Research Directions Areas

1
Research Directions / Areas

• Near-term (most important)
• Integration How do we really put things
  together?
• Organization What roles are appropriate, useful,
  clear?
• Sociology How to understand complexity of HRI in
  the world?
• Safety issues How do we make physical
  interaction work?
• Evaluation How do we know if we are making
  progress?
• Representation How do robots make sense of the
  world?
• Explainability / Fault recovery Being able to
  trust a robot?

2
Research Directions / Areas

• Far-term (top three)
• Standards What do you need to make HRI really
  work?
• Issues related to ontology, representation,
  cognitive skills
• Common and fixed tools
• NOT a standard interface
• Sustained interaction autonomy
• Robots that work for days and months and years
• Long-term relationships
• Mixed-initiative interaction
• Dynamic / adjustable / mixed (choose your term)
• Truly joint goals

3
Roadmap

• Near-term
• Funding, funding, funding !!! (especially for
  Alan)
• Community building
• Common language methodology (so we can talk to
  each other)
• Websites wiki (do these really build
  community?)
• Mailing list
• Journal, conference, etc.
• Sharing
• Common platforms
• Software repository

4
Roadmap

• Long-term
• Funding, funding, funding !!! (especially for
  Alan )
• Need commitment from funding sources
• HRI researchers need to take their turns as
  program managers
• Commercialize what we do
• If it sells, then (perhaps) we did something
  right
• Technology diffusion and acceptance
• HRI Handbook (built from the wiki)
• Things to know
• Lessons learned
• Methods, metrics, etc.

5
Experiments

• How to show achievement in HRI?
• Acceptance by
• End-user
• Organization
• Success on some scale
• Productivity
• Usability
• Sustainability
• Show repeatable results from longitudinal studies

6
Other Fields?

• Disciplines not currently in the fold
• Organizational psychology
• Core computer vision
• Cognitive neuroscience
• Domain experts (e.g., gerontologists)
• How can this be fostered?
• Funding, funding, funding (top-down)
• Targeted funding for interdisciplinary research
• Grass roots (bottom-up)
• Talking to people, giving colloquia, going to
  other conferences
• Require interdisciplinary authors
• Common language

7
Our questions

• to NSF program managers
• What HRI application domains are most important
  to NSF?
• How long will NSF fund HRI research?
• How will NSF put together appropriate review
  panels?
• To which cluster should HRI related proposals be
  submitted?
• to everyone
• What things do we believe are impossible to
  solve?
• What will be a deal-breaker for HRI?
• Why should NSF fund HRI?

8
Breakout B2 Summary

• Scientific Merit
• Standards methodology, tools, education /
  training
• What do you need to make HRI really work ?
• Integration modules, interfaces, methods
• How do we work together ?
• How do we really put things together ?
• Evaluation metrics, measures, procedures,
  analysis
• How do we know if we really are making progress ?
• Show repeatable results from longitudinal studies

9
Breakout B2 Summary

• Broader Impact
• Community building
• Common language
• Common methodology
• Regular meta-level meetings
• Sharing
• Lessons learned
• HRI Handbook
• NSF (long-term) support of HRI will lead to
• Robust interaction in everyday settings
• Supporting the work goals of other government
  agencies
• Opportunities for commercialization

10
Human-Robot Interaction Conference

• 2007
• March 9-11 (Fri-Sun)
• Arlington, Virginia
• http//hri2007.org
• 2008
• mid-March
• Amsterdam, The Netherlands
• Details to come