Modelagem de um motor de indução trifásico operando com tensões desequilibradas por meio de redes neurais artificiais

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Determinação da freqüência de ressonância de antenas tipo microfita triangular e retangular utilizando redes neurais artificiais

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

SIRCo: uma arquitetura para Sensoriamento Inteligente de Rádios Cognitivos

The increased demand for using the Industrial, Scientific and Medical (ISM) unlicensed frequency spectrum has caused interference problems and lack of resource availability for wireless networks. Cognitive radio (CR) have emerged as an alternative to reduce interference and intelligently use the spectrum. Several protocols were proposed aiming to mitigate these problems, but most have not been implemented in real devices. This work presents an architecture for Intelligent Sensing for Cognitive Radios (ISCRa), and a spectrum decision model (SDM) based on Artificial Neural Networks (ANN), which uses as input a database with local spectrum behavior and a database with primary users information. For comparison, a spectrum decision model based on AHP, which employs advanced techniques in its spectrum decision method was implemented. Another spectrum decision model that considers only a physical parameter for channel classification was also implemented. Spectrum decision models evaluated, as well as ISCRa's architecture were developed in GNU-Radio framework and implemented on real nodes. Evaluation of SDMs considered metrics of: delivery rate, latency (Round Trip Time - RTT) and handoff. Experiments on real nodes showed that ISCRa architecture with ANN based SDM increased packet delivery rate and presented fewer frequency variation (handoff) while maintaining latency. Considering higher bandwidth as application's Quality of Service requirement, ANN-SDM obtained the best results when compared to other SDM for cognitive radio networks (CRN).

Análisis del uso de las redes neuronales artificiales en el diseño de filtros antispam : una propuesta basada en arquitecturas neuronales no supervisadas

[ES]El spam, o correo no deseado enviado masivamente, es una amenaza que afecta al correo electrónico y otros medios de comunicación telemática. Su alto volumen de circulación genera pérdidas temporales y económicas considerables. Se presenta una solución a este problema: un sistema inteligente híbrido de filtrado antispam, basado en redes neuronales artificiales (RNA) no supervisadas. Consta de una etapa de preprocesado y de otra de procesado, basadas en distintos modelos de computación: programada (con 2 fases: manual y computacional) y neuronal (mediante mapas autoorganizados de Kohonen, SOM), respectivamente. Este sistema ha sido optimizado usando, como cuerpo de datos, ham de “Enron Email” y spam de dos fuentes diferentes. Se analiza la calidad y el rendimiento del mismo mediante diferentes métricas.

Global macrozoobenthos production and energy budget data base V150731, with link to database

I developed a new model for estimating annual production-to-biomass ratio P/B and production P of macrobenthic populations in marine and freshwater habitats. Self-learning artificial neural networks (ANN) were used to model the relationships between P/B and twenty easy-to-measure abiotic and biotic parameters in 1252 data sets of population production. Based on log-transformed data, the final predictive model estimates log(P/B) with reasonable accuracy and precision (r2 = 0.801; residual mean square RMS = 0.083). Body mass and water temperature contributed most to the explanatory power of the model. However, as with all least squares models using nonlinearly transformed data, back-transformation to natural scale introduces a bias in the model predictions, i.e., an underestimation of P/B (and P). When estimating production of assemblages of populations by adding up population estimates, accuracy decreases but precision increases with the number of populations in the assemblage.

Dysfunctional profile for patients in physical neurorehabilitation of upper limb

This paper proposes a first approach to Objective Motor Assessment (OMA) methodology. Also, it introduces the Dysfunctional profile (DP) concept. DP consists of a data matrix characterizing the Upper Limb (UL) physical alterations of a patient with Acquired Brain Injury (ABI) during the rehabilitation process. This research is based on the comparison methology of UL movement between subjects with ABI and healthy subjects as part of OMA. The purpose of this comparison is to classify subjects according to their motor control and subsequently issue a functional assessment of the movement. For this purpose Artificial Neural Networks (ANN) have been used to classify patients. Different network structures are tested. The obtained classification accuracy was 95.65%. This result allows the use of ANNs as a viable option for dysfunctional assessment. This work can be considered a pilot study for further research to corroborate these results.

