Dynamical capacity drop in a nonlinear stochastic traffic model

TL;DR

This paper demonstrates that a simple nonlinear stochastic traffic model can dynamically produce the inverse-$\lambda$ shape in the fundamental diagram, explaining traffic state coexistence and sudden transitions.

Contribution

It introduces a stochastic mesoscopic model based on gas-kinetic theory that captures the inverse-$\lambda$ shape and traffic state transitions due to noise effects.

Findings

The model reproduces the inverse-$\lambda$ shape in traffic flow data.

Stochastic noise influences the stability and transitions between traffic states.

Qualitative agreement between simulations and empirical data is achieved.

Abstract

In this work, we show that the inverse-$\lambda$ shape in the fundamental diagram of traffic flow can be produced dynamically by a simple nonlinear mesoscopic model with stochastic noises. The proposed model is based on the gas-kinetic theory of the traffic system. In our approach, the nonlinearity leads to the coexistence of different traffic states. The scattering of the data is thus attributed to the noise terms introduced in the stochastic differential equations and the transition among the various traffic states. Most importantly, the observed inverse-$\lambda$ shape and the associated sudden jump of physical quantities arise due to the effect of stochastic noises on the stability of the system. The model parameters are calibrated, and a qualitative agreement is obtained between the data and the numerical simulations.