Empirical Studies

Encoding of Spatial Information

We are interested in how people mentally encode spatial information for wayfinding. In one of our route-following studies, subjects either got a map or a verbal route direction for a given route, had to memorize this, and then to follow the route. It turns out that subjects memorize the route as a sequence of propositional instructions. They performed better for simple routes using the verbal directions; for more complex routes being able to recall a mental image of the map was beneficial. In another study employing a secondary task setting, subjects were distracted with either a verbal, a spatial, or a visual secondary task during the learning phase and then had to find their way in a virtual environment (see the image). This study supports the interpretation of the first study: environmental information for wayfinding seems to be stored in a propositional format employing abstract spatial knowledge rather than a visual image.

The VR-study has been done in collaboration with the MPI Tübingen.

Wayfinding Strategies

In collaboration with R5-[ArchWay] we conducted several studies on wayfinding in a complex, 3D indoor environment. Based on the subjects' familiarity with the environment, we could identify different strategies: the central point strategy of finding one's way by sticking as much as possible to well-known parts of the environment, the direction strategy of choosing routes that head towards and lead to the horizontal position of the destination as directly as possible, and the floor strategy of first finding one's way to the floor of the destination. The different strategies lead to differences in performance; the appropriateness of a given strategy seems to depend on the environmental setting.

Evaluation of Assistance Means

We elicited people's conceptualization of turning actions in wayfinding. The theory of wayfinding choremes is based on a seven-direction model (excluding the back-direction) that stems from subjects' sketches of turning situations at intersections. This model has been further refined in an experiment based on the card-sorting paradigm. Here, subjects had to group icons depicting turning situations however they felt it to be appropriate. A cluster analysis lead to the depicted direction model for turning actions.

Currently, we perform a study that compares different kinds of maps in a wayfinding study. This will result in further information with respect to the different maps' performances. We expect to see results in favor of schematic maps.