1RM prediction: a data mining case study

In the last few years there has been a heightened interest in data treatment and analysis with the aim of discovering hidden knowledge and eliciting relationships and patterns within this data. Data mining techniques (also known as Knowledge Discovery in Databases) have been applied over a wide range of fields such as marketing, investment, fraud detection, manufacturing, telecommunications and health. In this study, well-known data mining techniques such as artificial neural networks (ANN), genetic programming (GP), forward selection linear regression (LR) and k-means clustering techniques, are proposed to the health and sports community in order to aid with resistance training prescription. Appropriate resistance training prescription is effective for developing fitness, health and for enhancing general quality of life. Resistance exercise intensity is commonly prescribed as a percent of the one repetition maximum. 1RM, dynamic muscular strength, one repetition maximum or one execution maximum, is operationally defined as the heaviest load that can be moved over a specific range of motion, one time and with correct performance. The safety of the 1RM assessment has been questioned as such an enormous effort may lead to muscular injury. Prediction equations could help to tackle the problem of predicting the 1RM from submaximal loads, in order to avoid or at least, reduce the associated risks. We built different models from data on 30 men who performed up to 5 sets to exhaustion at different percentages of the 1RM in the bench press action, until reaching their actual 1RM. Also, a comparison of different existing prediction equations is carried out. The LR model seems to outperform the ANN and GP models for the 1RM prediction in the range between 1 and 10 repetitions. At 75% of the 1RM some subjects (n = 5) could perform 13 repetitions with proper technique in the bench press action, whilst other subjects (n = 20) performed statistically significant (p < 0:05) more repetitions at 70% than at 75% of their actual 1RM in the bench press action. Rate of perceived exertion (RPE) seems not to be a good predictor for 1RM when all the sets are performed until exhaustion, as no significant differences (p < 0:05) were found in the RPE at 75%, 80% and 90% of the 1RM. Also, years of experience and weekly hours of strength training are better correlated to 1RM (p < 0:05) than body weight. O'Connor et al. 1RM prediction equation seems to arise from the data gathered and seems to be the most accurate 1RM prediction equation from those proposed in literature and used in this study. Epley's 1RM prediction equation is reproduced by means of data simulation from 1RM literature equations. Finally, future lines of research are proposed related to the problem of the 1RM prediction by means of genetic algorithms, neural networks and clustering techniques. RESUMEN En los últimos años ha habido un creciente interés en el tratamiento y análisis de datos con el propósito de descubrir relaciones, patrones y conocimiento oculto en los mismos. Las técnicas de data mining (también llamadas de \Descubrimiento de conocimiento en bases de datos\) se han aplicado consistentemente a lo gran de un gran espectro de áreas como el marketing, inversiones, detección de fraude, producción industrial, telecomunicaciones y salud. En este estudio, técnicas bien conocidas de data mining como las redes neuronales artificiales (ANN), programación genética (GP), regresión lineal con selección hacia adelante (LR) y la técnica de clustering k-means, se proponen a la comunidad del deporte y la salud con el objetivo de ayudar con la prescripción del entrenamiento de fuerza. Una apropiada prescripción de entrenamiento de fuerza es efectiva no solo para mejorar el estado de forma general, sino para mejorar la salud e incrementar la calidad de vida. La intensidad en un ejercicio de fuerza se prescribe generalmente como un porcentaje de la repetición máxima. 1RM, fuerza muscular dinámica, una repetición máxima o una ejecución máxima, se define operacionalmente como la carga máxima que puede ser movida en un rango de movimiento específico, una vez y con una técnica correcta. La seguridad de las pruebas de 1RM ha sido cuestionada debido a que el gran esfuerzo requerido para llevarlas a cabo puede derivar en serias lesiones musculares. Las ecuaciones predictivas pueden ayudar a atajar el problema de la predicción de la 1RM con cargas sub-máximas y son empleadas con el propósito de eliminar o al menos, reducir los riesgos asociados. En este estudio, se construyeron distintos modelos a partir de los datos recogidos de 30 hombres que realizaron hasta 5 series al fallo en el ejercicio press de banca a distintos porcentajes de la 1RM, hasta llegar a su 1RM real. También se muestra una comparación de algunas de las distintas ecuaciones de predicción propuestas con anterioridad. El modelo LR parece superar a los modelos ANN y GP para la predicción de la 1RM entre 1 y 10 repeticiones. Al 75% de la 1RM algunos sujetos (n = 5) pudieron realizar 13 repeticiones con una técnica apropiada en el ejercicio press de banca, mientras que otros (n = 20) realizaron significativamente (p < 0:05) más repeticiones al 70% que al 75% de su 1RM en el press de banca. El ínndice de esfuerzo percibido (RPE) parece no ser un buen predictor del 1RM cuando todas las series se realizan al fallo, puesto que no existen diferencias signifiativas (p < 0:05) en el RPE al 75%, 80% y el 90% de la 1RM. Además, los años de experiencia y las horas semanales dedicadas al entrenamiento de fuerza están más correlacionadas con la 1RM (p < 0:05) que el peso corporal. La ecuación de O'Connor et al. parece surgir de los datos recogidos y parece ser la ecuación de predicción de 1RM más precisa de aquellas propuestas en la literatura y empleadas en este estudio. La ecuación de predicción de la 1RM de Epley es reproducida mediante simulación de datos a partir de algunas ecuaciones de predicción de la 1RM propuestas con anterioridad. Finalmente, se proponen futuras líneas de investigación relacionadas con el problema de la predicción de la 1RM mediante algoritmos genéticos, redes neuronales y técnicas de clustering.

