Automating the Evaluation of the Scalability, Flexibility, and Robustness of Collective Behaviors for Robot Swarms

• Gabriel Mauricio Madroñero Pachajoa Universidad de Nariño
• Wilson Achicanoy Universidad de Nariño
• David Garzón Ramos University of Bristol

Resumo

Robot swarms are large self-organizing groups of robots that coordinate to collectively perform missions that escape their individual capabilities. In a robot swarm, the individuals operate using local perception and communication, without relying on a leader or external infrastructure for coordination. These characteristics help robot swarms achieve desirable properties, such as scalability, flexibility, and robustness. In this paper, we use recently proposed experimental protocols to evaluate these properties in various typical collective behaviors for robot swarms. We focus on demonstrating that the joint application of defined protocols can help streamline and automate evaluation and reporting processes in swarm robotics research. In our experiments, instances of robot control software for typical collective behaviors are systematically simulated and evaluated across different environmental configurations and swarm sizes. We show how automated tools can facilitate the evaluation of swarm robotics properties.

Palavras-chave: Protocols, Robot kinematics, Scalability, Conferences, Robot control, Education, Swarm robotics, Robustness, Software, Robots, swarm performance indicators, collective behavior, system properties