Olga Ramos

Prediction of Solar Radiation through the Anfis Algorithm

The constant evolution of the energy industry, has introduced the need for ongoing research studies about climate change due to its direct action on the production of alternative energies. Thus, they have focused on developing predictive algorithms in order to resolve, in an early way, the climate action on each point of energy production. In the development of this work, the ANFIS algorithm and information from the NASA Langley research center virtual database were implemented. They being oriented to the analysis and prediction of solar radiation over the geographic area of the Nueva Granada Military University campus, Cajicá, Colombia, with the purpose of making appropriate use of the power generating system located in the zone. The development of such systems, would allow the early identification of solar radiation that can be present in different geographical areas of Colombia, in order to provide the necessary power to cover the electricity demand required in each region, achieving as results an approximation error less than 1%.

FragariaVesca Dehydration by flat solar collector with natural flow in Bogota

Actually optimization concepts in agricultural process are try to involve with major presence over this economical sector worldwide. This, due to use in an efficient way the resources since the cultivation to post-harvest. The last one, provide large amounts of food which are lost due to accelerated maturity and overproduction. In order to minimize the lost and supply the growing demand of natural and health products, process with a low nutritional impact are implemented to preserve the food on the time, solar dehydration is a trend which involve natural resources as primary source in the process related. In this work is presented the results obtained in FragariaVesca dehydration, known as strawberry by a flat solar collector working under natural circulation in Bogota.

MODEL AND SIMULATION OF A HOTAIR DEHYDRATION SYSTEM OF FRUITS

Abstract The most used practice for food preservation is the dehydration, in particular, the heat-air dehydration which is theoldest method used nowadays. In this work, a mathematical model was developed to describe this behavior, always working in the linear region of the process of dehydration. The model proposed was implemented in MATLAB software with the Simulink tool and evaluated with a simulation CFD in SolidWorks. The behavior of the analytical model compared with the CFD simulation, showed a high variation for the values of final moisture in the food. This is due to the initial restrictions made for the model, especially, assume that the temperature inside the chamber is homogeneous. The model allows to vary all the features of the system and identify the behavior during the process.

Reconocimiento de patrones de habla usando MFCC y RNA

En este trabajo se presentan los resultados del diseno y desarrollo de un algoritmo basado en inteligencia artificial para el reconocimiento de patrones de vocablos del idioma espanol, utilizando Coeficientes Cepstrales en las Frecuencias de Mel o (MFCC), para representar el habla a traves de la percepcion auditiva del ser humano. La utilizacion de MFCC permitio caracterizar las senales de voz teniendo en cuenta el posible ruido presente en el ambiente de grabacion, lo cual ayudo a la obtencion de patrones comunes entre estas senales cuando presentan alteraciones. Como resultado se obtuvo un reconocimiento superior al 95% de las tres vocales escogidas, en este caso la /a/,/e/,/o/, entre un grupo de 22 muestras por vocal para el entrenamiento y 11 muestras para la validacion. Las muestras fueron obtenidas de 11 personas, todas del genero masculino.

Augmented Reality for Assistance of Total Knee Replacement

The aim of this work was the development of a surgical assistance system based on augmented reality to support joint replacement procedures and implantation of prosthetic knee. Images of the scene were captured in order to detect the visual markers located on the lateral surface of the patient’s leg for overlapping the 3D models of the prosthesis and the joint, as well as the tool used by the medical specialist. With the marker identification, it was possible to compute its position and orientation for locating the virtual models, obtaining a monitoring system for giving accurate information about the procedure. Also it can be used as training platform for surgeons, without having volunteers or patients for performing real surgeries; instead they can train in a virtual environment. The results have shown an efficient system in terms of cost-benefit relation, taking into account the materials used for developing the system; nevertheless, the accuracy of the algorithm decreases according to the distance between the markers.

Reconocimiento de Patrones Vocálicos mediante la implementación de una red Neuronal Artificial Utilizando Sistemas Embebidos

In this work the results of the design and development of a neural network type feed-forward in an embedded system, to identify patterns associated with the pronunciation of vowels of Spanish language are presented. For the training of the network Mel frequencies cepstrals coefficients extracted from audio signals related to the pronunciation of the open vowels (/ a /, /e/ and /o/) were used. The capture of new samples was performed by using a computer and an embedded system to evaluate, test and compare the performance of the neural network. A percentage of correct identification above 98% was obtained indicating that the error in the class separation was less than 2% for patterns obtained from any of the three vowels. Based on the results it is concluded that implementation of artificial intelligence algorithms for classification tasks in embedded systems is feasible, and presents similar results to those that would have the same system working with the resources of a computer.