Artificial intelligence to help Lisbon firefighters

in News · 05-03-2020 11:36:00 · 0 Comments

Lisbon firefighters may soon reduce the response time to incidents, such as floods or fires, with the help of an artificial intelligence model, that predicts the likelihood of serious occurrences in certain locations.

The LX Analytics Hub model, presented in Lisbon, in an initiative of the City Council, will now be integrated into the Lisbon Intelligent Management Platform.

After this step, it will be available daily so that the Regiment of Sapadores Firefighters of Lisbon (RSB) and other municipal services can make decisions about the means to approach the places where the probability of occurrences is verified.

According to Miguel de Castro Neto, deputy director of NOVA IMS and Coordinator of NOVA Cidade - Urban Analytics Lab, one of the partners who developed the project, the model allows to predict, up to one day in advance, if there is a probability that a certain occurrence will occur in a given location and time period, based on the historical data available between 2013 and 2019.

Thus, the authorities can allocate emergency means flexibly to areas with the highest risk index, with the aim of allowing firefighters to respond to the emergency ideally in five minutes, but at the most by 10 minutes, he added.

The question that led to the model arose because the location of the RSB means in the barracks does not always guarantee a quicker response time, “because city conditions change every day [like] traffic, precipitation, events and traffic cuts”, he said.

The model developed analyzed a history of climatic variables (with the support of data from the Portuguese Institute of the Sea and Atmosphere - IPMA), of sociodemographic variables and of events recorded in Lisbon between 2013 and 2018.

Altogether, there are more than five million records analyzed, which, according to Miguel Castro Neto, allow 70 percent accuracy in the prediction of occurrences, that is, predict seven out of 10 occurrences.

To build the model, 52,560 records from three IPMA stations (such as temperature, precipitation, humidity, wind intensity and direction and solar radiation) were analyzed, as well as 3,662 socio-economic statistical data (gender, age, education, employment, building) and 22,229 occurrences.

Among the events are fires, car accidents, incidents in infrastructures (such as falling structures or floods) and other services provided by firefighters, from opening doors to searching and rescuing people.

The model also allows an approximate number of people who, at any given moment, are in risk areas, through the number of mobile phones connected to 2,769 Altice antennas.

The project, led by the Lisbon City Council (CML), was developed by Nova Information Management School (NOVA IMS), in partnership with ISEL, Microsoft and SAP (company that creates management software) and with collaboration of Altice and IPMA.

The model started from a master's thesis by Leonor Brás Teixeira, student at NOVA IMS and firefighter, in partnership with Pedro Sarmento, also a student at the institution.


Comments:

Be the first to comment on this article
Interactive Topics, send us your comments/opinion on this article.

Please note that The Portugal News may use selected comments in the printed edition of the newspaper.