Multi-Objective Optimization for Wind Estimation and Aircraft Model Identification

Abstract

In this paper, a novel method for aerodynamic model identification of a micro-air vehicle is proposed. The principal contribution is a technique of wind estimation that provides information about the existing wind during flight when no air-data sensors are available. The estimation technique employs multi-objective optimization algorithms that utilize identification errors to propose the wind-speed components that best fit the dynamic behavior observed. Once the wind speed is estimated, the flight experimentation data are corrected and utilized to perform an identification of the aircraft model parameters. A multi-objective optimization algorithm is also used, but with the objective of estimating the aerodynamic stability and control derivatives. Employing data from different flights offers the possibility of obtaining sets of models that form the Pareto fronts. Deciding which model best adjusts to the experiments performed (compromise model) will be the ultimate task of the control engineer.