Advanced search
1 file | 857.09 KB
Author
Organization
Project
AMARSI (Adaptive Modular Architecture for Rich Motor Skills)
Abstract
Using data retrieved from the Puppy II robot at the University of Zurich (UZH), we show that machine learning techniques with non-linearities and fading memory are effective for terrain classification, both supervised and unsupervised, even with a limited selection of input sensors. The results indicate that most information for terrain classification is found in the combination of tactile sensors and proprioceptive joint angle sensors. The classification error is small enough to have a robot adapt the gait to the terrain and hence move more robustly.
Keywords
proprioception, reservoir, computing, classification, quadruped robot, INDEPENDENT COMPONENT ANALYSIS, terrain

Downloads

  • ICML A 4 .pdf
    • full text
    • |
    • open access
    • |
    • PDF
    • |
    • 857.09 KB

Citation

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

Chicago
Degrave, Jonas, Robin Van Cauwenbergh, Francis wyffels, Tim Waegeman, and Benjamin Schrauwen. 2013. “Terrain Classification for a Quadruped Robot.” In 2013 12TH INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND APPLICATIONS (ICMLA 2013), VOL 1, 185–190. IEEE.
APA
Degrave, J., Van Cauwenbergh, R., wyffels, F., Waegeman, T., & Schrauwen, B. (2013). Terrain classification for a quadruped robot. 2013 12TH INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND APPLICATIONS (ICMLA 2013), VOL 1 (pp. 185–190). Presented at the 12th International Conference on Machine Learning and Applications (ICMLA), IEEE.
Vancouver
1.
Degrave J, Van Cauwenbergh R, wyffels F, Waegeman T, Schrauwen B. Terrain classification for a quadruped robot. 2013 12TH INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND APPLICATIONS (ICMLA 2013), VOL 1. IEEE; 2013. p. 185–90.
MLA
Degrave, Jonas, Robin Van Cauwenbergh, Francis wyffels, et al. “Terrain Classification for a Quadruped Robot.” 2013 12TH INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND APPLICATIONS (ICMLA 2013), VOL 1. IEEE, 2013. 185–190. Print.
@inproceedings{4215673,
  abstract     = {Using data retrieved from the Puppy II robot at the University of Zurich (UZH), we show that machine learning techniques with non-linearities and fading memory are effective for terrain classification, both supervised and unsupervised, even with a limited selection of input sensors. The results indicate that most information for terrain classification is found in the combination of tactile sensors and proprioceptive joint angle sensors. The classification error is small enough to have a robot adapt the gait to the terrain and hence move more robustly.},
  author       = {Degrave, Jonas and Van Cauwenbergh, Robin and wyffels, Francis and Waegeman, Tim and Schrauwen, Benjamin},
  booktitle    = {2013 12TH INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND APPLICATIONS (ICMLA 2013), VOL 1},
  isbn         = {9780769551449},
  keyword      = {proprioception,reservoir,computing,classification,quadruped robot,INDEPENDENT COMPONENT ANALYSIS,terrain},
  language     = {eng},
  location     = {Miami, Florida, USA},
  pages        = {185--190},
  publisher    = {IEEE},
  title        = {Terrain classification for a quadruped robot},
  url          = {http://dx.doi.org/10.1109/ICMLA.2013.39},
  year         = {2013},
}

Altmetric
View in Altmetric
Web of Science
Times cited: