SIMULATION OF MTXED TRAFFIC FLOW ON TWO LANE-ROADS

Abstract

Two-lane highways compose the predominant portion of most National and State Highway system in India. Two-third length of the National Highways and more than 50 percent of the State Highways are still two-lane wide. Two-lane roads have one lane for use by traffic in each direction and the faster vehicles when approach the slower vehicles use the opposing lane to overtake the slower vehicles depending on the sight distance available and o the gaps in opposing stream. Overtaking involves lane-change maneuvers, possibly of the acceleration and deceleration actions and estimation of relative speed of the overtaking and overtaken vehicles. Road capacity, Level of Service and safety are all affected by the passing ability of faster vehicles. The quality of traffic operation on a two-lane highway is described by the average travel speed; the percent time delay and capacity utilization. Traffic flow behaviour has been extensively studied in the developed countries to evolve capacity for twolane roads. Studies abroad have always been with the advantage of homogeneous traffic with very little interference from slow moving vehicles. Traffic on Indian roads is heterogeneous in nature with vehicles of diverse static and dynamic characteristics sharing the same roadway space without any physical segregation making it difficult to apply the results of studies conducted abroad to Indian conditions. Literature on two-lane roads is mostly available from developed countries where the traffic conditions are more or less homogeneous. ! r Various ideas flowing from the research conducted in US have been adopted in India also with minor modifications. Evidently, these findings do not take into account the peculiar behaviour of mixed traffic present in India. The arrival pattern has been modeled through headway data in most of the studies. Conflicting results are reported for the choice of mathematical model for modeling the headway data at 11 different volume levels. It is needed to model the arrival pattern ofvehicles directly without resorting to headway data. Very scanty data are available on speed-flow relationships on different categories of rural roads in India. New technology cars and light commercial vehicles have come on Indian roads in a big way. The speed and acceleration characteristics of these vehicles are completely different from old technology vehicles. It has demanded A update of speed-flow relationships developed in India during 90's and earlier./Few attempts have been made in India to study the effect of influencing parameters on capacity of a twolane road. These attempts are based on the analysis oftraffic flow data collected in the field. Field studies suffer from the drawbacks that more than one parameter are present at a time and it is not possible to completely isolate the effect of individual parameters on capacity. Simulation is the best option in such cases. The present study was undertaken with the following objectives. a) To model the arrival characteristics of mixed traffic at different volume levels on twolane roads. b) To study the speed and placement characteristics ofmixed traffic on two-lane roads and thereby to address the problem ofbimodality in speed and placement data. c) To study the overtaking behavior ofdifferent types ofvehicles on two-lane roads and to estimate the total overtaking time ofa vehicle as a function ofspeed difference between the overtaken and overtaking vehicle and acceleration characteristics ofthe later. d) To develop a simulation program incorporating the behavior of mixed traffic on a twolane road and thereby to estimate the capacity under different conditions of traffic composition, directional split and slow moving vehicles. In order to study the various characteristics of mixed traffic, data were collected at different sections of two-lane roads in different parts of India. All the sections were on in intercity portions of two-lane highways with 7.0 m wide pavement and 3.6 m wide earthen shoulders on either side. The test sections were free from the effect of intersection, bus stops, parked vehicles, curvature, and pedestrianj.The pavement surface condition was fair at all the test sections. Video recording technique was used to collect the data for arrival pattern of vehicles and speed studies. Data were collected at 15 sections for arrival pattern and at 18 sections for speed studies. The recording was done for about two hours on a typical weekday. For the lateral placement of vehicles, the data were collected at 19 sections by dividing the lane width into segments of 250 mm each numbered seriatim from the pavement edge to the centre line and the placement of rear left wheels (left hand drive conditions prevail in India) was recorded manually. The overtaking behaviour was studied by the moving carmethod and data were collected at 3 sections for seven days. From the analysis of data, it is observed that modeling of vehicle arrival pattern on two-lane roads can be done on the basis of coefficient of variation (CV) alone. Poisson's distribution defines the vehicle arrival pattern when CV lies between 0.85 and 1.30, inclusive of both values. For a value of 1.30 or more, negative binomial distribution provided a good fit. At CV equal to 0.85, both binomial and Poisson distributions fit the observed data equally satisfactorily, indicating a transition state. Similarly, at CV equal to 1.30, both negative binomial and Poisson distributions fit the observed data equally satisfactorily, indicating another transition state. Analysis ofspeed data indicated that the speed ofall vehicles combined at a section of two-lane road under mixed traffic conditions do not always follow the normal distribution. It deviates from the normal curve as the traffic becomes heterogeneous. Mathematical equations are suggested to address the problem of bimodality in the speed data. One statistical parameter, spread ratio (SR), is introduced to explain this phenomenon and the speed data at a iv section would follow a unimodal distribution if the SR is in the range of 0.69 and 1.35 otherwise, it is a case of bimodality. The speed data of individual category of vehicles are however normally distributed. It has been further established in this research that the vehicular placement across the lane width under mixed traffic conditions do not always follow a normal distribution. It may follow a unimodal or bimodal distribution depending on the traffic composition. Two statistical parameters; placement factor (PF) and skewness range (SR) control this phenomenon. The placement data would follow a unimodal curve if PF and SR are less than 1.30 and 0.54 respectively. Otherwise, a bimodal curve would better represent the placement data. The overtaking behaviour of different types of vehicles as observed in the field was analysed and it is found that the acceleration rate is variable for relatively shorter overtaking times whereas, it was more or less constant for moderate overtaking times and vehicles traveled at constant speed when the overtaking time was sufficiently high. The relationships between the acceleration rate and the overtaking time are established for different types of vehicles. The maximum acceleration was observed for new technology cars (7.40 km/hr/s) due to their engine power and high braking efficiency whereas, theminimum acceleration was observed for 3-wheeler (0.16 km/hr/s). The various relationships between the relative speed of overtaking and overtaken vehicle and the overtaking time, required gap, decision rules for various types ofovertaking are developed for their use in the simulation program. Finally, a computer program was developed to simulate the traffic flow on a two-lane highway incorporating all the above aspects. The simulation program has been coded in Visual Basic language and has been animated also. A number of experimental runs were performed to determine the capacity of a two-lane road, effect of traffic mix on capacity and speed, to suggest the PCU values for individual category of vehicles and the effect of directional split on capacity. Capacity of a two-lane road under all cars situation is estimated as 2860 PCU/h.). It increases to 5600 vph in case of all 2-wheelers and reduces to 580 vph in case of all tractors situation. The design service volumes (DSV) at different compositions of traffic stream are also suggested. The capacity ofa two-lane road reduces as the directional split moves away from an even split of 50/50. Astraight-line relationship is proposed between the capacity and the directional split and it was used to derive the adjustment factors for directional split. It is further observed that the capacity of a two-lane decreases as the proportion of 3-wheeler increases in the traffic stream. The similar trend was also observed for tractor and heavy vehicles.' is;... But, in case of 2-wheeler the opposing trend was observed and it is due to the small size of 2-wheelers and their maneuverability. Speed of a traffic stream of cars reduces as another type of vehicle is introduced in the stream. The amount of reduction in speed depends upon the type and proportion of second category of vehicle and total traffic volume. PCU values for each category ofvehicles were also calculated. The PCU for a vehicle type (2- wheeler, 3-wheeler, and tractor) decreases with increase in its proportion in the traffic stream and the volume capacity ratio. This trend is however not observed for heavy vehicle where PCU value reduces with volume to capacity ratio but increases with its own proportion in the traffic stream. It may be attributed to the large size of these vehicles and their operational efficiency. VI