Skip navigation

Distributed intelligent systems for a swarm of robots

Distributed intelligent systems for a swarm of robots

Eissa, Hazem Mohamed Fawzy Zakaria (2021) Distributed intelligent systems for a swarm of robots. PhD thesis, University of Greenwich.

Hazem Eissa 2021.pdf - Published Version
Available under License Creative Commons Attribution.

Download (4MB) | Preview


Area exploration is a task where a robot tries to gain information about an unknown environment. Exploring an unknown area is a challenging task for a group of robots as no pre-made map exists, leading to setting a suitable swarm formation compatible with the area to be explored. Having a suitable swarm formation allows the swarm to preserve the overall exploration time, by distributing sub-tasks for each robot, and collecting relevant data. Current swarm formations such as biologically inspired formations or Probabilistic RoadMap (PRM) tend to have a fixed shape, where robots are positioned in a fixed location point within the swarm, preventing the swarm from adjusting its formation to adapt to the unknown area, thus, are not suitable to explore unknown areas. One needs a more flexible formation, where each robot can change its position within the swarm. Consequently, this research aims to build a distributed robotic swarm formation using fractals.

Fractals have the properties of self-similarity, allowing for an equal distribution of the robots, and recursiveness, allowing for a gradual expansion of a swarm formation. Utilising the properties of fractals allow for a robotic swarm to develop a fractal as a swarm formation. Additionally, changing the parameters of each fractal formation, such as a number of branches, will provide the swarm with the flexibility to adjust the fractal formation and to continue exploring an unknown area. In order to determine both advantages and disadvantages of using fractals as a swarm formation, the first step is to classify each selected fractal into either a line or curve-based formation class to distinguish the similarities and differences in each fractal’s behaviour. The second step is to implement the growth rule of each fractal formation using robots to explore an unknown area. The last step is to study the effect of changing the parameters of the of implemented fractal formations toward exploring unknown areas.

The research’s outcome shows that using fractals as a swarm formation achieved near the amount of area covered by a traditional exploration method, such as PRM, with 88% less use of robots. Furthermore, fractal formations balances between the number of robots used, and the amount of area covered as each fractal uses only the robots needed to develop specific iterations. The effect of changing the parameters of a fractal formation increases the chance of covering more areas.

Item Type: Thesis (PhD)
Uncontrolled Keywords: Mathematical models, robotics
Subjects: Q Science > Q Science (General)
Faculty / School / Research Centre / Research Group: Faculty of Engineering & Science
Faculty of Engineering & Science > School of Engineering (ENG)
Last Modified: 10 Sep 2023 16:58

Actions (login required)

View Item View Item


Downloads per month over past year

View more statistics