Estratégia geral de ondas de mensagens para desenvolvimento de tarefas em enxames de robôs

Abstract

This thesis proposes a general methodology for a systematic approach of tasks in swarms of robots. Swarm robotics investigates collective behaviors in multiple robot systems. The study and development of robotic systems, in general, aims to assist the human being in difficult, heavy or unhealthy jobs. In contrast to robotic systems based on a single robot, the use of robotic swarms still does not translate into everyday applications. One possible explanation is the difficulty in implementing distributed controllers for such applications. Many research works are directed towards the development of techniques to assist in the elaboration of collective tasks. In the message wave strategy proposed in this work, the swarm operates as a network of robots, which interact through wireless communication. A distributed message wave algorithm is designed to be used in robot swarms, serving as the basis for the control of different tasks. In this thesis, the control software for commonly nown as basic tasks of recruitment, alignment, navigation, aggregation, and dispersion, are implemented. The operation of these tasks is evaluated through a simulated swarm of robots. Experiments show that the time required to complete these tasks is dependent on the size of the swarm, the propagation of messages and the characteristics of each task. The design and execution of such tasks serve as proof of concept for the proposed strategy, indicating its generality so as to be used to implement other collective tasks.