Nonlinear system identification of a UAV model with distributed aerodynamics and flexible structure

Abstract

This paper presents the nonlinear system identification of a slightly flexible 25 kg fixed-wing UAV in the time-domain using a computationally efficient distributed aerodynamics model and a linear structural dynamics representation. The equations of motion are formed by making use of the free vibration modes of the structure and the mean axes formulation. The structural modes and mode shapes are determined from ground vibration tests. The distributed aerodynamics, accounting for elastic deformations, are modeled using a quasi-steady stability and control derivative approach and by applying strip theory. Initial distributions for the derivatives are obtained from vortex-lattice-method calculations. For matching the model response to the measured response, parameters for scaling the initial derivative distributions are introduced. The flexible model is subsequently identified based on flight test data using the output error method in the time-domain and maximum-likelihood estimation. A good overall identification result is achieved with a close match of the fast aircraft dynamics. Finally, an evaluation is given on the suitability of the identified model for real-time simulation, loads’ estimation, and active load control law design.