Aplicação da iPNRD com feromônio de formiga para controle autônomo e distribuído de robôs móveis.

Abstract

Swarm robotics has several challenges in controlling social behavior when specific goals need to be achieved. Biologically inspired approaches, such as ant colony algorithms, are at the top of solutions that do not use artificial intelligence. In these algorithms local information scattered around the environment is used instead of a centralized local database. This thesis proposes a solution for the autonomous and distributed control of terrestrial robots through ant pheromone emulation. To this end, the iPNRD approach, which uses the data structure-based RFID element database as a Petri net is implemented. In this approach the trigger vector is saved in the environment tags, while the marking vector and the incidence matrix are saved in the robot reader. During tag data acquisition the robot updates its marking vector, changing its behavior. Robot-environment communication is performed through pseudoboxes, which represent control information external to the components of Petri nets. For experimental validation, a leaf cutting ant-based robot (Attabot) is designed, and for the simulation environment the V-REP platform was used. Operational tests indicate that the proposed method can coordinate ground robots to perform a specific task without any centralized control. This requires that the environment be modeled and have intelligence.