Analysis of driving behavior on sharp curves based on latent class model

Abstract

Aiming at the problem of distinguishing the driving characteristics of different types of drivers in sharp bends, a driver behavior classification method based on vehicle driving parameters was proposed. By analyzing the real vehicle driving data, six variables were selected, namely, vehicle type, speed before entering the bend, speed after exiting the bend, deceleration behavior before entering the bend, whether to follow the car, and speed difference between entering and exiting the bend. A latent class model was constructed, and the characteristics of different types of drivers were determined based on the model. The results show that the model can classify drivers in sharp bends into three types: stable type, restricted type, and free type. The characteristic parameters of stable type drivers are in the median of the three types of drivers, and they tend to enter and exit the bend at a constant speed; restricted type drivers have the most obvious deceleration or following driving characteristics when entering the bend; and free type drivers have the highest speeds entering and exiting the bend, and the deceleration behavior after exiting the bend is the most obvious.