CPG Driven RBF Network Control with Reinforcement Learning for Gait Optimization of a Dung Beetle-Like Robot

Matheshwaran Pitchai*, Xiaofeng Xiong, Mathias Thor, Peter Billeschou, Peter Lukas Mailänder, Binggwong Leung, Tomas Kulvicius, Poramate Manoonpong

*Kontaktforfatter for dette arbejde

Publikation: Bidrag til bog/antologi/rapport/konference-proceedingKonferencebidrag i proceedingsForskningpeer review

Resumé

In this paper, we employ a central pattern generator (CPG) driven radial basis function network (RBFN) based controller to learn optimized locomotion for a complex dung beetle-like robot using reinforcement learning approach called “Policy Improvement with Path Integrals (PI $$^2$$ )”. Our CPG driven RBFN controller is inspired by rhythmic dynamic movement primitives (DMPs). The controller can be also seen as an extension to a traditional CPG controller, which usually controls only the frequency of the motor patterns but not the shape. Our controller uses the CPG to control the frequency while the RBFN takes care of the shape of the motor patterns. In this paper, we only focus on the shape of the motor patterns and optimize those with respect to walking speed and energy efficiency. As a result, the robot can travel faster and consume less power than using only the CPG controller.

OriginalsprogEngelsk
TitelArtificial Neural Networks and Machine Learning – ICANN 2019 : Theoretical Neural Computation - 28th International Conference on Artificial Neural Networks, 2019, Proceedings
RedaktørerIgor V. Tetko, Pavel Karpov, Fabian Theis, Vera Kurková
Antal sider13
ForlagSpringer VS
Publikationsdato1. jan. 2019
Sider698-710
ISBN (Trykt)9783030304867
DOI
StatusUdgivet - 1. jan. 2019
Begivenhed28th International Conference on Artificial Neural Networks, ICANN 2019 - Munich, Tyskland
Varighed: 17. sep. 201919. sep. 2019

Konference

Konference28th International Conference on Artificial Neural Networks, ICANN 2019
LandTyskland
ByMunich
Periode17/09/201919/09/2019
NavnLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Vol/bind11727 LNCS
ISSN0302-9743

Fingeraftryk

Radial basis function networks
Reinforcement learning
Robots
Controllers
Energy efficiency

Citer dette

Pitchai, M., Xiong, X., Thor, M., Billeschou, P., Mailänder, P. L., Leung, B., ... Manoonpong, P. (2019). CPG Driven RBF Network Control with Reinforcement Learning for Gait Optimization of a Dung Beetle-Like Robot. I I. V. Tetko, P. Karpov, F. Theis, & V. Kurková (red.), Artificial Neural Networks and Machine Learning – ICANN 2019: Theoretical Neural Computation - 28th International Conference on Artificial Neural Networks, 2019, Proceedings (s. 698-710). Springer VS. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Bind. 11727 LNCS https://doi.org/10.1007/978-3-030-30487-4_53
Pitchai, Matheshwaran ; Xiong, Xiaofeng ; Thor, Mathias ; Billeschou, Peter ; Mailänder, Peter Lukas ; Leung, Binggwong ; Kulvicius, Tomas ; Manoonpong, Poramate. / CPG Driven RBF Network Control with Reinforcement Learning for Gait Optimization of a Dung Beetle-Like Robot. Artificial Neural Networks and Machine Learning – ICANN 2019: Theoretical Neural Computation - 28th International Conference on Artificial Neural Networks, 2019, Proceedings. red. / Igor V. Tetko ; Pavel Karpov ; Fabian Theis ; Vera Kurková. Springer VS, 2019. s. 698-710 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Bind 11727 LNCS).
@inproceedings{904d019f8ac5469e8dd8204d1de30d03,
title = "CPG Driven RBF Network Control with Reinforcement Learning for Gait Optimization of a Dung Beetle-Like Robot",
abstract = "In this paper, we employ a central pattern generator (CPG) driven radial basis function network (RBFN) based controller to learn optimized locomotion for a complex dung beetle-like robot using reinforcement learning approach called “Policy Improvement with Path Integrals (PI $$^2$$ )”. Our CPG driven RBFN controller is inspired by rhythmic dynamic movement primitives (DMPs). The controller can be also seen as an extension to a traditional CPG controller, which usually controls only the frequency of the motor patterns but not the shape. Our controller uses the CPG to control the frequency while the RBFN takes care of the shape of the motor patterns. In this paper, we only focus on the shape of the motor patterns and optimize those with respect to walking speed and energy efficiency. As a result, the robot can travel faster and consume less power than using only the CPG controller.",
keywords = "Artificial neural networks, Brain inspired computing, Reinforcement learning",
author = "Matheshwaran Pitchai and Xiaofeng Xiong and Mathias Thor and Peter Billeschou and Mail{\"a}nder, {Peter Lukas} and Binggwong Leung and Tomas Kulvicius and Poramate Manoonpong",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-3-030-30487-4_53",
language = "English",
isbn = "9783030304867",
pages = "698--710",
editor = "Tetko, {Igor V.} and Pavel Karpov and Fabian Theis and Vera Kurkov{\'a}",
booktitle = "Artificial Neural Networks and Machine Learning – ICANN 2019",
publisher = "Springer VS",

