A New Hybrid Model Based On Neuro Fuzzy Network Soft Switching Mechanism For System Identification

Yıl 2019, Cilt: 12 Sayı: 1, 1 - 8, 31.01.2019

Selçuk Mete , Hasan Zorlu , Şaban Özer

https://doi.org/10.17671/gazibtd.459399

Cited By: 3

Öz

This
paper aims to improve a new hybrid model for system identification area. The
proposed hybrid model consists of an adaptive Hammerstein model, an adaptive
Wiener model, and a Neuro-Fuzzy (NF) network based soft-switching mechanism
(SSM). SSM structure in hybrid model increases the success of block model by
selecting the best results of Hammerstein and Wiener model outputs. In
literature, there are various studies about NF based on Hammerstein or Wiener
model types applied to system identification. In the proposed model,
Hammerstein and Wiener models with NF network are used together different from
the literature. In simulation studies, five different type of systems are identified
with different models (Hammerstein, Wiener and the proposed hybrid model)
optimized by Recursive Least Square (RLS). Then the performances of these
models are compared. Simulation results reveal the effectiveness and robustness
of the proposed identification model. 

Anahtar Kelimeler

Fuzzy neural networks, hybrid intelligent systems, optimization, soft switching, system identification

Sistem Kimliklendirme İçin Bulanık Sinir Ağı Esnek Anahtarlama Mekanizması Temelli Yeni Bir Karma Model

Öz

Bu çalışmanın amacı sistem kimliklendirme
alanında yeni bir karma model geliştirmektir. Önerilen karma model
uyarlanabilen bir Hammerstein model, bir Wiener model ve esnek anahtarlama
mekanizmasına dayanan bulanık sinir ağını içermektedir. Karma modeldeki esnek
anahtarlama mekanizması Hammerstein ve Wiener model çıkışlarının en iyi
sonuçlarını seçerek blok model başarısını arttırmaktadır. Literatürde, sistem
kimliklendirmede uygulanan bulanık sinir ağı temelli Hammerstein ya da Wiener
modellerle ilgili birçok çalışma vardır. Önerilen modelde, bulanık sinir ağıyla
birlikte Hammerstein ve Wiener modelleri literatürden farklı olarak bir arada
kullanılmıştır. Simülasyon çalışmalarında, farklı tipteki beş sistem tekrarlayan
en küçük kare ile optimize edilmiş olan farklı modeller (Hammerstein, Wiener ve
önerilen model) ile kimliklendirilmiştir. Daha sonra bu modellerin
performansları karşılaştırılmıştır. Simülasyon çalışmaları önerilen modelin
etkinliğini ve sağlamlığını ortaya koymaktadır.

Anahtar Kelimeler

Bulanık sinir ağı, karma zeki sistemler, optimizasyon, esnek anahtarlama, sistem kimliklendirme

