SBIR-STTR Award

Limitations of many disabled individuals' production efficiency, communication rate and vocabulary organization impact negatively on their ability to communicate competently using present-day Augmentative and Alternative Communication (AAC) devices. This project proposes a reconceptualization in the way language units are organized and retrieved in present-day AAC technologies, by operationalizing recent linguistic research in frame theory. A frame-based organization for utterances is proposed by exploiting natural category systems found in the context, culture and conceptual structures of the individual. A method and system for implementing these knowledge structures on a computer system will be developed. Specifically, the purpose of Phase l of this project is to provide a proof of concept for the Frametalker technology design. During the course of the year we will (a) develop a software prototype of the Frametalker, (b) integrate it with a database of communication frame- structured utterances and (c) submit Frametalker to laboratory and field- testing. Proposed Commercial Applications:Frametalker is targeted at the entire population of augmentative communicators. Although Frametalker will be implemented as a stand-alone device for the STTR effort, its broad functional specification will render it amenable to integration into current AAC technologies (such as Dynavox(TM) and Liberator(TM)), either as the primary communication technique, or in a more supplemental role.

Thesaurus Terms:
biomedical equipment development, clinical biomedical equipment, communication disorder aid, language, speech synthesizer artificial intelligence, computer program /software, computer system design /evaluation, vocabulary medical rehabilitation related tagNATIONAL CANCER INSTITUTE

Continued research into an utterance-based communication aid for persons with communicative disorders is proposed. Frametalker is a patented system for providing AAC device users with context-specific, natural language constructions through a series of dynamically generated communication frames. The Phase I effort comprised incorporation of the Frametalker design into the Windows-based IMPACT augmentative communication package, construction of a prototype database, and completion of various tests to ensure the operability, efficiency, and usability of the invention. The communication rates and selection efficiencies achieved with Frametalker in preliminary field studies surpassed those provided by more traditional augmentative devices. In Phase II, the software will be refined and extended to provide additional input/output methods, handle more general frame specifications, support additional pragmatic shift methods, and perform automatic syntax correction. Dr. John Todman's TALK system and Enkidu Research's Foresight word prediction engine will be integrated into Frametalker to facilitate non-situated and unanticipated conversations. New methods for extending and editing communication frames wilt also be developed. A comprehensive data collection regime will yield an extensive database designed for use by non-speaking adults in employment contexts. Finally, the entire system will be evaluated through a series of field tests performed in actual employment settings