Development of Robust Control Laws for Disturbance Rejection in Rotorcraft UAVs

Inherent stability inside the flight envelope must be guaranteed in order to safely introduce private and commercial UAV systems into the national airspace. The rejection of unknown external wind disturbances offers a challenging task due to the limited available information about the unpredictable and turbulent characteristics of the wind. This thesis focuses on the design, development and implementation of robust control algorithms for disturbance rejection in rotorcraft UAVs. The main focus is the rejection of external disturbances caused by wind influences. Four control algorithms are developed in an effort to mitigate wind effects: baseline nonlinear dynamic inversion (NLDI), a wind rejection extension for the NLDI, NLDI with adaptive artificial neural networks (ANN) augmentation, and NLDI with L1 adaptive control augmentation. A simulation environment is applied to evaluate the performance of these control algorithms under external wind conditions using a Monte Carlo analysis. Outdoor flight test results are presented for the implementation of the baseline NLDI, NLDI augmented with adaptive ANN and NLDI augmented with L1 adaptive control algorithms in a DJI F330 Flamewheel quadrotor UAV system. A set of metrics is applied to compare and evaluate the overall performance of the developed control algorithms under external wind disturbances. The obtained results show that the extended NLDI exhibits undesired characteristics while the augmentation of the baseline NLDI control law with adaptive ANN and L1 output-feedback adaptive control improve the robustness of the translational and rotational dynamics of a rotorcraft UAV in the presence of wind disturbances