Data-based traffic profile generation tool for electric vehicle charging stations

Abstract

This paper presents a tool to generate realistic traffic profiles in Electric Vehicle (EV) charging stations. The tool emulates non-deterministic traffic cases based on data from similar applications. This obtained data does not often follow a normal distribution function, so the tool uses the Kernel Density Estimation (KDE) data-based technique to obtain the probability functions for the arrival and departure of the vehicles along with their missing energy at arrival. Scenarios without traffic data availability but fixed schedules like in private companies are also considered. For these cases the user may define expected schedules and shift types to generate possible traffic cases based on normal distributions around the rush hours. Based on the probability distribution analysis performed, the user obtains information of individual cases of vehicles using the charging station, which follows the trend of a real scenario.