Third International Conference on Spoken Language Processing (ICSLP 94)
In this paper we describe a system that provides speech-based access to complex information sources over a telephone. The main objective of this work is to provide information access with a simple spoken dialogue interface by integrating various technologies such as speech recognition, speech synthesis, natural language processing and dialogue management. The specific task chosen for feasibility testing of such an integration is getting street-map directions over the telephone. While the task presented in this paper is voice navigation, many of the concepts used in this system can be extended to other complex information access tasks. The prototype system, VoiceMap, requests the user's destination and starting location and provides directions for driving between these locations. All the components of the system are designed to run in a distributed architecture to increase performance. The job of controlling the interactions among the various components is an important aspect of the system. Since each component has different data requirements, a blackboard architecture is used to control flow of information. In this paper each of the components and the information flow among the components are discussed for the VoiceMap application domain.
Bibliographic reference. Bayya, Aruna / Durian, Michael / Meiskey, Lori / Root, Rebecca / Sparks, Randall / Terry, Mark (1994): "Voice map: a dialogue-based spoken language information access system", In ICSLP-1994, 1259-1262.