The Talk presents a Meta Study of 60 Papers about Social Robots in Museums. The specific scenario "Tour-Guide Robot" provides several challenges, social Robots have to accomplish for everyday life usability (e.g. "sponateneous short-term interaction"). Due to this between 1996 and 2005 more than 25 different Robots were built for museal Use.
I I evalueted these Robots concerning their attributes in Human Robot Interaction:
- How did the constructers modelled the User?
- What was the Project Goal / The Main Task of the Robot
- How could the robot 'interact' with visitors
- What results brought up the implementation in the real life environment Museum?
As conclusion the Presentation introduces three different Examples of Museum Robots to show two main principles for HRI: honesty and deception.
Robots in Museums - An Analysis of Best Practice in HRI
1. Robots in Museums
An Analysis of Best Practice in Human Robot Interaction
Andreas Bischof, M.A. Cultural Science
Graduate School CrossWorlds
University of Technology Chemnitz
06/14/2012
14.06.2012 Andreas Bischof, M.A.
2. Content
1. What Robots? From a Cultural Concept to „Robotics“, 3 Min.
Social Robots as border crossers
2. Robots in Museums Scenario „Tour Guide“; Timeline; 3 Examples: 8 Min.
Pioneer, robust & class winner
3. Designing HRI Having Ideas, lacking Methods; What can HRI 7 Min.
learn from Museums Robots?
4. What‘s up? Open Issues; Project Idea CrossWorlds 2 Min.
14.06.2012 Andreas Bischof, M.A.
3. 1. What Robots?
Cultural Concept “Robot”
What does „Robotics“ do?
Social Robots as border crossers
14.06.2012 Andreas Bischof, M.A.
4. 1. What Robots?
Cultural Concept „Robots“
ANCIENT WORLD
- artificial, man-like, subserving (i.e. Talus)
MIDDLE AGES & RENAISSANCE
- mechanic / automatic (DaVinci, Al-Jazari)
- souled, threatening (Golem)
18th & 19th CENTURY
- androids (Jaquet-Droz, E.T.A. Hoffmann )
- artificial life (Frankenstein‘s Monster)
20th CENTURY
- Robots as Members of Society
- Robots as Replacement for Humans
14.06.2012 Andreas Bischof, M.A.
5. 1. What Robots?
3 Strands from Cultural concepts of „Robots“
14.06.2012 Andreas Bischof, M.A.
6. 1. What Robots?
What is „Robotics“ doing?
14.06.2012 Andreas Bischof, M.A.
7. 1. What Robots?
Social Robots as border crossers
- Informational Gap due to Perspective Problem: users are not experts; creators
are not users
- People make sense about machines:
Theory of Mind
Uncanny Valley
- Robots in social and everyday scenarios are confronted with these cultural
concepts!
14.06.2012 Andreas Bischof, M.A.
8. 2. Robots in Museums
Scenario “Tour Guide”
Lock Back - Timeline
Minerva – The pioneer
Fraunhofer Robots – robust species
RoboX – Class Winner
14.06.2012 Andreas Bischof, M.A.
9. 2. Robots in Museums
Scenario „Tour-Guide Robots“
Challenges:
crowded ‘real’ environment, no laboratory
uninformed, untrained users
set of everyday tasks
short length of use „Spontaneous Short-Term Interaction“
Motivations:
Robots as Demonstrators
Robots as learning objects
Robots as attraction for museal context
Robots as Products
14.06.2012 Andreas Bischof, M.A.
10. 2. Robots in Museums
14.06.2012 Andreas Bischof, M.A.
11. 2. Robots in Museums
Minerva – The Pioneer
14.06.2012 Andreas Bischof, M.A.
12. 2. Robots in Museums
Minerva – The Pioneer
Features making her a Trendsetter:
- on the fly task planning
- constructors aim: „believable agent“ - display internal states to reach attended goals
14.06.2012 Andreas Bischof, M.A.
13. 2. Robots in Museums
Fraunhofer Robots – robust species
14.06.2012 Andreas Bischof, M.A.
14. 2. Robots in Museums
Fraunhofer Robots – robust species
- 40 km / day each
- working since 11 years
- simple Navigation (scratched map)
- simple Presentation (everything fixed)
- limited interactional potential
darlings of the public!
