Modeling Car-following Behavior along a Discrete Time Sequence and Its Applications in Mixed Traffic Flow Analysis

Abstract

This paper presents a new car-following model which focuses on car-following behavior along a discrete time sequence. We find an independent and identically distributed (IID) parameter ∆aj by performing autocorrelation tests on a field experiment driving dataset. We further find ∆aj follows a normal distribution and incorporate this parameter into Newell's car-following framework. The model is validated by successfully explaining the concave oscillation phenomenon observed in real-world vehicle platoon experiments. This paper also introduces two applications of the proposed model in mixed traffic flow analysis. We find that (i) Autonomous vehicles (AVs) can significantly stabilize their follower's car-following behavior and (ii) an increase in AV penetration rate can improve traffic safety without undermining traffic throughput.