ANÁLISIS Y RECONOCIMIENTO DE PATRONES DE CONDUCCIÓN TEMERARIA MEDIANTE ALGORITMOS DE APRENDIZAJE COMPUTACIONAL

LUGO HERNANDEZ, RAMIRO IVAN (2016) ANÁLISIS Y RECONOCIMIENTO DE PATRONES DE CONDUCCIÓN TEMERARIA MEDIANTE ALGORITMOS DE APRENDIZAJE COMPUTACIONAL. Maestría thesis, UNIVERSIDAD AUTONOMA DE CHIHUAHUA.

[img]
Vista Previa
Text
Tesis_Ramiro_Ivan_Lugo_Hernandez.pdf

Download (1881Kb) | Vista Previa

Resumen

During the period from August 2014 to September 2016, various machine learning models were implemented to address the problem of reckless driving patterns recognition. This problem has been constantly increasing over the last decade due to the large increase in the number of cars on the road, as well as the fast-paced driving style present nowadays. To deal with this problem, a data set was generated using signals obtained from the accelerometer of 5 smartphones placed in different locations and with an arbitrary position inside a car when driving. This set consists of 357 signals where each one represents a reckless driving event. A neural network was trained on these data, which yielded low classification percentages. With the analysis of these results, it is shown that due to the arbitrary position of the devices, there is no defined characterization of the signals, which does not allow a correct classification. For this a strategy is proposed combining pre-processing and classification, in order to generate an adequate characterization of the signals and to be able to classify them correctly. Also proposed is a third element of route scoring, after the classification, by the generation of a scoring function using genetic programming.

Tipo de Documento: Tesis (Maestría)
Palabras Clave: Keywords: neural network, genetic programming, machine learning, reckless driving.
Clasificación temática: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Usuario Remitente: Admin Administrador del Respositorio
Depositado: 06 Nov 2017 14:06
Ultima Modificación: 06 Nov 2017 14:06
URI: http://repositorio.uach.mx/id/eprint/129

Actions (login required)

Ver Objeto Ver Objeto

Universidad Autonoma de Chihuahua