Análise de modelos adversariais aplicados à classificação de sinais de Sonar Passivo

Julio De Castro Vargas Fernandes, José de Seixas, Natanael Nunes de Moura Junior

DOI: 10.14209/sbrt.2019.1570558846

Keywords:

Abstract

In naval warfare, several techniques have been developed for the detection and classification of ships. Given the confidential nature of the data it is extremely difficult to get a hold of large quantities of data which makes it extremely hard to use techniques that rely on abundant data, like deep learning. This paper proposes the use of generative adversarial neural networks for the classification of passive sonar signals. Since the proposed technique trains an synthetic image generator, which will later be used for the training of a classifier, its quality will be evaluated. The proposed method achieved an efficiency of 99,2 +- 1,0.

Download

EXEHDA-SF: an IoT Based Proposal for the Irrigation Management

Rogério da Costa Albandes, Adenauer Yamin, Roger Machado, Joao Lopes, Jorge Victória Barbosa

DOI: 10.14209/sbrt.2019.1570558849

Keywords:

Abstract

According to the National Water Agency (ANA), approximately 46.2% of all clean water used in Brazil is destined for irrigation, which has been motivating the evaluation of alternatives for water supply agricultural crops. Considering this scenario, this article presents the EXEHDA-SF proposal, which aims to explore Situation Awareness in the decision making process, with the perspective of minimizing socio-environmental impacts. For evaluation of EXEHDA-SF a prototype was developed that integrates open-source IoT technologies to the EXEHDA middleware and explores a weather prediction service. The results achieved were promising, reaching a success rate of approximately 94% regarding the irrigation decision.

Download

Abordagem I3VSM: Acompanhamento Autônomo de Pacientes Explorando Ciência de Situação na IoT

Rogério da Costa Albandes, Adenauer Yamin, Jorge Victória Barbosa, Roger Machado

DOI: 10.14209/sbrt.2019.1570558850

Keywords:

Abstract

- Mobility has become a daily practice of physicians, so it is possible that they remain periods of time without contact with the teams that support them in the treatment of patients. Longer periods between communications can cause delays in performing procedures, drug prescribing, etc. Considering this scenario, this work has as objective the conception an approach, called I3VSM, which integrates: (i) a platform for acquisition of vital signs, (ii) an environment for contextual processing, which through customizable rules builds the Situation Awareness of the patients; and (iii) a textual and graphic display interface for these signals, which can be accessed by IoT. As a source of vital signs, the MIMIC-III database is being used, which has been widely accepted by the international community for this purpose. In turn, for the evaluation of I3VSM together with health professionals, the Technology Acceptance Model (TAM) was used, and promising results were obtained.

Download

Extração de Sinal de ECG Fetal Compressado Utilizando Sistema de Inferência Neuro-Fuzzy Adaptativo

Felipe Meneguitti Dias, Thales Wulfert Cabral, Rayen Naji, Eduardo Pestana de Aguiar, Marcelo Lima, Leandro Manso, Leonardo Honorio

DOI: 10.14209/sbrt.2019.1570558851

Keywords:

Abstract

This paper presents a methodology for fetal electrocardiogram extraction in signals sampled using a technique known as compressive sampling. Two meters are used, one located on the chest and the other on the abdomen of a pregnant woman. In order to obtain the fetal electrocardiogram signal, an adaptive neuro-fuzzy inference system is used to estimate the interference of the maternal electrocardiogram in the abdomen meter. In this way, it is possible to subtract this interference and, consequently, to obtain the desired signal.

Download

Correção para Perda por Caminho em Ambientes Interiores com o Método FDTD-2D

Bruno Wallacy Martins, Rodrigo M. S. Oliveira, Victor Dmitriev, Lucas S. Almeida, Fabricio Barros

DOI: 10.14209/sbrt.2019.1570558852

Keywords:

Abstract

The Finite-Difference Time-Domain Method (FDTD) is consolidated in the literature for a prediction of electromagnetic signals. In two dimensions (FDTD-2D), it is simple to implement and does not require large computational capacity in terms of memory and processing time. However, with two-dimensional methods, such as FDTD-2D, the attenuation of the electromagnetic signal with low agreement with experiments. In this work, is proposed a correction factor applied to the curve of Path Loss, within a building, which increases the conformity between the results obtained with the FDTD-2D method and experimental data.

Download

Sobre o uso de GCC-PHAT na estimação de DoA para arranjos de dimensão reduzida

Felipe G Serrenho, Luiz C S Ramos, José Antonio Apolinário Jr.

