Author

Daniel Stella

Abstract

Swarm robotics pertains to the use of many simplistic robots working together to complete complex tasks. These robots behave autonomously and collaboratively by utilizing local sensors and/or communication mechanisms. To investigate behaviors of large robot swarms, simulations are often used, yet mostly assume unrealistic 2D terrain. While representing the exact physical terrain is ideal, it may not be desirable and necessary to do so for the understanding of robot behavior. To this end, this work aims at developing an efficient terrain model specifically compatible with robot swarms. In order to reduce physical terrain into a manageable data set, Triangulated Irregular Network (TIN) is chosen to capture the key factors for robot swarms, namely, the terrain slopes, breaks, and the robot line of sight. Since a TIN is a reduced data set composed of separate triangular planes, each plane contains its own slope and a shared edge between planes can signify a terrain break. This representation presents the key factors without overly complicated calculations. These features are designed so that the individual robots can visualize the terrain using Artificial Potential Fields (APF), a common technique used for robot swarms. The terrain model is implemented in both Matlab and Java to demonstrate (1) how individual robots perceive other objects and the terrain, and (2) the computational complexity and storage overhead. Creating TIN incurs O(n2) computational complexity, where n is the number of triangles. The complexity overhead to have all robots examine the terrain and surrounding robot is O(m(d1.5 + l*sqrt(d), where m is the number of robots, d is the TIN density and l is the object density. The results show that the proposed method is sustainable for generic computers.

Library of Congress Subject Headings

Autonomous robots--Motion--Computer simulation; Mobile robots; Swarm intelligence; Relief models--Data processing; Digital mapping

Publication Date

4-1-2010

Document Type

Thesis

Advisor

Not listed

Comments

Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works in December 2013. Physical copy available through RIT's The Wallace Library at: TJ211.49 .S74 2010

Campus

RIT – Main Campus

Share

COinS