Spectral decomposition is a modern tool that utilizes seismic data to generate additional useful information in seismic exploration for hydrocarbon detection, lithology identification, stratigraphic interpretation, filtering and others. Different spectral decomposition methods with applications to seismic data were reported and investigated in past years. Many methods usually do not consider the non-stationary features of seismic data and, therefore, are not likely to give satisfactory results. S-transform developed in recent years is able to provide time-dependent frequency analysis while maintaining a direct relationship with the Fourier spectrum, a unique property that other methods of spectral decomposition may not have. In this thesis, I investigated the feasibility and efficiency of using S-transform for hydrocarbon detection and time-varying surface wave filtering. S-transform was first applied to two seismic data sets from a clastic reservoir in the North Sea and a deep carbonate reservoir in the Sichuan Basin, China. Results from both cases demonstrated that S-transform decomposition technique can detect hydrocarbon zones effectively and helps to build the relationships between lithology changes and high frequency variation and between hydrocarbon occurrence and low-frequency anomaly. However, its time resolution needs to be improved. In the second part of my thesis, I used S-transform to develop a novel Time-frequency-wave-number-domain (T-F-K) filtering method to separate surface wave from reflected waves in seismic records. The S-T-F-K filtering proposed here can be used to analyze surface waves on separate f-k panels at different times. The method was tested using hydrophone records of four-component seismic data acquired in the shallow-water Persian Gulf where the average water depth is about 10m and Scholte waves and other surfaces wave persistently strong. Results showed that this new S-T-F-K method is able to separate and sttenuate surface waves and to improve greatly the quality of seismic reflection signals that are otherwise completely concealed by the aliased surface waves.

O presente trabalho apresenta uma comparação de duas técnicas para a análise tempo - frequência em análise de qualidade de energia elétrica para sinais de tensão que contenham distúrbios individuais ou simultâneos. Dessa forma, o objetivo, desta dissertação, é encontrar uma ferramenta que forneça as características e parâmetros para a localização, identificação e classificação de tais distúrbios. O estudo consiste na análise do desempenho da Transformada Wavelet Discreta e da Transformada-S, principalmente, quando os sinais são analisados na presença de múltiplos distúrbios. Ambas as transformadas fornecem informação importante nos domínios do tempo e da frequência. No entanto, essas ferramentas não tem sido amplamente exploradas para análise de múltiplos distúrbios. Neste contexto, ambas as transformadas são testadas para conhecer seus desempenhos e suas capacidades de identificação e localização de eventos de qualidade de energia elétrica. Para finalizar, é projetado um sistema classificador baseado em arvore de decisão capaz de reconhecer quinze tipos de distúrbios diferentes. / This work presents a comparison of two methods for time-frequency analysis applied in Power Quality signals containing single or multiple disturbances. In this way, the aim of this work is to apply tools that supply the parameters and characteristics to identify, locate and classify Power Quality disturbances. For that, the proposed method analyzes the performance of the Wavelet and S transforms, mainly when the signals are with more than one disturbance type. Both mathematical tools supply important information on the time and frequency domain. However, these tools have not been thoroughly used to analyze multiple events locate Power Quality events. In this contest, both transforms are tested in order to assess their performance to identify and locate electrical power quality events. According to a decision tree classifier, fifteen types of single and combined power disturbances are well recognized.

