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

Abstract

Relationships between macroscopic traffic level and time headway distribution are discussed. Several authors have studied time headway distribution as a function of aggregate traffic volume and have used various probabilistic models. The gamma-generalized queuing model (gamma-GQM) has been shown to have very high performance and effectiveness in modeling real-life time headway data in most cases. Consequently, the parameters of the gamma-GQM are studied to investigate the links between macroscopic information and time headway variable. The systematic data collected in 2005 from the RN 118 national roadway in France allowed study of the effects not only of traffic volume but also of traffic occupancy and space mean speed on time headway distribution. These macroscopic variables were obtained on the basis of the calculation aggregated over 6 min of data. A study of time headway distribution following the empirical fundamental diagram was also implemented. Exogenous conditions such as weather and lane type were taken into account. As a result, the variations of the gamma-GQM parameters were established and interpreted according to the various contexts. Two exogenous factor types influencing time headway distribution in the standard condition are proposed: impeding and propulsive factors. These two factors are differentiated by the tendencies of the scale parameters of the gamma-GQM compared with the corresponding values in the standard condition.