|
The shrinking size of personal computer systems now makes it possible to wear a functional computer with display goggles available for visualization, but data entry requires the use of keyboards and mice. Our work has focused on eliminating keyboards and joysticks by monitoring the electrical activity in the muscles known as Electromyograms (EMG). We have created software that translates the EMG signals into computer commands such as keystrokes and joystick movement. This software works in a way similar to speech recognition. When a gesture from a fixed set of gestures is performed, the resulting EMG signals are recognized and the computer command corresponding to that gesture is issued to the computer.
We have demonstrated two uses of this technology. The first demonstration involved using the virtual joystick interface to control a class IV flight simulation to land a 757 transport aircraft at San Francisco airport. This activity represents coarser grained movements. The acting pilot reached into the air, pretended to hold and manipulate a joystick. Four movements were used: banking to the left, banking to the right, pitching up, and pitching down. These gestures could be held for an arbitrary amount of time so that rolls could be performed. The strength of the gesture was translated into the rate of the banking and pitching movements.
To demonstrate finer grained gestures we measured finger movements involved in pretending to type on a number pad. We then were able to "type" on the surface of a table and even on our pants. Figure 1 depicts typing on the participant's knee as if it were a numerical keypad, the resulting numbers are shown on the screen.
This technology was demonstrated using movements similar to those that computer users are familiar with from day to day such as typing and moving joysticks. Obviously the ideal interface should evolve away from the qwerty keyboard and into one that is more natural and incorporates speech recognition. However speech alone is not adequate. For example consider rotating and zooming a three dimensional complex object, or numerical entry applications such as inventory. In particular we are focusing on applications specific to NASA such as being able to perform data entry while constrained by a space suit, controlling remote robotics, and focusing on wearable computing applications.
Point of Contact: Kevin Wheeler
(650) 604-3807
kwheeler@mail.arc.nasa.gov
Back To Top
Previous Paper
Return to Pioneer Technology Innovation
Next Paper |
|
Fig. 1. Demonstration of bioelectric keypad using wet electrodes. The left computer screen shows the recognized keystrokes, the right screen shows the raw EMG signals.
|
|
