A Two-Condition Continuous Asymmetric Car-Following Model for Adaptive Cruise Control Vehicles

Adaptive cruise control (ACC) vehicles have the potential to impact traffic flow dynamics. To better understand the impacts of ACC vehicles on traffic flow, an accurate microscopic car-following model for ACC vehicles is essential. Most of the ACC car-following models utilize a continuous function to describe vehicle acceleration and braking, e.g., the optimal velocity relative velocity (OVRV) model. However, these models do not necessarily describe car-following behavior with sufficient accuracy. Recent studies have proposed switching models to better describe realistic ACC dynamics. However, they often fail to accurately capture the driving behavior around the switching points, where a vehicle switches between acceleration and deceleration. In this study, we develop a two-condition, continuous asymmetric car-following (TCACF) model to capture ACC driving behavior in a physically interpretable manner, while preserving numerical soundness. The proposed TCACF model and multiple other car-following models are calibrated based on a real-world ACC trajectory dataset. The results show that the TCACF model better describes the asymmetric driving behavior of ACC vehicles than any of the commonly used car-following models, especially at switching points. The results indicate that the TCACF model considerably increases model accuracy by up to 32.46% when compared with other switching models and by up to 36.98% when compared to commonly used car-following models. The TCACF model is expected to offer new insights into modeling and simulating emerging ACC car-following dynamics with a higher degree of accuracy and can be used in applications where correctly simulating acceleration behavior is important.