A Combination of Machine Learning and Cerebellar-like Neural Networks for the Motor Control and Motor Learning of the Fable Modular Robot

Authors
Ismael Baira Ojeda, Silvia Tolu, Moisés Pacheco, David Johan Christensen, Henrik Hautop Lund
Corresponding Author
Ismael Baira Ojeda
Available Online 1 June 2017.
DOI
https://doi.org/10.2991/jrnal.2017.4.1.14How to use a DOI?
Keywords
Motor control, cerebellum, machine learning, modular robot, internal model, adaptive behavior.
Abstract
We scaled up a bio-inspired control architecture for the motor control and motor learning of a real modular robot. In our approach, the Locally Weighted Projection Regression algorithm (LWPR) and a cerebellar microcircuit coexist, in the form of a Unit Learning Machine. The LWPR algorithm optimizes the input space and learns the internal model of a single robot module to command the robot to follow a desired trajectory with its end-effector. The cerebellar-like microcircuit refines the LWPR output delivering corrective commands. We contrasted distinct cerebellar-like circuits including analytical models and spiking models implemented on the SpiNNaker platform, showing promising performance and robustness results.

Copyright
© 2013, the Authors. Published by ALife Robotics Corp. Ltd.
Open Access
This is an open access article distributed under the CC BY-NC license