An Improved Car-Following Model Based on Internal Heterogeneity of the Driver

Abstract

Experiment and simulation are two commonly used methods to study dynamic traffic flow. In Newell (2002), it is proposed that the car-following behavior is controlled by two factors, i.e., response time and distance offset. Each driver has two fixed factors in the following behavior. However, we find that there is internal heterogeneity of drivers, that is, the response time and distance offset cannot keep constant all the time, and show the characteristics of lognormal distribution. Thus, it is necessary to consider the heterogeneity within the driver into the model. This paper establishes a stochastic Newell’s model based on the probability density distributions of response time and distance offset from experimental data. The results show that the speed standard deviation in simulation is qualitatively and quantitatively consistent with the experimental results.