficult it is to engineer medical-grade monitoring safely and responsibly.

We learned that building in healthcare requires: • Precision • Empathy • Scientific grounding • Ethical awareness

And we learned that technology must also reduce burden, not add complexity.

What's next for SleepGuardian: Our vision goes beyond a prototype.

Pediatric Care Integration Connecting directly with clinical teams for seamless remote oversight.
Predictive Respiratory Insights Advancing our AI model to forecast respiratory events before they happen.
Expanded Disorder Coverage Adapting SleepGuardian for other rare respiratory conditions.
Nationwide Adoption Making intelligent overnight monitoring the standard of care, not a luxury.

We envision a future where no child with a rare respiratory disorder sleeps unprotected.

SleepGuardian is just the beginning.