Important disclosure
LifeGuard Orbit is a research prototype. It is NOT clinically validated and is not a medical device. All metrics below come from our own testing and should not be read as clinical performance.
Inspiration
Health crises such as atrial fibrillation episodes often start quietly, and people usually notice too late. Astronauts face the same problem, with limited access to doctors. We wanted an early-warning system that learns what is normal for one person and flags deviations before a crisis. This is our project for the NASA Space Apps 2026 Health Monitoring challenge.
What it does
LifeGuard Orbit is a smartwatch-first health-crisis early-warning system. It builds a personal baseline from wearable signals and adds contactless WiFi-CSI (channel state information) home sensing to cross-check what the watch sees. It is extended for space/astronaut health monitoring.
Reported results in our testing (research prototype, not clinically validated):
- AUROC 0.992 for atrial fibrillation detection
- About 13 minutes of early-warning lead time
- 252 automated tests
- 39 crisis states modelled
How we built it
Python, signal processing (filtering, feature extraction, CSI processing) and machine learning for personal-baseline anomaly detection and crisis-state classification.
Challenges
Separating real physiological deviation from noise, keeping the personal baseline robust, and fusing watch and WiFi-CSI signals.
What's next
Prospective studies with clinical partners and proper validation before any real-world use.
Links
Built With
- machine-learning
- python
- signal-processing
