DEVELOPMENT AND CALIBRATION OF ENTRY CAPACITY MODEL OF MODERN ROUNDABOUTS

Abstract

Traffic conditions in India are highly heterogeneous due to variety of vehicles with different static and dynamic characteristics sharing and operating at same road space. Roundabouts are a type of intersections which commonly used as a means of intersection control for moderate traffic flows. These adapt to junctions having variations in the intersection geometry. Efficiently designed roundabouts can handle traffic very smoothly without causing any delay. It facilitates an orderly movement of traffic in a circular motion around a central island whose shape is decided based on intersection geometry. The circulating traffic is considered to be the priority stream and entering traffic shall wait for a suitable gap in the circulating traffic. In this fashion, it reduces the stopped delays as observed on the signalized intersections, and thus improves the operational efficiency. The analysis and design of roundabouts in India is governed by IRC:65-1976. The conversion factors for heterogeneous traffic to homogeneous one are quite old and need validation. The capacity of the roundabout is based on weaving section. This approach has already been replaced by estimation of entry capacity. A look on the literature in this respect indicates that most of the studies are from US and other developed nations where homogeneity of traffic stream and lane discipline of drivers are two important characters of traffic flow. Very few attempts have been made in India to evaluate the entry capacity on roundabouts under mixed traffic and untidy behavior of drivers. Data were collected at eleven roundabouts spread across cities of Chandigarh, Noida and Panchkula. Geometric parameters were measured manually with measuring tape while traffic data was collected by means of videography. It required four types of analysis, namely estimation of passenger car units, estimation of critical gap for different categories of vehicles, calibration of HCM model based on estimated values of critical gaps under heterogeneous traffic conditions and, estimation of entry capacity model based on the field data (traffic and geometric). The estimation of passenger car unit (PCU) for different vehicles to convert heterogeneous traffic into homogeneous traffic is a well-accepted procedure. But the parameters used for mid-blocks may not be helpful on roundabouts as traffic flow iv characteristics on the two locations are different. Indian Roads Congress code (IRC-65) recommends static PCU values which is based on studies prior to 1976. Since then vehicle technology has changed. This study re-looks on the PCU values based on field studies and suggest the modified ones. The PCU for a vehicle is estimated based on lagging headway and width of the vehicle, to account for vehicle size and untidy flow conditions. It is also not clear whether to use static or dynamic PCU values on account of possible temporal and spatial variations across locations. It was found that PCU value for motorized two-wheeler reduced by half of its value given in IRC:65-1976. The car category got divided as small and big car. The PCU value of heavy vehicle got increased marginally. The problem to deal with re-estimation of PCU values at different locations, due to possible traffic flow variations, is dealt with by proposing a Heterogeneity Equivalency Factor (H-Factor). The factor is multiplicative and converts heterogeneous traffic (veh/h) into homogeneous traffic (pcu/h). Estimation of critical gap for a vehicle type under mixed traffic conditions prevailing in developing countries has been always a challenging task. This is due to the poor lane discipline and very limited priority being followed by the vehicles at priority intersections like roundabouts. A simple procedure, which is based on minimization of the sum of absolute difference between a gap value and accepted / rejected gap, is proposed in this study. The iterative procedure provides a value of gap that is termed as the critical gap under mixed traffic conditions. The method is different from maximum likelihood method (MLM) in two aspects. First, it does not assume any predefined distribution for the critical gap and second, it does not fail even when there are very few rejected gaps. Prominent methods available in literature to estimate critical gap are compared for different categories of vehicles. Based on the results of consistency test, the MLM and the proposed method are found to be the most acceptable estimation methods. It has been further observed that the proposed method is better than MLM when working with low sample size, as well as, in no-priority conditions, which arise due to heterogeneous traffic flow prevailing in developing countries. The entry capacity of a roundabout against variation in circulating flows is being assessed. Queue formation in the approach is taken as an indicator that the approach is operating at the capacity. The normal notion considered is that as circulating flow v decreases, the entry flow should increase. This may be due to the higher opportunities being made available to the vehicles desiring to enter the circulation area. Linear and exponential functions are found to be showing goodness-of-fit between entry capacity and circulating traffic flow. To find relation between entry capacity and geometric parameters, entry capacities have been plotted against central-island diameter, circulating roadway width and entry width. Power function is found to provide the best fit between entry capacity and central-island diameter, circulating roadway width and entry width. The variation or dispersion of data is found to be quite low for entry capacity and central-island diameter and is a bit high for the other two relationships. The Highway Capacity Manual of US (HCM 2010) has given gap acceptance model of entry capacity for single-lane and two-lane roundabouts in U.S. This manual is extensively used in different parts of the world for estimating the capacity of a traffic facility. However, the direct transferability of the HCM (2010) entry capacity model to Indian traffic flow conditions was doubtful as the manual do not consider driver behavior under mixed traffic flow. Therefore, the parameters of the equation were calibrated for its adaptation to heterogeneous traffic conditions, using critical gap values obtained from the field data. The modified HCM model for entry capacity was still found differing from the field entry capacity. Therefore, multiplicative adjustment factors for different size of roundabouts have been developed for modified HCM (2010) equations to satisfy the traffic flow condition prevailing in the developing countries like India. A regression model for estimating roundabout entry capacity was developed based on the traffic flow and roundabout geometric data. The analysis indicated that the widths of circulating roadway and central-island diameter have a significant influence on the entry capacity. The developed regression model is validated on another roundabout and only ±6 percent difference was observed between the field entry capacities and those predicted by the proposed model. Sensitivity analysis is used to see the effect of physical parameters of the roundabout on entry capacity. It was found that the central-island diameter has the greatest effect on entry capacity while the circulating roadway width has the smallest effect. The circulating traffic flow versus entry capacity charts were made with the purpose of comparing the results of the proposed model with the existing models available in the literature, namely Jordanian, Malaysian and Indian vi (Prakash 2010). The comparison indicated that the proposed entry capacity model gave the highest entry capacity, whereas, the Malaysian model gave the lowest entry capacity. The results of the proposed model have shown very good relationship with the field entry flow data, as compared to other models. The range of entry capacity has been found varying between 3000 to 4000 pcu/h for different size of roundabouts. Lower limit for entry capacity is expected to be ranging between 800 to 1200 pcu/h.