Learning to exploit passive compliance for energy-efficient gait generation on a compliant humanoid
dc.contributor.author | Kormushev, P. | |
dc.contributor.author | Uğurlu, Regaip Barkan | |
dc.contributor.author | Caldwell, D. G. | |
dc.contributor.author | Tsagarakis, N. G. | |
dc.date.accessioned | 2020-09-02T06:19:27Z | |
dc.date.available | 2020-09-02T06:19:27Z | |
dc.date.issued | 2019-01 | |
dc.identifier.issn | 0929-5593 | en_US |
dc.identifier.uri | http://hdl.handle.net/10679/6874 | |
dc.identifier.uri | https://link.springer.com/article/10.1007/s10514-018-9697-6 | |
dc.description.abstract | Modern humanoid robots include not only active compliance but also passive compliance. Apart from improved safety and dependability, availability of passive elements, such as springs, opens up new possibilities for improving the energy efficiency. With this in mind, this paper addresses the challenging open problem of exploiting the passive compliance for the purpose of energy efficient humanoid walking. To this end, we develop a method comprising two parts: an optimization part that finds an optimal vertical center-of-mass trajectory, and a walking pattern generator part that uses this trajectory to produce a dynamically-balanced gait. For the optimization part, we propose a reinforcement learning approach that dynamically evolves the policy parametrization during the learning process. By gradually increasing the representational power of the policy parametrization, it manages to find better policies in a faster and computationally efficient way. For the walking generator part, we develop a variable-center-of-mass-height ZMP-based bipedal walking pattern generator. The method is tested in real-world experiments with the bipedal robot COMAN and achieves a significant 18% reduction in the electric energy consumption by learning to efficiently use the passive compliance of the robot. | en_US |
dc.description.sponsorship | EU project AMARSi | |
dc.language.iso | eng | en_US |
dc.publisher | Springer Nature | en_US |
dc.relation.ispartof | Autonomous Robots | |
dc.rights | openAccess | |
dc.title | Learning to exploit passive compliance for energy-efficient gait generation on a compliant humanoid | en_US |
dc.type | Article | en_US |
dc.description.version | Publisher version | |
dc.peerreviewed | yes | en_US |
dc.publicationstatus | Published | en_US |
dc.contributor.department | Özyeğin University | |
dc.contributor.authorID | (ORCID 0000-0002-9124-7441 & YÖK ID 241209) Uğurlu, Barkan | |
dc.contributor.ozuauthor | Uğurlu, Regaip Barkan | |
dc.identifier.volume | 43 | en_US |
dc.identifier.issue | 1 | en_US |
dc.identifier.startpage | 79 | en_US |
dc.identifier.endpage | 95 | en_US |
dc.identifier.wos | WOS:000456142200005 | |
dc.identifier.doi | 10.1007/s10514-018-9697-6 | en_US |
dc.subject.keywords | Bipedal walking | en_US |
dc.subject.keywords | Energy efficiency | en_US |
dc.subject.keywords | Reinforcement learning | en_US |
dc.subject.keywords | Passive compliance | en_US |
dc.identifier.scopus | SCOPUS:2-s2.0-85041920871 | |
dc.contributor.authorMale | 1 | |
dc.relation.publicationcategory | Article - International Refereed Journal - Institution Academic |
Files in this item
This item appears in the following Collection(s)
Share this page