NIST Awards Funds to Three-Year Project to Bring Speech Technology to Automotive Market

GAITHERSBURG, MD and BOULDER, CO - Four parties formed a joint venture and propose to develop a next-generation interactive natural-language dialogue system (NLDS) that allows drivers to operate various complex in-car devices using conversational speech, thus demanding less cognitive effort and attention from drivers. The project will develop three increasingly sophisticated prototypes. When completed, the system will understand the meaning, tone of voice and intent of a user's speech and respond intelligently.

Technical difficulties to overcome include managing multiple dialogue threads (conversation topics), coping with changing noise levels and noise frequencies and implementing untested strategies for solving various speech-recognition problems. Robert Bosch Corporation will lead this three-year project and bring the technology to the automotive market. Volkswagen of America will demonstrate the functionality of the technology in its vehicles. Stanford University's Center for the Study of Language and Information and SRI International's Speech Technology and Research Laboratory will provide expertise in language technology. Automotive industry practices exclude the use of venture capital or bank financing to support risky multiyear R&D efforts. ATP funding is needed to enable intimate cooperation between the automotive partners and the research institutes that would not otherwise occur. ATP funding also allows research on a larger scale and of a broader scope, and accelerates the research. A recent study shows that operating in-car devices is even more distracting to drivers than using cell phones; therefore, an excellent NLDS can be expected to improve automobile safety. Such technology should stimulate the automotive industry, thereby strengthening a large part of the U.S. economy. The technology, developed in American English, could help U.S. equipment manufacturers retake the lead from foreign competitors in command-operated in-car devices. Additionally, the technology would likely have applications beyond the automotive industry, such as for automated telephone-assistance services and mobile communications.