GPS2GR: Formação otimizada de comboios de veículos em redes urbanas com base em trajetórias de GPS

Abstract

This work proposes a computationally feasible optimization method (offline) that optimizes the global flow of vehicles through traffic signal programming that encourages the formation of vehicle convoys traveling in a green wave by frequent routes using as input the estimated flow density by trajectories collected from GPS’s. For this, a schedule was used that generates a green wave movement in the most frequent flows of the city, not being limited to avenues or changes of streets of a set of trajectories. Frequent route discovery uses the principle of minimum traffic flow conservation per time window, time windows of heterogeneous widths allow you to obtain specific solutions for each for each moment throughout the day of a city. The street segment flow estimation is done by mapping the GPS coordinates to the street segment and counting each trajectory that passes in the street segment, in order to have greater precision in estimating traffic by street segments and by the time window. The temporal discretization is done by grouping the trajectories in the timeline, grouping the schedules that naturally concentrate many trajectories. In this work, we are using for the experiments the trajectories of a taxi company in the city of Fortaleza, but the solution can use trajectories from other sources, such as Waze. We understand that this model can be a centralized alternative of urban traffic orchestration, using the same equipment to allow a greater flow of daily urban traffic, maximizing it.