Towards a Framework for Developing Audio-Centric Interfaces for Field Data Collection Applications

Abstract

Speech technologies (speech recognition and text to speech translation) have greatly improved in the past several years. Audio-centric interfaces built using these technologies provide hands-free or nearly hand-free operation. Unfortunately, it is not always true that a speech interface is an ideal interface in a field environment. In order to take advantage of a speech interface, and all that it offers over other interface modalities, the application developer must have an understanding of available speech technologies, the human operators who will use the speech interface, and the physical and application environments in which such an interface will be employed. A true understanding of the advantages and limitations of speech interfaces can only be obtained by prototyping the system and evaluating it in the field. However, integrating a speech interface into an application is not a trivial task. The process is time-consuming and knowledge intensive because of the complexities associated with the technology. The need exists for a tool that can assist an application developer in integrating a speech interface into an inspection application and thus reduce the prototyping effort and knowledge required during the development process. This paper discusses the need for, and our approach in developing, a framework to support the development of audio-centric interfaces for field data collection applications.