Modeling Pattern Formation in a Swarm of Self-Assembling Robots
Trianni Vito, Labella Thomas Halva, Groß Roderich, Sahin Erol, Dorigo Marco, Deneubourg Jean-Louis

Abstract:

Self-assembly, the self-organized creation of structures composed of independent entities, represents a challenging class of problems for swarm intelligence. In this report, we present some preliminary results on pattern formation carried out by a group of simulated robots. We have defined a control architecture based on probabilistic choices of different basic behaviors. This architecture allows us to define the behavior of the robots, such that their interactions lead the group to form a desired pattern. In order to better understand the dynamics of the pattern formation task, we have applied a methodology composed of the following steps: first, the self-organizing process that leads to the formation of a particular pattern is investigated, producing a suitable mathematical model. Then, systematic experiments are conducted in simulation to estimate the parameters of the model. Finally, the predictions of the mathematical model are compared with the simulation results. This methodology is applied to the chain and cluster formation tasks, and the obtained results are reported.