Check the written project here.

Check the code foundations here.

CFT – Community Fire Tracker

19th September 2019

OVERVIEW

Wildfires are becoming a rising problem to the communities and their ecosystems, in fact, each year wildfires are devastating tens of million hectares around the world. Whether controlled, accidental or natural, unhandled forest fires has long-term consequences like threats to human and animal life, deterioration of the health and loss of livelihoods, besides threatens the survival of endangered species, weaken the structure and composition of the ecosystem and provide conditions for entry of invasive species, all of these costing billions of dollars.

The governments are good at reacting to short-term solutions but they lack the tools to offer sustainable solutions on a long-term, indeed science is oversimplified and distorted by politics usually, thus it’s important that people take care of themselves, their environment and play a proactive role preventing, mitigating or supporting the firefighting.

Communications are one of the most valuable elements on a workplace disaster. Precise, timely and relevant information is critical during emergencies, under the risk of imminent conflagrations technology solutions might offer a valuable tool to the people. This project is aiming to empower people and helping them to cooperate and protect themselves.

Briefly, what is CFT - Community Fire Tracker?

It’s a Social Bot that is powered by communities, enhanced with Artificial Intelligence (AI) algorithms and supplied with data analysis processed from the open datasets provided by NASA.

Government and Relief agencies will get geo-located facts in real-time provided from whether active users or satellite imagery offered on an open website.

Forecasts based on Data Analysis can be performed using big data techniques. In this way, the conflagration will be prevented, mitigated or attended promptly by both civilians and relief agencies.

When a single user reports a possible fire incident, the bot will trigger alerts to people nearby the threatened place, these people will be notified, able to confirm the fire event, and invited to join a temporal chat group created to host them encouraging cooperation.

False alarms can be managed by the same community as long as they confirm the incident.

On countryside zones equipped with Internet of Things (IoT) infrastructure, people can link their devices to this bot and send reports automatically during fire events, besides if there is no internet available, SMS over GSM/LTE networks can be sent with ease.

Uninhabited zones can be monitored using the forecasts performed by the data analysis along with satellite imagery.

GOALS

1. Empower and supply the community with an AI-enhanced bot to assist catastrophic events like conflagrations or others.
2. Supply to the people, the government and relief agencies with accurate data that allows them to monitor and attend emergencies with information submitted in real-time, this goes on an open website.
3. Make an exploratory analysis based on Big Data to prevent future conflagrations.

BENEFITS AND PERKS

CFT – Community Fire Tracker is a highly scalable solution with small investments, the initial prototype is planned to be developed with Telegram, but it could be easily ported or linked with similar platforms like Whatsapp, Messenger, Signal, and others.

The geo-located reports offer accurate data that communities can leverage with precision to spot the threatened zone.

When the bot receives an atypical report it can dig over other users, trying to confirm the emergent event; this feature is supported by the same community to avoid false alerts. A blockchain approach could be useful on this topic.

Reporting a conflagration emergency is a simple task that is not time-consuming, the users just need to send a text, audio, video or image stream, immediately the bot will propagate the alert notifying users around the reported zone trying to confirm the first report requesting multimedia reports.

Users can get audible instructions that can easily be followed to arrive at safe zones.

Chat apps are ubiquitous on mobile phones, this way the users don’t need to install any extra app that can easily be forgotten or uninstalled when it’s no longer used.

DEVELOPMENT

Supplies

Community Reports

The bot is enabled to receive and process different sort of data that will be analyzed with artificial intelligence to process audio and images mainly.

IoT Infrastructure Reports

The bot is capable to get reports from different sorts of devices that can be enabled to automatically send reports through chat messages or SMS.

Data Analysis

Making use of Apache Flink API and the NASA’s dataset we are aiming to forecast events to avoid potential conflagrations. This is possible through the analysis of different sources like IoT data streams, community reports, satellite imagery, weather behaviour, terrain characteristics and others.

Technical Requirements

• Main DB
  • To store information related to people like main locations (Home/Work/School), phone numbers, etc.
  • To store geo-locations of important places like shelters, water sources, transport terminals, etc.
• Realtime API
  • To manage the stream of data served on the open website.
• Open Website
  • To render relevant data in a thoughtful and usable graphical user interface (GUI).
• A served Chat Bot enhanced with Machine Learning (ML)
  • Google Vision AI API to process and parse data from still images.
  • Google Video AI API to process, parse and gain insights from video messages.
  • Google SPEECH TO TEXT API to process audio and video messages from reports to parse keywords from the audio stream that implies real threats or relevant data.
  • Google TEXT TO SPEECH API to generate audible guided routes or alerts to people with vision disabilities.
• Big Data
  • NASA’s Earth Observation data sets
  • Resource Definition Framework (RDF)
  • Sparql Engine
  • Cqels Engine
  • Apache Flink API

MILESTONES

• System API

Develop an open REST API to offer a basic CRUD that will be connected to the bot in its first set up and store the minimal required data. (User ID, Work/Home locations, valuable skills)

• CHAT BOT

Deploy a basic chatbot that is capable of receiving and sending messages to users under a certain ratio around a threatened location.

• AI FEATURES

Integrate Artificial Intelligence in the core of the bot to understand and interpret the content of the messages via audio, image and video mainly as those are the most simple and straightforward ways to send messages during an emergency.

• Empower Communities

Spread the voice and empower communities with this chat bot. Get registrants and active users via gamification.

• Integrate Big Data features

Build a framework based on Data Analysis using Apache Flink API to get forecasts based on weather data streams and IoT real-time logged data.

KEYWORDS

#Crowdsource #artificialIntelligence #computerVision #PhysicalComputing #dataScience #NASAforecasts #SocialBot #imageRecognition #naturalLanguageProcessing #OSS #OpenSource #IoT