Implementação de uma matriz de neurônios dinamicamente reconfigurável para descrição de topologias de redes neurais artificiais multilayer perceptrons

The Artificial Neural Networks (ANN), which is one of the branches of Artificial Intelligence (AI), are being employed as a solution to many complex problems existing in several areas. To solve these problems, it is essential that its implementation is done in hardware. Among the strategies to be adopted and met during the design phase and implementation of RNAs in hardware, connections between neurons are the ones that need more attention. Recently, are RNAs implemented both in application specific integrated circuits's (Application Specific Integrated Circuits - ASIC) and in integrated circuits configured by the user, like the Field Programmable Gate Array (FPGA), which have the ability to be partially rewritten, at runtime, forming thus a system Partially Reconfigurable (SPR), the use of which provides several advantages, such as flexibility in implementation and cost reduction. It has been noted a considerable increase in the use of FPGAs for implementing ANNs. Given the above, it is proposed to implement an array of reconfigurable neurons for topologies Description of artificial neural network multilayer perceptrons (MLPs) in FPGA, in order to encourage feedback and reuse of neural processors (perceptrons) used in the same area of the circuit. It is further proposed, a communication network capable of performing the reuse of artificial neurons. The architecture of the proposed system will configure various topologies MLPs networks through partial reconfiguration of the FPGA. To allow this flexibility RNAs settings, a set of digital components (datapath), and a controller were developed to execute instructions that define each topology for MLP neural network.

Progetto e sviluppo di un automa per la gestione ed il controllo di sistemi a pilotaggio remoto

L’attenzione di questa tesi si pone sulla progettazione e lo sviluppo di un automa per il controllo di sistemi a pilotaggio remoto. Punto di partenza dell'intero progetto è stata l'analisi del quadrirotore programmato ed utilizzato dai ricercatori del progetto SHERPA. Le funzionalità evidenziate dall'analisi sono state poi elaborate secondo l'approccio con Attuatore Generalizzato ed implementate in Codesys. Questo software si compone di una parte di controllo, una di simulazione ed una o più interfacce grafiche dedicate all'interazione con l'utente. L'utilizzo di Codesys ha permesso di implementare un automa modulare e riusabile, base di partenza per i futuri progetti di automi aerei. L'automa realizzato consente infine di simulare il comportamento di un generico sistema a pilotaggio remoto e di osservarne la sequenza logica con cui le azioni vengono eseguite.

Serious Game no tratamento de obesidade infantil: uma proposta de interação natural com monitoramento remoto

The number of overweight people has increased in the last few years. Factors such as attention to diet and changes in lifestyle are crucial in the prevention and control of obesity and diseases related to it. Experts believe that such actions are most effective when initiated during childhood, and that children raised in an environment that encourages physical activity ultimately become healthier adults. However, to arouse and maintain interest in such activities represent a major challenge, which are initially perceived as repetitive and boring, and, thus, soon abandoned. Computer games, traditionally seen as stimulants to a sedentary lifestyle are changing this perception using non-conventional controls that require constant movement of the player. Applications that combine the playfulness of such games to physical activity through devices, like Microsoft Kinect, might become interesting tools in this scenario, by using the familiarity of Natural User Interfaces along with the challenge and the fun of video games, in order to make attractive exercise routines for schoolchildren. The project carried out consists of an exergame composed of several activities designed and implemented with the participation of a Physical Educator, aimed at children between eight and ten years old, whose performance and progress can be remotely monitored by a professional via web interface. The application arising from this work was accompanied by tests with a group of graduating Physical Education students from the University of Rio Verde GO, and subsequently validated through questionnaires whose results are shown on this work.

Classificação de nódulos pulmonares em imagens tomográficas utilizando redes neurais artificiais em cascata

Lung cancer is the most common of malignant tumors, with 1.59 million new cases worldwide in 2012. Early detection is the main factor to determine the survival of patients affected by this disease. Furthermore, the correct classification is important to define the most appropriate therapeutic approach as well as suggest the prognosis and the clinical disease evolution. Among the exams used to detect lung cancer, computed tomography have been the most indicated. However, CT images are naturally complex and even experts medical are subject to fault detection or classification. In order to assist the detection of malignant tumors, computer-aided diagnosis systems have been developed to aid reduce the amount of false positives biopsies. In this work it was developed an automatic classification system of pulmonary nodules on CT images by using Artificial Neural Networks. Morphological, texture and intensity attributes were extracted from lung nodules cut tomographic images using elliptical regions of interest that they were subsequently segmented by Otsu method. These features were selected through statistical tests that compare populations (T test of Student and U test of Mann-Whitney); from which it originated a ranking. The features after selected, were inserted in Artificial Neural Networks (backpropagation) to compose two types of classification; one to classify nodules in malignant and benign (network 1); and another to classify two types of malignancies (network 2); featuring a cascade classifier. The best networks were associated and its performance was measured by the area under the ROC curve, where the network 1 and network 2 achieved performance equal to 0.901 and 0.892 respectively.

