Synchronizing sound from different devices over a TCP network

Nowadays, we can send audio on the Internet for multiples uses like telephony, broadcast audio or teleconferencing. The issue comes when you need to synchronize the sound from different sources because the network where we are going to work could lose packets and introduce delay in the delivery. This can also come because the sound cards could be work in different speeds. In this project, we will work with two computers emitting sound (one will simulate the left channel (mono) of a stereo signal, and the other the right channel) and connected with a third computer by a TCP network. The last computer must get the sound from both computers and reproduce it in a speaker properly (without delay). So, basically, the main goal of the project is to synchronize multi-track sound over a network. TCP networks introduce latency into data transfers. Streaming audio suffers from two problems: a delay and an offset between the channels. This project explores the causes of latency, investigates the affect of the inter-channel offset and proposes a solution to synchronize the received channels. In conclusion, a good synchronization of the sound is required in a time when several audio applications are being developed. When two devices are ready to send audio over a network, this multi-track sound will arrive at the third computer with an offset giving a negative effect to the listener. This project has dealt with this offset achieving a good synchronization of the multitrack sound getting a good effect on the listener. This was achieved thanks to the division of the project into several steps having constantly a good vision of the problem, a good scalability and having controlled the latency at all times. As we can see in the chapter 4 of the project, a lack of synchronization over c. 100μs is audible to the listener. RESUMEN. A día de hoy, podemos transmitir audio a través de Internet por varios motivos como pueden ser: una llamada telefónica, una emisión de audio o una teleconferencia. El problema viene cuando necesitas sincronizar ese sonido producido por los diferentes orígenes ya que la red a la que nos vamos a conectar puede perder los paquetes y/o introducir un retardo en las entregas de los mismos. Así mismo, estos retardos también pueden venir producidos por las diferentes velocidades a las que trabajan las tarjetas de sonido de cada dispositivo. En este proyecto, se ha trabajado con dos ordenadores emitiendo sonido de manera intermitente (uno se encargará de simular el canal izquierdo (mono) de la señal estéreo emitida, y el otro del canal derecho), estando conectados a través de una red TCP a un tercer ordenador, el cual debe recibir el sonido y reproducirlo en unos altavoces adecuadamente y sin retardo (deberá juntar los dos canales y reproducirlo como si de estéreo de tratara). Así, el objetivo principal de este proyecto es el de encontrar la manera de sincronizar el sonido producido por los dos ordenadores y escuchar el conjunto en unos altavoces finales. Las redes TCP introducen latencia en la transferencia de datos. El streaming de audio emitido a través de una red de este tipo puede sufrir dos grandes contratiempos: retardo y offset, los dos existentes en las comunicaciones entre ambos canales. Este proyecto se centra en las causas de ese retardo, investiga el efecto que provoca el offset entre ambos canales y propone una solución para sincronizar los canales en el dispositivo receptor. Para terminar, una buena sincronización del sonido es requerida en una época donde las aplicaciones de audio se están desarrollando continuamente. Cuando los dos dispositivos estén preparados para enviar audio a través de la red, la señal de sonido multi-canal llegará al tercer ordenador con un offset añadido, por lo que resultará en una mala experiencia en la escucha final. En este proyecto se ha tenido que lidiar con ese offset mencionado anteriormente y se ha conseguido una buena sincronización del sonido multi-canal obteniendo un buen efecto en la escucha final. Esto ha sido posible gracias a una división del proyecto en diversas etapas que proporcionaban la facilidad de poder solucionar los errores en cada paso dando una importante visión del problema y teniendo controlada la latencia en todo momento. Como se puede ver en el capítulo 4 del proyecto, la falta de sincronización sobre una diferencia de 100μs entre dos canales (offset) empieza a ser audible en la escucha final.

