Why STAR HEALTH?
On deep-space missions the most critical problem is not hardware but the time it takes to reach information. This project proposes a software design that removes that time.
Imagine an astronaut on Mars suffering a neurological crisis. The signal needs 22 minutes one way to reach mission control on Earth. So the ask-and-answer cycle takes more than 45 minutes. In a crisis, that window is irreversible.
Besides, not everyone on the crew is a doctor. Under microgravity, cosmic radiation and severe sleep deprivation, neurological disorders begin; reaction time, focus and decision-making ability erode fast. The astronaut may not even be able to assess their own condition correctly.
That is exactly why I built STAR HEALTH: an autonomous health assistant that needs no internet or external AI server and produces decisions in milliseconds.
How does it work?
Astronauts sign in with face recognition and gain access to personalized baseline neuro-signals and reference ranges specific to space physiology. Instead of waiting on an external AI API, the system analyzes data in milliseconds through a zero-latency (offline) smart rule engine running in the background and triggers pre-defined space-medicine protocols.
The critical point: the system does not decide by looking at a single data point. When sleep quality drops while Theta waves in the brain rise, it reads them together and says "micro-sleep and EVA abort risk" — that is, it performs a cross risk analysis. And it clearly shows the reason, triggers and confidence of every decision on screen.
Simulation or real?
This is a decision-support simulation platform. All astronaut telemetry it shows is simulated; NASA research is used with its source cited. The system is not a medical device, it makes no diagnosis, gives no dose and has passed no clinical validation. Its purpose is to demonstrate with a working prototype how such a system could be designed.
A short summary.
| Project name | STAR HEALTH — Autonomous Astronaut Health & Crisis Monitoring System |
|---|---|
| Event | NASA Space Apps Challenge 2026 |
| Category | Deep-space mission health monitoring · decision-support simulation |
| Technology | React 18 · TypeScript · Three.js / R3F · Firebase Auth · FaceNet · Tailwind · PWA |
| Module count | 12 modules, 4 cross rules (R-01…R-04), 10 3D anatomy models |
| Languages | English (default) · Turkish · Spanish · German |
| Connectivity | Runs fully offline; no external AI API is used |
| Developer | Uras AKAS · urasakas.com |
Legal Notice and Scope Limits
Decision-support simulation only. This software is not a certified medical device, not flight hardware, not clinically validated and not endorsed by NASA. It is an independent project entered in the NASA Space Apps Challenge.
All crew telemetry shown in the application is SIMULATED. NASA research data is shown only with a source (provenance) label, and NASA has no relationship with this project.
The system gives no dose, writes no prescription and makes no diagnosis. Drug stock information never means medical authorization. Consult a professional for medical advice or diagnosis.
Face verification is only a second factor; it can never bypass authentication. The single authority for authentication is Firebase Authentication.