Mobile Robot Intelligence Based SLAM Features Learning and Navigation

Abstract

For efficient, and knowledge based navigation, it is essential to blend mobile robot navigation details with information and details from navigation paths-localities. In this respect, the presented scheme was focused towards building intelligence for mobile robot navigation. Intelligence was achieved by considering the navigation capabilities while the mobile robot was in motion. The adopted learning paradigm was a five layers Neuro-Fuzzy learning architecture, with to ability to create an FIS inference for enhanced navigation. To capture the enormous visual and non-visual sensory data, the mobile robot platform has fully computerinterfaced stereo vision, and reliable 3D perception system onboard the mobile platform. A Neuro-Fuzzy intelligence paradigm was used to learn navigation maps (SLAM) main visual features, distances, nature of localities as it travels within spaces. Blinding intelligence with visual maps and non-visual sensory data, has indeed resulted in improved navigation capabilities.