Search Strategies and their Success in a Virtual Maze

Search Strategies and their Success in a Virtual Maze Simon J. Buechner (buechner@cognition.uni-freiburg.de) Christoph Holscher (hoelsch@cognition.uni-freiburg.de) Jan M. Wiener (mail@jan-wiener.net) University of Freiburg Center for Cognitive Science, Friedrichstrase 50 79098 Freiburg, Germany Abstract The study investigates strategies for search tasks in a virtual environment. Participants had to find and retrieve an object in an unfamiliar environment. Trajectories as well as search performance were analyzed with respect to three applicable strategies when the target was hidden at two different locations. Prior knowledge about the environment was varied within participants and between environments. The results show a strong preference for the perimeter strategy. Prior knowledge improved search performance only to some extent. Keywords: spatial cognition, navigation, search strategy, virtual environments Introduction Imagine you enter an environment and you are searching for an object. What would be a good strategy to quickly find it? How successful would that strategy be depending on the exact location of the object? The selection of the strategy as well as your search performance will certainly depend on whether you have been to the environment before or not. Although the search for objects is a task we frequently encounter, surprisingly little is known about the strategies people apply when performing tasks like this. Strategies A variety of studies investigated the strategies humans apply when navigating between locations (e.g. Golledge, 1995; Hochmair & Frank, 2002; Conroy Dalton, 2003; Wiener, Schnee & Mallot, 2004; Buchner, Holscher & Strube, 2007). In all these studies participants had some information about the location of the target place. Either they have been there before, they could permanently see it, or they could infer the location from other information such as background knowledge about the functional organization of buildings. The studies listed above have been conducted in the context of wayfinding (in its literal meaning) - participants had to find a way from one location to another having some knowledge about the target location. In a search task as we define it here, in contrast, the agent has no information about the location of the target and thus, strategies may play an even more important role for the success and efficiency of the search. Different scales of space may require different search strategies. When searching for an object in a room sized environment in which all potential target locations are visible, people may apply different strategies than in a building in which walls block visibility. Montello (1993) conceptualized space of different scales. He distinguished between vista spaces, environmental spaces, and geographical spaces. A Vista space is the space that is visible from a single observation point (e.g. a room). An Environmental space is a space in which an agent has to move in order to comprehend its entire structure (e.g., a maze or a building). Geographical spaces are even larger and include, for example, landscapes and other spaces that can only be experienced with some transportation device. Search strategies have been subject to empirical studies in different scales of space. Ruddle, Payne and Jones (1999) asked participants to search for nine objects in a virtual seascape. Participants were ‘flying’ over the seascape experiencing the world from a bird’s eye view. The authors found that in such a geographical space people tend to use an anchor strategy. They start searching at a visible object (in this case an island) and then move along the edge of that object? The authors also observed the lawn mower strategy (searching the area in regular back-and-forth pattern from one end to the other). In vista spaces, search strategies have been investigated for visually impaired people and blindfolded participants (Tellevik, 1992; Hill & Rieser, 1993). Tellevik (1992) identified three major strategies: grid-line, perimeter and reference-point. The grid-line strategy is equivalent to the lawn-mower strategy, i.e. people followed a regular pattern from one wall to the other, slowly moving from one end of the room to the other. The perimeter strategy involves a path along the outside walls of the room and a subsequent exploration of the center. When following the reference- point strategy people select a reference-point (often the starting location), explore one part of the space, return to the reference-point, explore the next part of the space and so on. A similar strategy has also been observed for navigation in abstract spaces like the World Wide Web: Users of websites often navigate between ‘hub’ page and several subordinate pages (Cockburn & McKenzie, 2001). Most of the aforementioned strategies, however, are only applicable in spaces in which the observer can oversee the whole space. This is different in environmental spaces in which walls and other objects limit visibility and impose constraints on the ways people move through the