Evolving Swarm Behavior for Simulated Spiderino Robots
In swarm robotics research, simulation is often used to avoid the difficulties of designing swarm behavior in the real world. However, designing the controller of swarm members remains a non-trivial task due to the complex interactions between members and the emerging behavior itself. In this paper, FRamework for EVOlutionary design (FREVO) is used as tool for the design, simulation and optimization of swarm behavior for a given problem. This paper demonstrates how FREVO can be used to develop and optimize the behavior of the entire swarm simultaneously by applying an evolutionary algorithm. A case study consisting of twenty robots given the task of gathering near a light source while keeping a minimum distance from each other based solely on local information from simplistic sensors is presented. Considering the need of a fitness function to guide any evolutionary process which is a problem-specific function, the robots' controller is evolved according to a fitness function depending on two factors: the robots' proximity to the light source and the ability to keep a minimum distance between the robots. We examine in particular how the problem can be modeled and how evolution can be applied to create a suitable controller for the swarm members.
Hier werden alle mit dieser Publikation in Zusammenhang stehenden Forschungsaktivitäten angezeigt. Mit dem untenstehenden Link können sie sich diese Forschungsaktivitäten in der Suche anzeigen lassen und gegebenenfalls exportieren.
(Achtung: Externe Aktivitäten werden im Suchergebnis nicht mitangezeigt)