Skip to content
AhqafCoder
AquaSense
Public

About

No description, website, or topics provided.

Resources

Readme
Activity

Stars

1 star

Watchers

0 watching

Forks

0 forks
Report repository
BranchesTags

Repository files navigation

🌊 Autonomous Water Quality Monitoring Boat

Smart India Hackathon 2025 Team Gandharvas Invertis University

An intelligent autonomous robotic boat designed for real-time water quality monitoring, plastic pollution detection, and algae bloom identification using advanced AI/ML models and hyperspectral sensing technology.

Boat System

📋 Table of Contents

🎯 What This Project Does

This autonomous boat system:

  1. Navigates autonomously through water bodies using GPS waypoints
  2. Monitors water quality in real-time using an 18-channel hyperspectral sensor
  3. Detects plastic pollution using computer vision (YOLOv8)
  4. Identifies algae blooms and predicts chlorophyll-A levels using machine learning
  5. Operates offline-first with automatic cloud synchronization when internet is available
  6. Provides real-time dashboard for monitoring and remote control
  7. Collects samples at specified locations using an automated pump system

🚨 Problem Statement & Solution

Problem

Water pollution is a critical environmental challenge affecting lakes, rivers, and oceans globally:

  • Plastic Pollution: 8+ million tons of plastic enter oceans annually, harming marine life
  • Algae Blooms: Harmful algal blooms (HABs) contaminate drinking water and kill aquatic life
  • Manual Monitoring: Traditional water quality testing is time-consuming, expensive, and covers limited areas
  • Limited Data: Lack of continuous, real-time water quality data for effective decision-making

Our Solution

An autonomous robotic boat that combines:

✅ AI-Powered Detection: Dual AI system for plastic detection (YOLOv8) and chlorophyll prediction (CatBoost)
✅ Hyperspectral Sensing: 18-channel spectral sensor (410-940nm) for detailed water analysis
✅ Offline-First Architecture: Works without internet, syncs data when connected
✅ Cost-Effective: Uses affordable hardware (Raspberry Pi + ESP32) compared to commercial solutions
✅ Scalable: Can deploy multiple boats to cover large water bodies
✅ Real-Time Monitoring: Live dashboard with GPS tracking and data visualization

Impact

  • Early detection of harmful algal blooms
  • Mapping plastic pollution hotspots
  • Continuous water quality monitoring for environmental agencies
  • Data-driven decision making for water resource management

✨ Key Features

🤖 Autonomous Navigation

  • GPS-based waypoint navigation
  • IMU and magnetometer for orientation
  • Obstacle avoidance capabilities
  • Return-to-home functionality
  • Mission planning with custom routes

🔬 Advanced Sensing

  • AS7265x Spectral Sensor: 18 channels from 410nm to 940nm
  • Pi Camera NoIR: Visual documentation and plastic detection
  • GPS NEO-6M: Position tracking with 2.5m accuracy
  • MPU6050 IMU: Pitch/roll compensation
  • HMC5883L Magnetometer: Heading calculation
  • TDS Sensor: Total dissolved solids measurement

🧠 Dual AI System

Macro AI (Plastic & Algae Detection)

  • YOLOv8n model trained on 5000+ images
  • Detects plastic bottles, bags, and algae blooms
  • Real-time inference on camera feed
  • 70+ training epochs for high accuracy

Micro AI (Chlorophyll Prediction)

  • CatBoost regression model
  • Predicts chlorophyll-A concentration from spectral data
  • SHAP explainability for model interpretability
  • Trained on calibrated spectral measurements

🌐 Connectivity & Communication

  • MQTT protocol for real-time updates
  • Offline SQLite database with cloud sync
  • WebSocket for live dashboard updates
  • RESTful API for data access
  • MongoDB cloud storage

📊 Web Dashboard

  • Real-time boat tracking on interactive map
  • Live sensor data visualization
  • Mission management interface
  • Historical data analysis with charts
  • Remote control capabilities
  • Alert system for anomalies

🔋 Power Management

  • Battery voltage monitoring
  • Low-power modes
  • Automatic shutdown on critical battery
  • Solar charging support (optional)

