Banca de QUALIFICAÇÃO: MARIA FERNANDA CABRAL RIBEIRO

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : MARIA FERNANDA CABRAL RIBEIRO
DATE: 30/11/2023
TIME: 14:30
LOCAL: Virtual - https://meet.google.com/wdf-bcdy-xqb
TITLE:

Smart and Fault Tolerant Embedded System Aimed at Women’s Safety

KEY WORDS:

Violence Against Women, Embedded Systems, Artificial Intelligence, Artificial Neural Networks, FPGA, Fault Tolerance.

PAGES: 60
BIG AREA: Ciências Exatas e da Terra
AREA: Ciência da Computação
SUBÁREA: Sistemas de Computação
SUMMARY:

Every year different bureaus and observatories provide studies on the rates of violence against women. This year, with the end of social distancing practices imposed by the Covid-19 pandemic, an increase was observed in all forms of violence suffered by Brazilian women. These data highlight the still unresolved need to think about developing solutions aimed at reduce violence against women. It is in this scenario that Salve Todas is inserted, an embedded system that, through an Artificial Neural Network (ANN), is capable of identifying possible risk situations for its user. Since the Artificial Neural Network plays a critical role in identifying risk situations for the user of the Salve Todas System, it is necessary to analyze techniques aimed at increasing its reliability in order to ensure safety for the user. Therefore, this work proposes the implementation of the ANN model of the Salve Todas System in hardware and the implementation of fault tolerance techniques in the embedded ANN. The purpose of this work includes: working on the weights of the ANN model, searching for values with smaller magnitudes and standard deviations; applying Triple Modular Redundancy in neurons of the input layer; and addition of specific neurons from the ANN hidden layer.

COMMITTEE MEMBERS:
Presidente - 1882699 - MONICA MAGALHAES PEREIRA
Interno - 1694485 - MARCIO EDUARDO KREUTZ
Externa à Instituição - FERNANDA GUSMÃO DE LIMA KASTENSMINDT