O uso de redes neuronais artificiais na estimação do preço das habitações na ilha do sal, em Cabo Verde

Esta dissertação tem como objetivo desenvolver um modelo de redes neuronais artificiais para aferir as caraterísticas das habitações que mais influenciam o preço na Ilha do Sal, em Cabo Verde. Foram consideradas caraterísticas como: área, número de quartos, existência de varandas, existência de terraços, número de casas de banho, localização do imóvel, número de andares e proximidade com instituições públicas. A amostra utilizada considerou 1092 habitações no período de 2009 a 2014. Para além da análise baseada no desenvolvimento do modelo de redes neuronais, efetuou-se a análise pela estimação do modelo dos preços hedónicos. Os resultados do modelo de redes neuronais artificiais permitiram verificar que o preço das habitações é fortemente influenciado pela área, e em seguida pela localização. A existência de caraterísticas, tais como a proximidade com a câmara municipal e finanças e existência de varandas, são as variáveis que menos influenciam o preço das habitações na Ilha do Sal. Os resultados da estimação com o modelo dos preços hedónicos indicam que o preço das habitações é fortemente influenciado por algumas variáveis representativas de características estruturais, localização e de vizinhança. Algumas dessas variáveis têm efeito estatisticamente significativo positivo no preço tais como, a localização do imóvel em Algodoeiro- Santa Maria, o número de quartos e a área. Outras variáveis têm efeito estatisticamente significativo negativo no preço, tais como a localização do imóvel no Bairro Novo e a proximidade com o hospital. Os resultados mostram que comparativamente com o modelo de preços hedónicos, o modelo de redes neuronais artificiais representa uma melhor alternativa para a previsão dos preços das habitações na Ilha do Sal, isto considerando a comparação dos erros estimados entre os modelos e as medidas de desempenho comumente utilizadas.

Utilizando a porta paralela para controle remoto de um dispositivo

VALENTIM, R. A. M.; NOGUEIRA, I. A.;ROCHA NETO, A. F. Utilizando a porta paralela para controle remoto de um dispositivo. Revista da FARN, Natal, v. 2, p. 103-114, 2002.

Acceso remoto (2016)

[ES]Tutorial en el que se explica cómo a través del acceso remoto se podrá acceder en línea a los recursos electrónicos de la Biblioteca de acceso restringido (bases de datos, revistas-e, libros-e...

Progetto di un'applicazione mobile per il controllo remoto di funzionalità hardware e software su piattaforma Raspberry

In questa tesi si è progettata una applicazione Android che permettesse di controllare da remoto funzionalità hardware e software offerte da una piattaforma Raspberry Pi. Si sono infine svolte alcune misure atte a testare le performance di rete di questa scheda.

Controlo remoto e autónomo de um sistema de aquacultura

Esta dissertação apresenta um sistema de aquacultura que possibilita a sua monitorização e o seu controlo autónomo. O sistema é capaz de ler grandezas como nível, pH, luminosidade, pressão e temperatura, e, posteriormente, ativar ou desativar os atuadores, como as eletroválvulas e as bombas peristálticas, até que se encontrem os valores pretendidos. O GSM é uma parte relevante do sistema desenvolvido, que permite uma interação à distância com o sistema, utilizando um telemóvel. Em qualquer momento, o utilizador pode pedir informações ao sistema, através de uma solicitação efetuada por mensagem de texto. As informações requeridas podem ser relativas ao estado dos atuadores e aos valores que os sensores apresentam. O utilizador pode, ainda, modificar o estado de um atuador, através de mensagem de texto enviada para o sistema. A resposta ao utilizador é uma mensagem com a verificação do que foi pedido. O sistema permite também ser configurado através de mensagem de texto, definindo os limites de cada parâmetro. Tem, ainda, a capacidade de enviar alertas, via mensagem de texto, quando algum valor estiver fora do pretendido. Apresenta uma aplicação em Visual Basic, que permite interagir com o sistema, controlar os limites e os atuadores, visualizar os alertas emitidos e registá-los em base de dados, para guardar o historial e compreender o comportamento através de gráficos.