Predicción espacio-temporal de la irradiancia solar global a corto plazo en España mediante geoestadística y redes neuronales artificiales

El enriquecimiento del conocimiento sobre la Irradiancia Solar (IS) a nivel de superficie terrestre, así como su predicción, cobran gran interés para las Energías Renovables (ER) - Energía Solar (ES)-, y para distintas aplicaciones industriales o ecológicas. En el ámbito de las ER, el uso óptimo de la ES implica contar con datos de la IS en superficie que ayuden tanto, en la selección de emplazamientos para instalaciones de ES, como en su etapa de diseño (dimensionar la producción) y, finalmente, en su explotación. En este último caso, la observación y la predicción es útil para el mercado energético, la planificación y gestión de la energía (generadoras y operadoras del sistema eléctrico), especialmente en los nuevos contextos de las redes inteligentes de transporte. A pesar de la importancia estratégica de contar con datos de la IS, especialmente los observados por sensores de IS en superficie (los que mejor captan esta variable), estos no siempre están disponibles para los lugares de interés ni con la resolución espacial y temporal deseada. Esta limitación se une a la necesidad de disponer de predicciones a corto plazo de la IS que ayuden a la planificación y gestión de la energía. Se ha indagado y caracterizado las Redes de Estaciones Meteorológicas (REM) existentes en España que publican en internet sus observaciones, focalizando en la IS. Se han identificado 24 REM (16 gubernamentales y 8 redes voluntarios) que aglutinan 3492 estaciones, convirtiéndose éstas en las fuentes de datos meteorológicos utilizados en la tesis. Se han investigado cinco técnicas de estimación espacial de la IS en intervalos de 15 minutos para el territorio peninsular (3 técnicas geoestadísticas, una determinística y el método HelioSat2 basado en imágenes satelitales) con distintas configuraciones espaciales. Cuando el área de estudio tiene una adecuada densidad de observaciones, el mejor método identificado para estimar la IS es el Kriging con Regresión usando variables auxiliares -una de ellas la IS estimada a partir de imágenes satelitales-. De este modo es posible estimar espacialmente la IS más allá de los 25 km identificados en la bibliografía. En caso contrario, se corrobora la idoneidad de utilizar estimaciones a partir de sensores remotos cuando la densidad de observaciones no es adecuada. Se ha experimentado con el modelado de Redes Neuronales Artificiales (RNA) para la predicción a corto plazo de la IS utilizando observaciones próximas (componentes espaciales) en sus entradas y, los resultados son prometedores. Así los niveles de errores disminuyen bajo las siguientes condiciones: (1) cuando el horizonte temporal de predicción es inferior o igual a 3 horas, las estaciones vecinas que se incluyen en el modelo deben encentrarse a una distancia máxima aproximada de 55 km. Esto permite concluir que las RNA son capaces de aprender cómo afectan las condiciones meteorológicas vecinas a la predicción de la IS. ABSTRACT ABSTRACT The enrichment of knowledge about the Solar Irradiance (SI) at Earth's surface and its prediction, have a high interest for Renewable Energy (RE) - Solar Energy (SE) - and for various industrial and environmental applications. In the field of the RE, the optimal use of the SE involves having SI surface to help in the selection of sites for facilities ES, in the design stage (sizing energy production), and finally on their production. In the latter case, the observation and prediction is useful for the market, planning and management of the energy (generators and electrical system operators), especially in new contexts of smart transport networks (smartgrid). Despite the strategic importance of SI data, especially those observed by sensors of SI at surface (the ones that best measure this environmental variable), these are not always available to the sights and the spatial and temporal resolution desired. This limitation is bound to the need for short-term predictions of the SI to help planning and energy management. It has been investigated and characterized existing Networks of Weather Stations (NWS) in Spain that share its observations online, focusing on SI. 24 NWS have been identified (16 government and 8 volunteer networks) that implies 3492 stations, turning it into the sources of meteorological data used in the thesis. We have investigated five technical of spatial estimation of SI in 15 minutes to the mainland (3 geostatistical techniques and HelioSat2 a deterministic method based on satellite images) with different spatial configurations. When the study area has an adequate density of observations we identified the best method to estimate the SI is the regression kriging with auxiliary variables (one of them is the SI estimated from satellite images. Thus it is possible to spatially estimate the SI beyond the 25 km identified in the literature. Otherwise, when the density of observations is inadequate the appropriateness is using the estimates values from remote sensing. It has been experimented with Artificial Neural Networks (ANN) modeling for predicting the short-term future of the SI using observations from neighbor’s weather stations (spatial components) in their inputs, and the results are promising. The error levels decrease under the following conditions: (1) when the prediction horizon is less or equal than 3 hours the best models are the ones that include data from the neighboring stations (at a maximum distance of 55 km). It is concluded that the ANN is able to learn how weather conditions affect neighboring prediction of IS at such Spatio-temporal horizons.

