Distributed online learning of central pattern generators in modular robots

David Johan Christensen, Alexander Spröwitz, Auke Jan Ijspeert

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Abstract

In this paper we study distributed online learning of locomotion gaits for modular robots. The learning is based on a stochastic approximation method, SPSA, which optimizes the parameters of coupled oscillators used to generate periodic actuation patterns. The strategy is implemented in a distributed fashion, based on a globally shared reward signal, but otherwise utilizing local communication only. In a physics-based simulation of modular Roombots robots we experiment with online learning of gaits and study the effects of: module failures, different robot morphologies, and rough terrains. The experiments demonstrate fast online learning, typically 5-30 min. for convergence to high performing gaits (≈ 30 cm/sec), despite high numbers of open parameters (45-54). We conclude that the proposed approach is efficient, effective and a promising candidate for online learning on many other robotic platforms.
Original languageEnglish
Title of host publicationFrom Animals to Animats 11
EditorsR. Goebel, J. Siekmann, W. Wahlster
PublisherSpringer
Publication date2010
Pages402-412
DOIs
Publication statusPublished - 2010
Externally publishedYes
Event11th International Conference on Simulation of Adaptive Behavior - Paris, France
Duration: 25 Aug 201028 Aug 2010
Conference number: 11

Conference

Conference11th International Conference on Simulation of Adaptive Behavior
Number11
CountryFrance
CityParis
Period25/08/201028/08/2010
SeriesLecture Notes in Computer Science
Number6226
ISSN0302-9743

Fingerprint

Dive into the research topics of 'Distributed online learning of central pattern generators in modular robots'. Together they form a unique fingerprint.

Cite this