🌲 SmartPark IoT

Environmental Monitoring System for I Giganti della Sila – Riserva Naturale Biogenetica di Fallistro

SmartPark IoT is a full-stack Internet of Things (IoT) system developed as part of an IoT university course project. The system focuses on real-time environmental monitoring and visitor support for I Giganti della Sila – Riserva Naturale Biogenetica di Fallistro, a protected ancient forest located in Sila National Park, Calabria, Italy.

The reserve contains 58 monumental Laricio pine trees (Pinus nigra subsp. laricio), some over 350 years old and reaching heights of up to 45 metres. The area is managed by FAI – Fondo Ambiente Italiano and represents one of the most ecologically significant forest ecosystems in southern Italy.

The platform integrates IoT sensing devices, environmental data collection, backend services, a visitor-facing web application, and an administrative monitoring dashboard.

🌍 Project Overview

The SmartPark IoT platform collects environmental and meteorological data from IoT sensors deployed across the reserve. Sensor readings are ingested by a backend API, stored in a time-series database, and made accessible through two dedicated interfaces.

Key Features

📡 Real-time sensor data ingestion from IoT devices deployed in the field
🌡️ Environmental monitoring — temperature, humidity, air quality, and weather conditions
🗺️ Trail recommendations based on live weather data and visitor preferences
🔐 User authentication via Firebase for personalised visitor experience
📊 Administrative dashboard for park staff to monitor system and sensor status
📈 Time-series visualisation via Grafana for historical data analysis

🏗 System Architecture

The system follows a two-tier architecture: an Edge Tier (sensors and gateway with on-device ML inference),and a Cloud Tier (remote management, cloud sync, and web applications).

╔══════════════════════════════════════════════════════════════╗
║                        EDGE TIER                             ║
║                                                              ║
║   S6000U Sensor  ──(Modbus RTU / RS485)──►  Robustel EG5120 ║
║   (Temp, Humidity, Pressure, Light,          │               ║
║    Noise, ToF, GPS)                          │ App Center    ║
║                                              ├─ Node-RED     ║
║                                              │   (orchestration,║
║                                              │    Modbus poll,  ║
║                                              │    GPS parsing,  ║
║                                              │    ML exec)      ║
║                                              ├─ InfluxDB     ║
║                                              │   (local time-║
║                                              │    series DB) ║
║                                              ├─ Grafana      ║
║                                              │   (on-device  ║
║                                              │    dashboard) ║
║                                              └─ 
╚══════════════════════════════╤═══════════════════════════════╝
                               │ HTTPS / 4G-5G Cellular
╔══════════════════════════════▼═══════════════════════════════╗
║                        CLOUD TIER                            ║
║                                                              ║
║  ┌─────────────────┐        ┌──────────────────────────┐    ║
║  │  InfluxDB Cloud │        │  Firebase Firestore       │    ║
║  │  (AWS region)   │        │  - User auth & prefs      │    ║
║  │  Long-term      │        │  - Trail tags & metadata  │    ║
║  │  telemetry      │        │  - Recommendations        │    ║
║  └────────┬────────┘        └──────────┬───────────────┘    ║
║           │                            │                     ║
║           ▼                            ▼                     ║
║  ┌──────────────────────────────────────────────────────┐    ║
║  │               FastAPI Backend                        │    ║
║  │  - Data ingestion & validation                       │    ║
║  │  - REST endpoints for both frontends                 │    ║
║  └──────────────┬───────────────────────────────────────┘    ║
║                 │                                            ║
║      ┌──────────┴──────────┐                                 ║
║      ▼                     ▼                                 ║
║  ┌───────────────┐   ┌──────────────────────────────────┐   ║
║  │ Admin Dashboard│   │     Visitor Web Application       │   ║
║  │ (Vue.js 3)    │   │     (HTML5 / CSS3 / Vanilla JS)   │   ║
║  │               │   │                                   │   ║
║  │ - Station map │   │  - Weather dashboard              │   ║
║  │ - Live sensor │   │  - Interactive map (Leaflet.js)   │   ║
║  │   readings    │   │  - Firebase Auth                  │   ║
║  │ - Historical  │   │  - Trail recommendations (CBF)    │   ║
║  │   trends      │   │  
║  │ - RAG Chatbot │   │                                   │   ║
║  └───────────────┘   └──────────────┬────────────────────┘   ║
║                                     │                        ║
║                             ┌───────▼──────────┐            ║
║                             │  ML Engine        │            ║
║                             │  (Node.js/Express)│            ║
║                             │  Content-Based    │            ║
║                             │  Filtering (CBF)  │            ║
║                             └───────────────────┘            ║
╚══════════════════════════════════════════════════════════════╝

Architecture Layers

1. Hardware Layer — The Robustel EG5120 (ARM Cortex-A7, 512 MB RAM, 4G LTE) acts as the central edge computing hub. The S6000U multi-purpose sensor connects via RS485/Modbus RTU and measures temperature, humidity, pressure, light, noise, ToF distance, and orientation.

