Quadrotor UAV control : online learning approach

Abstract

Quadrotor unmanned aerial vehicles (UAVs) are recognized to be capable of various tasks including search and surveillance for their agilities and small sizes. This thesis proposes a simple and robust trajectory tracking controller for a quadrotor UAV utilizing online Iterative Learning Control (ILC) that is known to be effective for tasks performed repeatedly. Based on a nonlinear model which considers basic aerogynamic and gyroscopic effects, the quadrotor UAV model is simulated to perform a variety of maneuvering such as take-off, landing, smooth translation and horizontal and spatial circular trajectory motions, PD online ILCs wirh switching gain (SPD ILCs) are studies, tested and compared. Simulation results prove the ability of the online ILCs to successfully perform certain missions in the presence of considerably large disturbances and SPD ILCs can obtain faster convergence rates.