In some virtual training environments, movement is accomplished using vehicles ranging in complexity from a bicycle to an F15 aircraft. Typically, the purpose of these virtual environments is to train for competence in the operation of the vehicle in the real world.

However, in other instances, operation of a vehicle is unrelated to the training goal. A more appropriate interaction would allow the trainee to walk through the virtual environment, because he or she walks when performing the task in the real world. In these cases, expending mental resources to control a vehicle-based interface may demand attention which would better be devoted to the task to be trained.

Furthermore, for some real-world tasks, we grasp and manipulate objects with our hands while we walk. If the interface requires the use of the hands to control walking, then it can interfere with the control of manual sub-tasks. For these applications, a hands-free, body-operated walking interface is ideal; immersive environments designed to train firefighters or foot-soldiers are good examples.

One solution to a pedestrian interface is a 360-degree treadmill, which would allow the user to interact in the virtual world using real-world kinesthetic motions. We call this approach a "full-motion" interface. In contrast, the HIT Lab has been developing prototypes of "sufficient-motion" interfaces, which allow the user to interact by using a subset of the real-world kinesthetic inputs. Though the ranges of motion are less than full, these inputs are sufficient to convince the user that he or she is moving in the virtual world. Our development of these interface devices is called the Virtual Motion Controller (VMC) Project.

The HIT Lab's VMC working prototype measures body position over the working surface with an arrangement of four weight sensors. The curved working surface provides important feedback to the user about his or her physical location, and therefore body locomotion input to the device. A head-tracker worn by the user measures yaw axis motion of the head, which controls the view orientation. This combination of inputs facilitates natural and independent control of travel direction and gaze direction. Other conceptual prototypes will be capable of operating without the head tracker, working instead with panel-mounted display screens (projector screens).

Currently, we are building virtual worlds in which to evaluate participant performance when using a VMC. We will compare the quality of the user's navigation and spatial awareness when using a VMC to that demonstrated using other interaction techniques.

Also see the VMC Experiment done by Barry Peterson.