Abstract
What geometric properties of the environment do humans depend upon for navigation? Last year (VSS 01), we reported that participants relied on ordinal structure rather than metric distances and angles, when they walked to learned locations in a virtual hedge maze. By ordinal structure, we mean the sequence of places, junctions and landmarks along a route. Previously, we stretched or sheared the maze to manipulate metric structure. In the present experiment, we add or delete paths to manipulate ordinal structure, and test whether subjects rely on the number of junctions along a known route. Participants walked in an immersive virtual environment (12 m × 12 m), while wearing a head mounted display (60 × 40 deg, 50 ms latency). The environment consisted of a virtual hedge maze containing eleven occluded places (statue, bird bath, etc.). During training, participants learned the layout of places by freely exploring the environment for 8 min. During testing, they were asked to walk to specified locations, starting from the same home position on each trial. On probe trials, a side path was either added to or deleted from the maze. If participants rely on the number of junctions from home, this would lead them to take an early or late turn to a different location. On the other hand, if they rely on the ordering of paths from another reference point (e.g. a unique junction, the far end of the path), or on metric structure, they would turn onto the correct path. The results indicate that participants rely on junction number on a significant percentage of trials, demonstrating an influence of ordinal sequence along a known route. However, other structural properties of the environment, such as distinctive junctions, path endpoints, or landmarks, may also provide reference points for an ordinal navigation strategy.
NSF LIS IRI-9720327