Research Cluster 7

Measuring and forecasting the quality of multi-modal telecommunications products/services

The work in Research Cluster 7 deals with the design of multi-modal telecommunication services, like entertainment portals, ambient intelligence applications and secure portals for authentification and specific administrative purposes. One of the goals here is to measure, and by means of simulation to predict, both the quality and the usability of such applications, without having to rely too much on the efforts of human participants used in the research. In such an undertaking, certain domain characteristics (i.e. large non-homogeneous user-groups and fast development cycles) must be considered and addressed by adequate methods.

This Research Cluster is supervised by Prof. Dr.-Ing. Sebastian Möller

Currently supervised colleagues/scholarship holders and projects:

Modeling of Modality Selection in Multimodal Human-Computer Interaction

Stefan Schaffer, M.A.

1st Supervisor: Prof. Dr.-Ing. Sebastian Möller
2nd Supervisor: Prof. Dr. Manfred Thüring

In this research project by means of cognitive architectures models for the simulation of multimodal human-computer interaction are formalized, implemented and validated for various prototypical applications. The resulting user models will be deployed for the usage within model based evaluation tools. Concretely, the issue of selection and preference of particular input modalities is investigated using prototypical applications of Deutsche Telekom Laboratories. Established tools for the model based evaluation of unimodal systems will be extended by aspects of multimodal interaction. The so provided methods allow to carry out the rating of quality und usability of multimodal system models in early stages of product development, were no prototype is available.

More Information in German {$ext.ml_links.fileSymbol} (202 KB)

Poster der 8. Berliner Werkstatt MMS {$ext.ml_links.fileSymbol} (862 KB)

Influence of System Feedback on the Usability of Multimodal Human-Machine Systems

Julia Seebode, M.A.

1st Supervisor: Prof. Dr.-Ing. Sebastian Möller
2nd Supervisor: Prof. Dr. Hartmut Wandke

Aim of this research is the development of methods to measure the subjectively perceived quality of multimodal interaction, with special interest in methods that are likely to evaluate aspects beyond conventional usability measure (e.g. hedonic aspects). Main focus of the research project is system feedback in different modalities like nonverbal auditory information as well as speech, tactile and graphical feedback. As a first step experiments are conducted for the different kinds of unimodal feedback to establish a suitable experimental paradigm. Based on the experiments certain feedback messages will be selected and implemented in a multimodal mobile prototype to study the influence of those feedback messages on the interaction of users with such systems and the perceived quality. Methods developed with this approach serve as a tool to collect user ratings and interaction data, which can be used for quality prediction in future.

More Information in German {$ext.ml_links.fileSymbol} (14 KB)

Poster der 8. Berliner Werkstatt MMS {$ext.ml_links.fileSymbol} (402 KB)

Design and evaluation of mobil multimodal interactions

Dipl.-Math. Ralf Reichmuth

1st Supervisor: Prof. Dr.-Ing. Sebastian Möller

One important field of application of multimodal interactions are mobile telecommunication services like the ones used on smartphones and tablet-computers. To ensure a high quality of use, user-centered prototyping and development cycles are essential. Besides that knowledge on how users use input- and output modalities and what kind of modality they prefer is important. Central aspects here are the mobile application context and associated influence factors, such as purpose of use, surrounding conditions and distraction by parallel activities. Therefore, this dissertation project focuses on new methods to design and evaluate multimodal apps with the help of typical apps used in a university context.

Auswirkungen von verzögertem Systemfeedback im mobilen Kontext (Bachelorarbeit)
Arne Denneler
Julia Seebode, M.A. | FSP 7

Modalitätenwahl bei multimodaler Interaktion {$ext.ml_links.fileSymbol} (1.2 MB)
Michael Minge
Stefan Schaffer, M.A.; Prof. Dr. Manfred Thüring; FSP 7