Page 172 - Neural Network Modeling and Identification of Dynamical Systems

162 4. NEURAL NETWORK BLACK BOX MODELING OF AIRCRAFT CONTROLLED MOTION

◦
tack lies now within ±2 , and in some cases even [10] Roskam J. Airplane flight dynamics and automatic
◦
within ±4 . flight control. Part I. Lawrence, KS: DAR Corporation;
Thus, the adaptation mechanisms allow to re- 1995.
duce the tracking error in a much larger region [11] Roskam J. Airplane flight dynamics and automatic
flight control. Part II. Lawrence, KS: DAR Corporation;
of state space (in the case of a nonlinear system) 1998.
and to expand the frequency band in which the [12] Cook MV. Flight dynamics principles. Amsterdam: El-
tracking error does not exceed a specific prede- sevier; 2007.
termined value. That is, endowing the system [13] Hull DG. Fundamentals of airplane flight mechanics.
with adaptive properties allows it to deal with a Berlin: Springer; 2007.
[14] Nguyen LT, Ogburn ME, Gilbert WP, Kibler KS,
much broader class of parametric uncertainties
Brown PW, Deal PL. Simulator study of stall/post-stall
in controlled objects. characteristics of a fighter airplane with relaxed longi-
We can draw some conclusions from the re- tudinal static stability. NASA TP-1538, Dec. 1979.
sults obtained in Section 4.3.4.The methodsof [15] Sonneveld L. Nonlinear F-16 model description. The
adaptive-robust modeling and control in vari- Netherlands: Control & Simulation Division, Delft Uni-
versity of Technology; June 2006.
ants with a reference and predictive model
[16] Shaughnessy JD, Pinckney SZ, et al. Hypersonic vehicle
are the powerful and promising tools that al- simulation model: Winged-cone configuration. NASA–
low solving the problems of fault-tolerant air- TM–102610, November 1990.
craft motion control under uncertainty condi- [17] Boyden RP, Dress DA, Fox CH. Subsonic static and
tions. dynamic stability characteristics of the test technique
demonstrator NASP configuration. In: 31st Aerospace
Sciences Meeting & Exhibit, January 11–14, 1993, Reno,
NV, AIAA-93-0519.
REFERENCES [18] Boyden RP, Dress DA, Fox CH, Huffman JK, Cruz CI.
Subsonic static and dynamic stability characteristics of
[1] Ljung L. System identification: Theory for the user. 2nd a NASP configuration. J Aircr 1994;31(4):879–85.
ed. Upper Saddle River, NJ: Prentice Hall; 1999. [19] Davidson J. et al., Flight control laws for NASA’s Hyper-
[2] Narendra KS, Parthasarathy K. Identification and con- X research vehicle. AIAA–99–4124.
trol of dynamic systems using neural networks. IEEE [20] Engelund WC. Holland SD, et al., Propulsion system
Trans Neural Netw 1990;1(1):4–27. airframe integration issues and aerodynamic database
[3] Chen S, Billings SA. Neural networks for nonlinear dy- development for the Hyper-X flight research vehicle.
namic systems modelling and identification. Int J Con- ISOABE–99–7215.
trol 1992;56(2):319–46. [21] Engelund WC, Holland SD, Cockrell CE, Bittner RD,
[4] Heister F, Müller R. An approach for the identification Aerodynamic database development for the Hyper-X
of nonlinear, dynamic processes with Kalman-filter- airframe integrated scramjet propulsion experiments.
trained recurrent neural structures. Research report se- In: AIAA 18th Applied Aerodynamics Conference, Au-
ries, Report No. 193, University of Würzburg, Institute gust 14–17, 2000, Denver, Colorado. AIAA 2000–4006.
of Computer Science; April 1999. [22] Holland SD, Woods WC, Engelund WC. Hyper-X re-
[5] Haykin S. Neural networks: A comprehensive founda- search vehicle experimental aerodynamics test program
tion. 2nd ed. Upper Saddle River, NJ, USA: Prentice overview. J Spacecr Rockets 2001;38(6):828–35.
Hall; 1998. [23] Morelli, E.A. Derry, S.D. Smith, M.S. Aerodynamic pa-
[6] Hagan MT, Demuth HB, Beale MH, De Jesús O. Neural rameter estimation for the X-43A (Hyper-X) from flight
network design. 2nd ed. PSW Publishing Co.; 2014. data. In: AIAA Atmospheric Flight Mechanics Confer-
[7] Gorban AN. Generalized approximation theorem and ence and Exhibit, August 15–18, 2005. San Francisco,
computational capabilities of neural networks. Sib J Nu- CA. AIAA 2005-5921.
mer Math 1998;1(1):11–24 (in Russian). [24] Davis MC, White JT. X-43A flight-test-determined aero-
[8] Etkin B, Reid LD. Dynamics of flight: Stability and con- dynamic force and moment characteristics at Mach 7.0.
trol. 3rd ed. New York, NY: John Wiley & Sons, Inc.; J Spacecr Rockets 2008;45(3):472–84.
2003. [25] Morelli EA. Flight test experiment design for character-
[9] Boiffier JL. The dynamics of flight: The equations. izing stability and control of hypersonic vehicles. J Guid
Chichester, England: John Wiley & Sons; 1998. Control Dyn 2009;32(3):949–59.