Developing a Dividual Model Using a Modular Neural Network for Human-Robot Interaction

Authors
Toshiyuki Tanaka, Kunikazu Kobayashi
Corresponding Author
Toshiyuki Tanaka
Available Online 1 June 2015.
DOI
https://doi.org/10.2991/jrnal.2015.2.1.9
Keywords
Model of dividual, Human-robot interaction, Robot, Modular neural network, Reinforcement leaning
Abstract
Currently, in the field of human-robot interaction (HRI), robots have a problem that can only interact the same at all times with humans. In this paper, therefore, we introduce the concept called a dividual and build a model of the dividual to grow through interactions with others. In addition, using a modular neural network and reinforcement learning (actor-critic), we confirmed the process to choose an appropriate dividual out of plural dividuals.

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 (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)