Estimation of Saturation Flow at Signalized Intersections Under Heterogeneous Traffic Conditions

Abstract

This research aims to develop a simulation-based methodology in the estimation of saturation flow at signalized intersections under heterogeneous traffic conditions in India. Field surveys have been carried out particularly during peak periods at two signalized intersections, namely Vijay Char Rasta intersection situated in Ahmedabad and Rangila Park intersection situated in Surat. Videographic technique is employed to capture the data at macro- and microlevel. The videographic approach provides the queue building, moving, and queue dissipation at microlevel. The field data pertains to traffic volume, compositions, turning movements, headway, cycle time, and phasing at the approach. The survey also concentrates on noting the queue length of each approach. Apart from videographic survey, road inventory survey and spot speed surveys are also carried out in order to use this data in VISSIM 7.0 for the development of network and defining speed distributions. The present work mainly focused on the development of PCUs, dynamic saturation flows, VISSIM modeling and validation, regression modeling and validation, and sensitivity analysis. Passenger car unit for each vehicle category is calculated by two methods—headway method and optimization method. Saturation flow is estimated with the help of PCU values estimated by different methods, and these values are compared. TRL method is adopted in the estimation of saturation flow. A 5-s interval is considered in this method. Nearly, 8842 VPH is the saturation flow against to 5460 PCU of headway method and 5071 PCU of optimization method for 3 lane approach of Vijay Char Rasta intersection. Saturation flow is almost double in terms of mixed traffic. Therefore, saturation flow cannot be static in present study but is dynamic. Dynamic saturation flow values are provided for varied composition and approach widths. Base network is created in VISSIM 7.0. Model is calibrated as per Indian driving behavior for both Vijay Char Rasta and Rangila Park. Average queue length is considered for validating the model. Sensitivity analysis is carried out after validating the simulation model. It is observed that change in road width is having significant effect on saturation flow when compared to change in turning movements. Regression analysis is carried out for developing multilinear regression saturation flow model (MLR–SFM). The model is developed between saturation flow and proportion of 2W, 3W, small car, big car, and road width. 65–35 combination is observed to be best fitted combination in developing and validating the model. Similar to the observations in VISSIM modeling, it is observed that change in road width is having a significant effect on saturation flow when compared to change in traffic compositions.