Search

We present a study of the effect of prolonged use on the way users interact with a multimodal form-filling system. The system accepts spoken input as well as pointing input and provides output both in speech and in graphics. We measured the usability of the system in a pre-test / post-test design and analysed in detail the changes in interaction patterns due to exposure. The study shows that with practice users learn to develop interaction patterns that ensure reliable and efficient interaction, resulting in decreased dialogue duration and more user satisfaction.

With the technical advances and market growth in the field, the issues of evaluation and usability of spoken dialogue systems, unimodal as well as multimodal, are as crucial as ever. This chapter discusses those issues by reviewing a series of European and US initiatives which have produced major results on evaluation and usability. Whereas significant progress has been made on unimodal spoken dialogue systems evaluation and usability, the emergence of, among others, multimodal, mobile, and non-task-oriented systems continues to pose entirely new challenges to research in evaluation and usability.

This chapter discusses a series of four user-oriented design analysis problems in a research prototype multimodal spoken language dialogue system for supporting drivers whilst driving. The problems are: (a) when should the system (not) listen to the speech and non-speech acoustics in the car; (b) how to make use of the in-car display in conjunction with spoken driver-system dialogue; (c) how to identify the present driver as a basis for building user models of the driver; and (d) how to create useful online adaptive user modelling of the driver.