Banca de QUALIFICAÇÃO: Jesus Noel Suarez Rubi

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
STUDENT : Jesus Noel Suarez Rubi
DATE: 04/08/2022
TIME: 14:00
LOCAL: Sala de Seminários (Espaço Sérgio Barroso)
TITLE:

A platform and ontologies for environment data sharing and the use of Machine Learning models for wildfire ignition and prediction.

KEY WORDS:

A platform and ontologies for environment data sharing and the use of Machine Learning models for wildfire ignition and prediction.

PAGES: 147
BIG AREA: Engenharias
AREA: Engenharia Elétrica
SUBÁREA: Telecomunicações
SPECIALTY: Sistemas deTelecomunicações
SUMMARY:

Ecosystems, settlements, and human lives are put at risk by forest fires every year, impacting economy and social-economic development. The Brazilian Federal District, inserted in the Cerrado biome, has showed an increase in such phenomenons. However, few studies have been conducted in the region. Several models have been proposed worldwide for the prediction of fire occurrence and behavior, and identification of their conditioning factors, risks, and post-effects. The direct application of such models in the Federal District region is challenging due to differences in data sources, geographic characteristics of the regions, and unavailability of data in some cases. On the other hand, technology and electronic equipment expansion have generated large volumes of data with substantial value for the development of smart cities. Particularly, Industry/- Forestry 4.0 and environmental smart city data can enrich wildfire studies. However, recent proposals have faced the same downside, since data are incomplete, follow different representation formats, and even have different semantic connotations. The heterogeneity of intelligent objects connected to the Internet (i.e., network interfaces, communication protocols, data structure, acquisition precision, and data semantics) has caused interoperability problems, hindering the effectiveness of decision-support systems closely related to the quality of data. The application of big data and machine learning algorithms for improving smart city-related processes are some of the examples negatively impacted by the lack of standards. Solutions for smart cities should grant interoperability from data capture to knowledge extraction and visualization through technologies such as Semantic Web and ontologies. Moreover, the components involved should include IoT devices, gateways, cloud, and fog computing for a better application of data analysis techniques. In this sense, this thesis proposes a smart city platform for environment quality monitoring based on semantic technologies and ontologies, enabling a multi-definition and multi-protocol data collection and sharing system. It also presents a methodology for the extraction of insights on the collected data and a mechanism for cloud- and fog-based computations. As a result, the environment monitoring platform has been developed and tested regarding the prediction of both spread and behavior of wildfires at a specific time and/or in specific regions for helping fire management agencies minimize the damages caused. A dataset was compiled from Brazilian governmental open data for the prediction of the wildfire behavior and used for the training of several Machine Learning models that consider the fire point of ignition to predict the areas that will be impacted. It includes observations on climate features from 5 monitoring stations and satellite data on fires that occurred over the past two decades and was enriched with other topographic, hydrographic, and anthropogenic features,such as urbanization index, distance to rivers/roads, and Normalized Difference Vegetation Index (NDVI).

BANKING MEMBERS:
Interno - 404943 - ADSON FERREIRA DA ROCHA
Externo à Instituição - JOEL JOSE PUGA COELHO RODRIGUES - UFPI
Interno - 1609344 - MARCELO MENEZES DE CARVALHO
Externo ao Programa - 1338972 - MAURO ELOI NAPPO
Interno - 2376576 - PAULO ROBERTO DE LIRA GONDIM