🏗️ System Architecture

┌─────────────────────────────────────────────────────────┐
│                   CLOUD BACKEND                         │
│  Express.js + MongoDB + MQTT Broker + WebSocket        │
│  • Mission management                                   │
│  • Data storage & analytics                             │
│  • Real-time updates                                    │
└───────────────────────┬─────────────────────────────────┘
                        │ MQTT/HTTPS
                        │ (when online)
┌───────────────────────▼─────────────────────────────────┐
│             RASPBERRY PI 4 (Master Controller)          │
│  Python + SQLite + MQTT Client                          │
│  • Navigation & Mission Control                         │
│  • Sensor Data Acquisition                              │
│  • AI Inference (Chlorophyll + Plastic)                 │
│  • Local Database & Sync Service                        │
│                                                          │
│  Sensors:                                               │
│  • AS7265x (I²C) - 18-channel spectral                 │
│  • Pi Camera NoIR - Image capture                       │
│  • NEO-6M GPS - Position                                │
│  • MPU6050 - IMU                                        │
│  • HMC5883L - Magnetometer                              │
└───────────────────────┬─────────────────────────────────┘
                        │ USB Serial
                        │ (115200 baud)
┌───────────────────────▼─────────────────────────────────┐
│              ESP32 (Motor Controller)                   │
│  Arduino C++ Firmware                                   │
│  • Controls 2× DC motors (L298N)                        │
│  • Controls 1× peristaltic pump                         │
│  • Battery voltage monitoring                           │
│  • Receives commands from RPi                           │
└───────────────────────┬─────────────────────────────────┘
                        │
                ┌───────┴────────┐
                │                │
          ┌─────▼─────┐    ┌────▼─────┐
          │  L298N #1 │    │ L298N #2 │
          │  Left +   │    │   Pump   │
          │  Right    │    │  Motor   │
          │  Motors   │    │          │
          └───────────┘    └──────────┘

🛠️ Technologies Used

Hardware

  • Raspberry Pi 4 (4GB RAM) - Main controller
  • ESP32 DevKit - Motor controller
  • AS7265x - 18-channel spectral sensor (Triad Spectroscopy)
  • Pi Camera NoIR - Near-infrared camera
  • NEO-6M GPS - Global positioning
  • MPU6050 - 6-axis IMU
  • HMC5883L - 3-axis magnetometer
  • 2× L298N - Dual H-bridge motor drivers
  • TDS Sensor - Water conductivity
  • 12V DC Motors - Propulsion
  • Peristaltic Pump - Water sampling

Software - Backend

  • Python 3.9+ - Raspberry Pi controller
  • Node.js + Express - Backend API server
  • MongoDB - Cloud database
  • SQLite - Local offline database
  • MQTT (Mosquitto) - Real-time communication
  • FastAPI - AI model inference API

Software - Frontend

  • Next.js 14 - React framework
  • TypeScript - Type-safe development
  • Tailwind CSS - Styling
  • Leaflet - Interactive maps
  • Chart.js - Data visualization
  • Socket.io - Real-time updates

AI/ML Models

  • YOLOv8n (Ultralytics) - Object detection
  • CatBoost - Gradient boosting for regression
  • LightGBM - Alternative ML model
  • SHAP - Model explainability
  • scikit-learn - Preprocessing & metrics
  • OpenCV - Image processing

Development Tools

  • Arduino IDE - ESP32 firmware
  • VS Code - Development environment
  • Git - Version control
  • Docker (optional) - Containerization

Communication Protocols

  • MQTT - Pub/Sub messaging
  • WebSocket - Real-time web updates
  • I²C - Sensor communication
  • UART/Serial - RPi ↔ ESP32 communication
  • REST API - HTTP endpoints

📦 Installation & Setup

Prerequisites

  • Raspberry Pi 4 (4GB+ RAM) with Raspberry Pi OS
  • ESP32 development board
  • Node.js 18+ (for backend server)
  • Python 3.9+ (for Raspberry Pi)
  • Arduino IDE (for ESP32 firmware)
  • MongoDB database (local or cloud)
  • MQTT broker (Mosquitto or cloud service)

Step 1: Hardware Assembly

Refer to detailed wiring guides:

Step 2: Setup Raspberry Pi

# SSH into your Raspberry Pi
ssh pi@raspberrypi.local

# Update system
sudo apt update && sudo apt upgrade -y

# Enable I2C, Serial, and Camera
sudo raspi-config
# Interface Options → Enable I2C, Serial, Camera
# Reboot after changes

