Modeling of E-Scooter Driving Behavior Adopting Tracked Trajectories

The increasing integration of electric scooters (ESs) into urban mobility, whether via shared fleets or private use, has introduced notable safety challenges, since ESs must coexist with cars on urban roadways. The scientific literature has extensively investigated the factors influencing the adoption of ESs and the characteristics of tracked trips, but a gap remains with regard to an in-depth understanding of ES drivers’ behavior. The aim of this study was to analyze ES trajectories collected from real driving experiments and to adapt and calibrate microscopic models for the simulation of trajectories in defined scenarios, with the vision of developing a tool for safety checks and planning purposes. The experimental phase was conducted from April to July 2023; a high-precision global navigation satellite system (GNSS) characterized by a sampling frequency of 1 Hz was adopted, which was placed on ESs and other involved vehicles. The GNSS recorded positional data over time, from which speed and acceleration were derived. Moreover, in cases of overtaking, the lateral distance between the follower and the leader was computed. From these databases, two different models were calibrated: a social force model to simulate single ES movements and a Gazis–Herman–Rothery model to describe leader–follower dynamics between two ESs or between a car following an ES. The insights derived from the trajectory analyses and model calibration were implemented in a Python microsimulator.