Page 9 - Neural Network Modeling and Identification of Dynamical Systems

PREFACE xi
optimal design of experiments for semiempirical acteristics for it (coefficients of lifting and lat-
models of controlled dynamical systems. eral force, coefficients of angles of roll, yaw and
Chapter 6 presents the results of compu- pitch), obtained as nonlinear functions of sev-
tational experiments obtained in solving real- eral variables, are presented. Then, modeling of
world applied problems related to the simu- the longitudinal trajectory and angular motion
lation of motion and the identification of the of the aircraft is considered, which allows iden-
aerodynamic characteristics of a maneuverable tifying the drag coefficient for it.
aircraft. These results show how efficient the The Appendix presents the results of numer-
semiempirical approach is to modeling nonlin- ous simulation results for adaptive systems of
ear controlled dynamical systems and to solv- the MRAC and MPC types, based on the use of
ing problems of identifying their characteristics. ANN modeling, as applied to aircraft of various
First, the simpler task of modeling the longi- classes.
tudinal short-period motion of a maneuverable
aircraft is considered. Then the results of solv- Yury Tiumentsev, Mikhail Egorchev
ing the problem of modeling the total angular Moscow, Russian Federation
motion of a maneuverable aircraft, as well as December 2018
the problem of identifying aerodynamic char-