Wireless communication enhancement methods for mobile robots in radiation environments

En entornos hostiles tales como aquellas instalaciones científicas donde la radiación ionizante es el principal peligro, el hecho de reducir las intervenciones humanas mediante el incremento de las operaciones robotizadas está siendo cada vez más de especial interés. CERN, la Organización Europea para la Investigación Nuclear, tiene alrededor de unos 50 km de superficie subterránea donde robots móviles controlador de forma remota podrían ayudar en su funcionamiento, por ejemplo, a la hora de llevar a cabo inspecciones remotas sobre radiación en los diferentes áreas destinados al efecto. No solo es preciso considerar que los robots deben ser capaces de recorrer largas distancias y operar durante largos periodos de tiempo, sino que deben saber desenvolverse en los correspondientes túneles subterráneos, tener en cuenta la presencia de campos electromagnéticos, radiación ionizante, etc. y finalmente, el hecho de que los robots no deben interrumpir el funcionamiento de los aceleradores. El hecho de disponer de un sistema de comunicaciones inalámbrico fiable y robusto es esencial para la correcta ejecución de las misiones que los robots deben afrontar y por supuesto, para evitar tales situaciones en las que es necesario la recuperación manual de los robots al agotarse su energía o al perder el enlace de comunicaciones. El objetivo de esta Tesis es proveer de las directrices y los medios necesarios para reducir el riesgo de fallo en la misión y maximizar las capacidades de los robots móviles inalámbricos los cuales disponen de almacenamiento finito de energía al trabajar en entornos peligrosos donde no se dispone de línea de vista directa. Para ello se proponen y muestran diferentes estrategias y métodos de comunicación inalámbrica. Teniendo esto en cuenta, se presentan a continuación los objetivos de investigación a seguir a lo largo de la Tesis: predecir la cobertura de comunicaciones antes y durante las misiones robotizadas; optimizar la capacidad de red inalámbrica de los robots móviles con respecto a su posición; y mejorar el rango operacional de esta clase de robots. Por su parte, las contribuciones a la Tesis se citan más abajo. El primer conjunto de contribuciones son métodos novedosos para predecir el consumo de energía y la autonomía en la comunicación antes y después de disponer de los robots en el entorno seleccionado. Esto es importante para proporcionar conciencia de la situación del robot y evitar fallos en la misión. El consumo de energía se predice usando una estrategia propuesta la cual usa modelos de consumo provenientes de diferentes componentes en un robot. La predicción para la cobertura de comunicaciones se desarrolla usando un nuevo filtro de RSS (Radio Signal Strength) y técnicas de estimación con la ayuda de Filtros de Kalman. El segundo conjunto de contribuciones son métodos para optimizar el rango de comunicaciones usando novedosas técnicas basadas en muestreo espacial que son robustas frente a ruidos de campos de detección y radio y que proporcionan redundancia. Se emplean métodos de diferencia central finitos para determinar los gradientes 2D RSS y se usa la movilidad del robot para optimizar el rango de comunicaciones y la capacidad de red. Este método también se valida con un caso de estudio centrado en la teleoperación háptica de robots móviles inalámbricos. La tercera contribución es un algoritmo robusto y estocástico descentralizado para la optimización de la posición al considerar múltiples robots autónomos usados principalmente para extender el rango de comunicaciones desde la estación de control al robot que está desarrollando la tarea. Todos los métodos y algoritmos propuestos se verifican y validan usando simulaciones y experimentos de campo con variedad de robots móviles disponibles en CERN. En resumen, esta Tesis ofrece métodos novedosos y demuestra su uso para: predecir RSS; optimizar la posición del robot; extender el rango de las comunicaciones inalámbricas; y mejorar las capacidades de red de los robots móviles inalámbricos para su uso en aplicaciones dentro de entornos peligrosos, que como ya se mencionó anteriormente, se destacan las instalaciones científicas con emisión de radiación ionizante. En otros términos, se ha desarrollado un conjunto de herramientas para mejorar, facilitar y hacer más seguras las misiones de los robots en entornos hostiles. Esta Tesis demuestra tanto en teoría como en práctica que los robots móviles pueden mejorar la calidad de las comunicaciones inalámbricas mediante la profundización en el estudio de su movilidad para optimizar dinámicamente sus posiciones y mantener conectividad incluso cuando no existe línea de vista. Los métodos desarrollados en la Tesis son especialmente adecuados para su fácil integración en robots móviles y pueden ser aplicados directamente en la capa de aplicación de la red inalámbrica. ABSTRACT In hostile environments such as in scientific facilities where ionising radiation is a dominant hazard, reducing human interventions by increasing robotic operations are desirable. CERN, the European Organization for Nuclear Research, has around 50 km of underground scientific facilities, where wireless mobile robots could help in the operation of the accelerator complex, e.g. in