2. Edge Processing Layer — Node-RED runs on the EG5120 and orchestrates the entire local pipeline: it polls the S6000U every 60 seconds via Modbus, parses GPS NMEA sentences, runs the Random Forest predictor (predictor.py) via an exec node, fuses sensor + GPS + ML results, writes to local InfluxDB, and syncs to the cloud.

3. ML Inference (Edge) — A Random Forest multi-label classifier trained on one year of historical weather data from Open-Meteo classifies current conditions into "Weather Vibe" categories (Frosty, Brisk, Crisp, Moody, Serene, Sun-Drenched). The model runs directly on the EG5120 and pushes trail tags to Firebase Firestore.

4. Storage Layer — Local InfluxDB on the gateway provides edge buffering. InfluxDB Cloud (AWS) stores long-term telemetry. Firebase Firestore holds user data, trail metadata, and recommendation outputs.

5. Application Layer — The FastAPI backend serves both frontends, handles data ingestion, and integrates a RAG-based AI chatbot (Groq API + Llama-3.3-70b + ChromaDB + LangChain). The Visitor Web App and Admin Dashboard consume these services. Trail recommendations are generated by a dedicated Node.js/Express CBF engine.

📂 Repository Structure

smart-park-iot/
│
├── Dockerfile                        # Root Dockerfile
├── compose.yaml                      # Main Docker Compose configuration
├── compose.debug.yaml                # Debug/development Compose override
├── nginx.conf                        # Nginx reverse proxy configuration
│
├── IoT_ProjectWeatherForcast/
│   ├── app/                          # FastAPI backend application
│   │   ├── main.py                   # Application entry point
│   │   ├── routers/                  # API route handlers
│   │   ├── models/                   # Data models / schemas
│   │   ├── services/                 # Business logic layer
│   │   └── database/                 # InfluxDB client & queries
│   ├── weather/                      # Admin dashboard (Vue.js)
│   │   ├── src/
│   │   │   ├── components/           # Reusable Vue components
│   │   │   ├── views/                # Page-level views
│   │   │   └── stores/               # State management (Pinia)
│   │   └── vite.config.js
│   ├── Robustel EG5120/              # IoT gateway configuration files
│   ├── Robustel EG5120 ML/           # Gateway ML integration config
│   ├── requirements.txt              # Python dependencies
│   └── .env.example                  # Environment variable template
│
├── web-app/                          # Visitor-facing web application
│   ├── index.html
│   ├── assets/
│   ├── public/
│   └── src/
│
├── ml-engines/                       # Machine learning recommendation engine
│   ├── generate_recom.js.js
│   ├── server.js
│   └── package.json
│
├── firebase-config/                  # Firebase project configuration
├── database/                         # Database schemas and seed data
└── docs/                             # Project documentation and diagrams

⚙️ Technology Stack

Layer	Technology	Purpose
IoT Sensor	Robustel S6000U (RS485 / Modbus RTU)	Multi-parameter environmental sensing
IoT Gateway	Robustel EG5120 (ARM Cortex-A7, 4G LTE, GPS)	Edge computing hub; data acquisition & forwarding
Edge Orchestration	Node-RED (on EG5120 App Center)	Modbus polling, GPS parsing, ML exec, cloud sync
Edge ML Inference	Python 3, scikit-learn Random Forest, joblib	Weather Vibe classification at the edge
Edge Monitoring	Grafana + InfluxDB (on EG5120)	Local real-time dashboards and data buffering
Backend API	Python 3.11+, FastAPI, Uvicorn	Data ingestion, REST API, RAG chatbot
Trail Recommendation	Node.js 18+, Express, Content-Based Filtering (CBF)	Personalised trail ranking from user preferences
Visitor Web Application	HTML5, CSS3 (Bootstrap 5), Vanilla JS, Leaflet.js	Visitor-facing UI: weather, map, recommendations
Admin Dashboard	Vue.js 3, Vite, Pinia	SPA for park administrators
Cloud Time-Series DB	InfluxDB Cloud (AWS)	Long-term telemetry storage and analytics
User Data & Sync	Firebase Firestore + Firebase Authentication	User profiles, trail tags, recommendations
Infrastructure	Docker, Docker Compose, Nginx	Containerisation, routing, static file serving
Remote Device Management	Robustel RCMS	OTA updates, remote Node-RED deploy, diagnostics

🛠 Prerequisites

Before running the project, ensure the following tools are installed on your machine:

Tool	Minimum Version	Installation
Docker	24.x	https://docs.docker.com/get-docker/
Docker Compose	2.x (plugin)	Included with Docker Desktop
Git	2.x	https://git-scm.com/

🚀 Getting Started

1. Clone the repository

git clone https://github.com/jhenals/smart-park-iot.git
cd smart-park-iot

2. Configure environment variables

Copy the environment variable template and fill in the required values:

cp IoT_ProjectWeatherForcast/.env.example IoT_ProjectWeatherForcast/.env

Open .env in your editor and configure the values. See the Environment Variables section for a full description of each variable.

