University of Salento · Internet of Things Exam Project

Raqeb Food راقب فود

IoT and AI system for intelligent school food delivery monitoring

A Raspberry Pi and React-based system that monitors food delivery conditions, tracks driver GPS from the phone browser, and supports administrators with AI-assisted driver recommendation.

View Backend Repository View Frontend Repository Explore Project
🌡️
Temperature 24.8°C
💧
Humidity 52%
📍
Phone GPS Live
🤖
AI Risk Low
Temperature & Humidity measured with DHT11
Tilt Detection using KY-017 module
GPS Tracking from driver phone browser
Edge Backend powered by Raspberry Pi 5

Project Overview

From simple delivery to monitored, traceable transportation

School meal transportation needs more than basic logistics. Raqeb Food combines IoT sensors, GPS tracking, role-based dashboards, delivery workflow management and AI recommendation to make each delivery more visible and controlled.

Problem

Condition loss can happen silently

Food deliveries for schools can be affected by temperature, humidity, unstable handling or missing traceability. Without monitoring, problems may only be discovered after the delivery is complete.

Solution

A connected delivery workflow

Raqeb Food links school requests, admin assignment, driver workflow, live sensor monitoring and delivery reports inside one prototype system designed for real-world IoT evaluation.

🌡️

Real-time Monitoring

Temperature, humidity and tilt data are collected from the delivery box through the Raspberry Pi backend.

📱

Driver Phone GPS

The driver browser sends live coordinates using the Geolocation API after delivery start.

🧠

AI Recommendation

A Decision Tree predicts delivery-condition risk and helps rank available drivers for the request.

📊

Delivery Reports

Important monitoring events are saved in the database and can be reviewed after completion.

System Architecture

A full-stack IoT prototype built around the delivery lifecycle

The following diagram presents the overall architecture of Raqeb Food, showing how the React dashboard, Flask backend, Raspberry Pi, SQLite database, IoT sensors, phone GPS and AI recommendation module interact.

Raqeb Food system architecture diagram
Architecture Diagram assets/architecture.png

Role-based Dashboards

Three interfaces, one connected workflow

Each user role has a focused dashboard. The administrator manages requests and users, the school creates and tracks delivery requests, and the driver executes the delivery from the phone.

Admin Dashboard

Control center for operations and platform users

  • Manage school delivery requests
  • Assign available drivers
  • Use AI recommendation and view reasons
  • Create and manage school and driver accounts
Operations

Request assignment and AI-assisted recommendation

The admin control center allows the administrator to review school requests, compare available drivers, display AI recommendation reasons, assign a driver, and access delivery reports.

Admin control center with AI recommendation
Admin Control Center Screenshotassets/admin-dashboard.png
Access control

User and role management

The user management page lets the administrator create accounts for drivers and schools, view registered users, and keep the role-based access structure organized.

Admin user management dashboard
Admin User Management Screenshotassets/admin-users.png

School Dashboard

Request creation and status tracking for schools

  • Create delivery requests
  • Select food items and quantities
  • Add requested delivery time and notes
  • Track submitted request statuses
Request creation

Food items, quantities and delivery time

Schools can build a delivery request by selecting food items, entering quantities, choosing the requested delivery time, and adding special notes for the administrator.

School create delivery request form
School Create Request Screenshotassets/school-create-request.png
Request tracking

Submitted requests and delivery status

The school can follow the delivery workflow by viewing submitted requests, food details, requested times and current statuses such as requested, assigned, in progress or completed.

School delivery request tracking dashboard
School My Requests Screenshotassets/school-my-requests.png

Driver Dashboard

Mobile workspace for monitored transportation

  • View assigned deliveries
  • Start and complete a delivery
  • Send GPS location from the phone browser
  • View live temperature, humidity, tilt and alerts
Driver mobile dashboard with active delivery monitoring

Live Monitoring

Important delivery events are tracked without flooding the database

The system monitors live data during active deliveries and stores meaningful event-based records such as sensor changes, tilt events, alerts and delivery start records.

🌡️

Temperature

Tracks box temperature and detects high-temperature situations.

💧

Humidity

Monitors humidity to protect sensitive food during transport.

📦

Tilt

Detects unstable handling or box movement using the KY-017 tilt module.

📍

GPS

Receives driver location from the phone browser during the active delivery.

⚠️

Alerts

Shows condition alerts when temperature, humidity or tilt becomes problematic.

Live monitoring dashboard screenshot
Live Monitoring Screenshot assets/live-monitoring.png

AI Driver Recommendation

Predicting delivery-condition risk, not judging driver quality

The AI module recommends the most suitable available driver for a request by combining request context, food sensitivity and driver monitoring history. A Decision Tree Classifier predicts the delivery-condition risk as low, medium or high. The backend then converts the prediction into a suitability score and ranks the available drivers.

Important distinction: the model predicts delivery-condition risk for a specific request context. It does not label a driver as good or bad in general.
Decision Tree Classifier Risk: Low · Medium · High
Prototype Dataset Training Saved Model New Request Driver Profiles Risk Prediction Score & Ranking Recommended Driver
AI recommendation screenshot
AI Recommendation Screenshotassets/ai-recommendation.png

Demo Flow

A complete monitored delivery scenario

The prototype demonstration follows the full path from a school request to AI-assisted assignment, live monitoring and final reporting.

01School creates request
02Admin asks AI recommendation
03Admin assigns driver
04Driver opens phone dashboard
05Driver starts delivery
06GPS and sensors update live
07Driver completes delivery
08Report is generated

Components & Repositories

Two main project parts plus the showcase website

The backend contains the Raspberry Pi, Flask, IoT, database and AI logic. The frontend contains the React dashboards for the three roles.

🧩

Backend / IoT / AI

Flask API, Raspberry Pi monitoring, SQLite database, authentication, delivery workflow and AI recommendation.

Open Backend Repo
🖥️

Frontend Dashboard

React role-based interface for admins, schools and drivers, including live monitoring and GPS workflow.

Open Frontend Repo
🌐

GitHub Pages Website

Static showcase website for presenting the project structure, demo flow, components and future work.

Open Website Repo

Limitations & Future Work

A prototype designed to grow into a stronger real deployment

Prototype Dataset

The current AI dataset is synthetic/prototype-based and can later be replaced by real deployment data.

Production GPS

A production deployment should use HTTPS hosting for reliable browser-based GPS tracking.

Real-time Map

A live map could be added to visualize the delivery route and current driver position.

More Sensors

Additional sensors could improve monitoring accuracy and food safety analysis.

Advanced AI

Future models could use more real examples and more advanced recommendation strategies.

Route History

Route tracking could be improved with richer historical GPS traces and reports.