conducting remote inspections and radiation surveys in different areas. The main challenges to be considered here are not only that the robots should be able to go over long distances and operate for relatively long periods, but also the underground tunnel environment, the possible presence of electromagnetic fields, radiation effects, and the fact that the robots shall in no way interrupt the operation of the accelerators. Having a reliable and robust wireless communication system is essential for successful execution of such robotic missions and to avoid situations of manual recovery of the robots in the event that the robot runs out of energy or when the robot loses its communication link. The goal of this thesis is to provide means to reduce risk of mission failure and maximise mission capabilities of wireless mobile robots with finite energy storage capacity working in a radiation environment with non-line-of-sight (NLOS) communications by employing enhanced wireless communication methods. Towards this goal, the following research objectives are addressed in this thesis: predict the communication range before and during robotic missions; optimise and enhance wireless communication qualities of mobile robots by using robot mobility and employing multi-robot network. This thesis provides introductory information on the infrastructures where mobile robots will need to operate, the tasks to be carried out by mobile robots and the problems encountered in these environments. The reporting of research work carried out to improve wireless communication comprises an introduction to the relevant radio signal propagation theory and technology followed by explanation of the research in the following stages: An analysis of the wireless communication requirements for mobile robot for different tasks in a selection of CERN facilities; predictions of energy and communication autonomies (in terms of distance and time) to reduce risk of energy and communication related failures during missions; autonomous navigation of a mobile robot to find zone(s) of maximum radio signal strength to improve communication coverage area; and autonomous navigation of one or more mobile robots acting as mobile wireless relay (repeater) points in order to provide a tethered wireless connection to a teleoperated mobile robot carrying out inspection or radiation monitoring activities in a challenging radio environment. The specific contributions of this thesis are outlined below. The first sets of contributions are novel methods for predicting the energy autonomy and communication range(s) before and after deployment of the mobile robots in the intended environments. This is important in order to provide situational awareness and avoid mission failures. The energy consumption is predicted by using power consumption models of different components in a mobile robot. This energy prediction model will pave the way for choosing energy-efficient wireless communication strategies. The communication range prediction is performed using radio signal propagation models and applies radio signal strength (RSS) filtering and estimation techniques with the help of Kalman filters and Gaussian process models. The second set of contributions are methods to optimise the wireless communication qualities by using novel spatial sampling based techniques that are robust to sensing and radio field noises and provide redundancy features. Central finite difference (CFD) methods are employed to determine the 2-D RSS gradients and use robot mobility to optimise the communication quality and the network throughput. This method is also validated with a case study application involving superior haptic teleoperation of wireless mobile robots where an operator from a remote location can smoothly navigate a mobile robot in an environment with low-wireless signals. The third contribution is a robust stochastic position optimisation algorithm for multiple autonomous relay robots which are used for wireless tethering of radio signals and thereby to enhance the wireless communication qualities. All the proposed methods and algorithms are verified and validated using simulations and field experiments with a variety of mobile robots available at CERN. In summary, this thesis offers novel methods and demonstrates their use to predict energy autonomy and wireless communication range, optimise robots position to improve communication quality and enhance communication range and wireless network qualities of mobile robots for use in applications in hostile environmental characteristics such as scientific facilities emitting ionising radiations. In simpler terms, a set of tools are developed in this thesis for improving, easing and making safer robotic missions in hostile environments. This thesis validates both in theory and experiments that mobile robots can improve wireless communication quality by exploiting robots mobility to dynamically optimise their positions and maintain connectivity even when the (radio signal) environment possess non-line-of-sight characteristics. The methods developed in this thesis are well-suited for easier integration in mobile robots and can be applied directly at the application layer of the wireless network. The results of the proposed methods have outperformed other comparable state-of-the-art methods.

