Modification of Webster's delay formula under non-lane based heterogeneous road traffic condition

Delay is one of the most important performance measures of signalised intersections. Various models including Webster’s classical delay formula have been developed in countries with car dominated traffic stream to estimate average delay per vehicle at signalised intersections. Webster’s classical delay formula has been formulated under UK situation where the road traffic condition is homogeneous as well as lane based and consequently the formula may not estimate delays accurately under heterogeneous road traffic condition. As a result, it is necessary to modify Webster’s delay formula to make it usable under non-lane based mixed road traffic condition. In this study, the Webster’s delay formula has been modified to suit the road traffic situation of Bangladesh. For this purpose, data have been collected using video camera at different signalised intersections of Dhaka city to measure average delay per vehicle at each cycle. Based on these data, a model in the form of multiple linear regression has been developed, which attempts to keep the first and second terms of Webster’s delay formula as it is but to modify the adjustment term. The model has been calibrated to form a ‘Modified Webster’s Delay Formula’, which is subsequently validated by comparing the expected delays with observed delays. The model provides a coefficient of correlation of 0.68 and all the independent variables are found to be statistically significant. In the comparison between expected and observed delays, although somewhat scatter has been found, the agreement between the two was found to be satisfactory so that the model has the potential to provide sufficiently accurate values of delay under Bangladesh condition.