Modeling the Gender Effects of Pedestrians and Calibration of the Modified Social Force Model

Abstract

The design of pedestrian sidewalks depends on pedestrian flow, which is related to the speed of pedestrians on the sidewalk. The social force model (SFM) is a microscopic pedestrian simulation model that has been able to reproduce many self-organization phenomena of pedestrian flow such as lane formation. Studies have shown that the SFM has been modified to model particular pedestrian behaviors in different situations by introducing new forces or introducing new factors in existing forces. Also, the literature shows that pedestrian speed varies because of pedestrian characteristics such as age, gender, group behavior, and so forth. There are no studies that model the effect of these pedestrian characteristics using the SFM. Therefore, in this study, we have modeled the effect of gender of pedestrians by introducing a gender factor ${\alpha }_{m/f}$ . A sidewalk in Mumbai, India has been chosen for this study. Pedestrian flow and speed were collected from the site. A base SFM containing the driving force, pedestrian–pedestrian interaction force, and pedestrian–boundary interaction force was coded in MATLAB. This model contains six parameters, which were calibrated using a genetic algorithm. Next, the SFM was modified to include different reaction times for the male and female pedestrians, ${\tau }_m$ and ${\tau }_f$ , respectively. Keeping other parameters as constant, ${\tau }_m$ and ${\tau }_f$ were calibrated and found. Gender factors ${\alpha }_m$ and ${\alpha }_f$ are found by dividing the reaction time ${\tau }_m$ and ${\tau }_f$ by ${\tau }_i$ , respectively. These gender factors could be found for the different male/female composition of pedestrians, which would help in analyzing the level of service of sidewalks.