Abstract
Previously, Stankiewicz, et al. (2006) used an ideal navigator to measure human navigation efficiencies with varying sizes of layout and limited visual information. The current studies will use the same approach to investigate the benefit of visual landmarks and layout topology on human navigation efficiencies.
In the current studies, participants were trained and tested in large-scale, virtual, indoor environments. Three viewing conditions were used: No-Landmarks, Landmarks+Fog and Landmarks. In the No Landmarks condition, the environment was visually sparse such that multiple places in the environment could produce the identical visual image (similar to Stankiewicz et al., 2006). In the Landmarks condition, pictures were placed in the environment such that every state generated a unique view. In the “fog” conditions, fog prevented participants from observing distal hallway structure.
During testing, participants navigated to a goal state trying to make as few actions (rotations and translations) as possible. We calculated navigation efficiency (Number_of_Actions_Ideal/Number_of_Actions_Human). For the current analysis we also included results from Experiment 2 of Stankiewicz et al. (2006).
Landmarks 0.836(SEM=0.027)
Landmarks+Fog 0.724(SEM=0.037)
No-Landmarks 0.598(SEM=0.040)
Stankiewicz et al. (2006)
No-Landmarks 0.604(SEM=0.055)
No-Landmarks+Fog 0.485(SEM=0.038)
To evaluate the benefit of landmarks we calculated the difference in performance in the Landmarks vs. No-Landmarks and the Landmarks+Fog vs. No-Landmarks+Fog conditions. The benefit of landmarks was not significantly different in these conditions (0.239 (Fog); 0.235 (No Fog)). To evaluate the benefit of hallway topology we compared the No-Landmarks vs. No-Landmarks+Fog and the Landmarks vs. Landmarks+Fog conditions. Again, there was no significant difference (0.112 (Landmarks); 0.117 (No Landmarks)). These additive effects suggest that landmarks and hallway topology provide independent sources of information when navigating and orienting.
This research was supported by funding from: AFOSR (FA09550-04-1-0236), AFOSR MURI (FA09550-05-1-0321) and NIH (EY016089).