Representações topológicas de ambientes assistidas por técnicas de robótica colaborativa

Abstract

Mapping activity is one of the most relevant topics for the field of robotics, as maps provide inputs for carrying out various tasks, such as navigation. Topological maps are one of several ways to represent a scene, and they use computational structures for that. This type of representation presents characteristics that collaborate with its use in the context of swarm robotics, since its agents usually have limited computational resources. Within this context, the present work proposes a topological mapping approach using a multi-agent system to carry out the exploratory process, in which the scene is divided into hexagonal structures that will represent the points of interest on the map and will delimit the movement of the robots. Throughout the work, investigations were carried out regarding the quality of the artifacts produced, in which batteries of tests were carried out to measure the influence of parameters with dimensions of the scene, number of agents and number of obstacles for the correct representation of the environment. When comparing the efficiency of collaboration with a work in the literature that presents a similar technique, the present work presented values up to 58% higher in the proposed evaluation measure, as well as presented results up to 20% higher with the inclusion of new agents. Finally, the work proposed investigations on the efficiency of collaboration during the exploratory process, in which it was possible to conclude that a composition of movement rules stands out from the others and allows a better performance in environments in which all regions to be explored are accessible. to all agents.