🌍 Green Tunisia — Smart Multi-Compartment Waste Management
📌 Overview
Green Tunisia is a smart waste management system developed for Sfax, Tunisia. It combines:
IoT-enabled multi-compartment bins (Plastic, Metal, Paper, Organic, Bread, Glass).
A web application for real-time monitoring and route optimization.
An AI model (LSTM) to forecast bin fullness and schedule garbage truck trips efficiently.
ESP32 firmware to convert ultrasonic sensor data into fill-level percentages.
The project addresses key challenges such as overflowing bins, static collection schedules, high fuel consumption, and lack of real-time data.
🗂 Project Components
1. 📡 IoT Firmware (ESP32)
Reads distance using an ultrasonic sensor.
Converts distance measurements into bin fill percentage.
Sends the fill-level data to the server.
2. 🌐 Web Application
Live dashboard with map of bins and fill levels.
Route optimization (Nearest Neighbor + 2-Opt).
Backend (Flask + DB) for data storage and API.
Operator interface + Driver PWA with collection checklist.
3. 🤖 AI Model
LSTM neural network for short-term fill-level forecasting.
Inputs: last 6h data (fill % history, time encoding, zone, waste type).
Outputs: next 6h predicted fill levels.
Trigger rule: dispatch a truck if predicted bins (≥80%) match truck capacity.
Metrics: MAE, RMSE, R², classification accuracy for ≥80% bins.