Transformación del espacio cinemático Laban para la evaluación del movimiento

Este proyecto tiene como objetivo la implementación de un sistema capaz de analizar el movimiento corporal a partir de unos puntos cinemáticos. Estos puntos cinemáticos se obtienen con un programa previo y se captan con la cámara kinect. Para ello el primer paso es realizar un estudio sobre las técnicas y conocimientos existentes relacionados con el movimiento de las personas. Se sabe que Rudolph Laban fue uno de sus mayores exponentes y gracias a sus observaciones se establece una relación entre la personalidad, el estado anímico y la forma de moverse de un individuo. Laban acuñó el término esfuerzo, que hace referencia al modo en que se administra la energía que genera el movimiento y de qué manera se modula en las secuencias, es una manera de describir la intención de las expresiones internas. El esfuerzo se divide en 4 categorías: peso, espacio, tiempo y flujo, y cada una de estas categorías tiene una polaridad denominada elemento de esfuerzo. Con estos 8 elementos de esfuerzo un movimiento queda caracterizado. Para poder cuantificar los citados elementos de esfuerzo se buscan movimientos que representen a alguno de ellos. Los movimientos se graban con la cámara kinect y se guardan sus valores en un archivo csv. Para el procesado de estos datos se establece que el sistema más adecuado es una red neuronal debido a su flexibilidad y capacidad a la hora de procesar entradas no lineales. Para la implementación de la misma se requiere un amplio estudio que incluye: topologías, funciones de activación, tipos de aprendizaje, algoritmos de entrenamiento entre otros. Se decide que la red tenga dos capas ocultas, para mejor procesado de los datos, que sea estática, siga un proceso de cálculo hacia delante (Feedforward) y el algoritmo por el que se rija su aprendizaje sea el de retropropagación (Backpropagation) En una red estática las entradas han de ser valores fijos, es decir, no pueden variar en el tiempo por lo que habrá que implementar un programa intermedio que haga una media aritmética de los valores. Una segunda prueba con la misma red trata de comprobar si sería capaz de reconocer movimientos que estuvieran caracterizados por más de un elemento de esfuerzo. Para ello se vuelven a grabar los movimientos, esta vez en parejas de dos, y el resto del proceso es igual. ABSTRACT. The aim of this project is the implementation of a system able to analyze body movement from cinematic data. This cinematic data was obtained with a previous program. The first step is carrying out a study about the techniques and knowledge existing nowadays related to people movement. It is known that Rudolf Laban was one the greatest exponents of this field and thanks to his observations a relation between personality, mood and the way the person moves was made. Laban coined the term effort, that refers to the way energy generated from a movement is managed and how it is modulated in the sequence, this is a method of describing the inner intention of the person. The effort is divided into 4 categories: weight, space, time and flow, and each of these categories have 2 polarities named elements of effort. These 8 elements typify a movement. We look for movements that are made of these elements so we can quantify them. The movements are recorded with the kinect camera and saved in a csv file. In order to process this data a neural network is chosen owe to its flexibility and capability of processing non-linear inputs. For its implementation it is required a wide study regarding: topology, activation functions, different types of learning methods and training algorithms among others. The neural network for this project will have 2 hidden layers, it will be static and follow a feedforward process ruled by backpropagation. In a static net the inputs must be fixed, this means they cannot vary in time, so we will have to implement an intermediate program to calculate the average of our data. A second test for our net will be checking its ability to recognize more than one effort element in just one movement. In order to do this all the movements are recorded again but this time in pairs, the rest of the process remains the same.

Glucose-Insulin regulator for type 1 diabetes using high order neural networks

In this paper a Glucose-Insulin regulator for Type 1 Diabetes using artificial neural networks (ANN) is proposed. This is done using a discrete recurrent high order neural network in order to identify and control a nonlinear dynamical system which represents the pancreas? beta-cells behavior of a virtual patient. The ANN which reproduces and identifies the dynamical behavior system, is configured as series parallel and trained on line using the extended Kalman filter algorithm to achieve a quickly convergence identification in silico. The control objective is to regulate the glucose-insulin level under different glucose inputs and is based on a nonlinear neural block control law. A safety block is included between the control output signal and the virtual patient with type 1 diabetes mellitus. Simulations include a period of three days. Simulation results are compared during the overnight fasting period in Open-Loop (OL) versus Closed- Loop (CL). Tests in Semi-Closed-Loop (SCL) are made feedforward in order to give information to the control algorithm. We conclude the controller is able to drive the glucose to target in overnight periods and the feedforward is necessary to control the postprandial period.

Overtopping of harbour breakwaters: a comparison of semi-empirical equations, neural networks, and physical model tests

This paper reports extensive tests of empirical equations developed by different authors for harbour breakwater overtopping. First, the existing equations are compiled and evaluated as tools for estimating the overtopping rates on sloping and vertical breakwaters. These equations are then tested using the data obtained in a number of laboratory studies performed in the Centre for Harbours and Coastal Studies of the CEDEX, Spain. It was found that the recommended application ranges of the empirical equations typically deviate from those revealed in the experimental tests. In addition, a neural network model developed within the European CLASH Project is tested. The wind effects on overtopping are also assessed using a reduced scale physical model

