Sistema de alerta y notificación de variaciones en el patrón de conducción

• Autor: Franco Caldas, Kevyn Pool
• Formato: tesis de grado
• Fecha de Publicación: 2021
• Institución: Universidad de Lima
• Repositorio: ULIMA-Institucional
• Lenguaje: español
• OAI Identifier: oai:repositorio.ulima.edu.pe:20.500.12724/13855
• Enlace del recurso: https://hdl.handle.net/20.500.12724/13855
• Nivel de acceso: acceso abierto
• Materia: Mobile apps; Traffic accidents; Aplicaciones móviles; Accidentes de tránsito; https://purl.org/pe-repo/ocde/ford#2.02.04

Each year, the number of deaths from accidents in urban areas increases by 3% (INEI, 2018), due to reckless driving and inefficient safety systems to avoid car accidents, therefore, in Peru, the Need for a system that allows drivers to be audibly alerted when their driving pattern differs from usual and to avoid car accidents. The objective of this research is to develop a system that allows to generate alerts when driving that differs from the usual patterns is detected, applying the Kohonen Neural Network Model methodologies and the hidden Markov Model and implementing them in a mobile application for Android itself. . In the experimentation, two driving states were determined: "No accident" and "Previous Accident", where drivers between 20 and 30 years old, within a controlled area, showed on average that 86% of the time they accelerated when they were on the road. “Previous Accident” status, but it is also pointed out that the average probability of changing from “Previous Accident” to “No accident” status was 91%, which indicates a driving trend within the usual patterns. The results obtained for the generation of driving patterns allow the driver to be alerted when his driving style differs from the usual patterns and to avoid car accidents. This finding contributes to improving our understanding of the links between variability in driving patterns and the incidence of motor vehicle accidents.