Improving clarification dialogs in speech command systems with the help of user modeling: A conceptualization for an in-car user interface

Abstract

Modern cars contain increasingly complex driver information systems. To handle such a system in a safe and comfortable way new approaches for in-car user interfaces have to be developed. A special focus lies on speech input because it promises the most comfortable operation of the system without distracting the driver from traffic. But because speech input is not reliable enough, an error management for handling recognition errors is needed. This paper introduces a framework for an in-car user interface, presents requirements for an error management and an approach based on user modeling with Bayesian networks.