O presente trabalho apresenta uma comparação de duas técnicas para a análise tempo - frequência em análise de qualidade de energia elétrica para sinais de tensão que contenham distúrbios individuais ou simultâneos. Dessa forma, o objetivo, desta dissertação, é encontrar uma ferramenta que forneça as características e parâmetros para a localização, identificação e classificação de tais distúrbios. O estudo consiste na análise do desempenho da Transformada Wavelet Discreta e da Transformada-S, principalmente, quando os sinais são analisados na presença de múltiplos distúrbios. Ambas as transformadas fornecem informação importante nos domínios do tempo e da frequência. No entanto, essas ferramentas não tem sido amplamente exploradas para análise de múltiplos distúrbios. Neste contexto, ambas as transformadas são testadas para conhecer seus desempenhos e suas capacidades de identificação e localização de eventos de qualidade de energia elétrica. Para finalizar, é projetado um sistema classificador baseado em arvore de decisão capaz de reconhecer quinze tipos de distúrbios diferentes. / This work presents a comparison of two methods for time-frequency analysis applied in Power Quality signals containing single or multiple disturbances. In this way, the aim of this work is to apply tools that supply the parameters and characteristics to identify, locate and classify Power Quality disturbances. For that, the proposed method analyzes the performance of the Wavelet and S transforms, mainly when the signals are with more than one disturbance type. Both mathematical tools supply important information on the time and frequency domain. However, these tools have not been thoroughly used to analyze multiple events locate Power Quality events. In this contest, both transforms are tested in order to assess their performance to identify and locate electrical power quality events. According to a decision tree classifier, fifteen types of single and combined power disturbances are well recognized.

Unmanned Aerial Vehicles (UAVs) may develop cracks, erosion, delamination or other damages due to aging, fatigue or extreme loads. Identifying these damages is critical for the safe and reliable operation of the systems. Structural Health Monitoring (SHM) is capable of determining the conditions of systems automatically and continually through processing and interpreting the data collected from a network of sensors embedded into the systems. With the desired awareness of the systems’ health conditions, SHM can greatly reduce operational cost and speed up maintenance processes. The purpose of this study is to develop an effective, low-cost, flexible and fault tolerant structural health monitoring system. The proposed Index Based Reasoning (IBR) system started as a simple look-up-table based diagnostic system. Later, Fast Fourier Transformation analysis and neural network diagnosis with self-learning capabilities were added. The current version is capable of classifying different health conditions with the learned characteristic patterns, after training with the sensory data acquired from the operating system under different status. The proposed IBR systems are hierarchy and distributed networks deployed into systems to monitor their health conditions. Each IBR node processes the sensory data to extract the features of the signal. Classifying tools are then used to evaluate the local conditions with health index (HI) values. The HI values will be carried to other IBR nodes in the next level of the structured network. The overall health condition of the system can be obtained by evaluating all the local health conditions. The performance of IBR systems has been evaluated by both simulation and experimental studies. The IBR system has been proven successful on simulated cases of a turbojet engine, a high displacement actuator, and a quad rotor helicopter. For its application on experimental data of a four rotor helicopter, IBR also performed acceptably accurate. The proposed IBR system is a perfect fit for the low-cost UAVs to be the onboard structural health management system. It can also be a backup system for aircraft and advanced Space Utility Vehicles.

This work describes basics of electroencephalography, measuring methods of electroen- cephalographic signals, their processing and especially the interpretation of EEG signal in frequency and time-frequency domains for mental fatigue detection purposes. Mental fatigue, its sources, consequences and connection with sensory-cognitive system and link to memory is discussed. The most basic normalized international system for measuring EEG from the scalp as well as some of the experiments that ultimately lead to mental fatigue are described. With this knowledge in mind, an experiment was prepared for inducing such a state. Ten subjects participated in the test which was conducted in la- boratory with EEG machine GES 410MR by EGI. The data were analyzed mainly with S-transform and Hilbert-Huang transform. These two transforms represent two distinct state of the art time-frequency methods of spectral analysis. The result of this work lies in evaluating the relationship between mental fatigue, errors accumulated during the task and with time.

A consensus model combines statistical methods with signal processing to create a better picture of the family of related signals. In this thesis, we will consider 32 signals produced by a single electroencephalogram (EEG) recording session. The consensus model will be produced by using the S-Transform of the individual signals and then normalized to unit energy. A bootstrapping process is used to produce a consensus spectrum. This leads to the consensus model via the inverse S-Transform of the consensus spectrum. The method will be applied to both a control and experimental EEG to show how the results can be used in clinical settings to analyze experimental outcomes.

