Zurich University of Applied Sciences Winterthur - PowerPoint PPT Presentation

1 / 21
About This Presentation
Title:

Zurich University of Applied Sciences Winterthur

Description:

Implement on PDA. Concept for Generation of MMUI. COST 278 / H.-P. ... Goals: What does the user want to do? Use Cases, scenarios. How does he want to do it? ... – PowerPoint PPT presentation

Number of Views:25
Avg rating:3.0/5.0
Slides: 22
Provided by: hanspete2
Category:

less

Transcript and Presenter's Notes

Title: Zurich University of Applied Sciences Winterthur


1
Zurich University of Applied Sciences Winterthur
  • Project GENUIN
  • GENeric multimodal User INterface
  • Prof. Dr. H.-P. Hutter
  • Roger Toenz
  • Michel Müller
  • Beat Tscharner

2
Outline
  • Objectives
  • Environments
  • Design Guidelines
  • Sample Application E-Mail-Client
  • Generation of MMUIs

3
Objectives
  • Develop a concept for a generic multimodal user
    interface for mobile applications
  • Supports different environments
  • At least two input and two output modalities
  • The different modalities cooperate and support
    each other.
  • The layout of the different user interface
    modalities adapt to the different interaction
    environments and to the user needs.
  • There is a general description of the user
    interface logic, from which the actual user
    interface is derived for a specific environment.

4
Environments Office
  • Controlled environment (quiet, stable)
  • User pays his full attention to the application
  • Broadband, high-quality interaction facilities
    (normal keyboard, large high-resolution displays)
  • High-quality multimedia facilities

5
Environments Passive Mobile
  • The environment can only partially be controlled
    (noise, movement).
  • The user attention is mainly focused on
    application
  • User sits or stands, possibly in a public
    transport facility
  • Limited varying bandwidth access
  • Small low-to-medium-resolution displays with no
    or small keyboard, pen, mircophone, camera.

6
Environments Active Mobile
  • Same situation as in the passive mobile
    environment
  • But user can only partially devote attention to
    the user interface, i.e., one or both hands are
    busy and/or eyes are busy and/or ears are busy.
  • Typical situation when user is driving a car or
    walks around.

7
Major Design Guidelines
  • 1. Minimize overall task completion time
  • ? Is dependent on user and environment

8
Consequences of Guideline 1
  • The more frequent a task is the lower its
    completion time should be
  • one click / one short command
  • accessible in all views
  • Less frequent tasks allow more indirect access
  • via menus
  • only accessible in some of the views

9
Major Design Guidelines
  • 2. The system must be transparent to the user
  • at any time, in all environments, about
  • what is the current state of the dialogue?
  • what ist the system currently doing?
  • what tasks are currently allowed?
  • what is the state of the different modalities?
  • which commands are currently available?
  • where he can find help

10
Major Design Guidelines
  • The different modalities should support each
    other as much as possible
  • What you see is what you can say
  • What you hear is how you have to say it
  • Your hear what you can see
  • Speech input You see and hear what is
    recognized and what is being done

11
Major Design Guidelines
  • Modalities should be under full user control
  • User can turn off each modality separately
  • User can interrupt speech output by barging-in
    or pen input

12
Sample Application E-Mail-Client
  • Standard E-Mail-Client-Functionality
  • Plus Voice Mails
  • Modalities
  • Pen/wheel/button- Input
  • Graphical output (130x170 Pixel)
  • Speech Input
  • Synth. Speech Output

13
E-Mail-Client MMUI-Design
speech output indicator
Taskbar
speech input indicator (on, off, recognizing,
garbage)
account disconnected
menubar pop-up
active window
direct help, speech help
context menu
preview window
14
Demo...
  • Demonstrator running on a PC
  • Developed with JAVA and JNI
  • Speech Recognition HTK/HAPI
  • Speech Synthesis SVOX

15
Next Steps
  • Collect user experience with prototype
  • Optimize MMUI of Prototype
  • Implement on PDA
  • Concept for Generation of MMUI

16
Implementation on PDA
  • What is needed?
  • robust speech recognition on PDA
  • noise
  • OOV-detection
  • vocabulary independent
  • speech synthesis
  • running on PDA
  • mixed language
  • output speech and phonemic transcription

17
Generation of MMUIs
  • Starting Points
  • Domain Model
  • concepts
  • attributes
  • associations
  • Task Model
  • Goals What does the user want to do?
  • Use Cases, scenarios
  • How does he want to do it? How often?
  • Which alternatives does he need?

18
Concept of MMUI Generation
19
E-Mail-Client Domain Model
20
Example E-Mail-Client Task Model
  • E-Mail-ConceptsAndTasks.xls

21
Questions
  • ?
Write a Comment
User Comments (0)
About PowerShow.com