Multidimensional classification and diagnosis of leukemia in flow cytometry

This research is to establish new optimization methods for pattern recognition and classification of different white blood cells in actual patient data to enhance the process of diagnosis. Beckman-Coulter Corporation supplied flow cytometry data of numerous patients that are used as training sets to exploit the different physiological characteristics of the different samples provided. The methods of Support Vector Machines (SVM) and Artificial Neural Networks (ANN) were used as promising pattern classification techniques to identify different white blood cell samples and provide information to medical doctors in the form of diagnostic references for the specific disease states, leukemia. The obtained results prove that when a neural network classifier is well configured and trained with cross-validation, it can perform better than support vector classifiers alone for this type of data. Furthermore, a new unsupervised learning algorithm---Density based Adaptive Window Clustering algorithm (DAWC) was designed to process large volumes of data for finding location of high data cluster in real-time. It reduces the computational load to ∼O(N) number of computations, and thus making the algorithm more attractive and faster than current hierarchical algorithms.

Artificial neural network nonlinear equalizer for coherent optical OFDM

We propose a novel low-complexity artificial neural network (ANN)-based nonlinear equalizer (NLE) for coherent optical orthogonal frequency-division multiplexing (CO-OFDM) and compare it with the recent inverse Volterra-series transfer function (IVSTF)-based NLE over up to 1000 km of uncompensated links. Demonstration of ANN-NLE at 80-Gb/s CO-OFDM using 16-quadrature amplitude modulation reveals a Q-factor improvement after 1000-km transmission of 3 and 1 dB with respect to the linear equalization and IVSTF-NLE, respectively.

Artificial neural network application in short-term prediction in an oscillating water column

Oscillating Water Column (OWC) is one type of promising wave energy devices due to its obvious advantage over many other wave energy converters: no moving component in sea water. Two types of OWCs (bottom-fixed and floating) have been widely investigated, and the bottom-fixed OWCs have been very successful in several practical applications. Recently, the proposal of massive wave energy production and the availability of wave energy have pushed OWC applications from near-shore to deeper water regions where floating OWCs are a better choice. For an OWC under sea waves, the air flow driving air turbine to generate electricity is a random process. In such a working condition, single design/operation point is nonexistent. To improve energy extraction, and to optimise the performance of the device, a system capable of controlling the air turbine rotation speed is desirable. To achieve that, this paper presents a short-term prediction of the random, process by an artificial neural network (ANN), which can provide near-future information for the control system. In this research, ANN is explored and tuned for a better prediction of the airflow (as well as the device motions for a wide application). It is found that, by carefully constructing ANN platform and optimizing the relevant parameters, ANN is capable of predicting the random process a few steps ahead of the real, time with a good accuracy. More importantly, the tuned ANN works for a large range of different types of random, process.

