Made available in DSpace on 2014-10-09T12:37:25Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T14:06:55Z (GMT). No. of bitstreams: 1 02036.pdf: 4983733 bytes, checksum: 00af5aaa0a2a8be9f23ba8c5d8abb48d (MD5) / Dissertacao (Mestrado) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

Made available in DSpace on 2014-10-09T12:43:51Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T14:10:10Z (GMT). No. of bitstreams: 1 06767.pdf: 4710608 bytes, checksum: 0df801946c2e3b5907a9d77d5ff698e5 (MD5) / Dissertacao (Mestrado) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

Made available in DSpace on 2014-12-17T14:55:44Z (GMT). No. of bitstreams: 1 AlanPOS_DISSERT.pdf: 1600832 bytes, checksum: d14cb41b3377752326e77580d9b32dba (MD5) Previous issue date: 2010-02-19 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior / Blind Source Separation (BSS) refers to the problem of estimate original signals from observed linear mixtures with no knowledge about the sources or the mixing process. Independent Component Analysis (ICA) is a technique mainly applied to BSS problem and from the algorithms that implement this technique, FastICA is a high performance iterative algorithm of low computacional cost that uses nongaussianity measures based on high order statistics to estimate the original sources. The great number of applications where ICA has been found useful reects the need of the implementation of this technique in hardware and the natural paralelism of FastICA favors the implementation of this algorithm on digital hardware. This work proposes the implementation of FastICA on a reconfigurable hardware platform for the viability of it's use in blind source separation problems, more specifically in a hardware prototype embedded in a Field Programmable Gate Array (FPGA) board for the monitoring of beds in hospital environments. The implementations will be carried out by Simulink models and it's synthesizing will be done through the DSP Builder software from Altera Corporation. / A Separa??o Cega Fontes (BSS) refere-se ao problema de estimar sinais originais a partir de misturas lineares observadas sem nenhum tipo de conhecimento acerca das fontes ou do processo de mistura. A An?lise de Componentes Independentes (ICA) ? uma t?cnica aplicada principalmente ao problema do BSS e dentre os algor?tmos que implementam essa t?cnica, o FastICA ? um algor?tmo iterativo de alto desempenho e de baixo custo computacional que utiliza medidas de n?o-gaussianidade baseadas em estat?stica de alta ordem para estimar as fontes originais. O grande n?mero de aplica??es onde ICA se mostra ?til reflete a necessidade da implementa??o dessa t?cnica em hardware e o paralelismo natural do FastICA favorece a implementa??o desse algor?tmo em plataforma de hardware digital. Este trabalho prop?e a implementa??o do FastICA em uma plataforma de hardware reconfigur?vel para a viabiliza??o de sua utiliza??o em problemas de separa??o cega de fontes, mais especificamente em um prot?tipo de hardware embarcado em uma placa Field Programmable Gate Array (FPGA) para a monitora??o de leitos em ambientes hospitalares. As implementa??es ser?o realizadas atrav?s de modelos em Simulink e a sintetiza??o dos mesmos ser? feita com o aux?lio do software DSP Builder da Altera Corporation.

