Characterising driver heterogeneity within stochastic traffic simulation

Drivers’ heterogeneity and the broad range of vehicle characteristics on public roads are primarily responsible for stochasticity observed in road traffic dynamics. Understanding behavioural differences drivers (human or automated systems) reproducing behaviours microsimulation has lately attracted significant attention. Calibration car-following model parameters is prevalent way to chracterize different driving behaviours. However, most models do not realistically reproduce free-flow accelerations therefore, usually mainly result over-fitting with limited possibility realistic drivers’ simulation. To solve this problem, present study proposes a novel framework identify individual driver fingerprints based their acceleration them microsimulation. The paper also discusses unsuitability properly characterise alone aggressiveness driver. A large experimental campaign simulation results demonstrate robustness proposed method.