Masters Theses
Abstract
"While considerable progress has been made in recent years toward the development of multirobot teams, much work remains to be done before these teams are used widely in real-world applications. In many real-world problems robots have to move from an initial starting position to the specified location. This thesis proposes to construct an artificial system with autonomous Khepera robots capable of achieving simple collective task by traversing in an unpredictable environment. Robotic agents would be deployed to collectively move to a known area, which is the destination. This could be called collective locomotion. An interesting question is how can such behavior best be evolved? Neural network(s) that act as brain(s) control the robots and these brain(s) evolve interacting with the environment. A powerful method is to coevolve them in separate subpopulations, and test together in the common task.
An attempt is made to model a team of robots collectively moving towards a specified goal; a method called Enforced Subpopulation (ESP) is used. Autonomous agents (robots) and artificial life are typically said to 'learn', 'develop' and 'evolve' in interaction with their environments.
The results of the experiments performed in this work show that coevolution of the robots is necessary for such a task and autonomous control of the robots is more efficient and robust than central control. By implementing neural networks for collective locomotion, this thesis demonstrates the application of evolutionary algorithms in developing evolutionary robotic systems and generating neural networks for collective locomotion. Such a method could potentially be used for evolving neural networks for variety of behaviors in robots"--Abstract, page iii.
Advisor(s)
Dagli, Cihan H., 1949-
Committee Member(s)
St. Clair, Daniel C.
Enke, David Lee, 1965-
Department(s)
Engineering Management and Systems Engineering
Degree Name
M.S. in Engineering Management
Publisher
University of Missouri--Rolla
Publication Date
Spring 2003
Pagination
ix, 68 pages
Note about bibliography
Includes bibliographical references (pages 65-67)
Rights
© 2003 Vikram Aedula, All rights reserved.
Document Type
Thesis - Restricted Access
File Type
text
Language
English
Subject Headings
Robots -- Control systemsRobots -- MotionNeural networks (Computer science)
Thesis Number
T 8201
Print OCLC #
53149934
Recommended Citation
Aedula, Vikram, "Collective behavior in robots using evolutionary neural networks" (2003). Masters Theses. 2290.
https://scholarsmine.mst.edu/masters_theses/2290
Share My Thesis If you are the author of this work and would like to grant permission to make it openly accessible to all, please click the button above.