Made available in DSpace on 2014-10-09T12:34:37Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T14:00:48Z (GMT). No. of bitstreams: 0 / Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) / Este trabalho apresenta o desenvolvimento de uma plataforma para o estudo teórico e experimental de dispositivos amplificadores de micro-ondas do tipo válvula de ondas progressivas (TWT). A plataforma é composta por um modelo matemático e uma bancada de testes. O modelo matemático descreve a TWT como uma linha de transmissão acoplada a um feixe eletrônico unidimensional, onde as forças de carga espacial AC e DC são calculadas auto consistentemente, eliminando-se a necessidade de um cálculo separado para o fator de redução de carga espacial. O modelo matemático deu origem a dois códigos para a simulação da TWT. Ambos foram comparados com resultados experimentais e teóricos disponíveis na literatura especializada para uma pré-validação. O nível de concordância entre os presentes resultados e aqueles de referência foi acima de 90%, o que atendeu as expectativas de exatidão do modelo, tendo em vista que nem todos os parâmetros de entrada estavam disponíveis na referência. A bancada de testes construída é composta por uma TWT com banda de operação de 6,0 a 18 GHz e potência saturada máxima em torno de 55 dBm (316 W) em 13 GHz, um circuito de polarização para a mesma e a instrumentação necessária para a realização das medidas pertinentes aos amplificadores de potência. A TWT em questão foi caracterizada segundo seu comportamento mono-sinal e multi-sinal. As curvas de ganho e potência foram obtidas em função da frequência utilizando a voltagem de aceleração do feixe eletrônico e a potência de entrada como parâmetros. As curvas de transferência de potência, de fase e compressão de ganho foram obtidas para frequências escolhidas ao longo da banda, tendo novamente a voltagem de aceleração como parâmetro. Adicionalmente, a produção de produtos de intermodulação de terceira ordem foi caracterizada no ponto de 1 dB de compressão de ganho ao longo da banda analisada. Um teste de linearização por injeção de sinais, que estava previsto no plano de trabalho, não apresentou o desempenho esperado devido a problemas no funcionamento do circuito linearizador. Esses problemas foram analisados e listou-se uma série de passos para saná-los. / Tese (Doutoramento) / IPEN/T / Instituto de Pesquisas Energeticas e Nucleares - IPEN-CNEN/SP / FAPESP:08/05286-1

Made available in DSpace on 2014-10-09T12:37:25Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T14:06:55Z (GMT). No. of bitstreams: 1 02036.pdf: 4983733 bytes, checksum: 00af5aaa0a2a8be9f23ba8c5d8abb48d (MD5) / Dissertacao (Mestrado) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

Made available in DSpace on 2014-10-09T12:43:51Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T14:10:10Z (GMT). No. of bitstreams: 1 06767.pdf: 4710608 bytes, checksum: 0df801946c2e3b5907a9d77d5ff698e5 (MD5) / Dissertacao (Mestrado) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

