Characterizing driver heterogeneity within stochastic traffic simulation

TL;DR

This paper introduces a new framework to analyze and categorize driver behaviors based on acceleration patterns, improving the realism of traffic microsimulation by capturing individual driver heterogeneity.

Contribution

It presents a novel method to identify driver fingerprints, cluster driving styles, and accurately reproduce individual behaviors in traffic simulations.

Findings

Successfully clusters drivers into meaningful categories like mild, normal, and aggressive.

Reproduces observed driver acceleration behaviors in microsimulation with high fidelity.

Highlights limitations of existing car-following models in capturing free-flow acceleration.

Abstract

Drivers' heterogeneity and the broad range of vehicle characteristics on public roads are primarily responsible for the stochasticity observed in road traffic dynamics. Understanding the behavioural differences in drivers (human or automated systems) and reproducing observed behaviours in microsimulation attracts significant attention lately. Calibration of car-following model parameters is the prevalent way to simulate different driving behaviors through randomly injected variation around average parameter values. An issue is that, as shown in the literature, most car-following model do not realistically reproduce free-flow acceleration, that is in turn highly correlated with heterogeneity in driving styles. Furthermore, often, model parameters lose their physical interpretation upon calibration. The present study proposes a novel framework to analyse observed vehicle trajectories from various drivers, identify individual driver fingerprints based on their acceleration behavior, cluster drivers in categorically-meaningful driving styles (e.g. mild, normal and aggressive) and reproduce observed individual driver acceleration behaviors in microsimulation.