Abstract

Asset protection is a behavior in which a team of robots establishes a formation around a resource marked as an asset in a hostile environment in order to protect the asset from threats. The robots are assumed to be homogeneous and run a decentralized control algo- rithm and possess a repulsive quality to the threats. Previous works in this area have used centralized control or considered the use of many robots. This work aims at developing an algorithm that is both decentralized, and able to protect assets using only a few robots. In order to provide this behavior an algorithm coined the Asset Guarding Intelligent System (AeGIS), was developed and analyzed. Using AeGIS, each robot will detect an asset move towards it and form a protective formation around it. AeGIS utilizes Quadratic Artificial Potential Fields (QAPFs) as the robot's path planning module. As such the fields are designed to move the robots into formation, avoid collisions, and in turn protect assets. AeGIS is tested using Leviathan -- an event-driven simulator designed to test groups of autonomous swarm robots employing distributed control algorithms. The success rate of different variations of AeGIS were tested. Additionally, the number of threats, robots employing AeGIS, and the number and mobility of assets were varied to observe their effect on the success rate. The simulation results show that with sufficient number of robots, the assets, static or mobile are well protected against 20 modeled threats. Through these results it is shown that AeGIS is a solution to the asset protection problem.

Library of Congress Subject Headings

Robots--Control systems; Swarm intelligence

Publication Date

2-1-2010

Document Type

Thesis

Department, Program, or Center

Computer Engineering (KGCOE)

Advisor

Yang, Shanchieh

Comments

Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: TJ211.35 .L37 2010

Campus

RIT – Main Campus

Download

Share

COinS