A controlled object dynamical properties adjustment using artificial neural network

Аuthors

Morozov N. I., Tiumentsev Y. V.^*, Yakovenko A. V.

Moscow Aviation Institute (National Research University), 4, Volokolamskoe shosse, Moscow, А-80, GSP-3, 125993, Russia

*e-mail: yutium@gmail.com

Abstract

Artificial neural network (ANN) capabilities are analyzed for some problems associated with aircraft motion control. A compact and effective representation is built for nonlinear aircraft motion model using ANN. The representation is used to adjust dynamical properties for aircraft considered as controlled object. Desired aircraft properties are described by nonlinear reference model. Adjusting controller is synthesized to minimize the difference between the controlled system (control object + adjusting controller) behavior and the reference model. An approach offered is tested computationally by solving the longitudinal short-period aircraft motion control problem. Results obtained demonstrate high efficiency of the approach.