AXON
IASI · EDGE-AI · INFRASTRUCTURE ANOMALY DETECTION
AXON (Adaptive eXtended-bridge Operational Network) — A three-layer Edge-AI IIoT framework for real-time infrastructure anomaly detection in transportation systems.
Edge Node (Pi Zero 2W) → Android Termux Bridge → Central Server → Live Dashboard
⚠️ Prototype Notice
This is a Phase 1 simulation prototype.
- All sensor data (vibration, obstacle detection, IASI scores) is synthetically generated in the browser
- No physical hardware is connected in this phase
- The full system (Raspberry Pi Zero 2W + Android Termux Bridge + Ubuntu Server) will be integrated in Phase 2
- This prototype demonstrates the complete dashboard UI, data visualization, and IASI scoring pipeline
🚆 Project Overview
Transportation infrastructure — railway tracks, urban roads, industrial pathways — degrades continuously between inspection cycles. Manual railway track inspection is conducted only once every 3–6 months, leaving a significant window during which cracks, joint failures, loose fasteners, and other structural faults go undetected.
AXON is a domain-agnostic edge-AI IIoT framework that monitors infrastructure health on every vehicle run. A low-cost edge node (Raspberry Pi Zero 2W + MPU6050 vibration sensor) captures structural signals continuously. Data travels through a resilient Android Termux MQTT bridge — with local SQLite offline buffering — to a central server running FFT analysis, Isolation Forest anomaly detection, and Random Forest severity classification.
At the core of the system is the IASI (Infrastructure Anomaly Severity Index) — a generic two-component composite score that quantifies infrastructure risk in real time. The same architecture applies to railway track monitoring, urban road quality assessment, and fleet vehicle health monitoring.
🏗️ System Architecture
| Layer | Component | Role |
|---|
| Layer 1 | Raspberry Pi Zero 2W | Edge sensing node — MPU6050 vibration at 500Hz, MQTT publishing |
| Layer 2 | Android Phone (Termux) | Resilient MQTT bridge — SQLite offline buffer, automatic replay on reconnect |
| Layer 3 | Laptop / Ubuntu Server | ML processing — FFT, Isolation Forest, Random Forest, IASI engine, Flask REST API |
| Layer 4 | Vercel + Render | Dashboard + API hosting — live browser visualization and operator console |
In this Phase 1 prototype, Layers 1–3 are simulated synthetically inside the browser. Layer 4 (this dashboard) is fully functional and live.
📊 IASI — Infrastructure Anomaly Severity Index
The IASI is a generic two-component composite score computed as:
IASI = (Vibration Anomaly Score × W1) + (Signal Continuity × W2)
Where W1 and W2 are domain-specific weights derived via the Analytic Hierarchy Process (AHP). For railway track monitoring: W1 = 0.85, W2 = 0.15.
| Classification | IASI Range | Meaning |
|---|
| SAFE | 0 – 25 | Normal operating conditions |
| MONITOR | 26 – 50 | Early degradation detected — increase inspection frequency |
| ALERT | 51 – 75 | Significant anomaly — schedule maintenance within 48 hours |
| CRITICAL | 76 – 100 | Severe structural risk — halt traffic, dispatch field crew immediately |
🎮 Simulation Scenarios
Five fault severity scenarios with realistic IASI ranges and FFT frequency signatures:
| Scenario | Description | IASI Range | Classification |
|---|
| NORMAL | Healthy infrastructure, baseline condition | 5 – 18 | SAFE |
| JOINT_FAULT | Rail joint loosening — periodic impulse signature at 18Hz | 22 – 38 | MONITOR |
| LOOSE_FASTENER | Loose fastener — localized vertical oscillation at 14Hz | 15 – 30 | MONITOR |
| RAIL_CRACK | Rail crack forming — sub-surface defect signature at 26Hz | 38 – 54 | ALERT |
| SEVERE_DAMAGE | Major structural failure — multi-band anomaly at 40Hz + 46Hz | 60 – 92 | CRITICAL |
✨ Dashboard Features
- 🔐 Operator authentication — secure login with session persistence