Orientador: Rosana do Carmo Novaes Pinto / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Estudos da Linguagem / Made available in DSpace on 2018-08-13T02:39:07Z (GMT). No. of bitstreams: 1 Valiante_JulianaBrazolinGomes_M.pdf: 483151 bytes, checksum: 2438c25e088839458f919d23cefe1e4a (MD5) Previous issue date: 2009 / Resumo: Esta dissertação tem como tema central a oficialização da Língua Brasileira de Sinais (LIBRAS) como forma legítima de expressão e comunicação de comunidades surdas. São apresentados e discutidos os principais documentos que reconhecem e oficializam a língua (Lei 10.436/2002) e a regulamentam (Decreto 5626/2005). Dentre os temas abordados, destacamos os problemas relativos à formação e ao perfil de profissionais envolvidos na implantação dos projetos nas redes regulares de ensino: professor bilíngüe, tradutor/intérprete e instrutor de LIBRAS, bem como a proposta de educação bilíngüe, que institui a Língua de Sinais Brasileira como primeira língua de instrução e a modalidade escrita do português como segunda língua. O Decreto trata ainda da inserção da disciplina de LIBRAS nos cursos de graduação em Fonoaudiologia, Pedagogia, Educação Especial e Licenciaturas. Estes pontos são problematizados, no Capítulo 2, considerando-se a literatura atual sobre o tema e depoimentos de profissionais envolvidos com o debate e com a implantação dos projetos. Para que se possa melhor compreender como os movimentos ideológicos e sociais levaram ao reconhecimento das Línguas de Sinais e à elaboração dos documentos oficiais, no primeiro capítulo é apresentado um breve histórico da educação dos surdos. A luta travada entre as perspectivas oralistas e as que defendem as Línguas de Sinais, ao longo da história, ainda têm reflexos em nossa sociedade. Muitos dos preconceitos persistem, atualmente, devido à ignorância sobre a surdez e sobre o funcionamento das Línguas de Sinais. Alguns conceitos da sociolingüística, apresentados no último capítulo (Capítulo 3), visam esclarecer sobre alguns mitos a respeito do tema, enfatizando que não há fundamento científico em qualquer afirmação que se faça a respeito de línguas de comunidades minoritárias como sendo primitivas ou inferiores. LIBRAS, assim como qualquer língua ou variante lingüística, é totalmente adequada e suficiente para a comunicação entre os membros de uma comunidade e desempenha todas as funções das línguas naturais, dentre as quais a de mediar o desenvolvimento cognitivo pleno. Da mesma forma, é necessário que se discuta a surdez fora do âmbito médico-biológico que a patologiza que a tomam como um fenômeno anormal. As diferenças são constitutivas da normalidade e devem-se buscar alternativas para que a inclusão de fato aconteça em todos os setores sociais. As propostas de educação inclusiva ainda têm um longo caminho a percorrer, dada a grande distância observada entre os textos oficiais e as práticas. A promulgação de leis e decretos, entretanto, força os debates e as ações para que a inclusão possa efetivamente ocorrer. / Abstract: This work has as central theme the processes that turn the Brazilian Sign Language (LIBRAS) an official language, a legitimate form of expression and communication of deaf communities. The main documents that recognize the language (Law 10436/2002) and regulate the law (Decree 5626/2005) are presented and discussed. Among the several issues mentioned, problems related to the professionals involved in the implementation of projects in the regular schools (rede regular de ensino) such as the bilingual teacher, the translator/interpreter and the instructor of LIBRAS are highlighted. It is also discussed the proposal for bilingual education, establishing the Brazilian Sign language as the first language of instruction and the written Portuguese as the second language. The Decree also inserts LIBRAS as an obligatory discipline in graduate programs such as Speech Therapy, Education, Special Education and "Licenciaturas". These points are critically discussed in Chapter 2, considering the literature on the topic and reports given by professionals involved in the debate and the implementation of projects. In order to better understand how the social and ideological movements led to the recognition of sign languages and to the official documents, in the first chapter a brief history of education of the deaf people is given. The battle between the perspectives that defended oral training and the ones that argued in favor of the use of Sign languages still has effects on our society. Many of the prejudices are still present, due to ignorance about what deafness is and about the functioning of Sign languages. Some concepts of sociolinguistics, which are presented in the last chapter (Chapter 3), seek to clarify some myths about the theme, emphasizing that there is no scientific basis on which any statement is made regarding the languages of minority communities as primitive or inferior. LIBRAS - as any other language - is entirely appropriate and sufficient for communication between members of a community and it performs all the functions of natural languages, among which to mediate the full cognitive development. Similarly, it is necessary to discuss the concepts related to deafness outside the medical and biological environment, which conceives it as pathology, as an abnormal phenomenon. The differences constitute normality and we should seek for alternatives to effective inclusion of deaf people in all social sectors. The inclusive education still has a long way to go, given the large distance observed between the official documents and actual practices. The laws and decrees, however, force the discussions and actions in order to make the inclusion a fact. / Mestrado / Linguistica / Mestre em Linguística

Orientadores: Ricardo Ribeiro Gudwin, Álvaro João Magalhães de Queiroz / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-19T09:25:02Z (GMT). No. of bitstreams: 1 Loula_AngeloConrado_D.pdf: 2782656 bytes, checksum: 0b6a174e75f7657f6dd611dd9c157e25 (MD5) Previous issue date: 2011 / Resumo: Linguagem ainda é algo com que a Inteligência Artificial precisa lidar de forma mais apropriada. Processos relacionados como comunicação, interpretação e representação são indissociáveis de linguagem e devem ser considerados na construção de sistemas artificiais que irão usar linguagem. Diversos trabalhos já foram realizados sobre a emergência de comunicação entre agentes artificiais, mas estudos sobre os processos representacionais subjacentes encontram pouca discussão. Neste trabalho, propomos estudar as condições para emergência de comunicação baseada em representações de diferentes modalidades em uma comunidade de criaturas artificiais, avaliando os processos semióticos envolvidos. O projeto de nossos experimentos sintéticos segue inspirações biológicas e uma teoria de processos representacionais como fonte de requisitos e restrições, procedimento metodológico que defendemos. Nossos resultados mostram que comunicação baseada em diferentes modalidades de representação pode emergir a depender da arquitetura cognitiva, do custo de aquisição de competências cognitivas e da disponibilidade de atalhos cognitivos / Abstract: Language is still something that Artificial Intelligence must deal with in a more appropriate way. Related processes such as communication, interpretation and representation are inseparable from language and should be considered in the building of artificial systems that will use language. Many works have already investigated the emergence of communication among artificial agents, but studies about the underlying representational processes find little discussion. In this work, we propose to study conditions to the emergence of communication based on representations of different modalities in a community of artificial creatures, assessing the semiotic processes involved. The design of our synthetic experiments follows biological inspirations and a theory of representations as sources of requirements and constraints, methodological approach that we advocate. Our results show that communication based on different modalities of representations can emerge depending on the cognitive architecture, on the cost of cognitive competences acquisition and on the availability of cognitive shortcuts / Doutorado / Engenharia de Computação / Doutor em Engenharia Elétrica

