An Urban Pedestrian Flow Prediction Model for Establishing a New Walkway Width Design Method

Abstract

Pedestrians form a significant share of the road users and are highly susceptible to safety threats if ample infrastructure is not provided for their mobility. Supply of pedestrian infrastructure should match the demand that is correlated with pedestrian flow on urban walkways at acceptable service levels. This study yielded a statistical regression flow prediction model and a new design approach for urban walkways. The objectives of the study were to assess urban pedestrian movement characteristics using chosen walkways, establish a pedestrian flow model by utilizing statistical regression and principal component analysis, and to create an approach for designing walkway width at prescribed levels of service. The methods used included, walkway geometric measurements done using a measuring tape, digital cameras to capture pedestrian photographic data so as to determine walking speeds, density and flow at four locations, namely Makerere road, Bombo road, Entebbe road and Jinja road, and a stopwatch to acquire travel time. The findings show that the mean urban walking speed is 77.42±5.36 m/min, maximum density at 4.00 ped/m2 and flow at 66 peds/min/m, respectively. These characteristics were found to be significantly influenced by walkway location, gender and activity (p < 0.05). The collected data were used to develop the flow-speed-density relationships for pedestrians along walkways in Kampala city. A significant correlation was observed between pedestrian flow, walking speed and density. Moreover, pedestrian characteristics from various cities in the world were compared to those obtained in the study. A pedestrian flow prediction model was developed using statistical regression and principal component analysis with coefficients of density and speed variables highly significant. The study also revealed that pedestrian flow along walkways in Kampala city is more influenced by pedestrian density than pedestrian walking speed. This model was used to develop a new approach for designing walkways for urban roads based on LOS, whereby walkway width is approximated using flow and space. The developed design curves could be used as basis for the development of more efficient, adequate and safer facilities for the pedestrians in Kampala city. Nonetheless, the approach requires more investigations including other relevant variables (e.g. gender, age and group motion) for generalizability.