A conexão de Geradores Distribuídos (GDs) no sistema de distribuição vem se intensificando nos últimos anos. Neste cenário, o aumento de GDs pode trazer alguns benefícios, como a redundância da geração e redução das perdas elétricas. Por outro lado, o problema do ilhamento também vem se destacando. Atualmente, existem técnicas já consolidadas para a detecção do ilhamento, sendo que as técnicas passivas estão entre as mais utilizadas. Entretanto, as técnicas passivas são bastante dependentes do desbalanço de potência entre a geração e as cargas no momento de ocorrência do ilhamento para atuarem corretamente. Caso o desbalanço de potência seja pequeno, as técnicas passivas tendem a não identificar o ilhamento, gerando as chamadas Zonas de Não Detecção (ZNDs). Para mitigar este problema, a pesquisa por técnicas passivas inteligentes baseadas em aprendizagem de máquina vem se tornando cada vez mais comum. Neste trabalho foi modelada uma proteção anti-ilhamento baseada em Redes Neurais Artificiais (RNAs). A classificação do ilhamento é feita com base no espectro de frequência das tensões nos terminais do GD com o uso da Transformada de Stockwell, ou apenas Transformada S (TS). Outro ponto importante da metodologia foi a implementação de uma etapa de detecção de eventos, também baseada nas energias do espectro de frequência das tensões, para evitar a constante execução do classificador. Assim, a RNA apenas irá classificar o evento após receber um sinal de trigger da etapa de detecção de evento. Para o treinamento da RNA foram testados dois algoritmos diferentes, o clássico Backpropagation (BP) e a Máquina de Aprendizado Extremo, do inglês Extreme Learning Machine (ELM). Ressalta-se o melhor desempenho obtido com as redes treinadas pelo ELM, que apresentaram uma capacidade de generalização muito maior, logo, resultando em taxas de acerto mais elevadas. De modo geral, depois de comparada com métodos passivos convencionais para a detecção de ilhamento, a proteção proposta se mostrou mais precisa e com um tempo de detecção muito menor, sendo inferior a 2 ciclos. Por fim, ainda foi realizada a análise das ZNDs para a proteção proposta e as técnicas convencionais, por ser uma característica muito importante para a proteção antiilhamento, mas que não é comumente abordada para técnicas passivas inteligentes. Nesta análise, o método para a detecção de ilhamento proposto novamente se sobressaiu às técnicas convencionais, apresentado uma ZND muito menor. / The connection of distributed generators (DG) in the distribution system has been intensified in the recent years. In this scenario, the increase of DG can bring some benefits, such as generation redundancy and reduction of power losses. On the other hand, the problem of islanding is also been highlighted. Currently, there are already consolidated techniques for islanding detection, and passive techniques are among the most used ones. However, the passive techniques are very dependent of the power unbalance between the generation and the loads at the moment of the islanding in order to actuate properly. If the power mismatch is small, the passive techniques tend to not identify the islanding, generating the so called Non-Detection Zones (NDZ). To mitigate this issue, the research of intelligent passive techniques based in machine learning is becoming more common. In this study, an anti-islanding protection based on Artificial Neural Networks (ANN) was modelled. The islanding classification is done based on the frequency spectrum of the DG\'s terminal voltages using the Stockwell Transform, or just S-Transform (ST). Another important point of the methodology was the implementation of an event detection stage, also based on the energies of the voltages frequency spectrum, to avoid the constant execution of the classifier. Therefore, the ANN will only classify the event after receiving a trigger signal from the event detection stage. To train the ANN, two different algorithms were tested: the classic Backpropagation and the Extreme Learning Machine (ELM). It is noteworthy the better performance obtained with the neural networks trained by the ELM, which had a greater capacity of generalization, hence resulting in higher success rates. In general, after being compared with conventional passive techniques for islanding detection, the proposed protection was more accurate and with a much smaller detection time, being less than 2 cycles. Finally, the analysis of the NDZ for the proposed protection and the conventional techniques was carried out, as it is a very important feature for anti-islanding protection, but is not commonly addressed for intelligent passive techniques. In this analysis, the islanding detection method proposed again overcame the conventional techniques, presenting a much smaller NDZ.

