Return to search

Multirobot Lunar Excavation using an Artificial Neural Tissue Controller

Automated site preparation on the Moon using a group of autonomous rovers is a topic of great interest for the establishment of a lunar base. A potentially very useful system in which multiple, autonomous rovers clear soil to create a landing pad while simultaneously forming berms with the soil cleared will be described. An Artificial Neural Tissue (ANT) architecture was used as the control algorithm to accomplish these tasks. This scalable architecture encourages task decomposition of the main mission tasks and requires minimal human supervision. To solve these tasks, a single fitness function to measure the performance of the controller and a set of allowable basis behaviors was defined. Next, an evolutionary (Darwinian) selection process was used to generate controllers in simulation. The fittest controller was subsequently implemented on LEGO robots for additional validation and testing. The ANT controller was ultimately integrated with a team of three large-scale rovers.

Identiferoai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:OTU.1807/27329
Date30 May 2011
CreatorsFu, Terence Pei
ContributorsD'Eleuterio, Gabriele M. T.
Source SetsLibrary and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada
Languageen_ca
Detected LanguageEnglish
TypeThesis

Page generated in 0.0017 seconds