- 📡 Live device status — Pi Zero 2W, Android Bridge, Central Server status indicators
- 🎛️ Scenario control panel — switch between 5 fault conditions in real time
- 📈 IASI Gauge — animated semicircular SVG dial, 0–100, color-coded by risk class
- 🏷️ Risk Classification Badge — SAFE / MONITOR / ALERT / CRITICAL with live state
- 📉 IASI Trend Line — last 60 readings at 1Hz, rolling Recharts line chart
- 📊 FFT Power Spectrum — 51 frequency bins (0–50Hz), dominant frequency highlighted
- 🚧 Obstacle Detection Log — event log with class, confidence bars, and status flags
- 📡 Signal Continuity Gauge — Android bridge health indicator
- 🖥️ System Telemetry — Pi packet count, uptime, Android ping, server push latency
- 🌐 Responsive layout — desktop, tablet, and mobile
🛠️ Tech Stack
| Technology | Purpose |
|---|
| Next.js 14 | React framework, App Router |
| Tailwind CSS | Utility-first styling, responsive layout |
| Recharts | IASI trend line and FFT vibration spectrum charts |
| Lucide React | SVG icon system |
| Google Fonts — Inter + JetBrains Mono | Clean sans-serif + monospace typography |
| Vercel | Hosting and continuous deployment from GitHub |
🚀 Getting Started
Prerequisites: Node.js 18+
git clone https://github.com/SaiSugeet/GuardRail-Central.git
cd GuardRail-Central
npm install
npm run dev
Open http://localhost:3000
Login credentials:
Operator ID: SaiSugeet
Access Code: 2004
📦 Deploy on Vercel
- Push to GitHub
- Import repository at vercel.com
- Vercel auto-detects Next.js — click Deploy
- Every
git push to main auto-redeploys
📁 Project Structure
axon/
├── app/
│ ├── layout.tsx # Root layout — fonts, metadata
│ ├── page.tsx # Entry point — login gate / dashboard router
│ └── globals.css # CSS variables, Tailwind base, animations
├── components/
│ ├── LoginPage.tsx # Operator authentication screen
│ ├── Dashboard.tsx # Main dashboard shell and layout
│ ├── Navbar.tsx # Top navigation bar
│ ├── DeviceStatusStrip.tsx # Pi / Bridge / Server status cards
│ ├── ScenarioControl.tsx # Fault scenario selector
│ ├── IRMSGauge.tsx # IASI semicircular arc gauge
│ ├── RiskBadge.tsx # Risk classification badge
│ ├── SignalContinuity.tsx # Bridge health vertical gauge
│ ├── IRMSTrendChart.tsx # 60-point rolling IASI trend line
│ ├── FFTSpectrumChart.tsx # 51-bin FFT power spectrum
│ ├── ObstacleLog.tsx # Obstacle detection event table
│ └── SystemTelemetry.tsx # System stats readout panel
├── hooks/
│ └── useSimulation.ts # Central simulation hook — all synthetic data
├── lib/
│ └── constants.ts # Scenarios, IASI ranges, color maps, credentials
├── types/
│ └── index.ts # TypeScript interfaces
├── public/ # Static assets
├── vercel.json # Vercel deployment config
├── next.config.js # Next.js configuration
├── tailwind.config.js # Tailwind CSS configuration
└── README.md # Project documentation
🗺️ Roadmap
| Phase | Status | Description |
|---|
| Phase 1 | ✅ Complete | Simulation dashboard — full UI, IASI logic, 5 fault scenarios |
| Phase 2 | 🔄 In Progress | Pi Zero 2W + MPU6050 real sensor integration, axon-server backend |
| Phase 3 | 🔄 Planned | Android Termux CLI bridge, end-to-end pipeline validation |
| Phase 4 | 🔄 Planned | GPS fault tagging, road quality monitoring, multi-vehicle support |
👥 Team
| Name | Role |
|---|
| G Sai Sugeet | Lead Developer & System Architect |
| S D Mukhesh | Backend Development & Testing |
| John Benny J | Hardware Integration & Testing |
| Neetha Udupa G | Hardware Integration & Communication |
Guide: Dr. Agalya, Assistant Professor
Institution: Dept. of Electrical & Electronics Engineering, New Horizon College of Engineering, Bengaluru — 560103
Subject Code: 22EEE65 | Academic Year: 2025–26
📄 License
MIT License — free to use for educational and research purposes.
AXON v2.0 | Phase 1 Simulation Prototype
Built for IEEE Publication | New Horizon College of Engineering | 2025–26
