Navegação em robôs moveis por arbitragem e fusão em arquiteturas comportamentais

Abstract

Mobile robotics is an area that involves aspects of automated reasoning, perception and control. Navigation is a crucial area of study in autonomous robots. The present work explores the use of behavior-based architecture using two existing strategies in the literature: arbitration and fusion. The first was implemented using hybrid control (discrete event systems with continuous variable dynamic systems), while the second uses Fuzzy controllers for its execution. The prototype uses the differential drive model and a conversion to the unicycle model was carried out in order to make the solution scalable, because if other robots could have their models converted to the unicycle model, the control system of this work can control them. The achieved results were satisfactory, however, under the arbitration strategy, it was verified that the decision process (automaton) is sensitive to the presence of noise. Under the strategy of fusion, the robot showed more aggressive oscillations caused by the subsumption architecture.