Real-world wayfinding experiments: Individual preferences, decisions and the space syntax approach at street corners

This thesis investigates the role of spatial configuration on individual spatial decision-making. Over 100 participants take part in laboratory wayfinding experiments based on real-world images of street corners, using fixed and mobile eye trackers. Participants are asked to perform directed and undirected spatial tasks; stimulus-derived and task-related viewing patterns are accounted for. Responses to the spatial tasks are tested for task-related bias against responses in non-spatial tasks (recall, free viewing, and controlled search). The evidence reveals that, during wayfinding, participants choose the more connected street, measurable with space syntax variables of relative street connectivity. Four space syntax variables are used: integration and choice at global and local scales. The resulting measure allows decisions made by individuals to be related directly to the space syntax analysis of spatial morphology. The fixation data allows for an investigation of how wayfinding choices and gaze bias may be linked. Viewing behaviour during the spatial tasks reveals areas of particular interest at each path alternative; these correspond to structural information in the built environment. A measure for identifying the location of such areas is proposed: "choice zones'". Choice zones are computed algorithmically, and are based on space-geometric measures visible in the scene. Choice zones offer a greater scope than existing measures because they are based on information visible in the real world; it is therefore possible to compute choice zones for images of different reference classes (eg. those with varying horizon or sky lines). The resulting measure has important implications for optimal routing and urban design, identifying those areas of the visual field that contain the most relevant environmental information pertaining to wayfinding