Desenvolvimento de arquiteturas híbridas através de sistemas computacionais inteligentes aplicados à robótica móvel autônoma

Abstract

This paper aims to present the main contributions developed through mobile robot control architectures (agents) based on Fuzzy Logic and Fuzzy Cognitive Maps, in particular, Hierarchical Weighted Fuzzy (HWF) and Hybrid Dynamic Fuzzy Cognitive Maps (HD-FCM). These architectures will be applied in generating trajectory of a mobile agent, initially in virtual scenarios and a posteriori, in a real robotic platform. The two architectures simulations are compared. Initial results point to the choice of HD-FCM driver to control the robot. As an initial proposal, the robot has the capability to travel between two preset points. Controllers establish a viable path and there may be no obstacles on scenarios. Simulations in virtual environments are used to verify the autonomy of the developed controllers. Results, although still early, with simulations in near ideal conditions, suggest the feasibility of the proposals.