Uma abordagem para mapeamento e localização simultâneos utilizando informação topológica

Abstract

The exploration of unknown environments is a common topic in mobile robotics. Assuming that the robot does not know its initial position, it is necessary for the robot to be able to navigate through the environment generating a map while it localizes itself on this same map. This process is generally known as SLAM (Simultaneous Localization and Mapping).This dissertation presents a novel approach for augmenting simultaneous localization and mapping (SLAM) with simple plans that guide the exploration of the environment. We use dynamically generated topological maps in conjunction with a utility function to decide which actions the robot should perform in order to improve mapping efficiency.We execute a series of simulated and real experiments in order to evaluate the proposed approach and results show a significant improvement of mapping efficiency both in terms of map quality and execution performance when we compared with a random exploration. Also, we executed experiments to test each term of the utility function, showing the importance of each one to the exploration and the map quality