Digital Signal Processing Reference
In-Depth Information
Chapter 8
Cognitive Dialog Systems for Dynamic
Environments: Progress and Challenges
Felix Putze and Tanja Schultz
Abstract In this chapter, we present our existing setup and ongoing research on the
development of cognitive dialog systems for dynamic environments like cars,
including the main components that we consider necessary to build dialog systems
to estimate the user's mental processes (hence, cognitive) and adapt their behavior
accordingly. In conducting realistic testing and recording environment to produce
real-life data, a realistic driving simulator was used. We also needed to observe the
user during these interactions in a multimodal way to estimate the current user state
based on this data. This information is integrated with cognitive modeling
components that enrich the observational data. We finally needed a dialog manage-
ment system which is able to use this information for adapting its interaction
behavior accordingly. In this chapter, we report our progress in building these
components, give an overview over the challenges we identified during this work
and the solutions we aim for.
Keywords Cognitive dialog system Cognitive model
Human machine
interaction User state detection
8.1
Introduction
Spoken dialog systems have matured to a point where they find their way to many
real-world applications. However, their application in very dynamic scenarios remains
an open and very interesting task. Spoken dialog systems as an interface for in-car
services are very desirable and at the same time very challenging. On one hand, they
offer eyes-free and hands-free control without visual or manual distraction from the
primary driving task. On the other hand, this task uses the user's cognitive capacity, so
F. Putze (
*
) T. Schultz
Cognitive Systems Lab, University of Karlsruhe, Karlsruhe, Germany
e-mail:
felix.putze@kit.edu
;
tanja@ira.uka.de
Search WWH ::
Custom Search