Empirical aerodynamic modeling for robust control design of an oceanographic Uninhabited Aerial Vehicle

Abstract

This paper demonstrates a systematic procedure of system identification, flight control design and robustness analysis for an Uninhabited Aerial Vehicle (UAV). Unscented Kalman Filter (UKF) is used to estimate the aerodynamic parameters with uncertainty bounds and to update the nonlinear model. A linearized model with parametric uncertainties is extracted from the nonlinear uncertain dynamics of the UAV by a new approach. Next, an accurate, equivalent worst-case gain unmodeled dynamic uncertainty model is constructed for the purpose of simplifying the resulting synthesized controller. The system has to be robust against varying system parameters. Two different robust methodologies named H-infinity and Mu synthesis are adopted for control laws development. The robustness of these two controllers is assessed via real-Mu analysis, using a linear fractional transformation (LFT) model with detailed parametric uncertainties. Finally, the nonlinear model analysis shows system performance under uncertainty perturbation by using Monte-Carlo simulation.