Figure 13: Diagram of the software technology that contributed to the presentation of the
House of the Future.
The House of the Future is a unique simulation for it not only interacts with the user as the user interacts with the house, it also is the first time that there have been a combination of various techniques in one virtual reality simulation (figure 13). These techniques include:
Craig Barnes and I designed the software that acts as the controlling mechanism for the lights and other functions. Each controlled object for example, a light, was controlled by a controller which kept track of the proximity of a house tourist using a sensor. Once that person entered within the scope of the sensor, that sensor would
Figure 14: Diagram of the data structure that keeps track of the controller and controlled objects
and their behaviors.
tell the controller to send the appropriate action to the controlled object (figure 14). Such actions for a light include the dimming up or down of a light and the speed of that action. Is this way, lighting control is not just in an all or nothing state. Unfortunately, due to the dimness of the CAVE, such effects are not always obvious to the untrained eye. There are a few different geometries for the sensors, for example, a rectangular or circular range of sensitivity. The controllers, controlled objects, and sensors are preprogrammed into the house using a script which is quite literary and is parsed into the data structure when the house program is run.
A script was developed to control the ambient sound throughout the simulation as well as the ``guide" who told the users about each feature in the room of the house. A unique feature of the simulation, the house itself served as guide and further enforced the interactive feeling between user and environment. The guide, while never seen, interjected his own unique type of humor to the proceedings. Using the script, the sounds were preprogrammed so they did not just turn on or off, but contained timings to control fading the music in and out according to the room. In addition, the script allowed the guide and the ambient sound to overlap when necessary.
Finally, the CAVE to CAVE aspect was more of a brain child of Jason Leigh than myself. He designed the library using variables which could be shared by the multiple machines involved in communication. Here, his goal was keep the messy machine to machine operating system communication details (i.e. using sockets) hidden from the user. The use of the software was successful as evidenced by the touring Jason avatar that awaited the house guests in the activity room of the house. This part of the simulation was executed with Jason Leigh in the ICE lab next door
Figure 15: Photo of Jason on playing his avatar role in the ICE lab during my presentation. He is
viewing the house and tour with a CAVE simulator on a SGI Indigo Extreme workstation.
watching the events on a CAVE simulator. Jason and I communicated using the primitive but effective speaker phone (figure 15).
While developing the house of the future, we implemented a collision detection system that relies on a two dimensional map of the house, read at the initiation of the program to determine wall and door locations. While difficulties with wand navigation made use of the collision detection unwieldy and therefore impractical at the public presentation, it has been further developed and used in architectural environments that evolved from this work. This work will be discussed at the conclusion of this document.
In addition, the techniques used to develop the House of the Future were also innovative. The environment within which to build the house was developed using the Silicon Graphics Inventor library of graphical tools. In order to do this, Jason and I designed a series of commands that would allow the ease of conversion and importing of inventor models into our house environment. After deciding that there just was not enough time to create a CAVE modeler to design the house, we determined that we could use the commercial modeler called Softimage to create three dimensional models which were converted to DXF object format and then inventor. This (along with the fact that I could use Softimage at home) helped facilitate the creation of our environment. Many people have since become interested in these techniques and are currently using them so that they can use inventor objects in the CAVE.
Another tool that was used during development from within the environment was Jason Leigh's heads up Virtual Visor augmented by his input device called the InYerFace. These tools are a programmable set of functions for object manipulation or retrieval of environmental information. The display of this interface is determined using the user's head orientation. It's use was invaluable for finding the correct house coordinates for the placement of the sensors and furniture within the house. The visor interface appeared in front of a user's face and was used to return coordination information about the house.