On-line aircraft parameter identification using fourier transform regression with an application to NASA F/A-18 harv flight data

Abstract

This paper applies a recently developed on-line parameter identification (PID) technique to sets of real flight data and compares the results with those of a state-of-the-art off-line PID technique. The on-line PID technique takes Linear Regression from Fourier Transformed equations and the off-line PID is based on the traditional Maximum Likelihood method. Sets of flight data from the NASA F/A-18 High Alpha Research Vehicle (HARV) aircraft, which has been recorded from specifically designed maneuvers and used for off line parameter estimation, are used for this study. The emphasis is given on the accuracy and on-line measure of reliability of the estimates. The comparison is performed for both longitudinal and lateral-directional dynamics for maneuvers at angles of attack ranging α=20δ through α=40δ. Results of the two estimation processes are also compared with baseline wind tunnel estimates whenever possible.