Variations in Parameters of Time Headway Models According to Macroscopic Variables, Fundamental Diagram, and Exogenous Effects

Abstract

Des relations entre les variables macroscopiques du trafic et les temps inter-véhiculaires sont ici mises en évidence. Plusieurs auteurs ont étudié la distribution du temps inter-véhiculaire en fonction du volume du trafic, et ont pour cela utilisé différentes lois de probabilité ; le modèle Gamma-GQM qui mélange la loi gamma et le modèle de file d'attente généralisé (Generalized Queuing Model), permet de très bien modéliser, dans la plupart des cas, la distribution du temps inter-véhiculaire. L'aspect présenté ici de l'étude du temps inter-véhiculaire en fonction du trafic consiste à étudier les variations des paramètres de cette loi en fonction du trafic. A partir des données microscopiques du trafic recueillies en 2005 sur la RN118, on a pu établir les distributions de temps inter-véhiculaire dans différentes conditions d'analyse - une condition d'analyse croisant un régime particulier identifié sur le diagramme fondamental, avec le caractère (pluie/non pluie), et avec le rang de la voie (voie rapide/voie lente). Les paramètres d'échelle des lois de probabilité du temps inter-véhiculaire varient en fonction de ces conditions d'analyse - le sens de ces variations étant déterminé par le caractère (contraignant ou non) de la condition d'analyse

Parameters of Time Headway Distribution as Performance Indicators of Motorway Traffic and Driver Behavior

“…where f, TH U, and TH V are the probability density functions of overall headway, following headway, and free headway, respectively, and t, θ, u, d, and µ are parameters of the functions. Gamma-GQM is a GQM that uses gamma distribution as the g function and can be represented as follows (5,6,18):…”

“…Time headway studies are carried out for a number of important reasons. This may include analysis of traffic safety, computation of highway capacity, study of drivers' car-following behavior, setting optimal variable speed limits, optimization of intersection signals, and generation of vehicles in microscopic traffic simulation (3)(4)(5). Also, the reciprocal of mean time headways is the hourly traffic flow rate, which makes it an important variable to a traffic engineer.…”

“…However, they considered only light and medium rainfall intensities, because no data matching to heavy rain intensity were found. Ha et al (5) conducted another study on time headway distribution using gamma-GQM and tried to determine the relationship between this microscopic traffic variable and other macroscopic variables, such as volume, occupancy, and mean speed. The strength of their work is that they tried to interpret the variation of the four parameters of gamma-GQM, whereas the major flaw of their work was treating rain as a binary variable (yes/no) rather than segmenting it quantitatively based on its intensity.…”

“…Simple models have difficulty in modeling short headways and limited capacity of modeling wider traffic scopes. On the contrary, GQM has proved to have good performance in modeling headways of different traffic contexts with a sound goodness of fit to headway histogram densities as shown by statistical tests of chi-squared and Kolmogorov-Smirnov (6,5,17,18). GQM has been shown to provide insight on the motorway traffic stream and driver behavior for various traffic conditions.…”