Citation

Griffiths JD, Hunt JG. Traffic Eng. Control 1991; 32(10): 458-462.

Affiliation

Univ of Wales Coll of Cardiff, Wales

Copyright

(Copyright © 1991, Hemming Group)

DOI

unavailable

PMID

unavailable

Abstract

Most investigators seem to agree that no single distribution will adequately describe the headway distribution even at the same point on the same road at the same time on successive days - much less at different times on different roads. This is clearly an unsatisfactory situation for the practising road traffic engineer. With the advent of powerful PCs, and with the sophisticated software now available for micro- and mainframe computers, a highway engineer has at his disposal tools which were not part of his armoury a few years ago. However, this advanced technology relies on the input of appropriate data, and it is the aim of this paper to provide a methodology to help in the provision of headway data relevant to urban situations. A large volume of data relating to headways of vehicles in single lanes of traffic was collected at 45 sites in the U.K. These sites were located in urban areas, typically in busy High Streets; consequently headways were influenced a good deal by such factors as the interference of other vehicles, by the presence of pedestrian crossings and signalised junctions, by parking manoeuvres, etc. Despite the fact that modelling of congestion in inner-city areas is now much in vogue, there appears to be little published material relating to headway distributions in this type of situation. It was found that a simple distribution, the Double Displaced Negative Exponential Distribution, provided a good fit to the observed headways at the vast majority of sites. A method of estimating site-specific parameters for this distribution has been described for the situation where it is possible to observe the raw headways. A further method has been outlined for less precise estimation of parameters when only the vehicle flow rate is available, and a 'look-up' table of parameter values has been provided. An added advantage of the distribution chosen is the ease with which it can be used for drawing samples for use in simulation exercises.