The calibration of vehicle and pedestrian flow in Mangalore city using PARAMICS

Abstract

This paper presents a traffic simulation approach for evaluating the pedestrian behavior by developing a model which includes pedestrians in a vehicular micro-simulation model i.e. PARAMICS that provides a default mechanism for simulating pedestrian movements. This approach is demonstrated by using a case study of a signalized intersection having large pedestrian flow interacting with large vehicular flow in the city of Mangalore, India. In this paper pedestrian behavior is calibrated with pedestrian speed flow models by defining pedestrians as vehicles. This work provides different methods for evaluating policies that effect both pedestrian and vehicle flow. In this work different critical factors that affect the pedestrian and vehicle flow are considered (i.e. Vehicle queue discharge rate, Pedestrian queue discharge rate, Pedestrian gap acceptance, Pedestrian speed etc.) and finally the network was coded in PARAMICS. The major parameters defined of the driving behavior that control the car-following behavior, lane changing behavior, lateral position and driver reaction to traffic lights are considered and calibrated in PARAMICS. After calibration and validation of pedestrian and vehicle flows the difference between the simulated and observed data were analyzed. The high variability in data sets confines the simulation resulting in a constant queue discharge rate at flows approaching saturated conditions. This work finally concludes that PARAMIC’S car-following algorithm is very much helpful in reproducing vehicle and pedestrian flow in complex as well as heterogeneous urban traffic. Application of this car-following algorithm demonstrated by calibrating in PARAMICS will be very useful in predicting known pedestrian speed-flow relationship.