Advanced search
1 file | 12.10 MB Add to list

Transfer learning of gaits on a quadrupedal robot

Jonas Degrave (UGent) , Michaël Burm (UGent) , Pieter-Jan Kindermans (UGent) , Joni Dambre (UGent) and Francis wyffels (UGent)
(2015) ADAPTIVE BEHAVIOR. 23(2). p.69-82
Author
Organization
Abstract
Learning new gaits for compliant robots is a challenging multi-dimensional optimization task. Furthermore, to ensure optimal performance, the optimization process must be repeated for every variation in the environment, for example for every change in inclination of the terrain. This is unfortunately not possible using current approaches, since the time required for the optimization is simply too high. Hence, a sub-optimal gait is often used. The goal in this manuscript is to reduce the learning time of a particle swarm algorithm, such that the robot’s gaits can be optimized over a wide variety of terrains. To facilitate this, we use transfer learning by sharing knowledge about gaits between the different environments. Our findings indicate that using transfer learning new robust gaits can be discovered faster compared to traditional methods that learn a gait for each environment independently.
Keywords
LOCOMOTION, COMPLIANT, WALKING, transfer learning. PARTICLE SWARM OPTIMIZATION, Quadrupedal robot, evolutionary algorithms, gait optimization, particle swarm optimization

Downloads

  • (...).pdf
    • full text
    • |
    • UGent only
    • |
    • PDF
    • |
    • 12.10 MB

Citation

Please use this url to cite or link to this publication:

MLA
Degrave, Jonas, Michaël Burm, Pieter-Jan Kindermans, et al. “Transfer Learning of Gaits on a Quadrupedal Robot.” ADAPTIVE BEHAVIOR 23.2 (2015): 69–82. Print.
APA
Degrave, J., Burm, M., Kindermans, P.-J., Dambre, J., & wyffels, F. (2015). Transfer learning of gaits on a quadrupedal robot. ADAPTIVE BEHAVIOR, 23(2), 69–82.
Chicago author-date
Degrave, Jonas, Michaël Burm, Pieter-Jan Kindermans, Joni Dambre, and Francis wyffels. 2015. “Transfer Learning of Gaits on a Quadrupedal Robot.” Adaptive Behavior 23 (2): 69–82.
Chicago author-date (all authors)
Degrave, Jonas, Michaël Burm, Pieter-Jan Kindermans, Joni Dambre, and Francis wyffels. 2015. “Transfer Learning of Gaits on a Quadrupedal Robot.” Adaptive Behavior 23 (2): 69–82.
Vancouver
1.
Degrave J, Burm M, Kindermans P-J, Dambre J, wyffels F. Transfer learning of gaits on a quadrupedal robot. ADAPTIVE BEHAVIOR. Sage; 2015;23(2):69–82.
IEEE
[1]
J. Degrave, M. Burm, P.-J. Kindermans, J. Dambre, and F. wyffels, “Transfer learning of gaits on a quadrupedal robot,” ADAPTIVE BEHAVIOR, vol. 23, no. 2, pp. 69–82, 2015.
@article{5815948,
  abstract     = {Learning new gaits for compliant robots is a challenging multi-dimensional optimization task. Furthermore, to ensure optimal performance, the optimization process must be repeated for every variation in the environment, for example for every change in inclination of the terrain. This is unfortunately not possible using current approaches, since the time required for the optimization is simply too high. Hence, a sub-optimal gait is often used. The goal in this manuscript is to reduce the learning time of a particle swarm algorithm, such that the robot’s gaits can be optimized over a wide variety of terrains. To facilitate this, we use transfer learning by sharing knowledge about gaits between the different environments. Our findings indicate that using transfer learning new robust gaits can be discovered faster compared to traditional methods that learn a gait for each environment independently.},
  author       = {Degrave, Jonas and Burm, Michaël and Kindermans, Pieter-Jan and Dambre, Joni and wyffels, Francis},
  issn         = {1059-7123},
  journal      = {ADAPTIVE BEHAVIOR},
  keywords     = {LOCOMOTION,COMPLIANT,WALKING,transfer learning. PARTICLE SWARM OPTIMIZATION,Quadrupedal robot,evolutionary algorithms,gait optimization,particle swarm optimization},
  language     = {eng},
  number       = {2},
  pages        = {69--82},
  publisher    = {Sage},
  title        = {Transfer learning of gaits on a quadrupedal robot},
  url          = {http://dx.doi.org/10.1177/1059712314563620},
  volume       = {23},
  year         = {2015},
}

Altmetric
View in Altmetric
Web of Science
Times cited: