University of Surrey

Test tubes in the lab Research in the ATI Dance Research

Bio-inspired self-organizing swarm robotics.

Ramezan Shirazi, Ataollah (2017) Bio-inspired self-organizing swarm robotics. Doctoral thesis, University of Surrey.

[img]
Preview
Text
thesis.pdf - Version of Record
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (31MB) | Preview

Abstract

Swarm robotics is the collaboration of a large number of robots to accomplish a set of specified tasks. It has great potential for the generation of self-organizing adaptive systems, where simple behaviours at agent level result in complex behaviours at swarm level. These systems promise to be robust, flexible, and scalable, and have many innovative applications in the future. Elimination of a central controller and instead relying on local awareness and distributed decision making are the main distinguishing characteristics of such, systems which make them different from classical engineering and necessitate a different design methodology. The challenges to design a control mechanism for self-organizing swarm robotic systems mainly come from the difficulty of mapping between the macroscopic behaviours of a swarm and the microscopic behaviours of its individual agents, and also decision making based on local awareness. Nature presents the best examples of self-organising collective systems. They can be divided into two categories, animal collective behaviours, and cellular organs. Although studying animal collective behaviours paves the way for understanding the principles of self-organizing collective systems, cellular organs show more complex behaviours and structures. The goal of this research is to adapt cellular morphogenesis mechanisms for collective behaviours in a swarm of minimalist robots. The trade of between the size of a swarm and the complexity of involved robots necessitates using simpler and cheaper robots. In addition, miniaturization of robots for future micro-robotic applications require to minimize the number of on-board devices in robots. In this thesis, we focus on developing minimalist algorithms inspired by biological morphogenesis for collective swarm behaviours, including collective flocking, target following, and target enclosure. The proposed algorithms are applicable to highly restricted robots without global positioning, directional sensing, motion feedback, and long-range communication devices. At first, I show how morphogens can retain the integrity and original shape of a swarm of robots without directional sensing, while the swarm moves and interacts with the environment. Then, a coordinated motion strategy is presented in order to preserve connectivity of a real swarm of minimalist robots following a target in their environment. Finally, a new approach is presented for target enclosure with a control over the shape of aggregation around the target. In this approach, a morphogen gradient produced by a target reacts with a second one diffusing through the edge of aggregation in order to spot weak points of the aggregation. The last two algorithms implemented in a real swarm of Kilobots.

Item Type: Thesis (Doctoral)
Divisions : Theses
Authors :
NameEmailORCID
Ramezan Shirazi, Ataollah
Date : 21 December 2017
Funders : Seventh Framework Programme, European Union research and development funding programme, Department of Computer Science, University of Surrey
Contributors :
ContributionNameEmailORCID
http://www.loc.gov/loc.terms/relators/THSJin, YaochuYaochu.Jin@surrey.ac.uk
Depositing User : Ata Shirazi Ramezan Shirazi
Date Deposited : 08 Jan 2018 09:55
Last Modified : 08 Jan 2018 09:55
URI: http://epubs.surrey.ac.uk/id/eprint/844948

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year


Information about this web site

© The University of Surrey, Guildford, Surrey, GU2 7XH, United Kingdom.
+44 (0)1483 300800