A car-following model accounting for probability distribution

Abstract

Various stochastic factors (e.g., the driver’s individual properties) widely exist in the real traffic system, but the existing studies cannot completely describe the impacts of various stochastic factors on traffic flow (especially the driving behavior). In this paper, we introduce the driver’s three perceived errors into the car-following model, and construct a car-following model with the probability distributions of the three perceived errors to explore the effects of the three perceived errors on the driving behavior under three typical situations (i.e., uniform flow, shock and rarefaction waves, and a small perturbation). The numerical results show that the three perceived errors have significant impacts on the evolution of traffic flow (including the headway distribution), i.e., the distribution of density does not prominently change under the three traffic states. In addition, the impacts are directly related to the initial condition. The results can help drivers reasonably adjust their driving behaviors based on their current traffic state (especially when some stochastic factors exist).