|
|
|
|
|
• By Date • By Publication Type • By Research Category • By Author (Currently at CMU) • By Author (Past and Current CORAL Contributors) •
Sonia Chernova and Manuela Veloso. An Evolutionary Approach To Gait Learning For Four-Legged Robots. In In Proceedings of IROS'04, Sendai, Japan, September 2004.
Developing fast gaits for legged robots is a difficult task that requires optimizing parameters in a highlyirregular, multidimensional space. In the past, walk optimization for quadruped robots, namely the Sony AIBO robot, wasdone by hand-tuning the parameterized gaits. In addition to requiring a lot of time and human expertise, this processproduced sub-optimal results. Several recent projects have focused on using machine learning to automate the parametersearch. Algorithms utilizing Powell's minimization method and policy gradient reinforcement learning have shownsignificant improvement over previous walk optimization results. In this paper we present a new algorithm for walkoptimization based on an evolutionary approach. Unlike previous methods, our algorithm does not attempt to approximatethe gradient of the multidimensional space. This makes it more robust to noise in parameter evaluations and avoidsprematurely converging to local optima, a problem encountered by both of the previously suggested algorithms. Ourevolutionary algorithm matches the best previous learning method, achieving several different walks of high quality.Furthermore, the best learned walks represent an impressive 20\% improvement over our own best hand-tuned walks.
@inproceedings{Chernova04iros,
title="An Evolutionary Approach To Gait Learning For Four-Legged Robots",
author="Sonia Chernova and Manuela Veloso",
booktitle="In Proceedings of IROS'04",
place="Sendai, Japan", month="September",
year="2004",
abstract={Developing fast gaits for legged robots is a difficult task that requires optimizing parameters in a highly
irregular, multidimensional space. In the past, walk optimization for quadruped robots, namely the Sony AIBO robot, was
done by hand-tuning the parameterized gaits. In addition to requiring a lot of time and human expertise, this process
produced sub-optimal results. Several recent projects have focused on using machine learning to automate the parameter
search. Algorithms utilizing Powell's minimization method and policy gradient reinforcement learning have shown
significant improvement over previous walk optimization results. In this paper we present a new algorithm for walk
optimization based on an evolutionary approach. Unlike previous methods, our algorithm does not attempt to approximate
the gradient of the multidimensional space. This makes it more robust to noise in parameter evaluations and avoids
prematurely converging to local optima, a problem encountered by both of the previously suggested algorithms. Our
evolutionary algorithm matches the best previous learning method, achieving several different walks of high quality.
Furthermore, the best learned walks represent an impressive 20\% improvement over our own best hand-tuned walks.},
bib2html_pubtype = {Refereed Conference},
bib2html_rescat = {RoboCup Publications, Robot Learning, Motion Control}
bib2html_dl_pdf = {http://www.cs.cmu.edu/~soniac/files/ChernovaVelosoIROS04.pdf},
}
Generated by bib2html.pl (written by Patrick Riley ) on Tue Oct 09, 2007 00:00:13
|
|
 | |||
| [home] [news] [people] [robosoccer] [projects] [multimedia] [download] [publications] [contact us] | |||