}

Pitchai, M, Xiong, X, Thor, M, Billeschou, P, Mailänder, PL, Leung, B, Kulvicius, T & Manoonpong, P 2019, CPG Driven RBF Network Control with Reinforcement Learning for Gait Optimization of a Dung Beetle-Like Robot. i IV Tetko, P Karpov, F Theis & V Kurková (red), Artificial Neural Networks and Machine Learning – ICANN 2019: Theoretical Neural Computation - 28th International Conference on Artificial Neural Networks, 2019, Proceedings. Springer VS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), bind 11727 LNCS, s. 698-710, 28th International Conference on Artificial Neural Networks, ICANN 2019, Munich, Tyskland, 17/09/2019. https://doi.org/10.1007/978-3-030-30487-4_53

CPG Driven RBF Network Control with Reinforcement Learning for Gait Optimization of a Dung Beetle-Like Robot. / Pitchai, Matheshwaran; Xiong, Xiaofeng; Thor, Mathias; Billeschou, Peter; Mailänder, Peter Lukas; Leung, Binggwong; Kulvicius, Tomas; Manoonpong, Poramate.

Artificial Neural Networks and Machine Learning – ICANN 2019: Theoretical Neural Computation - 28th International Conference on Artificial Neural Networks, 2019, Proceedings. red. / Igor V. Tetko; Pavel Karpov; Fabian Theis; Vera Kurková. Springer VS, 2019. s. 698-710 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Bind 11727 LNCS).

Publikation: Bidrag til bog/antologi/rapport/konference-proceedingKonferencebidrag i proceedingsForskningpeer review

TY - GEN

T1 - CPG Driven RBF Network Control with Reinforcement Learning for Gait Optimization of a Dung Beetle-Like Robot

AU - Pitchai, Matheshwaran

AU - Xiong, Xiaofeng

AU - Thor, Mathias

AU - Billeschou, Peter

AU - Mailänder, Peter Lukas

AU - Leung, Binggwong

AU - Kulvicius, Tomas

AU - Manoonpong, Poramate

PY - 2019/1/1

Y1 - 2019/1/1

N2 - In this paper, we employ a central pattern generator (CPG) driven radial basis function network (RBFN) based controller to learn optimized locomotion for a complex dung beetle-like robot using reinforcement learning approach called “Policy Improvement with Path Integrals (PI $$^2$$ )”. Our CPG driven RBFN controller is inspired by rhythmic dynamic movement primitives (DMPs). The controller can be also seen as an extension to a traditional CPG controller, which usually controls only the frequency of the motor patterns but not the shape. Our controller uses the CPG to control the frequency while the RBFN takes care of the shape of the motor patterns. In this paper, we only focus on the shape of the motor patterns and optimize those with respect to walking speed and energy efficiency. As a result, the robot can travel faster and consume less power than using only the CPG controller.

AB - In this paper, we employ a central pattern generator (CPG) driven radial basis function network (RBFN) based controller to learn optimized locomotion for a complex dung beetle-like robot using reinforcement learning approach called “Policy Improvement with Path Integrals (PI $$^2$$ )”. Our CPG driven RBFN controller is inspired by rhythmic dynamic movement primitives (DMPs). The controller can be also seen as an extension to a traditional CPG controller, which usually controls only the frequency of the motor patterns but not the shape. Our controller uses the CPG to control the frequency while the RBFN takes care of the shape of the motor patterns. In this paper, we only focus on the shape of the motor patterns and optimize those with respect to walking speed and energy efficiency. As a result, the robot can travel faster and consume less power than using only the CPG controller.

KW - Artificial neural networks

KW - Brain inspired computing

KW - Reinforcement learning

UR - http://www.scopus.com/inward/record.url?scp=85072852620&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-30487-4_53

DO - 10.1007/978-3-030-30487-4_53

M3 - Article in proceedings

SN - 9783030304867

SP - 698

EP - 710

BT - Artificial Neural Networks and Machine Learning – ICANN 2019

A2 - Tetko, Igor V.

A2 - Karpov, Pavel

A2 - Theis, Fabian

A2 - Kurková, Vera

PB - Springer VS

ER -

Pitchai M, Xiong X, Thor M, Billeschou P, Mailänder PL, Leung B et al. CPG Driven RBF Network Control with Reinforcement Learning for Gait Optimization of a Dung Beetle-Like Robot. I Tetko IV, Karpov P, Theis F, Kurková V, red., Artificial Neural Networks and Machine Learning – ICANN 2019: Theoretical Neural Computation - 28th International Conference on Artificial Neural Networks, 2019, Proceedings. Springer VS. 2019. s. 698-710. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Bind 11727 LNCS). https://doi.org/10.1007/978-3-030-30487-4_53