43 lines (31 loc) · 1.32 KB

BLE Attendance Prototype Architecture

Scope locked for prototype

Single mobile app with teacher/student roles
FastAPI + PostgreSQL backend
Android-first delivery with iOS-safe constraints from start
Local login and end-of-session biometric confirmation

Flow

Teacher logs in and starts a lecture session.
Backend creates active session and issues token.
Teacher app advertises token over BLE.
Student app scans and detects token repeatedly.
Student app submits session_id, RSSI, and proximity flag to backend.
Session ends by teacher.
Student confirms identity with biometrics and finalizes attendance.

Data model

users (teacher or student)
sessions (single active session per teacher)
detections (time-stamped RSSI samples)
attendance (presence ratio + biometric status)

Presence ratio (prototype)

Window size: 30 seconds
ratio = valid detections / elapsed windows
present if ratio >= 0.6

iOS feasibility guardrails

Student app should be foreground/background active and not force-killed
BLE behavior when force-killed is not guaranteed by iOS
Early iOS spike should validate scan reliability on one physical device

Hosting options for testing

API: Render free tier
DB: Supabase or Neon free Postgres
Alternative: expose local API with Cloudflare Tunnel