(Table 1) Age control and temporal resolution of paleoceanographic records of the North Atlantic

Short-term changes in sea surface conditions controlling the thermohaline circulation in the northern North Atlantic are expected to be especially efficient in perturbing global climate stability. Here we assess past variability of sea surface temperature (SST) in the northeast Atlantic and Norwegian Sea during Marine Isotope Stage (MIS) 2 and, in particular, during the Last Glacial Maximum (LGM). Five high-resolution SST records were established on a meridional transect (53°N-72°N) to trace centennial-scale oscillations in SST and sea-ice cover. We used three independent computational techniques (SIMMAX modern analogue technique, Artificial Neural Networks (ANN), and Revised Analog Method (RAM)) to reconstruct SST from planktonic foraminifer census counts. SIMMAX and ANN reproduced short-term SST oscillations of similar magnitude and absolute levels, while RAM, owing to a restrictive analog selection, appears less suitable for reconstructing "cold end" SST. The SIMMAX and ANN SST reconstructions support the existence of a weak paleo-Norwegian Current during Dansgaard-Oeschger (DO) interstadials number 4, 3, 2, and 1. During the LGM, two warm incursions of 7°C water to occurred in the northern North Atlantic but ended north of the Iceland Faroe Ridge. A rough numerical estimate shows that the near-surface poleward heat transfer from 53° across the Iceland-Faroe Ridge up to to 72° N dropped to less than 60% of the modern value during DO interstadials and to almost zero during DO stadials. Summer sea ice was generally confined to the area north of 70°N and only rarely expanded southward along the margins of continental ice sheets. Internal LGM variability of North Atlantic (>40°N) SST in the GLAMAP 2000 compilation (Sarnthein et al., 2003, doi:10.1029/2002PA000771; Pflaumann et al., 2003, doi:10.1029/2002PA000774) indicates maximum instability in the glacial subpolar gyre and at the Iberian Margin, while in the Nordic Seas, SST was continuously low.

Artificial Neural Network and Neutron Aplication in a Volume Fraction Calculation in Annular and Stratified Multiphase System

Multiphase flows, type oil–water-gas are very common among different industrial activities, such as chemical industries and petroleum extraction, and its measurements show some difficulties to be taken. Precisely determining the volume fraction of each one of the elements that composes a multiphase flow is very important in chemical plants and petroleum industries. This work presents a methodology able to determine volume fraction on Annular and Stratified multiphase flow system with the use of neutrons and artificial intelligence, using the principles of transmission/scattering of fast neutrons from a 241Am-Be source and measurements of point flow that are influenced by variations of volume fractions. The proposed geometries used on the mathematical model was used to obtain a data set where the thicknesses referred of each material had been changed in order to obtain volume fraction of each phase providing 119 compositions that were used in the simulation with MCNP-X –computer code based on Monte Carlo Method that simulates the radiation transport. An artificial neural network (ANN) was trained with data obtained using the MCNP-X, and used to correlate such measurements with the respective real fractions. The ANN was able to correlate the data obtained on the simulation with MCNP-X with the volume fractions of the multiphase flows (oil-water-gas), both in the pattern of annular flow as stratified, resulting in a average relative error (%) for each production set of: annular (air= 3.85; water = 4.31; oil=1.08); stratified (air=3.10, water 2.01, oil = 1.45). The method demonstrated good efficiency in the determination of each material that composes the phases, thus demonstrating the feasibility of the technique.

Aplicação de métodos de calibração multivariada para a determinação simultânea de riboflavina (VB2), tiamina (VB1), piridoxina (VB6) e nicotinamida (VPP)

In this work, the artificial neural networks (ANN) and partial least squares (PLS) regression were applied to UV spectral data for quantitative determination of thiamin hydrochloride (VB1), riboflavin phosphate (VB2), pyridoxine hydrochloride (VB6) and nicotinamide (VPP) in pharmaceutical samples. For calibration purposes, commercial samples in 0.2 mol L-1 acetate buffer (pH 4.0) were employed as standards. The concentration ranges used in the calibration step were: 0.1 - 7.5 mg L-1 for VB1, 0.1 - 3.0 mg L-1 for VB2, 0.1 - 3.0 mg L-1 for VB6 and 0.4 - 30.0 mg L-1 for VPP. From the results it is possible to verify that both methods can be successfully applied for these determinations. The similar error values were obtained by using neural network or PLS methods. The proposed methodology is simple, rapid and can be easily used in quality control laboratories.