Files in this item



application/pdfECE499-Sp2019-yan.pdf (543kB)Restricted to U of Illinois
(no description provided)PDF


Title:Finger counting in a humanoid robot
Author(s):Yan, Yichong
Contributor(s):Levinson, Stephen E.
Subject(s):bidirectional associative memory, humanoid robot, recurrent neural network
Abstract:Counting is the fundamental mathematic and arithmetic skill, and fingers in particular have been shown to have a strong connection with one’s ability to count. In this thesis, the problem of finger counting is studied on a humanoid robot in an attempt to teach the robot the concept of number through sensorimotor experience. The experiments are conducted on iCub, a humanoid robot platform designed to facilitate research in cognitive development with the emphasis on the idea of embodiment. A human counting model is first constructed, and a similar counting model is built for the robot with bidirectional associative memory (BAM) and recurrent neural network (RNN). The robot learns to count in sequence first and is expected to show basic understanding of numbers through single digit addition and subtraction.
Issue Date:2019-05
Date Available in IDEALS:2019-06-19

This item appears in the following Collection(s)

Item Statistics