2. Humanoid-Human Interaction
Study of human factors related to the tasking and control of humanoid
robots
Goal is to make communication efficiently, accurately and conveniently
with human
Since humanoid are most of the time heavy and large, safety should be
ensured during communication
Make humanoid suitable for learning and adaptive behavior
2
3. Mode of Communications
Sound recognition and Response
Identifying speech
identifying the area from where the
sound comes
Visual Communications
Gesture identifications
detecting emotions of the sounds
detecting specific commands from
sounds
Detecting Human/Animal faces
Acting according to the image/video
detection
Finding emotions
Show Impression according to the
input received
Identifying and showing different
signs and emotions
Head Pose estimation
Pointing gesture detection
Performing arts such as dancing,
showing special body movements
3
4. Karlsruhe Humanoid-Analysis of a
multimodal communication-1
Perceptual Components
A. Speech recognition
B. Sound event classifications
C. Person localization and tracking
D. Face identification
E. Pointing gesture detection
F. Head-pose estimation
G. Dialogue processing and Fusion
H. Tight coupling of speech and dialogue
processing
Karlsruhe Humanoid Robot
source: Ref [1]
4
5. Karlsruhe Humanoid-Analysis of a
multimodal communication-2
Figure in left
shows the
overview of
the perceptual
components of
Karlsruhe
Humanoid
source: Ref [1]
5
6. Challenges in Humanoid-Human
interaction
Determining which of the incoming visual/audio signals
are relevant to the current tasks
Mapping between bodies
Recognizing success and identifying inadequate actions
Chaining between the actions should be identified
Generalizing the tasks
source: Ref [2]
6
7. Technology-Visual and Audio
Visual
High resolution camera to Kinect
Inventing computer vision techniques
360 degree movable eyes and head
Sound
Signal processing software developed
Better hardware implementation
7
8. Why advancements for Humanoid-
Human interaction?
Social robotics
Humanoid-Human Collaboration
Breaking the communication barrier between human and machine
More situational awareness for humanoid
More task handling capability
8
9. Human-Humanoid Interaction
Asimo
Identifying specific people and
talk accordingly
Identifying speech, tone, order
No facial expression
Showing different signs
Can identify and evaluate task
Can perform Greetings and
handshake
Photo courtesy:
http://phys.org/news2397.html
9
10. Human-Robot Interaction:
Jibo-Social Robotics
Identifying specific people and
talk accordingly
recognizing speech, tone,
order
Showing facial expression
Can identify visual objects
Can identify and evaluate task
Show expressions like human
Photo courtesy:
http://www.technewsworld.com/story/80754.html
10
11. Conclusion
Maximum utilization of a Humanoid robot is
possible only when the interaction with human
is accurate and smooth
11
12. References
12
1. Rainer Stifelhagen, Hazim Kemal Ekenel, Chistian Fuguen, petre gieselmann, hartwig
holzpfel, florian kraft, kai nickel, Micheal voit and Alex waibel “Enabling multimodal human-
Robot interaction for the Karlsruhu Humanoid Robot”. IEEE Transtactions on robotics VOL
23, NO. 5, October, 2007.
2. Rodney A. Brooks, Cynthia Breazeal, Brial Scassellati, Una-May O’reilly, “Technologies for
Human/Humanoid natural interactions”, MiT artificial intelligence laboratory.
3. Cynthia breazeal, Andrew Brooks, David chilongo, Jesse gray, Guy Hoffman, Corry Kidd,
Hans lee, Jeff Lieberman, Andrea lockerd, “Working Collaboratively with humanoid robots”.
MIT media lab, Robotic life group.
4. Hyun S.Yang, Yong-Ho seo, Yeong-nam chae, II-woong jeong, Won-Hyung Kang, ju-ho Lee,
“Design and development of Biped Humanoid Robot, AMI2, for social interaction with
humans”. AIM Lab, EECS dept, Korea advanced institute of science and technology.