3. Build and start all services

docker compose up --build

To run in detached (background) mode:

docker compose up --build -d

The first build may take several minutes as Docker pulls base images and installs dependencies.

4. Verify services are running

docker compose ps

All services should show a status of running. If any service has exited, check its logs:

docker compose logs <service-name>

5. Access the applications

Service	URL
Visitor Web Application	http://localhost:8081
FastAPI Backend	http://localhost:8000
Admin Dashboard	http://localhost:5173
InfluxDB	http://localhost:8086
Grafana Dashboard	http://localhost:3001

6. Stop the system

docker compose down

To also remove all volumes (database data will be lost):

docker compose down -v

🔐 Environment Variables

The following variables must be set in IoT_ProjectWeatherForcast/.env:

Variable	Description	Example
`INFLUXDB_URL`	URL of the InfluxDB instance	`http://influxdb:8086`
`INFLUXDB_TOKEN`	InfluxDB authentication token	`your-influxdb-token`
`INFLUXDB_ORG`	InfluxDB organisation name	`smartpark`
`INFLUXDB_BUCKET`	InfluxDB bucket for sensor data	`sensor_data`
`FIREBASE_PROJECT_ID`	Firebase project ID	`smartpark-iot`
`FIREBASE_CREDENTIALS`	Path to Firebase service account JSON	`./firebase-config/key.json`
`GRAFANA_ADMIN_USER`	Grafana admin username	`admin`
`GRAFANA_ADMIN_PASSWORD`	Grafana admin password	`changeme`

⚠️ Never commit the .env file or any Firebase credentials to version control.

🔧 Services

The system is composed of several containerised services managed through Docker Compose:

Service	Container Name	Description
`webapp`	smartpark-webapp	Visitor-facing web application served by Nginx
`fastapi`	smartpark-api	Backend API for data ingestion and REST endpoints
`adminfrontend`	smartpark-admin	Vue.js administrative dashboard
`mlengine`	smartpark-ml	Node.js trail recommendation engine
`influxdb`	smartpark-influx	Time-series database storing sensor telemetry
`grafana`	smartpark-grafana	Monitoring dashboards and data visualisation

🧩 Project Modules

Visitor Web Application (`web-app/`)

A lightweight, static web interface designed for park visitors. It provides:

Environmental conditions panel — live temperature, humidity, and air quality readings retrieved from the backend API
Weather information — current and forecast data relevant to the reserve
Trail recommendation system — suggests trails based on current weather conditions and visitor-selected preferences
User authentication — visitors can log in via Firebase to save preferences and view personalised suggestions

The application is served by Nginx and is intentionally lightweight to ensure fast load times, even on mobile networks.

Backend API (`IoT_ProjectWeatherForcast/app/`)

A FastAPI-based REST API responsible for:

Receiving and validating data from IoT sensors and the Robustel EG5120 gateway
Writing telemetry to InfluxDB using the InfluxDB Python client
Exposing REST endpoints consumed by both the visitor web app and the admin dashboard
Providing automatic interactive API documentation at /docs (Swagger UI) and /redoc (ReDoc)

Admin Dashboard (`IoT_ProjectWeatherForcast/weather/`)

A Vue.js 3 single-page application providing park administrators with:

Real-time and historical sensor data visualisation
System health and connectivity status for each IoT device
Chart-based exploration of environmental trends over time

Machine Learning Engine (`ml-engines/`)

A Node.js module that:

Analyses current and forecast weather data retrieved from the backend
Combines environmental data with visitor-submitted preferences (duration, difficulty, accessibility)
Produces ranked trail recommendations returned to the visitor web application

IoT Gateway (Robustel EG5120)

The Robustel EG5120 acts as an industrial-grade cellular IoT gateway deployed in the field. It aggregates data from connected environmental sensors and forwards readings to the backend over the mobile network. The Robustel EG5120 ML/ directory contains additional configuration for edge-side ML pre-processing.

📖 API Reference

The FastAPI backend automatically generates interactive documentation. Once the system is running, visit:

Swagger UI → http://localhost:8000/docs
ReDoc → http://localhost:8000/redoc

Key Endpoints (overview)

Method	Endpoint	Description
GET	`/api/weather/forecast/?minutes={minutes}&measurement={name}`	minutes={minutes}&measurement={name}Retrieve sensor data for the specified time window and measurement

Full request/response schemas are available in the interactive documentation.

