Quality and Usability

Usability Modelling and Prediction

Optimized usability engineering of interactive services by usability prediction      

Motivation & Project Description

  • State-of-the-art usability measurement methods still involve subjective tests with human participants, which come late in the service development cycle and at a high expense.
  • The aim of this project is to provide usability estimations for multimodal interactive services on the basis of algorithmic models, including
    • Models of human behavior during the interaction
    • Models for predicting usability problems (e.g. ISO 9241)
    • Models for predicting user judgments and hedonistic quality aspects
  • This way, services can be developed much quicker and cheaper.


Models for describing user interaction behavior (input to MeMo ID project)
for VUIs: DA Thorsten Kois, Sept. 2007
for GUIs: BA Andreas Nilkens, Dec. 2007

Deep analysis of conventional usability prediction model (PARADISE, Walker 1997), leading to better modelling approach in the future

MeMo Workbench (together with MeMo)

Installation of "INSPIRE" in SerCho (integration of ASR engine Janus, modality fusion (DA Lionel Maetz, Aug. 2008)) 

Time Frame: 05/2007 - 12/2008

T-labs Team Members: K.-P. Engelbrecht, C. Kühnel, S. Möller, A. Naumann, R. Schleicher, I. Wechsung

Funding by:Deutsche Telekom Laboratories