Reference frames in learning from maps and navigation

Abstract

In everyday life, navigators often consult a map before they navigate to a destination (e.g., a hotel, a room, etc.). However, not much is known about how humans gain spatial knowledge from seeing a map and direct navigation together. In the present experiments, participants learned a simple multiple corridor space either from a map only, only from walking through the virtual environment, first from the map and then from navigation, or first from navigation and then from the map. Afterwards, they conducted a pointing task from multiple body orientations to infer the underlying reference frames. We constructed the learning experiences in a way such that map-only learning and navigation-only learning triggered spatial memory organized along different reference frame orientations. When learning from maps before and during navigation, participants employed a map- rather than a navigation-based reference frame in the subsequent pointing task. Consequently, maps caused the employment of a map-oriented reference frame found in memory for highly familiar urban environments ruling out explanations from environmental structure or north preference. When learning from navigation first and then from the map, the pattern of results reversed and participants employed a navigation-based reference frame. The priority of learning order suggests that despite considerable difference between map and navigation learning participants did not use the more salient or in general more useful information, but relied on the reference frame established first.

Embodied Spatial Knowledge Acquisition in Immersive Virtual Reality: Comparison to Map Exploration

“…This was consistently found after spatial learning that provided a fixed reference frame like cardinal directions of a map or environmental features (Brunyé, Burte, Houck, & Taylor, 2015;Frankenstein et al, 2012;McNamara, 2003;McNamara et al, 2003;Montello et al, 2004). After direct experience in the environment, which might lead to multiple local reference frames (Meilinger, 2008;Meilinger et al, 2015Meilinger et al, , 2006Montello et al, 2004) an alignment effect was much less consistently found (Burte & Hegarty, 2014;Presson & Hazelrigg, 1984). In agreement, our results revealed no alignment effect of tested house orientations towards north after direct experience in the virtual city.…”

“…In line with previous research (Meilinger, Frankenstein, Watanabe, Bülthoff, & Hölscher, 2015), we hypothesized to find a better task accuracy with smaller distances between tested houses (distance effect) with direct experience in Seahaven. As we have to be able to compare the distance between two houses for this analysis, we only considered the relative orientation and the pointing task.…”

“…Investigating the influence of distance between tested houses, previous research indicated that when getting to know a large-scale environment by direct experience smaller distances between tested objects improved task performance (distance effect), whereas after learning with a map no distance effect was found (Frankenstein et al, 2012;Loomis et al, 1993;Meilinger et al, 2015). In contrast, our results revealed the same pattern of distance dependency after direct city exploration in VR and with a map.…”

“…Previous research revealed that landmark, route and survey knowledge may evolve by direct experience in real-world (Montello, 1998;Richardson et al, 1999;Siegel & White, 1975;Thorndyke & Hayes-Roth, 1982), even though there are large individual differences (Hegarty, Burte, & Boone, 2018;Ishikawa & Montello, 2006;Montello, 1998;Wiener, Büchner, & Hölscher, 2009). Direct experience supports route knowledge in relation to an egocentric reference frame whereas spatial learning with a map heightens survey knowledge in relation to an allocentric reference frame (Frankenstein, Mohler, Bülthoff, & Meilinger, 2012;Meilinger et al, 2013;Meilinger et al, 2015;Montello et al, 2004;Richardson et al, 1999;Shelton & McNamara, 2004;Taylor et al, 1999;Thorndyke & Hayes-Roth, 1982). Due to previous research using the feelSpace augmentation device that gives information about magnetic north (Kaspar et al, 2014;König et al, 2016), we designed the tasks investigated in this study with respect to an allocentric reference frame.…”

Qualitative differences in memory for vista and environmental spaces are caused by opaque borders, not movement or successive presentation

“…Such a viewer-spacealignment experienced later during learning was found to be more important for determining reference frame orientation than initial views on a room (Kelly & McNamara, 2008;Shelton & McNamara, 2001;Valiquette & McNamara, 2007). A relatively simple environmental space (e.g., few orthogonally interlinked corridors) may still be represented along a single main axis that spans the entire environmental space (axis presumably aligned with the first vista space encountered; Meilinger, Frankenstein, Watanabe, Bülthoff, & Hölscher, 2015;Tlauka, Carter, Mahlberg, & Wilson, 2011;Wilson, Wilson, Griffiths, & Fox, 2007). However, sufficiently complex environmental spaces seem to be represented within multiple local reference frames, with each local corridor or street occupying a distinct reference frame aligned with the respective corridor (Meilinger et al, 2014;Werner & Schmidt, 1999).…”