Overview of System Architecture

Abstract

Spoken language dialogue systems are evolving from task-specific systems that run on large servers in the public telephone network into personal multi-purpose interaction partners on handheld devices or in the home environment. This requires more flexible system architectures, which can enable multimodal interaction as well as integrate several independently developed applications into one coherent user interface. Together with a powerful set of widely accepted standards for interfaces and application logic, this is the prerequisite for the success of next-generation conversational user interfaces. In this chapter, we analyse the requirements for a system architecture of a multimodal conversational user interface and describe two existing frameworks, the Galaxy Communicator infrastructure and the SmartKom testbed. We also take a short look on emerging and existing standards for spoken language dialogue systems and their multimodal extensions.