by Glenna A. Satalich
Virtual Reality (VR) has been described as "a magical window onto other worlds, from molecules to minds," (Rheingold, 1991). VR been proclaimed to change the way we might learn by the way people visualize and interact with objects. A major component of this visualization is the ability to view the virtual environment from different perspecties. These perspectives include exploring the environment in an egocentric manner, flying above the environment to gain an exocentric viewpoint, or combining the exocentric view with the egocentric view. Obvious questions arise with so much visual orientation potential: Will the participants know where they are in a large-scale virtual environment? Will they know where other objects or locations are? Will they recognize what objects they are looking at? Finally, does experiencing a synthetic environment lead to similar behaviors a person would have in the real world under similar circumstances?
To answer these questions we have to look into the literature of human behavior in navigation and wayfinding in the real world, and compare it with behaviors exhibited in virtual reality. If the behaviors match, then we can apply navigational tools that are enlisted to aid in navigation awareness in the real environments, and apply them to virtual environments. If the behaviors are not alike then the differences must be understood so that appropriate actions can be undertaken to ease the navigation and wayfinding experience for the participant in the virtual environments.
In the real world, wayfinding behaviors can be affected by varying exploration conditions and varying navigational tools (Butler, Acquino, Hissong, and Scott, 1993; Sonnefield, 1985; Goldin and Thorndyke, 1984). Research in this field has been difficult in the past because of lack of control over the environment being examined, the expense in changing the environment, and lack of control over exposure time to the environment (Peponis, Zimring and Choi, 1990). It is because of these reasons that navigation and wayfinding have also been studied using simulations of real and imaginary environments, (Goldin and Thorndyke, 1982). These studies have also been hindered because of the media used, such still photographs, film and drawings (Thorndyke and Hayes-Roth, 1982; Infield, 1991). Peponis et. al., (1990) summarize some of the problems in stating that "direct observation of wayfinding is relatively rare and it is not always clear what is being recorded and what is being analyzed." Because the findings in these domains are problematic, an in depth look at the experimental studies is needed.
Chapter One defines the terms used in this thesis. Chapter Two reviews the literature pertinent to these areas and the metrics used in these studies. Chapters Three and Four will describe a research study conducted in virtual reality and further issues that must be addressed.
Navigational Awareness is defined as having complete navigational knowledge of an environment. There are two distinct types of navigational knowledge of an environment and each type affords different behaviors. The first type of navigational knowledge is called procedural knowledge or route knowledge. Procedural knowledge is ego-referenced and is usually gained by personal exploration of a new area. The characteristics of procedural knowledge are that the navigator can successfully go from one landmark to another on a known route, but does not recognize alternate routes, such as short-cuts. A person who has procedural knowledge may know the approximate distance between the landmarks along the route they traveled. The reason for better judgment of distance for a specific route traveled is because learning is formed by sequential travel (Allen and Kirasic, 1985). Knowledge of relationships of places along this route are unidimensional, a person will be better at recalling when it is in the direction they learned the route (Allen and Kirasic, 1985).
Survey knowledge, the second type of navigational knowledge, is attained by multiple explorations of an environment using multiple routes. Survey knowledge is characterized by the ability to take an exocentric viewpoint and is therefore world referenced. The mental representation of an area is seen from a bird's eye point of view. This is similar to having a mental representation of a physical map (Goldin & Thorndyke, 1982), which is sometimes referred to as a cognitive map. Having procedural knowledge does not guarantee survey knowledge, although it is very probable that this will happen (Moeser, 1988). When survey knowledge is built by personal experience through exploration, it is referred to as a "primary" experience (Presson & Hazelrigg, 1984). Survey knowledge can also be built in a "secondary" manner through map or picture study alone (Goldin & Thorndyke, 1982) (Thorndyke & Hayes-Roth, 1982). Presson & Hazelrigg (1984) and Scholl (1993), have shown survey knowledge gained in this manner is inferior to primary survey knowledge. The inferiority arises in the orientation and location of landmarks. When a person has both procedural knowledge and primary survey knowledge, they have complete navigational awareness. The characteristics of survey knowledge are that distances between, and location of, landmarks are known and routes can be inferred even though they have not been traveled before. The benefits of survey knowledge come into play when procedural knowledge is not sufficient. An example of this is when there is a large traffic back-up on the route someone travels daily. A person with survey knowledge may try a different route with success even though they have never traveled it before.
The research in navigational awareness has shown that to achieve complete navigational knowledge in a new large-space environment, a person has to go through a constructive dynamic process. This process is described in a model from Siegel & White, (1975), called the "Sequential and Hierarchical" model. The first two steps are necessary for procedural knowledge and the third and fifth for complete survey knowledge.
1. Landmark recognition: Objects become landmarks for two reasons; their distinctiveness and personal meaning (Lynch, 1960). The objects can be distinctive for example, because of their architectural style, their size, or color (Weisman, 1981). These distinctive objects become salient landmarks especially when they also give directional information, such as being on a corner of two perpendicular streets. Objects that have personal meaning also become salient landmarks, even though they many not have directional meaning (Infield 1991).
2. Routes or Links: Routes and links are formed when traveling between two landmarks. At this point relative distance between two landmarks on a traveled route is achieved. While forming route knowledge, images and landmarks are recalled, if asked for in the same manner as the route is traveled.
3. "Primary" Survey knowledge: This type of knowledge is achieved after significant traveling of routes and links to the point where alternate routes can be inferred and straight line distances between landmarks can be determined.
