Technical Specs

home_button.JPG (4171 bytes)

VoiCE.JPG (15082 bytes)

Technical Specs
(Overview)

Overview § System Concept § System Architecture
Voice Recognition Processors § HW/SW Components § Functional Specs CS410

		The Voice Controlled Entry (VoiCE) project will integrate "robust" computer speech recognition and automated door technology. VoiCE will apply research on robust computer speech recognition such as that done at the Oregon Grad Inst. by Ronald A. Cole (http://www.nsf.gov, 1999) and the Sphinx Project at Carnegie Mellon University (Acero, 1993, p.17-36). Noisy environments such as the inside of a moving car or the exterior of a home on a busy street require robust speech recognition. The VoiCE system will be speaker dependent and will use a small vocabulary. The vocabulary will use one word commands necessary to open, lock, and unlock the door. The system will be trained to these commands as spoken by specific individuals. The system will not need to recognize commands as spoken by random individuals.
	We are seeking project funding from the National Science Foundation's (NSF) Robotics and Human Augmentation directorate. The target consumers will be the disabled and elderly. This project will combine speech, sensor input/output, electro-mechanical door components, and decision making algorithms that will decide what door actions to initiate. The final designs will include both standalone and PC based systems.