A Scalable Neuro-inspired Robot Controller Integrating a Machine Learning Algorithm and a Spiking Cerebellar-like Network

Ismael Baira Ojeda, Silvia Tolu, Henrik Hautop Lund

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

508 Downloads (Pure)

Abstract

Combining Fable robot, a modular robot, with a neuroinspired controller, we present the proof of principle of a system that can scale to several neurally controlled compliant modules. The motor control and learning of a robot module are carried out by a Unit Learning Machine (ULM) that embeds the Locally Weighted Projection Regression algorithm (LWPR) and a spiking cerebellar-like microcircuit. The LWPR guarantees both an optimized representation of the input space and the learning of the dynamic internal model (IM) of the robot. However, the cerebellar-like sub-circuit integrates LWPR input-driven contributions to deliver accurate corrective commands to the global IM. This article extends the earlier work by including the Deep Cerebellar Nuclei (DCN) and by reproducing the Purkinje and the DCN layers using a spiking neural network (SNN) implemented on the neuromorphic SpiNNaker platform. The performance and robustness outcomes from the real robot tests are promising for neural control scalability.
Original languageEnglish
Title of host publicationProceedings of Living Machines 2017
PublisherSpringer
Publication date2017
Pages375-386
DOIs
Publication statusPublished - 2017
EventLiving Machines 2017 - Stanford University, Stanford, United States
Duration: 25 Jul 201728 Jul 2017

Conference

ConferenceLiving Machines 2017
LocationStanford University
CountryUnited States
CityStanford
Period25/07/201728/07/2017
SeriesLecture Notes in Computer Science
Volume10384
ISSN0302-9743

Keywords

  • Neuro-robotics
  • Bio-inspiration
  • Motor control
  • Cerebellum
  • Machine Learning
  • Compliant control
  • Internal model

Cite this

Baira Ojeda, I., Tolu, S., & Lund, H. H. (2017). A Scalable Neuro-inspired Robot Controller Integrating a Machine Learning Algorithm and a Spiking Cerebellar-like Network. In Proceedings of Living Machines 2017 (pp. 375-386). Springer. Lecture Notes in Computer Science, Vol.. 10384 https://doi.org/10.1007/978-3-319-63537-8 31
Baira Ojeda, Ismael ; Tolu, Silvia ; Lund, Henrik Hautop . / A Scalable Neuro-inspired Robot Controller Integrating a Machine Learning Algorithm and a Spiking Cerebellar-like Network. Proceedings of Living Machines 2017. Springer, 2017. pp. 375-386 (Lecture Notes in Computer Science, Vol. 10384).
@inproceedings{a568835d2b774a14824b2444a1f7062d,
title = "A Scalable Neuro-inspired Robot Controller Integrating a Machine Learning Algorithm and a Spiking Cerebellar-like Network",
abstract = "Combining Fable robot, a modular robot, with a neuroinspired controller, we present the proof of principle of a system that can scale to several neurally controlled compliant modules. The motor control and learning of a robot module are carried out by a Unit Learning Machine (ULM) that embeds the Locally Weighted Projection Regression algorithm (LWPR) and a spiking cerebellar-like microcircuit. The LWPR guarantees both an optimized representation of the input space and the learning of the dynamic internal model (IM) of the robot. However, the cerebellar-like sub-circuit integrates LWPR input-driven contributions to deliver accurate corrective commands to the global IM. This article extends the earlier work by including the Deep Cerebellar Nuclei (DCN) and by reproducing the Purkinje and the DCN layers using a spiking neural network (SNN) implemented on the neuromorphic SpiNNaker platform. The performance and robustness outcomes from the real robot tests are promising for neural control scalability.",
keywords = "Neuro-robotics, Bio-inspiration, Motor control, Cerebellum, Machine Learning, Compliant control, Internal model",
author = "{Baira Ojeda}, Ismael and Silvia Tolu and Lund, {Henrik Hautop}",
year = "2017",
doi = "10.1007/978-3-319-63537-8 31",
language = "English",
series = "Lecture Notes in Computer Science",
publisher = "Springer",
pages = "375--386",
booktitle = "Proceedings of Living Machines 2017",

