Traffic models from a velocity point of view and implementation of traffic conditions in excess fuel consumption estimates

Abstract

The estimation of excess fuel consumption due to pavement - vehicle interactions (PVI) is receiving an increasing amount of attention, for it has major economic and ecological implications for the design and maintenance of road networks. As any dissipative mechanisms, the ones we find in the interaction of the pavement and the vehicles are highly dependent on the velocity of the agents in the system, hence the travelling speed of the vehicles. In this work, we propose a coupling of mechanistic dissipative models with evaluation of velocity profiles which shall enable a higher accuracy in the prediction of excess fuel consumption. We focus our attention on the simulation of the probability density functions associated with the velocity of vehicles on a single lane road, for agent-based and cellular automata models. While our approach neglects the influence of lane interactions on the velocity probability distribution, this simulation-based method enables us to obtain predictions of the dissipation resulting from both deflection and roughness of the pavement on any highway segment as a function of the traffic flow. It is shown that taking into account the presence of traffic jams instead of assuming all vehicles traveling at maximum velocity, as is done to obtain usual estimates, increases the deflection-induced dissipation per vehicle per traveled length and reduces the roughness-induced one. This difference is, however, shown to lose in magnitude when the temperature of the environment increases.