Peptide design by artificial neural networks and computer-based evolutionary search

A technique for systematic peptide variation by a combination of rational and evolutionary approaches is presented. The design scheme consists of five consecutive steps: (i) identification of a “seed peptide” with a desired activity, (ii) generation of variants selected from a physicochemical space around the seed peptide, (iii) synthesis and testing of this biased library, (iv) modeling of a quantitative sequence-activity relationship by an artificial neural network, and (v) de novo design by a computer-based evolutionary search in sequence space using the trained neural network as the fitness function. This strategy was successfully applied to the identification of novel peptides that fully prevent the positive chronotropic effect of anti-β1-adrenoreceptor autoantibodies from the serum of patients with dilated cardiomyopathy. The seed peptide, comprising 10 residues, was derived by epitope mapping from an extracellular loop of human β1-adrenoreceptor. A set of 90 peptides was synthesized and tested to provide training data for neural network development. De novo design revealed peptides with desired activities that do not match the seed peptide sequence. These results demonstrate that computer-based evolutionary searches can generate novel peptides with substantial biological activity.

The neural development and organization of letter recognition: Evidence from functional neuroimaging, computational modeling, and behavioral studies

Although much of the brain’s functional organization is genetically predetermined, it appears that some noninnate functions can come to depend on dedicated and segregated neural tissue. In this paper, we describe a series of experiments that have investigated the neural development and organization of one such noninnate function: letter recognition. Functional neuroimaging demonstrates that letter and digit recognition depend on different neural substrates in some literate adults. How could the processing of two stimulus categories that are distinguished solely by cultural conventions become segregated in the brain? One possibility is that correlation-based learning in the brain leads to a spatial organization in cortex that reflects the temporal and spatial clustering of letters with letters in the environment. Simulations confirm that environmental co-occurrence does indeed lead to spatial localization in a neural network that uses correlation-based learning. Furthermore, behavioral studies confirm one critical prediction of this co-occurrence hypothesis, namely, that subjects exposed to a visual environment in which letters and digits occur together rather than separately (postal workers who process letters and digits together in Canadian postal codes) do indeed show less behavioral evidence for segregated letter and digit processing.

Self-organized phase transitions in neural networks as a neural mechanism of information processing.

Transitions between dynamically stable activity patterns imposed on an associative neural network are shown to be induced by self-organized infinitesimal changes in synaptic connection strength and to be a kind of phase transition. A key event for the neural process of information processing in a population coding scheme is transition between the activity patterns encoding usual entities. We propose that the infinitesimal and short-term synaptic changes based on the Hebbian learning rule are the driving force for the transition. The phase transition between the following two dynamical stable states is studied in detail, the state where the firing pattern is changed temporally so as to itinerate among several patterns and the state where the firing pattern is fixed to one of several patterns. The phase transition from the pattern itinerant state to a pattern fixed state may be induced by the Hebbian learning process under a weak input relevant to the fixed pattern. The reverse transition may be induced by the Hebbian unlearning process without input. The former transition is considered as recognition of the input stimulus, while the latter is considered as clearing of the used input data to get ready for new input. To ensure that information processing based on the phase transition can be made by the infinitesimal and short-term synaptic changes, it is absolutely necessary that the network always stays near the critical state corresponding to the phase transition point.

Aplicação de uma rede neural artificial para a avaliação da rugosidade e soprosidade vocal

