creators_name: De Nardi, Renzo
creators_name: Togelius, Julian
creators_name: Holland, Owen
creators_name: Lucas, Simon M.
type: confpaper
datestamp: 2006-10-15
lastmod: 2011-03-11 08:56:39
metadata_visibility: show
title: Evolution of Neural Networks for Helicopter Control: Why Modularity Matters
ispublished: pub
subjects: comp-sci-mach-learn
subjects: comp-sci-neural-nets
subjects: comp-sci-robot
full_text_status: public
abstract: The problem of the automatic development of controllers for vehicles for which the exact characteristics are not known is considered in the context of miniature helicopter flocking. A methodology is proposed in which neural network based controllers are evolved in a simulation using a dynamic model qualitatively similar to the physical helicopter. Several network architectures and evolutionary sequences are investigated, and two approaches are found that can evolve very competitive controllers. The division of the neural network into modules and of the task into incremental steps seems to be a precondition for success, and we analyse why this might be so.
date: 2006
date_type: published
publisher: IEEE Press
refereed: TRUE
citation:   De Nardi, Renzo and Togelius, Julian and Holland, Owen and Lucas, Simon M.  (2006) Evolution of Neural Networks for Helicopter Control: Why Modularity Matters.  [Conference Paper]     
document_url: http://cogprints.org/5222/1/DeNardi2006Evolution.pdf