Forest Fire Detection

Summary

1. Introduction: Problem introduction and a possible solution
2. Architecture: Architecture of the idea
3. Project structure: How the project is organized
4. Getting started: Guide to run the project

Introduction

This is a project for the exam of Serverless Computing for IoT.

Forest fire is one of the world's biggest calamities that produce significant damage to the atmosphere and also the injunction of danger environment.
Internet of things and drone technology are fested booming sectors, that evolves rapidly. This advanced technology can prevent such kind of incident and also making fast decision without the intervention of the humans.
The proposed service oriented architecture involves a series of communication between devices and distribute in real-time all the data to the Ground Station. The adoption of all this intelligent devices, with communication capability enable a new concept called Internet of Forest Things.

The architecture implements sensors which are installed on the tree that detects the presents of fire and immediately sents the right coordinate to the Unmanned Air Vehicles (UAVs). The UAV immediately takeoff and fly over the forest, reached the flame spot, the UAV releases the fire balls extinguisher and than return to launch position.

Architecture

From the architecture, in the block "Iot Forest Sensors" we have three types of sending fire alarm, on the top we have the HW prototype made with an ESP32 microcontroller properly connected to a flame detector sensor, led/buzeer and an LCD screen.
Initially to speed up development, the flame alarm was simulated by two different nodes, one through a sendfirealarm function created through the Nuclio Framework and the other through the Node-RED framework. Both simulated and real sensors use the MQTT protocol to publish their status.
When the sensor detects a flame sends a notification in the RabbitMQ queue forest/iot/fire, the foresthandler function through a trigger reads all messages sent on the previous topic and when the flame alert condition occurs, the function publish an alert message on the topic forest/iot/alert, then two events arise:

1. HTTP request to IFTTT Webhook, configured to to send an alert messages and flame coordinates to a telegram bot.
2. Activation of the alert status in the Forest Dashboard, publish the flame coordinates to the connected UAV unit that is subscribed to the topic.

The unit received the coordinates, initializes the flight parameters, loads the flight mission, arms the engines and start the mission. The unit in flight continues to publish on the topic drone/sensor the parameters necessary for geolocation and also the status of the drone.
The drone reached the fire spot drops the fire ball extinguisher and publish on the topic forest/drone/released the actual release, then sets his status to return to launch position, at the same time the nuclio function dronehandler is triggered, so a new messase is sent to update the flame status alarm.

Project Structure

1. Function

• foresthandler.js: Nuclio function that is executed when a fire is detected;
• dronehandler.js: Nuclio function that is executed when the drone released the fire ball extinguisher;
• sendfirealarm.js: Nuclio function to simulare a fire alarm;
• .yaml: function exported from Nuclio framework with all the deploying informations.

2. Sensor

• IoTForestSensor.ino: Sketch for the ESP32 MCU, used as flame detector.

3. Node-RED

• flow.json: Node-RED exported flow, in detail:
  • Forest IoT Sensors: flow used for the simulation of the different sensors usefull for fire detection;
  • Forest Dashboard: flow used to visualize the data stream from the forest sensors (with optional button for sensor simulation);
  • Drone Dashboard: flow used to visualize the status and the paramenter of the UAV.

4. Drone

• mqtt_drone.py: file containing all the function used by the drone to pub/sub and run specific corridor mission;
• mqtt_config.json: contains the parameters to configure the drone as mqtt client.

Getting Started

Important Notes

Install Node.js using the installer Download Install Docker following Docker installation guide Get Docker.

• In order to execute the drone python script you need Python 3.7+: run python --version or python3 --version in a terminal to check the installed version
• If not installed use the following link: Python.
• Install pip3 package manager or run pip3 instead of pip and python3 instead of python, depending of your system defaults
• Install: git, MAVSDK, aioconsole and aysncio-mqtt

IFTTT

Register on IFTTT and create a new applet by adding on IF clause WebHooks service > Receive a web request with the following parameter:

and with a THEN clause Telegram > Send Message with the following parameters:

Once done, by going into My Applets > Fire Notification Applets > WebHook > Documentation and copy the ifttt_event_key. That key need to be saved and used in the Nuclio Forest function and Nuclio Drone function

Nuclio

Start Nuclio using a docker container.

$ docker run -p 8070:8070 -v /var/run/docker.sock:/var/run/docker.sock -v /tmp:/tmp nuclio/dashboard:stable-amd64

Update and deploy Functions:

• Browse to http://localhost:8070 to open the homepage of Nuclio;
• Create new project and call it Forest Fire Detection;
• Press Create function, Import and upload the .yaml function that are in the Function folder;
• Change the ip, port, username and password according with your MQTT broker also update the ifttt_event_key generated from IFTTT;
• On Trigger tab also change the ip, port, username and password according with your MQTT Broker;
• Press Deploy.

RabbitMQ

Start RabbitMQ instance with MQTT enabled using docker.

$ docker run -p 9000:15672  -p 1883:1883 -p 5672:5672  cyrilix/rabbitmq-mqtt

Browse to http://localhost:9000. The default username is guest, and the password is guest

PX4 SITL (Software In The Loop)

Start PX4 SITL instance using docker.

$ docker run --rm -it jonasvautherin/px4-gazebo-headless:1.13.0

MAVSDK Python

Is a Python wrapper for MAVSDK. MAVSDK is a set of libraries providing a high-level API to MAVLink, it allow to write programs and to control a MAVLink-enabled drone.

Parameters

In order to execute Python script (mqtt_drone.py) you need to set the following parameters on (mqtt_config.json):

• ip: address of the broker;
• port: port of the broker;
• username: user used for the authentication;
• password: psw used for the authentication.

Once done execute the script:

$ python3 mqtt_drone.py

Node-RED

Start Node-RED instance using docker.

$ docker run -it -p 1880:1880 -v node_red_data:/data --name mynodered nodered/node-red

Installation and Configuration:

• Browse to http://localhost:1880 to open the homepage of Node-RED;
• 'Import' the ([flow.json])(Node-RED/flow.json);
• Press Manage palette and from Install tab, search for the following palette:
  • node-red-dashboard;
  • node-red-contrib-web-worldmap;
  • node-red-contrib-ui-led;
  • node-red-contrib-mqtt-broker;
• Press on one of the mqtt nodes, edit mqtt-broker node and change according with your MQTT Broker, Server and _Port in Connection and User and Password in Security;

Forest IoT Sensor

Connect all the components according with [ESP32_Connections.fzz])(IoTForestSensor/ESP32_Connections.fzz);

Installation and Configuration:

• Start Arduino IDE and Open [IoTForestSensor.ino])(IoT_Forest_Sensor/IoTForestSensor.ino);
• Configure mqtt_server and mqtt_port according with your MQTT Broker and set ssid and password;
• Press and select the right board: Tools > Board > ESP32 Arduino > DOIT ESP32 DEVKIT V1
• Uodate the Upload Speed to 115200 and Port with the correct one;
• Verify and Upload the Sketch on the ESP32 MCU board.
• Open the Serial monitor and check if connected.