Fuel consumption in urban traffic: A twenty car designed experiment

Abstract

This paper re-evaluates earlier methodology and derives a more reliable and accurate method of measuring the fuel consumption and related effects of a traffic management change. The study, involving a significantly different approach to the traditional before and after method, was carried out during the morning peak over 2 1 4 km of multi-lane arterial road in Sydney. The traffic signals were operated on the SCAT (Sydney Co-ordinated Adaptive Traffic) system. A comparison was made of two maximum cycle length settings. The experiment was formally designed enabling results to be obtained by analysis of variance. Twenty test cars were selected to represent vehicle types in the traffic stream and fitted with fuel meters. Questions addressed included the number of test cars required, the previous treatment of driver variation and the disparities arising from the neglect of factors which vary with time during the course of an experiment. Signal plan and lane type were found to cause significant changes to fuel consumption and other trip variables, the information overall being more coextensive than in prior studies. The significance of traffic volumes measured inside the experimental area in reference to other associated variables is dealt with. The paper questions the validity of currently accepted procedures for assessing traffic management changes.