Wayfinding in scene space: modelling transfers in public transport

One of the characteristics of public transport is the need for transfers. During transfers, travellers are pedestrian wayfinders and have to find their way from the place of arrival to the place of departure of another means of transport. Wayfinding in public transport takes place in two types of spaces: network space and scene space. Network space includes the transport network. Scene space consists of the halls, squares, and platforms at interchange nodes; it lacks an obvious network structure. Published timetables allow pre-trip wayfinding in network space, but wayfinding in scene space requires interaction with the actual environment, using the information provided therein. The theory of image schemata (cognitive patterns that structure our perceptions and actions) suggests that structural information is immediately usable. Other types of information, notably signage, require conscious effort to use. This research is about transfers in railway stations, i.e., wayfinding in scene space. I hypothesise that structural information provided by the architectural layout is enough to guide travellers through railway stations. Computational modelling is used to test this hypothesis. Image schemata form the cognitive basis for a formal model of scene space, called Schematic Geometry. A software agent is developed next and used to simulate wayfinding tasks in Schematic Geometry models of two railway stations. Results indicate that structural information supports wayfinding, but is not always enough to find optimal routes. Previews of parts of the environment are found to be highly effective wayfinding aids that build on human object recognition rather than on sign reading. Conclusions are that humans integrate information from various sources and of different types to achieve good wayfinding performance; well designed architecture can further support wayfinding by providing structural clues and previews. This thesis contributes to our understanding of wayfinding, especially the information needs for wayfinding in scene space, and provides a cognitively motivated but formal model of scene space, Schematic Geometry.