Cette thèse dans le domaine du traitement des signaux non-stationnaires, appliqué aux bruits du cœur mesurés avec un stéthoscope numérique, vise à concevoir un outil automatisé et « intelligent », permettant aux médecins de disposer d’une source d’information supplémentaire à celle du stéthoscope traditionnel. Une première étape dans l’analyse des signaux du cœur, consiste à localiser le premier et le deuxième son cardiaque (S1 et S2) afin de le segmenter en quatre parties : S1, systole, S2 et diastole. Plusieurs méthodes de localisation des sons cardiaques existent déjà dans la littérature. Une étude comparative entre les méthodes les plus pertinentes est réalisée et deux nouvelles méthodes basées sur la transformation temps-fréquence de Stockwell sont proposées. La première méthode, nommée SRBF, utilise des descripteurs issus du domaine temps-fréquence comme vecteur d’entré au réseau de neurones RBF qui génère l’enveloppe d’amplitude du signal cardiaque, la deuxième méthode, nommée SSE, calcule l’énergie de Shannon du spectre local obtenu par la transformée en S. Ensuite, une phase de détection des extrémités (onset, ending) est nécessaire. Une méthode d’extraction des signaux S1 et S2, basée sur la transformée en S optimisée, est discutée et comparée avec les différentes approches qui existent dans la littérature. Concernant la classification des signaux cardiaques, les méthodes décrites dans la littérature pour classifier S1 et S2, se basent sur des critères temporels (durée de systole et diastole) qui ne seront plus valables dans plusieurs cas pathologiques comme par exemple la tachycardie sévère. Un nouveau descripteur issu du domaine temps-fréquence est évalué et validé pour discriminer S1 de S2. Ensuite, une nouvelle méthode de génération des attributs, basée sur la décomposition modale empirique (EMD) est proposée.Des descripteurs non-linéaires sont également testés, dans le but de classifier des sons cardiaques normaux et sons pathologiques en présence des souffles systoliques. Des outils de traitement et de reconnaissance des signaux non-stationnaires basés sur des caractéristiques morphologique, temps-fréquences et non linéaire du signal, ont été explorés au cours de ce projet de thèse afin de proposer un module d’aide au diagnostic, qui ne nécessite pas d’information à priori sur le sujet traité, robuste vis à vis du bruit et applicable dans des conditions cliniques. / This thesis in the field of biomedical signal processing, applied to the heart sounds, aims to develop an automated and intelligent module, allowing medical doctors to have an additional source of information than the traditional stethoscope. A first step in the analysis of heart sounds is the segmentation process. The heart sounds segmentation process segments the PCG (PhonoCardioGram) signal into four parts: S1 (first heart sound), systole, S2 (second heart sound) and diastole. It can be considered one of the most important phases in the auto-analysis of PCG signals. The proposed segmentation module in this thesis can be divided into three main blocks: localization of heart sounds, boundaries detection of the localized heart sounds and classification block to distinguish between S1and S2. Several methods of heart sound localization exist in the literature. A comparative study between the most relevant methods is performed and two new localization methods of heart sounds are proposed in this study. Both of them are based on the S-transform, the first method uses Radial Basis Functions (RBF) neural network to extract the envelope of the heart sound signal after a feature extraction process that operates on the S-matrix. The second method named SSE calculates the Shannon Energy of the local spectrum calculated by the S-transform for each sample of the heart sound signal. The second block contains a novel approach for the boundaries detection of S1 and S2 (onset & ending). The energy concentrations of the S-transform of localized sounds are optimized by using a window width optimization algorithm. Then the SSE envelope is recalculated and a local adaptive threshold is applied to refine the estimated boundaries. For the classification block, most of the existing methods in the literature use the systole and diastole duration (systole regularity) as a criterion to discriminate between S1 and S2. These methods do not perform well for all types of heart sounds, especially in the presence of high heart rate or in the presence of arrhythmic pathologies. To deal with this problem, two feature extraction methods based on Singular Value Decomposition (SVD) technique are examined. The first method uses the S-Transform and the second method uses the Intrinsic Mode Functions (IMF) calculated by the Empirical Mode Decomposition (EMD) technique. The features are applied to a KNN classifier to estimate the performance of each feature extraction method. Nonlinear features are also tested in order to classify the normal and pathological heart sounds in the presence of systolic murmurs. Processing and recognition signal processing tools based on morphological, time-frequency and nonlinear signal features, were explored in this thesis in order to propose an auto-diagnosis module, robust against noise and applicable in clinical conditions.

