NH-SJC25HACK-IOT-020_GROVE_STREET_ENGINEERS

Silver Spectrum TechFest 2025 A National Level 48-Hour Hackathon - 28, 29, and 30 October 2025

# AI Intrusion Detection System

Adaptive home security system that learns your environment using neural networks.  
Runs on Raspberry Pi with camera, microphone, and door/window sensors.

---

## Prerequisites

- Basic knowledge of Raspberry Pi and GPIO setup  
- Python 3.11+ installed  
- Virtual environment (`venv`) installed  
- Internet connection for model training and updates  

---

## Requirements

### Hardware
- Raspberry Pi 4 (or 3B+)
- Pi Camera or USB webcam
- USB microphone
- Door/window contact sensors
- Buzzer (optional)

### Software
- Raspberry Pi OS
- Python 3.11+

---

## Installation

```bash
# Clone and setup
cd ~/MASTER_NHCE
python3 -m venv MASTER_NHCE
source MASTER_NHCE/bin/activate

# Install dependencies
pip install -r requirements.txt
sudo apt-get install python3-pyaudio portaudio19-dev

# Find audio device
arecord -l
# Update AUDIO_DEVICE in dotenv file

---

Wiring

GPIO 5  → Door sensor
GPIO 6  → Window sensor
GPIO 16 → Buzzer

---

Configuration

Edit .env file:

AUDIO_DEVICE=plughw:3,0
GPIO_DOOR_PIN=5
GPIO_WINDOW_PIN=6
GPIO_BUZZER_PIN=16
LEARNING_RATE=0.01

---

Usage

python3 main_system.py

Modes

• Learning Mode (0–100 clips): Records 15-second clips, sends for feedback
• Confidence Mode (100–250 clips): Refines detection
• Production Mode (250+ clips): Auto-detects intrusions, activates buzzer

---

Feedback

Send via WebSocket:

{"jsonType": "feedback_response", "requestId": "timestamp", "label": 0}

Label definitions:

• 0 = normal
• 1 = intrusion

---

Face Recognition

Add known faces to the images/ folder:

mkdir images
# Add images/John.jpg, images/Jane.png, etc.

---

Troubleshooting

• Camera error: Enable legacy camera in raspi-config
• Audio error: Check arecord -l and update .env
• High probability stuck: Delete models/ and system_state.json, then restart

---

Features

• Multi-sensor fusion (camera, audio, GPIO)
• Temporal pattern recognition (1-second windows)
• Face recognition for known persons
• Real-time alerts via WebSocket
• Buzzer activation on high threat
• 19-feature neural network