A New Generic Approach to the Pedestrian Fundamental Diagram

Abstract

The pedestrian fundamental diagram, which describes the relationship between speed, flow and density, is one of the key concepts for the design of pedestrian facilities. It allows to determine the capacity of facilities as well as to calculate estimated walking times, for example when changing trains. In the last decades, laboratory and field measurements have been conducted to obtain the fundamental diagram for different situations. Based on this it was revealed that the fundamental diagram differs significantly between different locations and pedestrian compositions. Various parameters which significantly influence the shape of the fundamental diagram curve have been determined. Thus, poor results might occur when using a general form of the fundamental diagram for design purpose. Still, only either fundamental diagrams measured in a single setting or ones aggregating these measurements are available. No comprehensive model was found which is able to describe the differences observed as well as being able to estimate fundamental diagrams for other pedestrian compositions. Currently, an estimation of a specific fundamental diagram without corresponding empirical data is not possible. The idea of this thesis is to address this research gap by creating a generic model, which describes the main walking principles and the interaction between pedestrians. Using this model allows to generate situation specific fundamental diagrams. In addition, the underlying principles which determine the shape of the fundamental diagram are determined and the effect of different influence factors on human walking can be studied. By applying situation specific input parameter values, it shall be possible to generate situation specific fundamental diagrams useful for the design of pedestrian facilities. The model creation was done in two steps. As a basis for the quantitative fundamental diagram model, a qualitative generic walking model was developed. A comprehensive overview of the basic principles of walking and the interaction of pedestrian was given in the generic walking model. This was done by combining individual findings available in literature. Starting from a simple case, where a pedestrian is walking alone at constant speed, the model is extended to include all relevant influences on a pedestrian. The generic pedestrian walking model was then used as a basis for the fundamental diagram model. Here, a microscopic model was established where also the distribution of input parameter values can be simulated. Again, first a simple model was set up and complexity was added step by step. In the end a model was created based on lane movement of pedestrians, where lane swapping is possible to enable overtaking. For the model setup, a thoroughly literature review was performed at the beginning of this work. A special focus was laid on the pedestrian walking speed. The different influences on the free flow walking speed were collected and their effect was estimated. In addition, a definition of the fundamental diagram useful for this work and a detailed discussion on the time aspect of the Level of Service concept, which is closely linked to the fundamental diagram, was made in connection to the literature review. For the calibration and validation of the model, empirical data already available from experiments and real-life measurements was used. This allowed to cover a wide range of different setting in the validation. In conclusion, the validation showed that the model is expected to provide adequately accurate and useful results for the expected scope of application. Using the model it was possible to accept the research question stating that it is possible to create a generic pedestrian fundamental diagram model. This can be used to estimate specific fundamental diagrams. However, it was found that although the model is based strongly on the basic principles of walking, the estimation of the input parameters can be challenging. Still, for experienced users it is expected that the model will provide better results than general fundamental diagrams currently used. The validated model is ready to be used to estimate situation specific fundamental diagrams. This was done for different scenarios, for example for the expected situation in 2050 or to determine the effect of the mixture of commuters and shoppers on the pedestrian flow. In the future scenario it was shown that the assumed trends result in reduced walking speeds, especially at higher densities. Here, the model allows for the first time to evaluate the effect of different changes and the simulation of fundamental diagrams for situations, where no empirical data is available.