TinyML & Edge Computing • ECE • IoT • EEE • Final Year Capstone (7th/8th Sem)
IoT Anomaly Detector
An autonomous IoT diagnostic device running an on-device quantized neural network to detect motor vibration, voltage spikes, and mechanical faults before catastrophic failure occurs.
Hardware Bench Preview
Component Bill of Materials (BOM)
Total: ₹1,440 1. NodeMCU ESP32 (Xtensa Dual-Core 240MHz)
Qty: 1 ₹420
2. MPU6050 6-Axis Accelerometer/Gyro
Qty: 1 ₹180
3. ACS712 Current Sensor Module
Qty: 1 ₹190
4. 0.96" I2C OLED Display (128x64)
Qty: 1 ₹240
5. Piezo Buzzer & Diagnostic Status LEDs
Qty: 1 ₹60
6. Custom FR4 PCB Test Bench & Jumper Wiring
Qty: 1 ₹350
Academic Deliverables Included
- ✓ IEEE formatted Synopsis approved for college guide submission
- ✓ 70+ page project report with circuit schematics and FFT analysis charts
- ✓ Fully commented Arduino / C++ source code and trained neural weights
- ✓ Viva defense presentation deck with expected examiner Q&A
Bench Testing Benchmarks
- • Edge impulse inference latency < 28ms on dual-core ESP32
- • Real-time FFT vibration spectrum analysis displayed on OLED
- • MQTT telemetry via Wi-Fi with automatic cloud fault logging
- • Zero external cloud dependency for critical shutdown triggers