14.06.2012 Andreas Bischof, M.A.
15. 2. Robots in Museums
RoboX – class winner
14.06.2012 Andreas Bischof, M.A.
16. 2. Robots in Museums
RoboX – class winner
secret of success: human-like interactional behaviour
- ‚natural‘ perception (Face, Voice)
- ‚natural‘ expression (facial expression, 7 emotional states)
- includes Reactive Presentation Scenarios
14.06.2012 Andreas Bischof, M.A.
17. 2. Robots in Museums
Comparing
Minerva Fraunhofer-Robots RoboX
built 1997 built 2000 built 2002
Main Task Guiding Attracting Guiding
Mapping live scratched live
Task Planning on the fly fixed on the fly, reactive
Perception obstacle avoidance obstacle avoidance Face, Voice, Buttons
Internal States 4 States 2 States 7 States
Expression Screen, Voice, Face Screen, Voice Screen, Voice, Face
Presentation of Reactive Scenarios:
Content* adapted: "crowded?" initialized: "people?" "yes/no?"
(canned)
14.06.2012 Andreas Bischof, M.A.
18. 3. Designing HRI
Colliding Term “Interaction”
Having ideas, lacking Methods
Results of Inquiry
2 principles
14.06.2012 Andreas Bischof, M.A.
19. 3. Designing HRI
Human Robot Interaction: Colliding Definitions of „Interaction“
Informatics: one-way; Human > Computer
i.e. dragging, pushing, zooming, manipulating
Average Joe: mixture of both; tends to anthropomorphise
technical artifacts
Social Sciences: both ways; Human <> Human (or others?)
i.e. talking, coordination on sidewalk, presenting a paper
14.06.2012 Andreas Bischof, M.A.
20. 3. Designing HRI
Having Ideas, lacking Methods
- a lot of (successfull) ideas in designing HRI, but no methodical reflected
procedures to modell user or specific interaction scenarios up front
- i.e. „What does it mean for the user to be in a Museum?“
- Newer developments: Axiomatic Design, Interaction Design
14.06.2012 Andreas Bischof, M.A.
21. 3. Designing HRI
What is the problem about this HRI trial and error?
What they say… … what happens
We thought… ? … this specific Voice /
Content / Menu would
be approperiate.
The Robot does… … recognizes
person, asks for
aternatives.
14.06.2012 Andreas Bischof, M.A.
22. 3. Designing HRI
Nevertheless they evaluated HRI
(3 out of 60 papers)
- "people are basically destructive" (Nourbakhsh et al. 2002a)
- most attracting: motion and awareness of human presence
- most effective: dialogical interactions, multi modality
- Physical design: height between 1.50 - 1.60 m for interactional robots
- Children are significantly higher attracted by robots than adults
14.06.2012 Andreas Bischof, M.A.
23. Which principles make HRI successfully?
(making the other 57 papers talking)
- Knowing about interactional structures makes robots better interaction
partners
- Form follows function: HCI input can be adopted easily
- Perception of humans in a human way raises potential for adaptive behaviour
(from „something?“ to „somebody!“)
14.06.2012 Andreas Bischof, M.A.
24. 3. Designing HRI
„Honesty“ „Deception“
transparent design: impressive design:
tutorial-welcoming character design
showing (real) states and goals use of professional Storytellers
allowing the user to understand using the (cultural and social pre-
the machine‘s (and it‘s formed) narratives and Typecasts
constructor‘s) intention of everyday-life
both principles can achieve trust (Reliability)!
both principles can lead to Uncanny Valley problems!
14.06.2012 Andreas Bischof, M.A.
25. 4. What‘s up?
Fields to Improve
Questions on our Project
Project Idea CrossWorlds
14.06.2012 Andreas Bischof, M.A.
26. 4. What‘s up?
Fields to improve
(technical & HRI)
Everything concerning human speech
more ‚natural‘ HRI
Multi-User Usability (i.e. age)
Interacting with Web Devices Locomotion Concepts
adjust robot‘s features to the Museum‘s goals
14.06.2012 Andreas Bischof, M.A.
27. Questions for our Project
- Motivation: learning object, demonstrator, museum feature?
- Scenario: What Problem does the robot solves in the Museum?
- Fame: What do we do better than others?
14.06.2012 Andreas Bischof, M.A.
28. 4. What‘s up
Just an Idea…
- foster interactions between visitors
- attach robot features to the museum‘s needs
- foster collaboration between graduate school members
14.06.2012 Andreas Bischof, M.A.