Simulation of pedestrians crossing a signalised refuge island using the link transmission mode

Abstract

As Singapore strives to transform into a car-lite society, stronger attention is given to cycling and walking. This entails creating safer roads and infrastructures for pedestrians, and foreseeing changing demands such as Singapore’s aging population. Singapore’s paradigm shift to focus on public transport, walking and cycling would later on translate to higher pedestrian volumes on sidewalks and on crosswalks, especially of those near main facilities such as an MRT station. Traffic models are often essential tools used for planning and operating road infrastructure. On the other hand, traffic simulation provides visual demonstrations of different scenarios where its output can be used to evaluate the performance capabilities of designed infrastructures. In 2007, Yperman proposed the Link Transmission Model. This model proposes a node-link structure of networks where a single link is used to characterise roads between two intersections (nodes). The model has been highly applauded for its low computation time whilst maintaining predictive accuracy. While extensive research has been given to LTM in the realm of vehicle modelling, no action is taken in pedestrian modelling. Hence, this study seeks to put forth a model that is capable of simulating pedestrians crossing a signalised crosswalk by utilising the concept of pedestrian rows. The simulation is done and run entirely on Excel sheets, keeping its computation time low. Simulation results are then compared against actual footage data to assess its effectiveness. This study pays most attention to pedestrians on a traffic refuge island, where it provides pedestrians with the waiting space to cross. Calibration is attempted in Chapter 6 to fine-tune the model to suit its location and demographic. The result is to obtain a simulation that replicates actual data. Hence, the aim of this study is to achieve minimum peak time difference and minimum peak pedestrian volume difference.