}

Baira Ojeda, I, Tolu, S & Lund, HH 2017, A Scalable Neuro-inspired Robot Controller Integrating a Machine Learning Algorithm and a Spiking Cerebellar-like Network. in Proceedings of Living Machines 2017. Springer, Lecture Notes in Computer Science, vol. 10384, pp. 375-386, Living Machines 2017, Stanford, United States, 25/07/2017. https://doi.org/10.1007/978-3-319-63537-8 31

A Scalable Neuro-inspired Robot Controller Integrating a Machine Learning Algorithm and a Spiking Cerebellar-like Network. / Baira Ojeda, Ismael; Tolu, Silvia; Lund, Henrik Hautop .

Proceedings of Living Machines 2017. Springer, 2017. p. 375-386 (Lecture Notes in Computer Science, Vol. 10384).

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

TY - GEN

T1 - A Scalable Neuro-inspired Robot Controller Integrating a Machine Learning Algorithm and a Spiking Cerebellar-like Network

AU - Baira Ojeda, Ismael

AU - Tolu, Silvia

AU - Lund, Henrik Hautop

PY - 2017

Y1 - 2017

N2 - Combining Fable robot, a modular robot, with a neuroinspired controller, we present the proof of principle of a system that can scale to several neurally controlled compliant modules. The motor control and learning of a robot module are carried out by a Unit Learning Machine (ULM) that embeds the Locally Weighted Projection Regression algorithm (LWPR) and a spiking cerebellar-like microcircuit. The LWPR guarantees both an optimized representation of the input space and the learning of the dynamic internal model (IM) of the robot. However, the cerebellar-like sub-circuit integrates LWPR input-driven contributions to deliver accurate corrective commands to the global IM. This article extends the earlier work by including the Deep Cerebellar Nuclei (DCN) and by reproducing the Purkinje and the DCN layers using a spiking neural network (SNN) implemented on the neuromorphic SpiNNaker platform. The performance and robustness outcomes from the real robot tests are promising for neural control scalability.

AB - Combining Fable robot, a modular robot, with a neuroinspired controller, we present the proof of principle of a system that can scale to several neurally controlled compliant modules. The motor control and learning of a robot module are carried out by a Unit Learning Machine (ULM) that embeds the Locally Weighted Projection Regression algorithm (LWPR) and a spiking cerebellar-like microcircuit. The LWPR guarantees both an optimized representation of the input space and the learning of the dynamic internal model (IM) of the robot. However, the cerebellar-like sub-circuit integrates LWPR input-driven contributions to deliver accurate corrective commands to the global IM. This article extends the earlier work by including the Deep Cerebellar Nuclei (DCN) and by reproducing the Purkinje and the DCN layers using a spiking neural network (SNN) implemented on the neuromorphic SpiNNaker platform. The performance and robustness outcomes from the real robot tests are promising for neural control scalability.

KW - Neuro-robotics

KW - Bio-inspiration

KW - Motor control

KW - Cerebellum

KW - Machine Learning

KW - Compliant control

KW - Internal model

U2 - 10.1007/978-3-319-63537-8 31

DO - 10.1007/978-3-319-63537-8 31

M3 - Article in proceedings

T3 - Lecture Notes in Computer Science

SP - 375

EP - 386

BT - Proceedings of Living Machines 2017

PB - Springer

ER -

Baira Ojeda I, Tolu S, Lund HH. A Scalable Neuro-inspired Robot Controller Integrating a Machine Learning Algorithm and a Spiking Cerebellar-like Network. In Proceedings of Living Machines 2017. Springer. 2017. p. 375-386. (Lecture Notes in Computer Science, Vol. 10384). https://doi.org/10.1007/978-3-319-63537-8 31