Forecasting the Onset of Traffic Congestions on Circular Roads

Abstract

Study of traffic flow dynamics has a long tradition. However, predicting traffic jams before they occur is still a challenge. In this paper, we introduce recently developed tools of tipping point forecasting in complex systems, namely early warning indicators and bifurcation forecasting methods, and investigate their application to predict traffic jams on a circular road. The main advantage of the proposed methods is that they are model-free. The methods are based on exploiting the phenomenon of critical slowing down which occurs in dynamical systems near certain types of bifurcations, such as traffic jams. One can forecast the onset of traffic jams and the dynamics of the traffic after the bifurcation by using a few traffic measurements before the tipping point occurs. The measurements required for forecasting are recorded dynamical features of the system such as headways between cars in traffic, velocity or accelerations of each car. Forecasting approaches are applied to several simulated and experimental traffic flow conditions. Results show that one can successfully predict the onset of traffic jams and the traffic dynamics after this critical point using the proposed approaches while no model of the system is required.