A avaliação perceptivo-auditiva tem papel fundamental no estudo e na avaliação da voz, no entanto, por ser subjetiva está sujeita a imprecisões e variações. Por outro lado, a análise acústica permite a reprodutibilidade de resultados, porém precisa ser aprimorada, pois não analisa com precisão vozes com disfonias mais intensas e com ondas caóticas. Assim, elaborar medidas que proporcionem conhecimentos confiáveis em relação à função vocal resulta de uma necessidade antiga dentro desta linha de pesquisa e atuação clínica. Neste contexto, o uso da inteligência artificial, como as redes neurais artificiais, indica ser uma abordagem promissora. Objetivo: Validar um sistema automático utilizando redes neurais artificiais para a avaliação de vozes rugosas e soprosas. Materiais e métodos: Foram selecionadas 150 vozes, desde neutras até com presença em grau intenso de rugosidade e/ou soprosidade, do banco de dados da Clínica de Fonoaudiologia da Faculdade de Odontologia de Bauru (FOB/USP). Dessas vozes, 23 foram excluídas por não responderem aos critérios de inclusão na amostra, assim utilizaram-se 123 vozes. Procedimentos: avaliação perceptivo-auditiva pela escala visual analógica de 100 mm e pela escala numérica de quatro pontos; extração de características do sinal de voz por meio da Transformada Wavelet Packet e dos parâmetros acústicos: jitter, shimmer, amplitude da derivada e amplitude do pitch; e validação do classificador por meio da parametrização, treino, teste e avaliação das redes neurais artificiais. Resultados: Na avaliação perceptivo-auditiva encontrou-se, por meio do teste Coeficiente de Correlação Intraclasse (CCI), concordâncias inter e intrajuiz excelentes, com p = 0,85 na concordância interjuízes e p variando de 0,87 a 0,93 nas concordâncias intrajuiz. Em relação ao desempenho da rede neural artificial, na discriminação da soprosidade e da rugosidade e dos seus respectivos graus, encontrou-se o melhor desempenho para a soprosidade no subconjunto composto pelo jitter, amplitude do pitch e frequência fundamental, no qual obteve-se taxa de acerto de 74%, concordância excelente com a avaliação perceptivo-auditiva da escala visual analógica (0,80 no CCI) e erro médio de 9 mm. Para a rugosidade, o melhor subconjunto foi composto pela Transformada Wavelet Packet com 1 nível de decomposição, jitter, shimmer, amplitude do pitch e frequência fundamental, no qual obteve-se 73% de acerto, concordância excelente (0,84 no CCI), e erro médio de 10 mm. Conclusão: O uso da inteligência artificial baseado em redes neurais artificiais na identificação, e graduação da rugosidade e da soprosidade, apresentou confiabilidade excelente (CCI > 0,80), com resultados semelhantes a concordância interjuízes. Dessa forma, a rede neural artificial revela-se como uma metodologia promissora de avaliação vocal, tendo sua maior vantagem a objetividade na avaliação.

Desenvolvimento artificial autônomo de um grafo sensório-motor auto-organizável.

A teoria de Jean Piaget sobre o desenvolvimento da inteligência tem sido utilizada na área de inteligência computacional como inspiração para a proposição de modelos de agentes cognitivos. Embora os modelos propostos implementem aspectos básicos importantes da teoria de Piaget, como a estrutura do esquema cognitivo, não consideram o problema da fundamentação simbólica e, portanto, não se preocupam com os aspectos da teoria que levam à aquisição autônoma da semântica básica para a organização cognitiva do mundo externo, como é o caso da aquisição da noção de objeto. Neste trabalho apresentamos um modelo computacional de esquema cognitivo inspirado na teoria de Piaget sobre a inteligência sensório-motora que se desenvolve autonomamente construindo mecanismos por meio de princípios computacionais pautados pelo problema da fundamentação simbólica. O modelo de esquema proposto tem como base a classificação de situações sensório-motoras utilizadas para a percepção, captação e armazenamento das relações causais determiníscas de menor granularidade. Estas causalidades são então expandidas espaço-temporalmente por estruturas mais complexas que se utilizam das anteriores e que também são projetadas de forma a possibilitar que outras estruturas computacionais autônomas mais complexas se utilizem delas. O modelo proposto é implementado por uma rede neural artificial feed-forward cujos elementos da camada de saída se auto-organizam para gerar um grafo sensóriomotor objetivado. Alguns mecanismos computacionais já existentes na área de inteligência computacional foram modificados para se enquadrarem aos paradigmas de semântica nula e do desenvolvimento mental autônomo, tomados como base para lidar com o problema da fundamentação simbólica. O grafo sensório-motor auto-organizável que implementa um modelo de esquema inspirado na teoria de Piaget proposto neste trabalho, conjuntamente com os princípios computacionais utilizados para sua concepção caminha na direção da busca pelo desenvolvimento cognitivo artificial autônomo da noção de objeto.

Representation of 2D objects with a topology preserving network

Comunicación presentada en el 2nd International Workshop on Pattern Recognition in Information Systems, Alicante, April, 2002.

Real time motion estimation using a neural architecture implemented on GPUs

This work describes a neural network based architecture that represents and estimates object motion in videos. This architecture addresses multiple computer vision tasks such as image segmentation, object representation or characterization, motion analysis and tracking. The use of a neural network architecture allows for the simultaneous estimation of global and local motion and the representation of deformable objects. This architecture also avoids the problem of finding corresponding features while tracking moving objects. Due to the parallel nature of neural networks, the architecture has been implemented on GPUs that allows the system to meet a set of requirements such as: time constraints management, robustness, high processing speed and re-configurability. Experiments are presented that demonstrate the validity of our architecture to solve problems of mobile agents tracking and motion analysis.

