Prototipo de un sistema para la detección de baches en vías vehiculares específicas de la ciudad de Bogotá a través del uso de un modelo basado en machine learning y sensores en dispositivos móviles

Abstract

Bogotá is one of the busiest cities in the country, both in terms of automobiles and different types of vehicles, as well as pedestrians who use the streets of this city daily, which is why the importance of road conditions is a crucial factor, since it can incur various catastrophic phenomena due to lack of attention or correct management regarding the continuous improvement of the paving found. Given one of the latest reports by the Ministry of Transportation corresponding to December 27, 2021 (Ministerio de Transporte, 2021), it can be seen that the Institute of Roads is constantly updating what is known as the status of the Road Network, within which we find classifications as paved and unpaved roads, in this way the government seeks to take preventive measures if required and thus generate effective control for users who transit these roads in the different media Of transport. By 2021, it was found in the middle of the year that the problem of the road network in Bogotá was more serious than previously thought, coming to perceive that 15% was full of potholes, holes, gaps, craters, quagmires, and land deformations. For years, given the creation of Transmilenio, the paved areas that were destined for its use have also been affected, although not in all, but there was a noticeable deterioration over time, it is stated according to various magazines that the trunks through which this means of transport is located or passes continuously, have 77% of the pavement that can be considered as in good condition, on the other hand 22% good to dry and 1% totally deteriorated. As can be seen, this process, as it requires a large intervention by a user and manually providing data, makes it an inefficient process to report gaps in a large city such as Bogotá. For this reason, the idea of make a prototype in real time that allows the identification of holes in the city of Bogotá through the use of a machine learning model and using accelerometers in mobile devices so that this process of identifying holes, obtaining data and generating pothole reports to be made automatically and which in turn allows the visualization of these through a web application.