The purpose of this research is to provide persons who have physical impairment and unintelligible speech with an access method for assistive devices that (i) significantly reduces the physical fatigue experienced during device access and, (2) enables multiple access methods to be used in tandem. This will be accomplished by developing an access method that enables people, who have the ability to reliably produce a small number of distinct vocalizations, to use their speech for command-and-control access to computers, alternate communication devices, and environmental control units. The central element of this Voice Access System will be speech recognition software that will serve as an interface between the user and the assistive device. This software will include a training mode that automatically """"""""learns"""""""" a set of vocalizations that the user can reliably produce. Once the system is trained, it will configure the assistive device for voice access. The Voice Access System will be a full-featured access method for people with disabilities. The opportunity to """"""""mix"""""""" access methods may significantly reduce the fatigue inherent in accessing assistive devices via a single access site. Reducing fatigue is likely to increase the productivity of assistive device users, making device access more rewarding and effective.
The Voice Access System will provide a low-cost (approximately $200), low physical demand, computer access method for people with limited physical and speech capabilities. It is anticipated that the access method will be appropriate for standard computer applications such as word processing, as well as augmentative communication and environmental control application. Long-term application to stand alone assistive devices is likely. The commercial potential is excellent for this market.
