The effects of 2D and 3D maps on learning virtual multi-level indoor environments

Abstract

It is known that people have problems when wayfinding in multi-level buildings. We propose that this challenge is largely due to development of inaccurate multi-level cognitive maps of the 3D building structure. We argue that better visualization of the layered structure of the building could facilitate multi-level cognitive map development and significantly improve spatial behaviors requiring cross-floor knowledge. To address this issue, we compare two viable mobile digital visualization methods (2D top-down view maps and 3D bird's eye view maps), each showing users' real-time position as they navigated. Participants first learned a multi-level virtual building using each of these conditions, as well as a third control (non-assisted) condition. Their task was to find and learn four targets situated at two different landmark types (e.g., contiguous landmarks that were vertically aligned on each floor and non-contiguous landmarks that had no obvious alignment between floors). Participants then took part in three cross-level testing tasks performed without the map assistant used during learning: pointing between targets, vertical navigation (e.g., navigate from point A on floor 1 to the corresponding vertically aligned point on floor 2) and paper-based drilling (e.g., drawing circles to indicate floor 1's target locations on floor 2's layout). Preliminary results showed that both 2D and 3D map significantly improved pointing and vertical navigation accuracy compared to the control condition with no map assistance. However, no significant differences were found for either condition between the two map conditions. By contrast, in the paper-based drilling task, users showed significantly higher accuracy in the 2D map condition than in both the 3D map and control conditions, giving 2D interactive maps the advantage for supporting multi-level spatial behaviors.