4. "Secondary" Survey knowledge: This step is not part of Siegel & White's model, but has been added because partial survey knowledge can be attained. It is also very unclear where secondary survey knowledge falls in this continuum. Secondary survey knowledge involves using only maps to learn about an environment. Distance between landmarks is achieved but their alignment or absolute location may be faulty. When distance or routes are curvy they may be underestimated also (Wickens, 1991). This short-cut to survey knowledge may inhibit further progression until one back tracks and gains procedural knowledge.
5. Chunking of the environment: If the environment is extremely large it becomes necessary to chunk the environment into smaller regions. Landmarks located in different regions maybe susceptible to slight distance distortions (Wilton & Pidcock, 1982) (Steven & Coupe, 1978). These smaller regions then become nested into larger regions, and so on. An example of this type of chunking is; several neighborhoods that could be nested under a town, a county, under a state, under a country. This nesting ability gives us the ability to mentally zoom in and out of representations to give distance or direction to another target. As we zoom out though, we lose granularity of detailed description (Wilton, 1977). According to Infield (1991), the representation that one would use will be the one which is similar to the level of the target space. To transition from one layer of representation to another, a common feature must be recognized in both representational layers. This common feature then becomes the link between the two.
Landmark knowledge is tested through the landmark recognition task, the landmark placement task, and landmark orientation. The landmark recognition task requires a person to distinguish between sights that have and have not been seen during exploration. The landmark placement task involves having subjects place landmarks in their proper position on a map. The orientation task is also referred to as the directional pointing task. This task can be executed in a variety of ways, but the essence of the task is test whether a subject knows where a landmark is in relation to other objects. The angle between where the subject points to and the angle to the landmark is calculated.
Route knowledge is commonly tested in two different manners. The first method is called the route distance estimation task. This involves asking a subject to give an estimation of distance between two objects or between themselves and an object. The second method is called landmark sequencing. The subjects in this type of task would be given pictures of two landmarks located on a route and then asked which landmark would be encountered first.
Survey knowledge is usually tested by the Euclidean distance estimation task. Similar to the route distance estimation task, the Euclidean distance estimation task involves giving distance estimates between two objects or between oneself and an object. The distance this time would be a straight line or "as the crow flies" distance. Survey knowledge can also be measured by having subjects infer a route between two objects if the common route is blocked.
The definition of spatial ability is tenuous at best, as it has been studied for most of this century, and to this date a consensus has not been reached and remains an "ill defined concept" (Pellegrino and Goldman, 1983). What is agreed upon is that spatial ability is comprised of various dimensions. The three major dimensions of spatial ability that are commonly addressed are spatial orientation, spatial visualization and spatial relations (Lohman, 1979). Spatial orientation involves the ability to mentally move or transform stimuli, while retaining their relationships. Spatial orientation also involves the mental manipulation of an object using oneself for reference. Spatial visualization goes further, in that the person can manipulate the relationships within an object. The third dimension, spatial relations, consists of the ability to imagine how an object will align from different perspectives.
Spatial ability is usually measured by psychometric tests. A problem with these tests is that they present problems using small scale objects (Eliot, 1987), whereas most of our interaction with the real world is in large-scale spaces. Small scale refers to an object or environment that can be seen in its entirety from at least one viewpoint. Large-scale spaces must be learned from sequential exposures (Infield, 1991). It was not until recently that one of these spatial ability tests, the Guilford-Zimmerman Spatial Orientation Test was shown to predict performance in a large scale space (Infield, 1991). At present, the other psychometric tests measuring spatial relations and visualization have not been validated for large-scale space performance.
Wayfinding as defined by Gluck (1990) is "the process used to orient and navigate. The overall goal of wayfinding is to accurately relocate from one place to another in a large-scale space". Peponis, et. al., (1990); describe wayfinding as "the ability to find a way to a particular location in an expedient manner and to recognize the destination when reached". Downs and Stea (1973), proposed that this is done in four steps.
1.Orientation: Determining where one is in respect to nearby objects and the target location.
2. Route Decision: Choosing a route that will get one to their destination.
3. Route Monitoring: Monitoring the route one has taken to confirm that one is on the correct route and is going in the right direction.
4. Destination Recognition: Recognizing that one has reached the correct destination, or at least a point nearby.
The first step of this model infers that some landmarks must have been distinguished and selected by the wayfinder. The person knows what the landmarks are, where they are, and their relative position in relation to their own location.
In summary, Spatial Ability is perceiving the environment through our senses, the cognitive process of how we learn our environment, and the relationships between objects. Spatial Awareness is how well we perform in the world, or under experimental conditions using our spatial ability. Navigational Awareness is the result of exploring an environment well enough to have both procedural and survey knowledge of it. Wayfinding is the dynamic process of using our spatial ability and navigational awareness of an environment to reach a desired destination.
Visual access pertains to whether different areas of the building can be seen from other areas. To increase successful navigation a person should be able see parts of the environment from many vantage points (Garling et. al., 1986). Good visual access in a building can be achieved by having an open area that can lead to many other areas of the building. Visual access can also include the ability to see outside the building by using windows or portals. For an outdoor environment this could mean access to a tall building or hills from which to view the environment.
Complexity refers to the layout of the building or environment. A simple layout would be one that has good visual access, right-angle intersections, and few segmented hallways (Peponis, et. al., 1990). A detailed description of how these three navigational cues were incorporated into the environment used in this study can be found in Chapter 3.