Algoritmo para el Reconocimiento Automático de la Posición de Acelerómetros en Vehículos M1

Abstract

This work was focused on determining an algorithm to automatically recognize the location of the OBD II connector in vehicles of category M1, using a protocol for obtaining data and, as a first point, the three-dimensional spaces were characterized, with respect to the accelerations obtained using the FREEMATICS ONE + device. In addition, for the statistical study, information was obtained, by means of the arithmetic mean, having selected the average for its greater contribution in the information, by facilitating the characterization of the experiment; these values ​​generated databases, which were applied for the manufacturing and training of the artificial neural network. Also, it was analyzed with different types of workouts to use the lowest error method, in which through Trainlm training, an error of 0.38265% is reached. The verification of the operation of the Neural Network of location was carried out in 3 different vehicles with which there were obtained results in favor of the investigation.