VISUAL FEEDBACK NAVIGATION AND CONTROL OF A QUADROTOR USING FUZZY EXPERT SYSTEMS

摘要

This paper proposes a novel tele-autonomous control method of a quadrotor miniature air vehicle (MAV) that utilizes visual feedback mechanisms. The controller navigates a commercially available quadrotor to different Region-Of-Interests (ROI) as determined by the user. Visual feedback and state estimation is accomplished via vanishing point estimation for corridor centering and navigation. Speeded-Up Robust Feature (SURF) algorithms are used for feature selection, distance estimation, and approaching the ROI. The results of these visual feedback mechanisms are then fed into multiple Fuzzy Expert Systems for navigation and control. Experimental results are presented validating the approach using an unmanned quadrotor in an unknown indoor environment.