CASE FILE // IOT / AI / MOBILE
AEROVIT.
Hybrid fitness ecosystem pairing a Flutter app with a custom ESP32-S3 smartwatch, MediaPipe BlazePose tracking, adaptive AI coaching, and an ERC-20 reward token on Ethereum Sepolia — complete at demonstration depth as a fully walkable prototype.
HIGHLIGHTS
- ▸ Custom smartwatch hardware: ESP32-S3, 1.69" touch LCD, MAX30102 HR/SpO2, QMI8658 IMU, BLE 5.0
- ▸ 33-landmark BlazePose pipeline with EMA smoothing, joint-angle scoring, and rep/form state machines
- ▸ Gamification: Hunter ranks (E → National), XP, quests, leaderboards, and a 20-floor turn-based dungeon RPG
- ▸ Web3 rewards: AERO ERC-20 on Sepolia — off-chain accrual, 100 AERO minimum, 24h cooldown, 5% burn
- ▸ AI coaching: RAG (ChromaDB), STT/TTS, low-latency audio, RL-style intensity adaptation
ARCHITECTURE
- ▸ Biometrics stream from the watch as 28-byte BLE packets into the Flutter app
- ▸ Camera frames feed the pose pipeline in parallel; the AI coach consumes both signals
- ▸ Firebase Auth, Firestore, and Cloud Functions form the backend plane
- ▸ Earnings accrue off-chain; withdrawal gates enforce minimum, cooldown, and burn before Sepolia transfer
STACK
MOBILE
- Flutter
- Dart
- Flame 2D
HARDWARE
- ESP32-S3
- MAX30102
- QMI8658
- BLE 5.0
AI/ML
- MediaPipe BlazePose
- ChromaDB RAG
- PPO coaching
BACKEND / WEB3
- Firebase
- Cloud Functions
- Ethereum Sepolia
- ERC-20
UPLINKS
GALLERY & EVIDENCE FILES
■ [ VIDEO DEMO ] AEROVIT SMARTWATCH & POSE TRACKING DEMO — REAL-TIME ESP32-S3 BIOMETRICS AND MEDIAPIPE BLAZEPOSE INTENSITY MAPPING
■ [ YOUTUBE FEED ] AEROVIT HYBRID FITNESS WALKTHROUGH — DETAILED DEMONSTRATION OF SMARTWATCH BLE BIOMETRICS AND MEDIAPIPE POSE RECOGNITION SYSTEM (YOUTUBE)
■ AEROVIT SMARTWATCH — CUSTOM ESP32-S3 BIOMETRICS WEARABLE WITH CUSTOM ENCLOSURE AND INTEGRATED PULSE SENSOR
■ AEROVIT RESEARCH & DEVELOPMENT POSTER — HARDWARE SCHEMATIC, MEDIAPIPE PROCESSING PIPELINE, AND CORE SYSTEMS ARCHITECTURE