Road safety is an aspect of major concern on road networks, not only on a local scale in South Africa, but also on a global scale. South Africa is experiencing an unacceptably high rate of fatal road traffic crashes and as a result has one of the highest per capita rate of road fatalities in the world. The Global Plan for the Decade of Action for Road Safety 2011 – 2020 was published in 2011 by the World Health Organisation (WHO) to guide and assist countries with the improvement of road safety. The management of vehicle speeds is one of numerous important interventions identified in the Global Plan. The technical and economic feasibility of using a specific Road safety is an aspect of major concern on road networks, not only on a local scale in South Africa, but also on a global scale. South Africa is experiencing an unacceptably high rate of fatal road traffic crashes and as a result has one of the highest per capita rate of road fatalities in the world. The Global Plan for the Decade of Action for Road Safety 2011 – 2020 was published in 2011 by the World Health Organisation (WHO) to guide and assist countries with the improvement of road safety. The management of vehicle speeds is one of numerous important interventions identified in the Global Plan. The technical and economic feasibility of using a specific technology, namely Vehicle Activated Signs (VAS), to contribute to the overall improvement of road safety on a national level in South Africa were investigated, with particular emphasis on the improvement of road safety through the reduction of fatal road traffic crashes. VAS are part of a range of measures that can be used to curb vehicle speeds, similar to conventional measures like speed cameras and traffic calming devices. VAS technology has seen limited use in South extrapolation of results from SCC’s VAS effectiveness study, and the application of these results to the South African road safety and road user statistics for the one year period March 2011 to March 2012. Key factors relating to speed and general driver obedience were also taken into consideration. It was established that the use of VAS would lead to a moderate reduction in the number of fatal road traffic crashes that occur annually, which in turn holds financial and socio-economic benefits for South Africa. A comparison of selected products available from two South African based electronic road sign manufacturing companies with conventional VAS was also carried out, focussing on product function, use and cost. It has been concluded that it would be technically and economically feasible to introduce VAS on the South African road network to improve the prevailing road safety situation.

