A Cerebellar Internal Models Control Architecture for Online Sensorimotor Adaptation of a Humanoid Robot Acting in a Dynamic Environment

Marie Claire Capolei*, Nils Axel Andersen, Henrik Hautop Lund, Egidio Falotico, Silvia Tolu

*Corresponding author for this work

Research output: Contribution to journalJournal articleResearchpeer-review

136 Downloads (Pure)

Abstract

Humanoid robots are often supposed to operate in non-deterministic human environments, and as a consequence, the robust and gentle rejection of the external perturbations is extremely crucial. In this scenario, stable and accurate behavior is mostly solved through adaptive control mechanisms that learn an internal model to predict the consequences of the outgoing control signals. Evidences show that brain-based biological systems resolve this control issue by updating an appropriate internal model that is then used to direct the muscles activities. Inspired by the biological cerebellar internal models theory, that couples forward and inverse internal models into the biological motor control scheme, we propose a novel methodology to artificially replicate these learning and adaptive principles into a robotic feedback controller. The proposed cerebellar-like network combines machine learning, artificial neural network, and computational neuroscience techniques to deal with all the nonlinearities and complexities that modern robotic systems could present. Although the architecture is tested on the simulated humanoid iCub, it can be applied to different robotic systems without excessive customization, thanks to its neural network-based nature. During the experiments, the robot is requested to follow repeatedly a movement while it is interacting with two external systems. Four different internal model architectures are compared and tested under different conditions. The comparison of the performances confirmed the theories about internal models combinatory action. The combination of models together with the structural and learning features of the network, resulted in a benefit to the adaptation mechanism, but also the system response to nonlinearities, noise, and external forces.
Original languageEnglish
JournalIEEE Robotics and Automation Letters
Volume5
Issue number1
Pages (from-to)80-87
Number of pages8
ISSN2377-3766
DOIs
Publication statusPublished - 2020

Keywords

  • Biomimetics
  • Model Learning for Control
  • Neurorobotics
  • Control Architectures and Programming

Cite this

@article{5f23144ffb1d42829bb903e4171a9342,
title = "A Cerebellar Internal Models Control Architecture for Online Sensorimotor Adaptation of a Humanoid Robot Acting in a Dynamic Environment",
abstract = "Humanoid robots are often supposed to operate in non-deterministic human environments, and as a consequence, the robust and gentle rejection of the external perturbations is extremely crucial. In this scenario, stable and accurate behavior is mostly solved through adaptive control mechanisms that learn an internal model to predict the consequences of the outgoing control signals. Evidences show that brain-based biological systems resolve this control issue by updating an appropriate internal model that is then used to direct the muscles activities. Inspired by the biological cerebellar internal models theory, that couples forward and inverse internal models into the biological motor control scheme, we propose a novel methodology to artificially replicate these learning and adaptive principles into a robotic feedback controller. The proposed cerebellar-like network combines machine learning, artificial neural network, and computational neuroscience techniques to deal with all the nonlinearities and complexities that modern robotic systems could present. Although the architecture is tested on the simulated humanoid iCub, it can be applied to different robotic systems without excessive customization, thanks to its neural network-based nature. During the experiments, the robot is requested to follow repeatedly a movement while it is interacting with two external systems. Four different internal model architectures are compared and tested under different conditions. The comparison of the performances confirmed the theories about internal models combinatory action. The combination of models together with the structural and learning features of the network, resulted in a benefit to the adaptation mechanism, but also the system response to nonlinearities, noise, and external forces.",
keywords = "Biomimetics, Model Learning for Control, Neurorobotics, Control Architectures and Programming",
author = "Capolei, {Marie Claire} and Andersen, {Nils Axel} and Lund, {Henrik Hautop} and Egidio Falotico and Silvia Tolu",
year = "2020",
doi = "10.1109/LRA.2019.2943818",
language = "English",
volume = "5",
pages = "80--87",
journal = "IEEE Robotics and Automation Letters",
issn = "2377-3766",
publisher = "IEEE",
number = "1",

}

TY - JOUR

T1 - A Cerebellar Internal Models Control Architecture for Online Sensorimotor Adaptation of a Humanoid Robot Acting in a Dynamic Environment

AU - Capolei, Marie Claire

AU - Andersen, Nils Axel

AU - Lund, Henrik Hautop

AU - Falotico, Egidio

AU - Tolu, Silvia

PY - 2020

Y1 - 2020

N2 - Humanoid robots are often supposed to operate in non-deterministic human environments, and as a consequence, the robust and gentle rejection of the external perturbations is extremely crucial. In this scenario, stable and accurate behavior is mostly solved through adaptive control mechanisms that learn an internal model to predict the consequences of the outgoing control signals. Evidences show that brain-based biological systems resolve this control issue by updating an appropriate internal model that is then used to direct the muscles activities. Inspired by the biological cerebellar internal models theory, that couples forward and inverse internal models into the biological motor control scheme, we propose a novel methodology to artificially replicate these learning and adaptive principles into a robotic feedback controller. The proposed cerebellar-like network combines machine learning, artificial neural network, and computational neuroscience techniques to deal with all the nonlinearities and complexities that modern robotic systems could present. Although the architecture is tested on the simulated humanoid iCub, it can be applied to different robotic systems without excessive customization, thanks to its neural network-based nature. During the experiments, the robot is requested to follow repeatedly a movement while it is interacting with two external systems. Four different internal model architectures are compared and tested under different conditions. The comparison of the performances confirmed the theories about internal models combinatory action. The combination of models together with the structural and learning features of the network, resulted in a benefit to the adaptation mechanism, but also the system response to nonlinearities, noise, and external forces.

AB - Humanoid robots are often supposed to operate in non-deterministic human environments, and as a consequence, the robust and gentle rejection of the external perturbations is extremely crucial. In this scenario, stable and accurate behavior is mostly solved through adaptive control mechanisms that learn an internal model to predict the consequences of the outgoing control signals. Evidences show that brain-based biological systems resolve this control issue by updating an appropriate internal model that is then used to direct the muscles activities. Inspired by the biological cerebellar internal models theory, that couples forward and inverse internal models into the biological motor control scheme, we propose a novel methodology to artificially replicate these learning and adaptive principles into a robotic feedback controller. The proposed cerebellar-like network combines machine learning, artificial neural network, and computational neuroscience techniques to deal with all the nonlinearities and complexities that modern robotic systems could present. Although the architecture is tested on the simulated humanoid iCub, it can be applied to different robotic systems without excessive customization, thanks to its neural network-based nature. During the experiments, the robot is requested to follow repeatedly a movement while it is interacting with two external systems. Four different internal model architectures are compared and tested under different conditions. The comparison of the performances confirmed the theories about internal models combinatory action. The combination of models together with the structural and learning features of the network, resulted in a benefit to the adaptation mechanism, but also the system response to nonlinearities, noise, and external forces.

KW - Biomimetics

KW - Model Learning for Control

KW - Neurorobotics

KW - Control Architectures and Programming

U2 - 10.1109/LRA.2019.2943818

DO - 10.1109/LRA.2019.2943818

M3 - Journal article

VL - 5

SP - 80

EP - 87

JO - IEEE Robotics and Automation Letters

JF - IEEE Robotics and Automation Letters

SN - 2377-3766

IS - 1

ER -