Regresar

Identificación de congestionamiento vehicular a través del análisis de agrupamiento de trayectorias GPS.

Abstract:

Currently, tra c congestion in the City of Guayaquil has been increasing due to the growth in sales in the automotive sector, generating a high demand for vehicles which circulate daily on the streets causing tra c, loss of time, environmental pollution, accidents; being necessary to carry out an analysis of GPS trajectories used in the identification of tra c congestion patterns. The DBSCAN algorithm was applied to determine the analysis of the clustering of GPS tracks; It is implemented with the use of a programming language, a module with some tra c congestion indicators based on the resulting groups. The results of the groupings and analysis of these indicators are presented graphically through maps and tables. The research carried out is nonexperimental with a quantitative approach. The processed data are from a set of GPS records of the city of Guayaquil, with which, and using the DBSCAN algorithm it was possible to carry out several experiments for the city with di erent subsets of data or cut-o periods, with these periods it was possible to test the tra c indicators of vehicular flow, in such a way that it was possible to test with di erent values for the epsilon parameters and minimum points and thus obtain the number of possible clusters. Then at the end to apply the tra c congestion indicators, a specific subset of Ave. Pedro Men´endez Gilbert was taken, from the exit of the Cerro Santa Ana Tunnel, direction S-E to N-W. The implemented algorithm allows the identification of vehicle speed levels through the analysis of GPS trajectories. Direct beneficiaries were defined as the people who transit in the city of Guayaquil and indirect beneficiaries to the transit agents.Currently, tra c congestion in the City of Guayaquil has been increasing due to the growth in sales in the automotive sector, generating a high demand for vehicles which circulate daily on the streets causing tra c, loss of time, environmental pollution, accidents; being necessary to carry out an analysis of GPS trajectories used in the identification of tra c congestion patterns. The DBSCAN algorithm was applied to determine the analysis of the clustering of GPS tracks; It is implemented with the use of a programming language, a module with some tra c congestion indicators based on the resulting groups. The results of the groupings and analysis of these indicators are presented graphically through maps and tables. The research carried out is nonexperimental with a quantitative approach. The processed data are from a set of GPS records of the city of Guayaquil, with which, and using the DBSCAN algorithm it was possible to carry out several experiments for the city with di erent subsets of data or cut-o periods, with these periods it was possible to test the tra c indicators of vehicular flow, in such a way that it was possible to test with di erent values for the epsilon parameters and minimum points and thus obtain the number of possible clusters. Then at the end to apply the tra c congestion indicators, a specific subset of Ave. Pedro Men´endez Gilbert was taken, from the exit of the Cerro Santa Ana Tunnel, direction S-E to N-W. The implemented algorithm allows the identification of vehicle speed levels through the analysis of GPS trajectories. Direct beneficiaries were defined as the people who transit in the city of Guayaquil and indirect beneficiaries to the transit agents.