CRITICAL GAP AND DELAY CHARACTERISTICS AT UNCONTROLLED INTERSECTIONS

Abstract

Studies on uncontrolled intersections have always been a subject of interest among traffic engineers throughout the world for over last three decades. Gap acceptance provides the basis of operations at uncontrolled intersections. Measures of gap acceptance enter into the calculation of capacity, warrants for stop-sign. etc. Critical gap estimation is the most important, but very complicated factor in the analysis of uncontrolled intersections. Delay experienced at an intersection is a measure of the performance of the intersection. Majority of the studies reported in literature on critical gap and delay estimation are confined to uniform traffic conditions. Traffic conditions at uncontrolled intersections under mixed traffic situations prevalent in developing countries are very complex and unique. A detailed review of earlier studies on uncontrolled intersections indicated that no suitable and comprehensive study is readily available for analysing mixed traffic prevailing at uncontrolled intersections in India. The following specific problems of mixed traffic flow at uncontrolled intersections need to be addressed. i. The traffic in India is highly heterogeneous consisting of a variety of both fast moving and slow moving vehicles. The static and dynamic characteristics of these vehicles vary significantly; their behaviour and requirements of gap to execute a movement are entirely different from those under uniform traffic conditions. These factors will have significant influence on the gap acceptance characteristics and capacity of uncontrolled intersections, ii. Small size vehicles often squeeze into any available gap between the large size vehicles and move into the intersection area in a zig-zag manner, a situation not analysed earlier. iii. The theoretical models developed to analyse traffic at uncontrolled intersections conditions are based on the concept that some of the movements have absolute priority and a low priority movement will always yield to the high priority movements. The rule of priority is often violated under mixed traffic conditions and minor street vehicles enter the intersection area even in small gaps forcing the major street vehicles to slow down and provide sufficient gaps for their movement, iv. Data extractions itself become challenging and reference point has to be re-defined keeping in mind that all drivers do not stop at the stop line, v. Due to chaotic behaviour, the delay experienced by lower priority movements will be different from that under homogeneous traffic conditions, and it would depend on traffic composition also. In order to estimate critical gap at uncontrolled intersections using widely used methods, two 3- legged intersections located in semi urban areas were selected. The intersections selected for the study had major streets with four lane divided highway, and minor streets were undivided two-lane road with 7.0 m carriageway width. All the selected intersections are located in semi-urban areas with minimum pedestrian activities. Video recording was done for about 2 hours on a typical clear weekday to collect the data at these intersections. These data were supplemented with manually collected information on geometry and layout of the intersection. Critical gap was estimated for different types of vehicles and for different priority movements by methods like lag, Harder's, logit, probit, modified Raff and Hewitt's method. The results showed large variability in the critical gap values estimated by the different methods. The values are found to be quite low also:. Though in a mixed traffic condition, the critical gap is in expected to be on the lower side, it cannot be as low as estimated by these methods. It shows that the methods developed for homogeneous types of traffic cannot be applied to mixed type of traffic. Anew concept of estimating the critical gap of a vehicle considering its behavioral aspect is developed. The proposed method requires cumulative curves of gap and lag acceptance (Fa) and clearing time (Fct) for a vehicle type. The point of intersection of these two curves indicates a situation when clearing time is just equal to the gap or lag accepted. This is the critical gap for the vehicle type. These two curves would necessarily intersect at a common point when traffic conditions at the intersections are near saturation. However in the case of light traffic on major road, the gaps and lags offered to the minor street vehicles will be larger and the two curves may not have a common point ofintersection. To address such situations, it is proposed to use (1- Fet) curve in place of Fc, to get adefinite point of intersection. The above concept is extended to six more intersections and critical gap values were estimated for 3 types of vehicles; car, three-wheeler and two-wheeler, for all movements. The proposed method gives quite reasonable values and varies with the type of maneuver and the size of the vehicle executing the maneuver. Critical gap values are found to be lowest for LT from minor street followed by RT from major and RT from minor. Among the various priority movements, through movement from the minor street has the highest critical gap followed by right turn from the minor street, which is followed by right turn from the major street. Troutbeck (1999) has reported that the critical gap for apriority movement at unsignalised intersection depends upon the amount of conflicting traffic. Critical gap values corresponding to different values of conflicting traffic are determined for different types of vehicles, executing different maneuvers using the IV rotation technique. In this technique, the two cumulative distribution curves corresponding to the highest conflicting traffic and the lowest conflicting traffic are rotated about their mean until they are parallel to each other. Then the horizontal distance between these two curves is interpolated to see the effect of conflicting traffic on critical gap. The method is highly iterative and was executed through a computer program. Entry capacity of a priority movement is determined using the Blunden's equation. Critical gap corresponding to different values of conflicting traffic is used for arriving at the capacity of various movements for three types of vehicles, viz; car, three-wheeler and two- wheeler. The mean value of follow up time measured in field is used for finding the capacity. The entry capacity of a movement is found to decrease sharply as the conflicting traffic increases and it is dependent on the type of the vehicle also. A simulation model was developed in Visual Basic to study various aspects of service delay at uncontrolled intersections considering all characteristics of mixed traffic. Placement of vehicles in a lane, speed characteristics, arrival pattern, acceleration and intersection clearing time distributions are studied in field and are incorporated into the program. Chaotic behavior of a vehicle is considered through its clearing time. The simulation model is validated using average service delay experienced by a vehicle at an intersection taking right turn from major street. The maximum variation between the observed and predicted average service delay is less than 8%. The model is further validated by examining the observed and simulated approach headways of vehicles. Chi-square test is conducted to test the goodness of fit. Using this simulation program, basic service delay models are developed for four categories of vehicles viz car, 2-wheeeler, 3-wheeler and heavy vehicle for all priority movements. F test revealed that service delay model for LT from minor street and RT from major and also for through and RT from minor street can be represented by single models. Higher priority movements like RT from major and LT from minor are found to have less service delay than lower priority movements like RT and through from minor streets, which is quite obvious. For a given priority movement, service delay to a vehicle increases with addition of a second category of vehicle in the conflicting traffic stream. It also increased with the increase in the proportion of the second category of vehicle as well. As expected, among the various types of vehicles, introduction of heavy vehicle in conflicting stream inflicted more service delay to various priority movements, while 2-wheelers are found to produce less delay. This trend is more indicative for lower priority like RT and through movement from minor. Models are developed relating service delay of a vehicle with conflicting traffic when the latter consists of cars and one of the remaining three categories of vehicles. These models are used to arrive at adjustment factors for various combinations of subject vehicle, priority movement and conflicting traffic. Adjustment factors are multiplicative in nature and aredefined as the ratio of service delay in a mixed traffic situation to that in an 'all car' situation. To estimate the average service delay under actual field conditions where traffic stream consist of more than 2 types of vehicles, multiple linear regression models are suggested to convert all vehicles in the traffic stream into equivalent percentage of heavy vehicles. Thus the mixed traffic stream is converted into one having cars and heavy vehicles only. Heavy vehicle is considered as standard vehicle for the reason that its effect on service delay is the highest among all the three categories of vehicles considered in this study. It would avoid the problem of equivalent percentage reaching more than 100%. A worked out example is also included to explain the procedure of evaluating service delay to a vehicle under given conditions of conflicting traffic. VI