Measuring Spatial Knowledge: Effects of the Relation between Acquisition and Testing

Abstract

Performance in human-robot interaction is related to the operator's mental map of the space in which the robot travels. Accordingly, accurate assessment of mental maps will be important for the design of human-robot interfaces. The present research used a factorial design experiment to examine two methods for acquiring spatial knowledge (reading a map vs. navigating in the space), three methods of testing spatial knowledge (drawing a map, navigating through the space, and estimating point-to-point distances. The results showed that performance in the navigation test was influenced by factors unrelated to the navigated distance, whereas map drawing especially was closely related to the actual distance. Map drawing resulted in a stronger relation between map distance and actual distance in the map training condition than in the navigation training condition. The results are interpreted in terms of transfer appropriate processing, and are applied to human-robot interface design.