# Clone repository
git clone https://github.com/AhqafCoder/GANDHARVAS_SIH25204.git
cd GANDHARVAS_SIH25204/raspberry-pi

# Install Python dependencies
pip3 install -r requirements.txt

# Configure settings
cp config.yaml.example config.yaml
nano config.yaml  # Edit with your settings

Step 3: Setup ESP32 Firmware

# On your development PC (Windows/Mac/Linux)

# 1. Install Arduino IDE from https://www.arduino.cc/

# 2. Add ESP32 board support:
#    File → Preferences → Additional Board Manager URLs:
#    https://dl.espressif.com/dl/package_esp32_index.json

# 3. Install ESP32 boards:
#    Tools → Board → Boards Manager → Search "ESP32" → Install

# 4. Open firmware
#    File → Open → firmware/esp32_motor_controller/esp32_motor_controller.ino

# 5. Configure
#    Tools → Board → ESP32 Dev Module
#    Tools → Port → (Select your COM port)

# 6. Upload
#    Click Upload button (→)

Step 4: Setup Backend Server

# On server machine or cloud instance

# Clone repository
git clone https://github.com/AhqafCoder/GANDHARVAS_SIH25204.git
cd GANDHARVAS_SIH25204/server

# Install dependencies
npm install

# Setup environment variables (see below)
cp .env.example .env
nano .env  # Configure your settings

# Start server
npm start

# Or with PM2 for production
npm install -g pm2
pm2 start index.js --name boat-server

Step 5: Setup Frontend Dashboard

# On your development machine

cd GANDHARVAS_SIH25204/client

# Install dependencies
npm install

# Configure API endpoint
# Edit .env.local
echo "NEXT_PUBLIC_API_URL=http://your-server:3000" > .env.local
echo "NEXT_PUBLIC_WS_URL=ws://your-server:3000" >> .env.local

# Development mode
npm run dev

# Production build
npm run build
npm start

Step 6: Setup AI Models

# Setup Macro AI (YOLOv8)
cd GANDHARVAS_SIH25204/ai\ macro/dataset-pipeline

# Install dependencies
pip install -r requirements.txt

# Download/train model
python train_70epochs.py

# Setup Micro AI (CatBoost)
cd ../../ai\ micro

# Install dependencies
pip install -r requirements.txt

# Train models
python train_improved.py

# Start inference API
cd api
uvicorn main:app --host 0.0.0.0 --port 8000

🔐 Environment Variables

Backend Server (.env)

# Server Configuration
PORT=3000
NODE_ENV=production

# MongoDB
MONGODB_URI=mongodb://localhost:27017/boat_monitoring
# Or MongoDB Atlas:
# MONGODB_URI=mongodb+srv://username:password@cluster.mongodb.net/boat_monitoring

# MQTT Broker
MQTT_BROKER_URL=mqtt://localhost:1883
# Or cloud broker:
# MQTT_BROKER_URL=mqtts://broker.hivemq.com:8883
MQTT_USERNAME=boat_user
MQTT_PASSWORD=your_secure_password
MQTT_CLIENT_ID=backend_server

# JWT Authentication (Optional)
JWT_SECRET=your_jwt_secret_key_here

# File Upload
UPLOAD_DIR=./uploads
MAX_FILE_SIZE=10485760  # 10MB

# API Keys (if using external services)
WEATHER_API_KEY=your_weather_api_key
MAPS_API_KEY=your_maps_api_key

Raspberry Pi (config.yaml)

# Boat Configuration
boat_id: "BOAT_001"

# Server Connection
mqtt:
  broker: "mqtt://your-server.com:1883"
  username: "boat_user"
  password: "your_password"
  topics:
    status: "boat/BOAT_001/status"
    data: "boat/BOAT_001/data"
    commands: "boat/BOAT_001/commands"

# Hardware Settings
hardware:
  esp32_port: "/dev/ttyUSB0"
  esp32_baud: 115200
  gps_port: "/dev/serial0"
  gps_baud: 9600
  camera_resolution: [1920, 1080]

# Sampling Configuration
sampling:
  interval_meters: 10  # Sample every 10 meters
  spectral_integration_time: 50  # ms
  pump_duration: 5  # seconds

# AI Models
ai:
  yolo_model: "models/yolov8n_plastic_algae.pt"
  chlorophyll_model: "http://your-server:8000/predict/chlorophyll"
  confidence_threshold: 0.6