👥 Project Contributions

This project was developed as a group assignment for a university IoT course. The system was built collaboratively, with team members contributing to different components of the stack.

Group contributions covered:

IoT device configuration and gateway setup (Robustel EG5120)
FastAPI backend development and InfluxDB integration
Admin dashboard (Vue.js)
System integration, Docker Compose orchestration, and deployment

My individual contribution focused on the visitor-facing layer of the system:

Designed and implemented the visitor web application (HTML, CSS, JavaScript)
Integrated the web interface with the FastAPI backend to display live environmental data
Connected the application to the ML engine for trail recommendations
Integrated Firebase Authentication for user login and personalised experience
Modified and adapted existing backend endpoints where necessary to meet frontend requirements
Contributed to system integration testing across the full stack

Parts of the backend and infrastructure in this repository originate from the original group project and have been adapted where needed to support the web application integration.

🐛 Troubleshooting

A service fails to start

Check the logs for that specific service:

docker compose logs fastapi
docker compose logs influxdb

InfluxDB connection errors

Ensure the INFLUXDB_TOKEN, INFLUXDB_ORG, and INFLUXDB_BUCKET values in your .env file match the values used when InfluxDB was initialised. If running for the first time, InfluxDB may take a few seconds to become ready — the backend will retry automatically.

Port conflicts

If a port is already in use on your machine, edit compose.yaml and change the host-side port mapping. For example, to change the web app from port 8081 to 9081:

ports:
  - "9081:80"

Admin dashboard not loading

The admin dashboard (adminfrontend) runs a Vite development server. If it is slow to start, wait 15–20 seconds after docker compose up and then refresh the browser.

Rebuilding after code changes

docker compose up --build

To force a full clean rebuild:

docker compose down
docker compose build --no-cache
docker compose up

📄 License

This project was developed for educational purposes as part of a university IoT course. It is not intended for production deployment without further security hardening.

🌱 About the Reserve

I Giganti della Sila – Riserva Naturale Biogenetica di Fallistro is a protected forest located in Sila National Park, near Camigliatello Silano in the province of Cosenza, Calabria, southern Italy.

The reserve takes its name from its most remarkable feature: a stand of ancient Laricio pine trees (Pinus nigra subsp. laricio) planted in the 17th century. Some trees reach 45 metres in height and nearly 2 metres in trunk diameter, making them among the largest and oldest individual trees in Italy.

The forest floor hosts a rich understorey of ferns, mosses, and flowering plants, while the broader Sila plateau supports a diverse ecosystem including wolves, deer, and a wide variety of bird species.

The reserve is managed by FAI – Fondo Ambiente Italiano (Italian Environment Fund) and is open to visitors from April to November.

📍 Location: Near Camigliatello Silano, Province of Cosenza, Calabria, Italy
🌐 More information: fondoambiente.it/i-giganti-della-sila-eng

Developed as part of an IoT course project — academic year 2025/2026.

Name		Name	Last commit message	Last commit date
Latest commit History 74 Commits
IoT_ProjectWeatherForcast		IoT_ProjectWeatherForcast
database		database
docs		docs
firebase-config		firebase-config
ml-engines		ml-engines
web-app		web-app
.dockerignore		.dockerignore
.gitignore		.gitignore
Dockerfile		Dockerfile
README.md		README.md
compose.debug.yaml		compose.debug.yaml
compose.yaml		compose.yaml
nginx.conf		nginx.conf
package-lock.json		package-lock.json

Folders and files

Latest commit

History

Repository files navigation

🌲 SmartPark IoT

Environmental Monitoring System for I Giganti della Sila – Riserva Naturale Biogenetica di Fallistro

📋 Table of Contents

🌍 Project Overview

Key Features

🏗 System Architecture

Architecture Layers

📂 Repository Structure

⚙️ Technology Stack

🛠 Prerequisites

🚀 Getting Started

1. Clone the repository

2. Configure environment variables

3. Build and start all services

4. Verify services are running

5. Access the applications

6. Stop the system

🔐 Environment Variables

🔧 Services

🧩 Project Modules

Visitor Web Application (web-app/)

Backend API (IoT_ProjectWeatherForcast/app/)

Admin Dashboard (IoT_ProjectWeatherForcast/weather/)

Machine Learning Engine (ml-engines/)

IoT Gateway (Robustel EG5120)

📖 API Reference

Key Endpoints (overview)

👥 Project Contributions

🐛 Troubleshooting

A service fails to start

InfluxDB connection errors

Port conflicts

Admin dashboard not loading

Rebuilding after code changes

📄 License

🌱 About the Reserve

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Visitor Web Application (`web-app/`)

Backend API (`IoT_ProjectWeatherForcast/app/`)

Admin Dashboard (`IoT_ProjectWeatherForcast/weather/`)

Machine Learning Engine (`ml-engines/`)

Packages