While many advances have been made in creating interactive worlds, techniques for human interaction with these worlds lag behind. In order to allow a user to navigate a three dimensional space, most commercial systems encumber the user with head-mounted displays, electro-magnetic or sonic position sensors, gloves, and/or body suits . While such systems can be extremely accurate, they limit the freedom of the user due to the tethers associated with the sensors and displays. Furthermore, the user must don or remove the equipment each time they want to enter or exit the environment. Some systems avoid this problem by passively or actively ``watching'' the user. These systems often modify the environment with specially colored or illuminated backdrops, require the user to wear special clothes, or involve special equipment like range finders or active floor tiles [11, 1, 19].
The ability to enter the virtual environment just by stepping into the sensing area is very important. The users do not have to spend time ``suiting up,'' cleaning the apparatus, or untangling wires. Furthermore, social context is often important when using a virtual environment, whether it be for game playing or designing aircraft. In a head mounted display and glove environment, it is very difficult for a bystander to participate in the environment or offer advice on how to use the environment. With unencumbered interfaces, not only can the user see and hear a bystander, the bystander can easily take the user's place for a few seconds to illustrate functionality or refine the work that the original user was creating. This section describes the methods we use to create such systems.