For time-varying signals such as speech and audio, short-time analysis becomes necessary to compute specific signal attributes and to keep track of their evolution. The standard technique is the short-time Fourier transform (STFT), using which one decomposes a signal in terms of windowed Fourier bases. An advancement over STFT is the wavelet analysis in which a function is represented in terms of shifted and dilated versions of a localized function called the wavelet. A specific modeling approach particularly in the context of speech is based on short-time linear prediction or short-time Wiener filtering of noisy speech. In most nonstationary signal processing formalisms, the key idea is to analyze the properties of the signal locally, either by first truncating the signal and then performing a basis expansion (as in the case of STFT), or by choosing compactly-supported basis functions (as in the case of wavelets). We retain the same motivation as these approaches, but use polynomials to model the signal on a short-time basis (“short-time polynomial representation”). To emphasize the local nature of the modeling aspect, we refer to it as “local polynomial modeling (LPM).” We pursue two main threads of research in this thesis: (i) Short-time approaches for speech enhancement; and (ii) LPM for enhancing smooth signals, with applications to ECG, noisy nonuniformly-sampled signals, and voiced/unvoiced segmentation in noisy speech. Improved iterative Wiener filtering for speech enhancement A constrained iterative Wiener filter solution for speech enhancement was proposed by Hansen and Clements. Sreenivas and Kirnapure improved the performance of the technique by imposing codebook-based constraints in the process of parameter estimation. The key advantage is that the optimal parameter search space is confined to the codebook. The Nonstationary signal enhancement solutions assume stationary noise. However, in practical applications, noise is not stationary and hence updating the noise statistics becomes necessary. We present a new approach to perform reliable noise estimation based on spectral subtraction. We first estimate the signal spectrum and perform signal subtraction to estimate the noise power spectral density. We further smooth the estimated noise spectrum to ensure reliability. The key contributions are: (i) Adaptation of the technique for non-stationary noises; (ii) A new initialization procedure for faster convergence and higher accuracy; (iii) Experimental determination of the optimal LP-parameter space; and (iv) Objective criteria and speech recognition tests for performance comparison. Optimal local polynomial modeling and applications We next address the problem of fitting a piecewise-polynomial model to a smooth signal corrupted by additive noise. Since the signal is smooth, it can be represented using low-order polynomial functions provided that they are locally adapted to the signal. We choose the mean-square error as the criterion of optimality. Since the model is local, it preserves the temporal structure of the signal and can also handle nonstationary noise. We show that there is a trade-off between the adaptability of the model to local signal variations and robustness to noise (bias-variance trade-off), which we solve using a stochastic optimization technique known as the intersection of confidence intervals (ICI) technique. The key trade-off parameter is the duration of the window over which the optimum LPM is computed. Within the LPM framework, we address three problems: (i) Signal reconstruction from noisy uniform samples; (ii) Signal reconstruction from noisy nonuniform samples; and (iii) Classification of speech signals into voiced and unvoiced segments. The generic signal model is x(tn)=s(tn)+d(tn),0 ≤ n ≤ N - 1. In problems (i) and (iii) above, tn=nT(uniform sampling); in (ii) the samples are taken at nonuniform instants. The signal s(t)is assumed to be smooth; i.e., it should admit a local polynomial representation. The problem in (i) and (ii) is to estimate s(t)from x(tn); i.e., we are interested in optimal signal reconstruction on a continuous domain starting from uniform or nonuniform samples. We show that, in both cases, the bias and variance take the general form: The mean square error (MSE) is given by where L is the length of the window over which the polynomial fitting is performed, f is a function of s(t), which typically comprises the higher-order derivatives of s(t), the order itself dependent on the order of the polynomial, and g is a function of the noise variance. It is clear that the bias and variance have complementary characteristics with respect to L. Directly optimizing for the MSE would give a value of L, which involves the functions f and g. The function g may be estimated, but f is not known since s(t)is unknown. Hence, it is not practical to compute the minimum MSE (MMSE) solution. Therefore, we obtain an approximate result by solving the bias-variance trade-off in a probabilistic sense using the ICI technique. We also propose a new approach to optimally select the ICI technique parameters, based on a new cost function that is the sum of the probability of false alarm and the area covered over the confidence interval. In addition, we address issues related to optimal model-order selection, search space for window lengths, accuracy of noise estimation, etc. The next issue addressed is that of voiced/unvoiced segmentation of speech signal. Speech segments show different spectral and temporal characteristics based on whether the segment is voiced or unvoiced. Most speech processing techniques process the two segments differently. The challenge lies in making detection techniques offer robust performance in the presence of noise. We propose a new technique for voiced/unvoiced clas-sification by taking into account the fact that voiced segments have a certain degree of regularity, and that the unvoiced segments do not possess any smoothness. In order to capture the regularity in voiced regions, we employ the LPM. The key idea is that regions where the LPM is inaccurate are more likely to be unvoiced than voiced. Within this frame-work, we formulate a hypothesis testing problem based on the accuracy of the LPM fit and devise a test statistic for performing V/UV classification. Since the technique is based on LPM, it is capable of adapting to nonstationary noises. We present Monte Carlo results to demonstrate the accuracy of the proposed technique.