Abstract
Programming of complex motor skills for humanoid robots can be a time intensive task, particularly within conventional textual or GUI-driven programming paradigms. Addressing this drawback, we propose a new programming-by-demonstration method called Kinesthetic Bootstrapping for teaching motor skills to humanoid robots by means of intuitive physical interactions. Here, "programming" simply consists of manually moving the robot's joints so as to demonstrate the skill in mind. The bootstrapping algorithm then generates a low-dimensional model of the demonstrated postures. To find a trajectory through this posture space that corresponds to a robust robot motion, a learning phase takes place in a physics-based virtual environment. The virtual robot's motion is optimized via a genetic algorithm and the result is transferred back to the physical robot. The method has been successfully applied to the learning of various complex motor skills such as walking and standing up.
| Original language | English (US) |
|---|---|
| Title of host publication | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
| Pages | 492-499 |
| Number of pages | 8 |
| Volume | 5803 LNAI |
| DOIs | |
| State | Published - 2009 |
| Externally published | Yes |
| Event | 32nd Annual German Conference on Artificial Intelligence, KI 2009 - Paderborn, Germany Duration: Sep 15 2009 → Sep 18 2009 |
Publication series
| Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
|---|---|
| Volume | 5803 LNAI |
| ISSN (Print) | 03029743 |
| ISSN (Electronic) | 16113349 |
Other
| Other | 32nd Annual German Conference on Artificial Intelligence, KI 2009 |
|---|---|
| Country | Germany |
| City | Paderborn |
| Period | 9/15/09 → 9/18/09 |
Fingerprint
ASJC Scopus subject areas
- Computer Science(all)
- Theoretical Computer Science
Cite this
Kinesthetic bootstrapping : Teaching motor skills to humanoid robots through physical interaction. / Ben Amor, Hani; Berger, Erik; Vogt, David; Jung, Bernhard.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5803 LNAI 2009. p. 492-499 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5803 LNAI).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Kinesthetic bootstrapping
T2 - Teaching motor skills to humanoid robots through physical interaction
AU - Ben Amor, Hani
AU - Berger, Erik
AU - Vogt, David
AU - Jung, Bernhard
PY - 2009
Y1 - 2009
N2 - Programming of complex motor skills for humanoid robots can be a time intensive task, particularly within conventional textual or GUI-driven programming paradigms. Addressing this drawback, we propose a new programming-by-demonstration method called Kinesthetic Bootstrapping for teaching motor skills to humanoid robots by means of intuitive physical interactions. Here, "programming" simply consists of manually moving the robot's joints so as to demonstrate the skill in mind. The bootstrapping algorithm then generates a low-dimensional model of the demonstrated postures. To find a trajectory through this posture space that corresponds to a robust robot motion, a learning phase takes place in a physics-based virtual environment. The virtual robot's motion is optimized via a genetic algorithm and the result is transferred back to the physical robot. The method has been successfully applied to the learning of various complex motor skills such as walking and standing up.
AB - Programming of complex motor skills for humanoid robots can be a time intensive task, particularly within conventional textual or GUI-driven programming paradigms. Addressing this drawback, we propose a new programming-by-demonstration method called Kinesthetic Bootstrapping for teaching motor skills to humanoid robots by means of intuitive physical interactions. Here, "programming" simply consists of manually moving the robot's joints so as to demonstrate the skill in mind. The bootstrapping algorithm then generates a low-dimensional model of the demonstrated postures. To find a trajectory through this posture space that corresponds to a robust robot motion, a learning phase takes place in a physics-based virtual environment. The virtual robot's motion is optimized via a genetic algorithm and the result is transferred back to the physical robot. The method has been successfully applied to the learning of various complex motor skills such as walking and standing up.
UR - http://www.scopus.com/inward/record.url?scp=76649103008&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=76649103008&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-04617-9_62
DO - 10.1007/978-3-642-04617-9_62
M3 - Conference contribution
AN - SCOPUS:76649103008
SN - 3642046169
SN - 9783642046162
VL - 5803 LNAI
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 492
EP - 499
BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
ER -