An ESP32 and LoRa based Smart Safety Vest and Smart Safety Helmet designed for industrial worker safety monitoring, fall detection, GPS tracking, environmental monitoring, emergency SOS alerts, and real-time IoT communication.
SafeGuard combines ESP32, LoRa, GPS, MPU6050, BME280, MQ-2, MAX30100, Firebase, solar charging, and piezoelectric energy harvesting to build a connected industrial worker safety system.
SafeGuard is an IoT-based industrial safety system consisting of:
- 🦺 Smart Safety Vest
- ⛑️ Smart Safety Helmet
- 📍 GPS Worker Tracking
- 🚨 Fall Detection & Emergency SOS
- 🌡️ Environmental Monitoring
- 📡 Long-Range LoRa Communication
- ☁️ Firebase Cloud Monitoring
- 📊 Real-Time Web Dashboard
- Real-time industrial worker monitoring
- MPU6050-based motion and fall detection
- BME280 temperature and humidity monitoring
- GPS worker location tracking
- LoRa long-range communication
- Emergency SOS alert system
- Firebase IoT cloud synchronization
- Real-time web monitoring dashboard
- Industrial worker helmet safety monitoring
- Ultrasonic obstacle and proximity detection
- MQ-2 harmful gas detection
- MAX30100 heart rate and SpO₂ monitoring
- Wi-Fi and GSM connectivity
- Firebase-based remote monitoring
- Solar-powered charging
- Piezoelectric energy harvesting
- Rechargeable Li-ion battery system
- Low-power wearable safety architecture
ESP32 Arduino Nano LoRa GPS MPU6050 BME280
MQ-2 MAX30100 Firebase IoT GSM Wi-Fi
Real-time worker safety monitoring dashboard for sensor data, worker status, alerts, GPS tracking and environmental monitoring.
The Smart Safety Vest focuses on worker fall detection, motion monitoring, environmental sensing, GPS tracking, LoRa communication and emergency safety features.
The Smart Safety Helmet includes obstacle detection, harmful gas monitoring, heart-rate and SpO₂ monitoring, and wireless communication.
Combined project materials, documentation and final-release resources.
- ESP32 DevKit V1
- Arduino NANO
- BME280
- MPU6050
- NEO-M8N GPS
- SX1278 RA-02
- Ultrasonic Sensor
- MQ-02
- GY-MAX30100
- ESP-01 WiFi
- SIM800L GSM
- 5V Flexible Solar Panel
- Piezo Elements
- DFRobot Solar Power Manager 5V
- 18650 Li-ion Battery
- Active Buzzer
- SOS Push Button
- On/Off Switch
SafeGuard
│
├── 🦺 Smart Safety Vest
│ ├── ESP32
│ ├── MPU6050
│ ├── BME280
│ ├── GPS
│ └── LoRa
│
├── ⛑️ Smart Safety Helmet
│ ├── Arduino Nano
│ ├── Ultrasonic Sensor
│ ├── MQ-2
│ ├── MAX30100
│ ├── Wi-Fi
│ └── GSM
│
└── ☁️ IoT Monitoring Platform
├── Firebase
└── Web Dashboard
SafeGuard
├── index.html ← Web dashboard (must stay at repo root for GitHub Pages)
├── sw.js ← Dashboard's service worker (root-scoped, must stay here)
│
├── vest/ ← Smart Safety Vest — ✅ Complete
│ ├── README.md
│ ├── VEST-CONNECTIONS.md (full hardware wiring reference)
│ ├── firmware/ (Phase-01 through Phase-10)
│ ├── diagrams/
│ ├── datasheets/
│ ├── docs/
│ ├── hardware/
│ ├── images/
│ ├── results/
│ └── thesis/
│
├── helmet/ ← Smart Safety Helmet — 🚧 In Progress
│ ├── README.md
│ ├── firmware/
│ │ └── Phase-01_Component_Verification/
│ ├── diagrams/
│ ├── datasheets/
│ └── docs/
│
└── final-project/ ← Combined final-release deliverables (report, final
presentation, consolidated diagrams, BOM, results)
for the project as a whole, once both sections
are complete
/vest— LoRa-based Smart Safety Vest: fall detection (MPU6050), environmental sensing (BME280), GPS tracking, LoRa communication, Firebase cloud sync, web dashboard, and Android app./helmet— Dual-Nano Smart Safety Helmet: near-miss obstacle detection (ultrasonic), gas monitoring (MQ-2), vitals (MAX30100), WiFi (ESP-01) + GSM (SIM800L) redundant connectivity to the same Firebase project./final-project— Combined final-release materials for the completed system.- Root (
index.html,sw.js) — the live web dashboard, kept at the repository root since GitHub Pages serves from root and the service worker's scope depends on its path.
| Section | Status |
|---|---|
| Vest | ✅ Complete |
| Helmet | 🚧 Phase 1 (Component Verification) nearly done |
| Final Project | ⏳ Pending both sections' completion |
Two AI-assisted features run entirely in the browser on the data the dashboard already reads from Firebase (/worker1, /worker1/logs, /helmet1/live, /helmet1/logs). Nothing is written to Firebase.
| File | Purpose |
|---|---|
ai-engine.js |
Pure logic: alert evaluation, incident grouping, anomaly detection, Bangla/English intent detection |
ai-ui.js |
"Ask SafeGuard" chat widget + "AI Analyze" modal |
ai.css |
Styles (dark/light theme aware) |
1. Ask SafeGuard (chat) — floating button, bottom-right. Answers questions such as "আজকে কয়টা অ্যালার্ট হয়েছে?" or "ওয়ার্কারের অবস্থা এখন কেমন?" from live data and logs. Every number is computed from real data (no hallucination). Optionally, a free Google Gemini key can be pasted under ⚙️ so free-form questions are answered by an LLM that is given the live data (the key stays in that browser only).
2. AI Anomaly Analysis — header button AI Analyze → choose 1H / 6H / 24H / 7D → press ANALYZE (manual, never automatic). Methods: robust z-score (median/MAD), sudden-jump detection, sustained-trend regression with time-to-threshold forecast, flat-line (stuck sensor), data gaps, prolonged inactivity, and temperature + heart-rate cross-signal. Each finding is compared with the fixed thresholds ("would the fixed limit have alarmed?"). Tunable constants: METRICS and CFG in ai-engine.js.
- Ajoy Das
- Shafiul Azam
- Ritu Biswas
- Abdullah Al Mamun
- Abdul Halim
