Investigating Service Performances of Signalized Intersections Operating under Mixed Traffic Condition

Abstract

This article proposes a new insight of modelling service quality offered by signalized intersections, which are the nodal focuses in a transportation network in developing countries. To achieve the objective of this research, a broad spectrum of geometrical, traffic operational, built-environmental and behavioral data sets are collected from 45 diversified signalized intersections with a widely varying driving environment through field investigations, videography techniques, and perception survey. Responses from around automobile drivers were gathered seeking socio-demographic information and overall satisfaction scores for respective approaches of intersections. Accordingly the six parameters exerting significant influences on driver’s satisfaction were highlighted by Spearman’s correlation analysis. Exceptionally reliable, and less complex Automobile Level of Service (ALOS) models were formulated considering these six variables with the assistance of a unique and widely used artificial intelligence technique in particular, Multi-Gene Genetic Programming (MGGP) and Differential Evolution (DE). DE model displayed incredible likelihood efficiencies with high coefficient of determination (R2) of 0.93 and 0.894 for training and testing datasets respectively. The sensitivity analysis showed that queue length plays a significant role in fixing ALOS standards of signalized intersections, having highest negative influence of 67.153%. Hence, optimizing traffic signalization timings and increasing effective green time for major approaches of an intersection in peak hours will significantly enhance the service quality of respective intersections. Similarly, other parameters are ranked in decreasing order of their relative importance to help the transportation administrators for making efficient resolutions for better infrastructural design.