Designing Speech-Controlled Media File Selection for Automotive Systems

Abstract

We present the prototype of an in-car speech-controllable MP3 player. In addition to basic commands such as “next”, “stop”, or “repeat”, one main feature of the system is the selection of titles, artists, albums, or genres by directly uttering them. We first describe the characteristics of automotive speech dialogue systems in general, their technology and design. We then motivate the need to design-implement-evaluate iterations as a proof-of-concept method and describe the design process of the implemented MP3 player dialogue. Finally, we discuss real-world challenges and future extensions for this application. The findings and considerations of the chapter straightforwardly extend to other embedded systems and general audio media.