Page 138 - Neural Network Modeling and Identification of Dynamical Systems

REFERENCES 127
[19] Potenza R, Dunne JF, Vulli S, Richardson D, King P. Mul- [36] Malki HA, Karayiannis NB, Balasubramanian M. Short-
ticylinder engine pressure reconstruction using NARX term electric power load forecasting using feedforward
neural networks and crank kinematics. Int J Eng Res neural networks. Expert Syst 2004;21(3):157–67.
2017;8:499–518. [37] Messai N, Riera B, Zaytoon J. Identification of a class
[20] Patel A, Dunne JF. NARX neural network modelling of hybrid dynamic systems with feed-forward neural
of hydraulic suspension dampers for steady-state and networks: About the validity of the global model. Non-
variable temperature operation. Veh Syst Dyn: Int J Veh linear Anal Hybrid Syst 2008;2:773–85.
Mech Mobility 2003;40(5):285–328. [38] Baek S, Park DS, Cho J, Lee YB. A robot endeffector
[21] Gaya MS, Wahab NA, Sam YM, Samsudin SI, Ja- tracking system based on feedforward neural networks.
maludin IW. Comparison of NARX neural network Robot Auton Syst 1999;28:43–52.
and classical modelling approaches. Appl Mech Mater [39] Janczak A. Identification of nonlinear systems using
2014;554:360–5. neural networks and polynomial models: A block-
[22] Siegelmann HT, Horne BG, Giles CL. Computa- oriented approach. Berlin, Heidelberg: Springer-Verlag;
tional capabilities of recurrent NARX neural net- 2005.
works. IEEE Trans Syst Man Cybern, Part B, Cybern [40] Giri F, Bai EW. Block-oriented nonlinear system identi-
1997;27(2):208–15. fication. Berlin, Heidelberg: Springer-Verlag; 2010.
[23] Kao CY, Loh CH. NARX neural networks for nonlinear [41] Janczak A. Comparison of four gradient-learning algo-
analysis of structures in frequency domain. J Chin Inst rithms for neural network Wiener models. Int J Syst Sci
Eng 2008;31(5):791–804. 2003;34(1):21–35.
[24] Pearson PK. Discrete-time dynamic models. New York– [42] Ozer S, Zorlu Y, Mete S. System identification ap-
Oxford: Oxford University Press; 1999. plication using Hammerstein model. Int J Syst Sci
[25] Cybenko G. Approximation by superposition of 2016;4(6):597–605.
a sigmoidal function. Math Control Signals Syst [43] Sut HT, McAvoy TJ. Integration of multilayer percep-
1989;2(4):303–14. tron networks and linear dynamic models: A Ham-
[26] Hornik K, Stinchcombe M, White H. Multilayer feed- merstein modeling approach. Ind Eng Chem Res
forward networks are universal approximators. Neural 1993;32:1927–36.
Netw 1989;2(5):359–66. [44] Peng J, Dubay R, Hernandez JM, Abu-Ayyad M. A
[27] Gorban AN. Generalized approximation theorem and Wiener neural network-based identification and adap-
computational capabilities of neural networks. Sib J Nu- tive generalized predictive control for nonlinear SISO
mer Math 1998;1(1):11–24 (in Russian). systems. Ind Eng Chem Res 2011;4:7388–97.
[28] Haykin S. Neural networks: A comprehensive founda- [45] Wills A, Schön TB, Ljung L, Ninness B. Identification
tion. 2nd ed.. Upper Saddle River, NJ, USA: Prentice of Hammerstein–Wiener models. Ind Eng Chem Res
Hall; 1998. 2012;49:70–81.
[29] Hagan MT, Demuth HB, Beale MH, De Jesús O. Neural [46] Peia JS, Smyth AW, Kosmatopoulos EB. Analysis and
network design. 2nd ed.. PSW Publishing Co.; 2014. modification of Volterra/Wiener neural networks for
[30] Sandberg IW, Lo JT, Fancourt CL, Principe JC, Kata- the adaptive identification of non-linear hysteretic dy-
giri S, Haykin S. Nonlinear dynamical systems: Feed- namic systems. J Sound Vib 2004;275:693–718.
forward neural network perspectives. Wiley; 2001. [47] Li S, Li Y. Model predictive control of an intensified con-
[31] Levin AU, Narendra KS. Recursive identification tinuous reactor using a neural network Wiener model.
using feedforward neural networks. Int J Control Neurocomputing 2016;185:93–104.
2013;61(3):533–47. [48] Lawry´nczuk M. Practical nonlinear predictive control
[32] Thibault J. Feedforward neural networks for the iden- algorithms for neural Wiener models. J Process Control
tification of dynamic processes. Chem Eng Commun 2013;23:696–714.
1991;105:109–28. [49] Tan AH, Godfrey K. Modeling of direction-dependent
[33] Kuschewski JG, Hui S, Zak SH. Application of feed- processes using Wiener models and neural networks
forward neural networks to dynamical system identi- with nonlinear output error structure. IEEE Trans In-
fication and control. IEEE Trans Control Syst Technol strum Meas 2004;53(3):744–53.
1993;1(1):37–49. [50] Michalkiewicz J. Modified Kolmogorov neural net-
[34] Rankovi´cVM, Nikoli´c IZ. Identification of nonlinear work in the identification of Hammerstein and
models with feedforward neural network and digital re- Wiener systems. IEEE Trans Neural Netw Learn Syst
current network. FME Trans 2008;36:87–92. 2012;23(4):657–62.
[35] Mironov K, Pongratz M. Applying neural networks for [51] Lin DT, Dayhoff JE, Ligomenides PA. Trajectory pro-
prediction of flying objects trajectory. Vestn UGATU duction with the adaptive time-delay neural network.
2013;17(6):33–7. Neural Netw 1995;8(3):447–61.