When we consider the nature of the scientific community in conjunction with a sense of typical economic circumstances we find that there are two distinct paths for development. One path involves hypothesis testing and evolution of strategies that are linked with iterations in equipment advances. A second, more complicated scenario, can involve external influences whether economic, political, or otherwise, such as the government closure of NASA's space program in 2011 which will no doubt influence research in associated fields. The following chapters are an account of examples of two statistical techniques and the importance of both on the two relatively unrelated geological fields of coastal geomorphology and ground water hydrology. The first technique applies a multi-dimensional approach to defining groundwater water table response based on precipitation in areas where it can reasonably be assumed to be the only recharge. The second technique applies a high resolution multi-scalar approach to a geologic setting most often restricted to either high resolution locally, or low resolution regionally. This technique uses time-frequency analysis to characterize cuspate patterns in LIDAR data are introduced using examples from the Atlantic coast of Florida, United States. These techniques permit the efficient study of beachface landforms over many kilometers of coastline at multiple spatial scales. From a LIDAR image, a beach-parallel spatial series is generated. Here, this series is the shore-normal position of a specific elevation (contour line). Well-established time-frequency analysis techniques, wavelet transforms, and S-Transforms, are then applied to the spatial series. These methods yield results compatible with traditional methods and show that it is useful for capturing transitions in cuspate shapes. To apply this new method, a land-based LIDAR study allowing for rapid high-resolution surveying is conducted on Melbourne Beach, Florida and Tairua Beach, New Zealand. Comparisons and testing of two different terrestrial scanning stations are evaluated during the course of the field investigation. Significant cusp activity is observed at Melbourne Beach. Morphological observations and sediment analysis are used to study beach cusp morphodynamics at the site. Surveys at Melbourne were run ~500 m alongshore and sediment samples were collected intertidally over a five-day period. Beach cusp location within larger scale beach morphology is shown to directly influence cusp growth as either predominantly erosional or accretional. Sediment characteristics within the beach cusp morphology are reported coincident with cusp evolution. Variations in pthesis size distribution kurtosis are exhibited as the cusps evolve; however, no significant correlation is seen between grain size and position between horn and embayment. During the end of the study, a storm resulted in beach cusp destruction and increased sediment sorting. In the former technique using multi-dimensional studies, a test of a new method for improving forecasting of surficial aquifer system water level changes with rainfall is conducted. The results provide a more rigorous analysis of common predictive techniques and compare them with the results of the tested model. These results show that linear interpretations of response-to-rainfall data require a clarification of how large events distort prediction and how the binning of data can change the interpretation. Analyses show that the binning ground water recharge data as is typically done in daily format may be useful for quick interpretation but only describes how fast the system responds to an event, not the frequency of return of such a response. Without a secure grasp on the nonlinear nature of water table and rainfall data alike, any binning or isolation of specific data carries the potential for aliasing that must be accounted for in an interpretation. The new model is proven capable of supplanting any current linear regression analysis as a more accurate means of prediction through the application of a multivariate technique. Furthermore, results show that in the Florida surficial aquifer system response-to-rainfall ratios exhibit a maxima most often linked with modal stage.

Knowledge of protein hot spots and the ability to successfully predict them while using only primary protein structure has been a worldwide scientific goal for several decades. This thesis describes the importance of hot spots and sums up advances achieved in this field of study so far. Besides that we introduce hot spot prediction algorithm using only a primary protein structure, based primarily on signal processing techniques. To convert protein sequence to numerical signal we use the EIIP attribute, while further processing is carried out via means of S-transform. The algorithm achieves sensitivity of more than 60 %, positive predictive value exceeds 50 % and the main advantage over competitive algorithms is its simplicity and low computational requirements.