# Database
database:
  local_path: "data/boat_data.db"
  sync_interval: 300  # seconds (5 minutes)

Frontend (.env.local)

NEXT_PUBLIC_API_URL=http://localhost:3000
NEXT_PUBLIC_WS_URL=ws://localhost:3000
NEXT_PUBLIC_MQTT_BROKER=ws://localhost:9001
NEXT_PUBLIC_MAP_TILE_URL=https://{s}.tile.openstreetmap.org/{z}/{x}/{y}.png

🚀 Usage Guide

Starting the System

  1. Start Backend Server

    cd server
npm start

  2. Start Raspberry Pi Controller

    cd raspberry-pi
python3 boat_controller.py

  3. Start Frontend Dashboard

    cd client
npm run dev

  4. Access Dashboard

    • Open browser: http://localhost:3000

Creating a Mission

  1. Navigate to Mission Planner tab
  2. Click on map to add waypoints
  3. Set mission parameters:
    • Start coordinates
    • Destination coordinates
    • Sampling interval (meters)
    • Waypoint sequence
  4. Click Save Mission
  5. Deploy boat and click Start Mission

Monitoring Real-Time Data

  • Map View: See boat position, track history, waypoints
  • Dashboard: Live sensor readings, water quality metrics
  • Charts: Time-series graphs of chlorophyll, TDS, etc.
  • Alerts: Notifications for anomalies (high chlorophyll, plastic detected)

Manual Control

Use the Command Panel to:

  • FORWARD - Move forward
  • BACKWARD - Move backward
  • LEFT - Turn left
  • RIGHT - Turn right
  • STOP - Stop all motors
  • PUMP_ON - Start water sampling
  • PUMP_OFF - Stop pump
  • RTH - Return to home

Data Analysis

  1. Navigate to Data Explorer
  2. Filter by date range, location, or parameters
  3. Export data as CSV/JSON
  4. View AI predictions and confidence scores
  5. Download captured images

📸 Screenshots

Dashboard Overview

Dashboard

Mission Planning

Mission Planner

Water Quality Charts

Charts

Plastic Detection

Plastic Detection

Boat Hardware

Hardware

Note: Replace placeholder images with actual screenshots from your deployed system.

👥 Team

Team Gandharvas - Smart India Hackathon 2025
Institute: Invertis University

Name Role
Sarthak Maurya R&D Lead & Hardware Lead
Aadya Saraswat R&D and Designing Lead
Ahqaf Ali Hardware & Backend Lead
Vivek Vishwakarma AI/ML Lead
Lal Shivam Singh UI/UX Lead
Trisha Bhardwaj Documentation Lead

📚 Documentation

For detailed documentation, refer to:

🔧 Troubleshooting

Boat not connecting to server

  • Check internet connectivity
  • Verify MQTT broker URL and credentials
  • Check firewall settings on server

Sensors not reading

  • Verify I2C connections: i2cdetect -y 1
  • Check power supply to sensors
  • Ensure proper pull-up resistors on I2C lines

Motors not responding

  • Test ESP32 serial communication: screen /dev/ttyUSB0 115200
  • Check motor driver connections
  • Verify battery voltage (should be 11-12.6V)

GPS not getting fix

  • Move to open area (away from buildings)
  • Wait 2-3 minutes for satellite lock
  • Check GPS antenna connection

AI models not working

  • Ensure model files are in correct paths
  • Check FastAPI server is running
  • Verify network connectivity to inference API

🤝 Contributing

We welcome contributions! Please follow these steps:

  1. Fork the repository
  2. Create a feature branch (git checkout -b feature/AmazingFeature)
  3. Commit your changes (git commit -m 'Add some AmazingFeature')
  4. Push to the branch (git push origin feature/AmazingFeature)
  5. Open a Pull Request

📄 License

This project is developed for Smart India Hackathon 2025.

🙏 Acknowledgments

  • Smart India Hackathon for the opportunity
  • Invertis University for support and resources
  • Ultralytics for YOLOv8 framework
  • CatBoost team for ML library
  • OpenStreetMap for map tiles

📞 Contact

For questions or support:

Made with ❤️ by Team Gandharvas

⭐ Star this repository if you find it helpful!

About

No description, website, or topics provided.

Resources

Readme
Activity

Stars

1 star

Watchers

0 watching

Forks

0 forks
Report repository

Releases

Packages

Contributors

Languages