TY - GEN
T1 - Application of an Artificial Neural Tissue controller to multirobot lunar ISRU operations
AU - Thangavelautham, Jekanthan
AU - Smith, Alexander
AU - Boucher, Dale
AU - Richard, Jim
AU - D'Eleuterio, Gabriele M.T.
PY - 2007
Y1 - 2007
N2 - Automation of mining and resource utilization processes on the Moon with teams of autonomous robots holds considerable promise for establishing a lunar base. We present an Artificial Neural Tissue (ANT) architecture as a control system for autonomous multirobot tasks. An Artificial Neural Tissue (ANT) approach requires much less human supervision and pre-programmed human expertise than previous techniques. Only a single global fitness function and a set of allowable basis behaviors need be specified. An evolutionary (Darwinian) selection process is used to train controllers for the task at hand in simulation and is verified on hardware. This process results in the emergence of novel functionality through the task decomposition of mission goals. ANT based controllers are shown to exhibit self-organization, employ stigmergy (communication mediated through the environment) and make use of templates (unlabeled environmental cues). With lunar in-situ resource utilization (ISRU) efforts in mind, ANT controllers have been tested on a multirobot resource gathering task in which teams of robots with no explicit supervision can successfully avoid obstacles, explore terrain, locate resource material and collect it in a designated area by using a light beacon for reference and interpreting unlabeled perimeter markings.
AB - Automation of mining and resource utilization processes on the Moon with teams of autonomous robots holds considerable promise for establishing a lunar base. We present an Artificial Neural Tissue (ANT) architecture as a control system for autonomous multirobot tasks. An Artificial Neural Tissue (ANT) approach requires much less human supervision and pre-programmed human expertise than previous techniques. Only a single global fitness function and a set of allowable basis behaviors need be specified. An evolutionary (Darwinian) selection process is used to train controllers for the task at hand in simulation and is verified on hardware. This process results in the emergence of novel functionality through the task decomposition of mission goals. ANT based controllers are shown to exhibit self-organization, employ stigmergy (communication mediated through the environment) and make use of templates (unlabeled environmental cues). With lunar in-situ resource utilization (ISRU) efforts in mind, ANT controllers have been tested on a multirobot resource gathering task in which teams of robots with no explicit supervision can successfully avoid obstacles, explore terrain, locate resource material and collect it in a designated area by using a light beacon for reference and interpreting unlabeled perimeter markings.
KW - Collective robotics
KW - Developmental systems
KW - Evolutionary algorithms
KW - ISRU
KW - Neural networks
UR - http://www.scopus.com/inward/record.url?scp=77957975193&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=77957975193&partnerID=8YFLogxK
U2 - 10.1063/1.2437478
DO - 10.1063/1.2437478
M3 - Conference contribution
SN - 9780735403864
T3 - AIP Conference Proceedings
SP - 389
EP - 399
BT - Space Technology and Applications International Forum, STAIF 2007, including Co-located Conferences
T2 - Space Technology and Applications International Forum: Space Renaissance: Inspiring the Next Generation, STAIF-2007
Y2 - 11 February 2007 through 15 February 2007
ER -