Trajectory Tracking of a Quadcopter Using Adaptive Neuro-Fuzzy Controller with Sliding Mode Learning Algorithm

Abstract

In this work, the trajectory tracking control of an unmanned aerial vehicle (UAV) has been accomplished using adaptive neuro-fuzzy controllers. The update rules of the proposed controller have been derived based on the sliding mode control theory, where a sliding surface has been generated utilizing the parameters of the neuro-fuzzy controller to direct the error towards zero in a stable manner. To assess the effectiveness of the proposed control scheme, Parrot AR.Drone 2.0 has been utilized as the test platform, on which conventional PID and fuzzy logic controllers have been also implemented to provide means for comparing the performance of the proposed controller. Different reference trajectories have been generated for the real-time experimental studies, in which the discrepancies from these trajectories are used to determine the input signals to be applied to the proposed controllers. The analytical claims have been justified by the obtained results from the real-time experiments in the presence of large nonzero initial errors.