Previsão do índice bovespa por meio de redes neurais artificiais: uma análise comparada aos métodos tradicionais de séries de tempo

Forecast is the basis for making strategic, tactical and operational business decisions. In financial economics, several techniques have been used to predict the behavior of assets over the past decades.Thus, there are several methods to assist in the task of time series forecasting, however, conventional modeling techniques such as statistical models and those based on theoretical mathematical models have produced unsatisfactory predictions, increasing the number of studies in more advanced methods of prediction. Among these, the Artificial Neural Networks (ANN) are a relatively new and promising method for predicting business that shows a technique that has caused much interest in the financial environment and has been used successfully in a wide variety of financial modeling systems applications, in many cases proving its superiority over the statistical models ARIMA-GARCH. In this context, this study aimed to examine whether the ANNs are a more appropriate method for predicting the behavior of Indices in Capital Markets than the traditional methods of time series analysis. For this purpose we developed an quantitative study, from financial economic indices, and developed two models of RNA-type feedfoward supervised learning, whose structures consisted of 20 data in the input layer, 90 neurons in one hidden layer and one given as the output layer (Ibovespa). These models used backpropagation, an input activation function based on the tangent sigmoid and a linear output function. Since the aim of analyzing the adherence of the Method of Artificial Neural Networks to carry out predictions of the Ibovespa, we chose to perform this analysis by comparing results between this and Time Series Predictive Model GARCH, developing a GARCH model (1.1).Once applied both methods (ANN and GARCH) we conducted the results' analysis by comparing the results of the forecast with the historical data and by studying the forecast errors by the MSE, RMSE, MAE, Standard Deviation, the Theil's U and forecasting encompassing tests. It was found that the models developed by means of ANNs had lower MSE, RMSE and MAE than the GARCH (1,1) model and Theil U test indicated that the three models have smaller errors than those of a naïve forecast. Although the ANN based on returns have lower precision indicator values than those of ANN based on prices, the forecast encompassing test rejected the hypothesis that this model is better than that, indicating that the ANN models have a similar level of accuracy . It was concluded that for the data series studied the ANN models show a more appropriate Ibovespa forecasting than the traditional models of time series, represented by the GARCH model

Utilizando a porta paralela para controle remoto de um dispositivo

VALENTIM, R. A. M.; NOGUEIRA, I. A.;ROCHA NETO, A. F. Utilizando a porta paralela para controle remoto de um dispositivo. Revista da FARN, Natal, v. 2, p. 103-114, 2002.

Sensoriamento óptico quase-distribuído aplicado ao monitoramento de trocadores de calor de usinas hidrelétricas

This work aims to develop optical sensors for temperature monitoring in hydroelectric power plant heat exchangers. The proposed sensors are based on the Fiber Bragg Gratings technology. First of all, a prototype with three sensors inscribed in a same fiber was developed. This fiber was then fixed to a conventional Pt100 sensor rod and inserted in a thermowell. The ensemble was then calibrated in a workbench, presenting a maximum combined uncertainty of 2,06 °C. The sensor was installed in one of the heat exchangers of the Salto Osório’s hydroelectric power plant. This power plant is situated in the Iguaçu river, at the Paraná state. Despite the satisfactory results, the sensor was improved to a second version. In this, fifteen optical Bragg sensors were inscribed in a same fiber. The fixation with a conventional sensor was no longer necessary, because the first version results comproved the efficiency and response time in comparison to a conventional sensor. For this reason, it was decided to position the fiber inside a stainless steel rod, due to his low thermal expansion coefficient and high corrosion immunity. The utilization of fifteen fiber Bragg gratings aims to improve the sensor spatial resolution. Therefore, measurements every ten centimeters with respect to the heat exchanger’s height are possible. This provides the generation of a thermal map of the heat exchanger’s surface, which can be used for determination of possible points of obstruction in the hydraulic circuit of the heat exchanger. The heat exchanger’s obstruction in hydroelectric power plants usually occur by bio-fouling, and has direct influence in the generator’s cooling system efficiency. The obtained results have demonstrated the feasibility in application of the optical sensors technology in hydroelectric power plants.

Características limnológicas e estrutura trófica das comunidades de peixes de alguns lagos naturais e artificiais do Rio Grande do Norte

Artificial lakes must differ from natural lakes in important structural and functional aspects that need to be understood so that these ecosystems can be properly managed. The aim of this work was to test the hypothesis that the artificial lakes (impoundments) in the semi-arid region of the Rio Grande do Norte State are more eutrophic and turbid and have different trophic structure when compared to the natural coastal lakes that occur in the humid eastern coast of the State. To test this hypothesis, 10 natural lakes and 8 artificial lakes with about 100 ha were sampled between September and November 2005 for the determination of some limnological variables and the abundance of the main fish species, which were grouped in three trophic guilds: facultative piscivores, facultative planktivores and omnivores. The results show that the artificial lakes had significantly higher concentrations of total nitrogen, total phosphorus, chlorophyll a , total and volatile suspended solids than the natural lakes. Results also show that the values of pH, total alkalinity, electric conductivity, turbidity as well as the coefficient of vertical attenuation of light were significantly higher in the artificial lakes than in the natural lakes. In the artificial lakes, the abundance of facultative planktivores was significantly higher, while the abundance of facultative piscivores significantly lower than in the natural lakes. There was no significant difference in the abundance of omnivorous fish between the two types of lakes. These results suggest that the increase in turbidity together with the other changes in the water quality of the artificial lakes, modifies the trophic structure of the fish communities reducing the importance of piscivores and the length of the food chains