Author(s)

SUBHADIP BISWAS, SATISH CHANDRA, INDRAJIT GHOSH

Abstract

Use of speed prediction model is multi-faceted and widely acknowledged in various aspects of Traffic Engineering. However, developing speed model is a challenging task particularly in the context of urban mixed traffic where vehicles possessing diverse static and dynamic characteristics use the same carriageway without any segregation. Here the traffic composition plays a vital role in determining the vehicular speed. Thus, the present study aims to develop a volume based speed model which will also afford the opportunity to monitor the individual influences of different vehicle categories present in the traffic stream. Classified traffic volume and speed data have been collected from different mid-block locations on two lane undivided urban roads. A novel Lambert W function based framework has been proposed in this work to model individual speed in a mixed traffic stream. Validation results affirm the accuracy of the model since a good agreement has been observed between the speeds observed and the speeds predicted by the model developed. Determination of Passenger Car Unit (PCU) also requires speed information of individual vehicle types. Therefore, the predicted speeds by the model developed have been utilized in estimation of PCU for individual vehicle type. Sensitivity analysis has been performed in order to observe the influences of traffic volume and its composition on speed and as well as PCU of individual vehicles.

Keywords

speed, traffic volume, PCU, Lambert W, macroscopic traffic flow model, mixed traffic