3D model reconstruction using neural gas accelerated on GPU

In this work, we propose the use of the neural gas (NG), a neural network that uses an unsupervised Competitive Hebbian Learning (CHL) rule, to develop a reverse engineering process. This is a simple and accurate method to reconstruct objects from point clouds obtained from multiple overlapping views using low-cost sensors. In contrast to other methods that may need several stages that include downsampling, noise filtering and many other tasks, the NG automatically obtains the 3D model of the scanned objects. To demonstrate the validity of our proposal we tested our method with several models and performed a study of the neural network parameterization computing the quality of representation and also comparing results with other neural methods like growing neural gas and Kohonen maps or classical methods like Voxel Grid. We also reconstructed models acquired by low cost sensors that can be used in virtual and augmented reality environments for redesign or manipulation purposes. Since the NG algorithm has a strong computational cost we propose its acceleration. We have redesigned and implemented the NG learning algorithm to fit it onto Graphics Processing Units using CUDA. A speed-up of 180× faster is obtained compared to the sequential CPU version.

Morphological classification of microtidal sand and gravel beaches

A new classification of microtidal sand and gravel beaches with very different morphologies is presented below. In 557 studied transects, 14 variables were used. Among the variables to be emphasized is the depth of the Posidonia oceanica. The classification was performed for 9 types of beaches: Type 1: Sand and gravel beaches, Type 2: Sand and gravel separated beaches, Type 3: Gravel and sand beaches, Type 4: Gravel and sand separated beaches, Type 5: Pure gravel beaches, Type 6: Open sand beaches, Type 7: Supported sand beaches, Type 8: Bisupported sand beaches and Type 9: Enclosed beaches. For the classification, several tools were used: discriminant analysis, neural networks and Support Vector Machines (SVM), the results were then compared. As there is no theory for deciding which is the most convenient neural network architecture to deal with a particular data set, an experimental study was performed with different numbers of neuron in the hidden layer. Finally, an architecture with 30 neurons was chosen. Different kernels were employed for SVM (Linear, Polynomial, Radial basis function and Sigmoid). The results obtained for the discriminant analysis were not as good as those obtained for the other two methods (ANN and SVM) which showed similar success.

Designing Regularizers and Architectures for Recurrent Neural Networks

Cette thèse contribue a la recherche vers l'intelligence artificielle en utilisant des méthodes connexionnistes. Les réseaux de neurones récurrents sont un ensemble de modèles séquentiels de plus en plus populaires capable en principe d'apprendre des algorithmes arbitraires. Ces modèles effectuent un apprentissage en profondeur, un type d'apprentissage machine. Sa généralité et son succès empirique en font un sujet intéressant pour la recherche et un outil prometteur pour la création de l'intelligence artificielle plus générale. Le premier chapitre de cette thèse donne un bref aperçu des sujets de fonds: l'intelligence artificielle, l'apprentissage machine, l'apprentissage en profondeur et les réseaux de neurones récurrents. Les trois chapitres suivants couvrent ces sujets de manière de plus en plus spécifiques. Enfin, nous présentons quelques contributions apportées aux réseaux de neurones récurrents. Le chapitre \ref{arxiv1} présente nos travaux de régularisation des réseaux de neurones récurrents. La régularisation vise à améliorer la capacité de généralisation du modèle, et joue un role clé dans la performance de plusieurs applications des réseaux de neurones récurrents, en particulier en reconnaissance vocale. Notre approche donne l'état de l'art sur TIMIT, un benchmark standard pour cette tâche. Le chapitre \ref{cpgp} présente une seconde ligne de travail, toujours en cours, qui explore une nouvelle architecture pour les réseaux de neurones récurrents. Les réseaux de neurones récurrents maintiennent un état caché qui représente leurs observations antérieures. L'idée de ce travail est de coder certaines dynamiques abstraites dans l'état caché, donnant au réseau une manière naturelle d'encoder des tendances cohérentes de l'état de son environnement. Notre travail est fondé sur un modèle existant; nous décrivons ce travail et nos contributions avec notamment une expérience préliminaire.