Estimation of Aerodynamic Derivatives using Neural Network based Method

Abstract

Abstract The paper presents the estimation of lateral-directional aerodynamic derivatives (parameters) using neural based method from real flight data of Hansa-3 aircraft. The conventional methods such as Least Squares (LS) & Maximum Likelihood (ML) require exact model postulation whereas the neural-based method such as Neural-Gauss-Newton (NGN) is an algorithm that utilizes Feed Forward Neural Network & Gauss-Newton optimization to estimate the parameters and does not require a priori postulation of mathematical model or solution of equations of motion. The results obtained in terms of lateral-directional aerodynamic parameters were reasonably accurate to establish NGN with an additional advantage of non-requirement of a priori mathematical model.