by Glenna A. Satalich
The literature on navigation and wayfinding points to research questions that need to be addressed. The first question is whether navigation awareness is best attained by self-exploration of an environment or is better to be actively or passively guided? The second question is what navigational tools would be beneficial to navigational awareness? Previous research has shown that studying a map before entering an environment is helpful for creating secondary survey knowledge, but that secondary knowledge is inferior compared with primary survey knowledge. Previous research has also shown that having a map may or may not be helpful during an exploration period. The third question is what effects do these conditions and navigational tools have on wayfinding in the same environment at a later time?
Figure 1 Bird's Eye View of Building (100ft x 100ft)
The walls, floors, and ceiling were modeled in Alias Studio (version 6), a 3D data modeling and animation software package. The virtual building was assembled and modified using Division's dVISE application software and custom user functions. The height of the building from floor to ceiling was 10 feet. The building contains 39 separate rooms with over 500 objects using approximately 50,000 polygons. Collision detection was incorporated so subjects could not walk through the walls, but was not incorporated for objects located throughout the building. Forty-four sensors were also created in Alias Studio and laid out on the floor. Stepping on a sensor would activate it and then the program would draw only those objects the subject could view from that sensor. This was instituted because Division software by default will draw every object in a world, which causes an increase of lag time between a subject's action and what they see as a result of their action.
Each room was populated using modified objects from Division demos. Paintings on the walls were either created using Adobe Photoshop or downloaded from WWW cites. The world was specifically designed to have some wide open spaces to allow visual access and distinctive areas (e.g., classrooms were located on one side of the building and offices on the other, Figure 3). Paintings and objects were located throughout the halls of buildings to be used as landmarks. No doors were used in this environment, as the interaction to open the doors was not intuitive and unnatural.
The helmet mounted display (HMD) was the Kaiser Electro-Optics Inc., VIM 1000HRpv. The VIM 1000HRpv has a 100deg horizontal by a 30deg vertical field of view. It has a full color multiple active matrix, the LCD's contain 720,00 color elements or 240,000 color groups. The IPD was self adjustable by the participant. The total weight of the helmet was 26 ounces. This HMD is not stereo enabled which did not present a concern for this particular type of study
The subjects moved in the direction their head was pointed. The subject's position and orientation in the world were afforded by the Polhemus Fastrak system, which monitored the position of the subject's head and of a 3-D joystick. A restriction on movement was placed so subjects could only move horizontally and not vertically. The 3-D joystick, also referred to as a wand, was manufactured by Division. The wand provided the subjects' movements in the virtual environment. The wand was configured so that only 3 buttons were active for this experiment. On top of the wand were three buttons laid out horizontally. The left button when pressed moved the subject forward and the right button when pressed moved the subject backwards. The middle button was disabled. On the front of the body of the wand were two buttons arranged vertically. The top button was not used in this study. The bottom button increased the speed of movement forwards or backwards, when pressed in conjunction with one of the directional buttons above.
The experimenter's workstation consisted of a computer keyboard and a VGA computer monitor. The monitor presented the real-time imaging of what the subject was seeing in the virtual environment.
A 3 X 2 X 2 between subject factors design was used (Table 1). The first factor was type of exploration; self-exploration (SE) where the subjects was free to explore the building in any manner they wanted, Active Guided (AG) where the subject would follow a pre-determined path using the wand, and Passive Guided (PG) a guided path where the subject would be moved through the environment at a constant speed with no interaction. (The path layout can be seen in figure 2.) The final condition was the Control group who did not explore the environment but only studied a map of the building.
Table 1 Experimental Design Layout
Self Exploration Active Guided Passive Guided Control (Map/noMap Before) MB nMB MB nMB MB nMB MB Map During N/A no Map During N/A
In the SE condition, subjects were given one half hour to explore the building, moving the environment in any manner they chose. In the AG condition, the subjects were instructed to follow the path drawn on the floor and to not deviate from it. The path was laid in 90 degree angles going into or passing the doorway of every room. The AG subjects were also given one half hour to explore the building. If they finished before the 30 minute time limit they were to begin the tour again. In the PG condition, the subjects were moved along the path at a constant velocity, so that tour took a total time of 30 minutes and 8 seconds. These subjects were allowed to move their head and look in any direction they wanted to, much like a passenger in a car.
Figure 2 Path layout for Active and Passive Guided Tours (begins with the green line)
The second factor was access to a map (figure 3) of the building for five minutes before entering the virtual building. There were two conditions; having a map vs. not having a map.
Figure 3 Representation of the Map given to the Subjects to Study Before they entered the Environment
The third factor is whether or not subjects were given a map during the exploration phase of the environment (figure 4). This map only showed the configuration of the building. The map was attached to the subject's view much like a heads up display. It was located in the lower right field-of-view and had a red crosshair that moved as the participant moved. The crosshair allowed the participant to see their position in the world at all times, but did not show their orientation. The map was a North-up map. Images of how the experimental conditions appeared in the actual virtual environment can be seen in figures 5-8.
Figure 4 Representation of the Map used During Exploration (crosshair showing position)
There was also one control group run in this experiment who performed the same tests as the experimental groups. This control group had only the map to study before entering the environment to be tested, without the benefit of exploration. This group was essential because it provided a baseline against which all the exploration groups could be compared.
Figure 5 Guided Tour & Map During
Figure 6 Self Exploration & Map During
Figure 7 Guided Tour (No Map)
Figure 8 Self Exploration (No Map)
Figure 9 Objects used in the Orientation Task
Figure 10 Distance estimation using self as reference: route and Euclidean
The second route distance estimation task was also given to the subjects after they came out of the environment (figure 11). The subjects were given three pairs of locations and asked to estimate the distance between them (Appendix B). Any estimation within 5 feet was recorded as an exact match. If the subjects reported that they do not know where one or both objects were located they were asked to guess the route distance.
Figure 11 Route distance estimations between 2 objects
The second Euclidean distance estimation task was given to the subjects after they completed the second route distance estimation task (Appendix B). They were asked to give Euclidean distance estimations for three pairs of locations (figure 12). Any estimation within 5 feet was recorded as an exact match. If the subjects reported that they did not know where one or both objects were located they were asked to guess.
Figure 12 Euclidean distance estimations between 2 objects
Figure 13 Path for Wayfinding Task 1 (solid line is the most efficient route)
In the second wayfinding task the subjects were placed back at the entrance door. If they had not been successful in returning in the first wayfinding task. They were then asked to locate another room where they were to imagine a friend was located. They were also told that the building was on fire in certain locations and that they could not walk through the fire (Appendix D, figure 14). The two most efficient routes were blocked by fire, so the subjects had to infer a third route to the location. The path taken, the amount of time and the number of errors were recorded. Again, the subjects had five minutes to find the room. If they were successful they then had five minutes to return to the entrance door.
Figure 14 Path for Wayfinding Task 2 (solid line is the most efficient route)
The subjects were first given a human subject consent form to sign. Following this they were given a detailed timeline of the experiment (Appendix A). The complete experiment took approximately 2 hours. The timeline for those in the experimental groups was as follows
1. Guilford-Zimmerman Test
2. A detailed description of the condition they were in (Appendix A)
3. Training on how to use the apparatus, and how to move in the virtual world
4. Exploration of the virtual building for one-half hour
5. Orientation Task (in VR)
6. Taken out of VR and asked to give Route distance and Euclidean distances between themselves and objects (Appendix B)
7. Asked for route and Euclidean distances between two pairs of objects. (Appendix B)
8. Put back into the virtual building and asked to do the wayfinding tasks. (Appendix C)