DOI: 10.14209/sbrt.2019.1570558859

Keywords: GCC-PHAT arranjo de microfones diferença de tempo de chegada

Abstract

Este trabalho apresenta a importância da utilização de pesos denominados Phase Transform (PHAT) na GCC para a correta estimação da TDoA em arranjos de dimensões reduzidas. Apresenta as variações GCC-NPHAT, GCC-\(\alpha\)PHAT e a composição destas GCC-\(\alpha\)NPHAT, que visam aumentar a robustez quando o sinal de interesse é espectralmente esparso e há presença de interferidores banda larga. Ainda é abordado a importância da utilização de algoritmos de interpolação como forma de compensar os efeitos da discretização da função de correlação cruzada. A estimação da DoA é realizada por meio de um algoritmo de busca exaustiva utilizando o critério de mínimos quadrados.

Download

Utilização de Aprendizado de Máquina na Demodulação de Sistemas OAM utilizando Multiplexação de Modos Orbitais

Cristhof Roosen Runge, Jaime Portugheis, Ivan Oliveira, Ulisses Dias, Bruno Freitas, Cecilio Pimentel

DOI: 10.14209/sbrt.2019.1570558862

Keywords:

Abstract

Este trabalho apresenta os resultados de simulação para um sistema de transmissão óptica OAM no espaço livre sob efeito de turbulência atmosférica. Os momentos angulares orbitais, correspondendo aos símbolos de transmissão, são utilizados para modular um feixe laser no modo Laguerre-Gauss. Na recepção as imagens capturadas por uma câmera CCD são analisadas por uma rede neural convolucional no processo de demodulação e decisão dos símbolos recebidos. Foram simuladas as transmissãos de alfabetos formados por tuplas obtidas por superposição simples, e por multiplexação de superposição simples de dois momentos orbitais. Os resultados de simulação apontam para desempenho similar para os casos de sistemas utilizando superposição simples e utilizando tuplas formadas por multiplexação de superposições simples de dois momentos, nas condições de turbulência leve e moderada. Este fato pode conduzir a possíveis ganhos de capacidade para sistemas OAM que utilizem alfabetos obtidos através da multiplexação de superposições simples de momentos angulares orbitais.

Download

Um Estimador de Envelope Temporal para Redução de Ruído em Implantes Cocleares

Rafael Attili Chiea, Marcio H Costa, Maria Madalena Pinheiro, Júlio Cordioli

DOI: 10.14209/sbrt.2019.1570558865

Keywords:

Abstract

This work presents a new noise reduction technique for cochlear implants based on speech time-envelope estimates. Numerical simulations using an objective intelligibility criterion, corroborated by a psychoacoustic experiment with a cochlear implant user, indicate a consistent increase in the intelligibility in relation to raw contaminated speech signals as well as to Wiener filter processed signals. These observations are valid for signal to noise ratios (S NR) less than 10 dB. In addition, the intelligibility gain increases with the decrease of the SNR.

Download

Uma Comparação entre Máscaras Tempo-frequência para Redução de Ruído em Implantes Cocleares

Rafael Attili Chiea, Marcio H Costa, Guillaume Barrault

DOI: 10.14209/sbrt.2019.1570558867

Keywords:

Abstract

Time-frequency filtering techniques have been used for noise reduction in cochlear implants with relative success. The Binary Mask and the Wiener Filter are among the most applied strategies, however they have limited conformation capability. In order to overcome this issue, alternatives have been proposed, such as the parametric Wiener Filter and variations. This work presents a performance comparison between four time-frequency masks in terms of speech intelligibility and quality. Numerical simulations were performed and objective measures correlated to the psychoacoustic perception of cochlear-implant users' were analyzed. Results indicate that parametric masks are more appropriate to this specific application, even though there are not well defined methods to determine parameters' values.

Download

Implantação de um Sistema Fotovoltaico Otimizado em Arquitetura de Rádio Centralizada

Paline A Saraiva, Marcela Alves Souza, Hugo Pereira Kuribayashi, Adam D. F. Santos, Joao Weyl Costa, Diego L Cardoso

DOI: 10.14209/sbrt.2019.1570558872

Keywords:

Abstract

Among the requirements of the next generation of mobile networks is the massive reduction in power consumption. The adoption of photovoltaic systems (SF) has been investigated in the literature, however the implementation costs are still challenging, and for an economically viable deployment it is necessary to size SF with minimum cost. Thus, this work proposes the use of an optimized SF in a Centralized Radio Architecture CRA). Based on the results, it is observed that the optimization technique used implies the reduction of SF ownership costs, with an average difference of 0.02 million reais.

Download