Analysis of Delay at an Un-Controlled Intersection under Heterogeneous Traffic Condition

Abstract

The non-lane disciplined traffic condition and variations in vehicle composition leads to the complex traffic behavior, resulting in to significant vehicular delays at uncontrolled intersections, especially with the rapid increase in the number of vehicles. The aim of the present research work is to investigate delay and propose a model for evaluating delay using regression analysis and artificial neural network (ANN) considering several parameters affecting it such as pedestrian crossing movements, turning moments, volume, and queue length under mixed traffic conditions at unsignalized intersections.  The data required was collected from two uncontrolled intersections in Bhuvanagiri City. An analysis of the statistical distribution using the observed speed data has validated in the presence of several well-known distributions. The Fisher-Tippett distribution has been identified as the most suitable for approaches handling higher traffic volumes i.e. more than 2000 veh/hr. Some other types of distribution such as Gamma and GEV are fitted commonly to the filed data where high proportion of smaller sized vehicle types such as 2W and 3W are comprising the traffic flow. Also, in the present study, the Level of Service of unsignalized intersections was assessed using developed delay model under mixed traffic conditions. It was found that the unsignalized intersections in Bhuvanagiri city are functioning at Level of Service (LOS) E, with delays for various approaches ranging from 32 to 43 seconds per vehicle.