CTM-BASED MACROSCOPIC TRAFFIC FLOW MODEL CALIBRATION AND VALIDATION FOR DHAKA CITY

Abstract

The road network system of Dhaka city is predominantly non-lane-based and heterogeneous in nature. Although microscopic model can simulate this traffic condition, it is difficult to simulate in any macroscopic traffic flow model. The choice of macroscopic model has become very popular these days, because these models consider aggregate behaviour of a set of vehicle i.e. traffic stream as a whole. These models are not only easier to observe and validate depending upon the traffic condition and roadway environment but also demand very less parameters, which are easier to achieve from the field. The main advantage of using these models is to predict and evaluate traffic over time and space in a very short time, including transient phenomena such as the building, propagation, and dissipation of queues, which is very useful in active traffic management and traffic control strategy. Considering all these, this research develops a new modified Cell Transmission Model (CTM) for Dhaka city, a first-order macroscopic model, which can simulate non-lane-based heterogeneous traffic accurately. The modified CTM takes the parameter values from Fundamental Diagram (FD), which has to be modelled prior to CTM. For this high-resolution data was collected from a stretch of 3.26KM roadway divided in 5 cells. Interestingly, FD is found variable in different cells. To confirm the structure of the FD, initially the speed-density plots of the field data for different links are fitted with four general structures: namely, the linear, logarithmic, exponential and polynomial forms. It is revealed that the 3rd degree polynomial structure is best suited for prevailing traffic condition. The boundary values of the experimental setup then fed into modified CTM i.e. length of each cells, free-flow speed, critical density, jam density and capacity. The non-linear equations of FD for heterogeneous traffic behaviour then incorporated in the computational algorithm to modify CTM. This new model, incorporating the new calibration methodology, simulate the traffic system of the experimental setup. The model is then validated comparing with the field data. The sensitivity analysis proves that the modified CTM can simulate the system of heterogeneous traffic with better accuracy than the original CTM.