Ingeniería en Electrónica y Telecomunicaciones-Pregrado

Permanent URI for this collectionhttps://dspace-test.ucuenca.edu.ec/handle/123456789/43

Browse

Browsing Ingeniería en Electrónica y Telecomunicaciones-Pregrado by Author "Arévalo Villacrés, Josué Marcelo"

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Sistema de detección de la intención de dos movimientos de la mano a través del procesamiento de señales EEG
    (Universidad de Cuenca, 2020-09-18) Zea Paredes, David Andrés; Arévalo Villacrés, Josué Marcelo; Minchala Ávila, Luis Ismael; Astudillo Salinas, Darwin Fabián
    This work is part of the partial results of the project “Robotic exoskeleton for functional assistance in walking of patients with incomplete spinal cord injuries: design and application”, which is currently under development by the University of Cuenca. The motivation for the development of this project has research purposes and to support people who, due to some condition or circumstance, have suffered the loss of some of their extremities, specifically their hands. This project aims to develop a system for detecting and classifying the opening and closing movements of the hand, and an inactive state by acquiring electroencephalographic (EEG) signals through the use of the Emotiv EPOC+ device. The methodology proposed in this project consists of an acquisition stage through the Emotiv EPOC+ device. This is followed by a stage of pre-processing the EEG signal by applying an oset filter, a bandpass filter and a Common Average Reference (CAR) filter. Subsequently, a processing stage is introduced for analysis in the frequency and frequency-time domains with the application of the Fast Fourier Transform (FFT) and the Discrete Wavelet Transform (DWT) respectively. In the final stage, the most relevant characteristics in each transformed domain are extracted to guarantee a classifier performance using neural networks. At the end of this project, a data set of 18 records corresponding to 8 patients is conformed, 3 of them shown pathology. It was determined that the result of the characterization of EEG signals through DWT provides greater accuracy and information compared to processing through FFT with percentages of 82% and 77% respectively. After the classification using (Artificial Neuronal Networks (ANN)) for the classes open hand, closed hand and inactivity, it is possible to move a robot hand via serial communication between the Arduino and Python programs.