Kaynakça

• [1] M. Peker, O. Özkaraca, B. Kesimal, “Enerji tasarruflu bina tasarımı için ısıtma ve soğutma yüklerini regresyon tabanlı makine öğrenmesi algoritmaları ile modelleme”, Bilişim Teknolojileri Dergisi, 10(4), 443-449, 2017.
• [2] U. Köse, E. Ülker, “Pareto zarflama-temelli seçim algoritması (PESA) ile B-spline eğri tahmini”, Bilişim Teknolojileri Dergisi, 5(2), 25-31, 2012.
• [3] S. Özden, A. Öztürk, “Yapay sinir ağları ve zaman serileri yöntemi ile bir endüstri alanının (ivedik OSB) elektrik enerjisi ihtiyaç tahmini”, Bilişim Teknolojileri Dergisi, 11(3), 255-261, 2018.
• [4] T. Schweickhardt, F. Allgöwer, "On system gains, nonlinearity measures, and linear models for nonlinear systems", IEEE Transactions on Automatic Control, 54, 62-78, 2009.
• [5] N.B. Hizir, M.Q. Phan, R. Betti, R.W. Longman, "Identification of discrete-time bilinear systems through equivalent linear models", Nonlinear Dynamics, 69, 2065-2078, 2012.
• [6] Y. Mao, F. Ding, Y. Liu, "Parameter estimation algorithms for Hammerstein time-delay systems based on the orthogonal matching pursuit scheme", IET Signal Processing, 11, 265-274, 2017.
• [7] F. Ding, X.P. Liu, G. Liu, "Identification methods for Hammerstein nonlinear systems", Digital Sig. Proc., 21, 215-238, 2011.
• [8] S. Ozer, H. Zorlu, "Identification of bilinear systems using differential evolution algorithm", Sadhana Academy Proceedings in Engineering Sciences, 36, 281-292, 2011.
• [9] F. Guo, A new identification method for wiener and hammerstein systems, Doktora Tezi, Karlsruhe Universitesi, Angewandte Informatik Bölümü, 2004.
• [10] L.A. Aguirre, M.C.S. Coelhoand, M.V. Correa, "On the interpretation and practice of dynamical differences between hammerstein and wiener models", IEE P-Contr. Theor. Ap., 152, 349-356, 2005.
• [11] J. Lee, W. Cho, T.F. Edgar, "Control system design based on a nonlinear first-order plus time delay model", J Process Contr., 7, 65-73, 1997.
• [12] H.X. Li, "Identification of hammerstein models using genetic algorithms", IEE P-Contr. Theor. Ap., 146, 499-504, 1999.
• [13] K.S. Narendra, P.G. Galman, "An iterative method for the identification of nonlinear systems using a hammerstein model", IEEE T. Automat. Contr., 11, 546-550, 1966.
• [14] L. Yu, J. Zhang, Y. Liao, J. Ding, "Parameter estimation error bounds for hammerstein nonlinear finite impulsive response models", Appl. Math. Comput., 202, 472-480, 2008.
• [15] A. Gotmare, R. Patidar, N.V. George, "Nonlinear system identification using a cuckoo search optimized adaptive hammerstein model", Expert Syst. Appl., 42, 2538-2546, 2015.
• [16] H.N. Al-Duwaish, "A genetic approach to the identification of linear dynamical systems with static nonlinearities", International Journal of Systems Science, 31, 307- 313, 2010.
• [17] D.L. Zhang, Y.G. Tang, J.H. Ma, X.P. Guan, "Identification of Wiener model with discontinuous nonlinearities using differential evolution", International Journal of Control, Automation and Systems, 11, 511-518, 2013.
• [18] H. Al-Duwaish, M.N. Karim, V. Chandrasekar, "Use of multilayer feedforward neural networks in identification and control of Wiener model", IEE Proc., part D., 143, 255-258, 1996.
• [19] S. Haykin, Neural Networks: A Comprehensive Foundation, Macmillan College Publishing Company, New York, 1994.
• [20] J.T.R. Jang, C.T. Sun, E. Mizutani, Neuro-Fuzzy and Soft Computing, Prentice Hall, PTR, 1997.
• [21] M.E. Yuksel, A. Basturk, "Efficient removal of impulse noise from highly corrupted digital images by a simple neuro-fuzzy operator", Int. J. Electron. Commun. (AEU), 57, 214-219, 2003.
• [22] M.E. Yuksel, E. Besdok, "A simple neuro-fuzzy impulse detector for efficient blur reduction of impulse noise removal operators for digital images", IEEE Trans. Fuzzy Syst., 12, 854-865, 2004.
• [23] H. Zorlu, Identification of nonlinear systems with soft computing techniques, Doktora Tezi, Erciyes Üniversitesi, Fen Bilimleri Enstitüsü, 2011.
• [24] L. Jia, M.S. Chiu, S.S. Ge, "A noniterative neuro-fuzzy based identification method for Hammerstein processes", Journal of Process Control., 15, 749-761, 2005.
• [25] J.S. Wang, Y.P. Chen, "A Hammerstein recurrent neuro-fuzzy network with an online minimal realization learning algorithm", IEEE Transactions on Fuzzy Systems, 16, 1597-1612, 2008.
• [26] J. Zhai, J. Zhou, L. Zhang, J. Zhao, W. Hong, "Dynamic behavioral modeling of power amplifiers using ANFIS based Hammerstein", IEEE Microwave and Wireless Compo. Letters, 18, 704-706, 2008.
• [27] E. Mohammadi, M. Montazeri-Gh, "A new approach to the Gray-Box identification of Wiener models with the application of gas turbine engine modelling", J. Eng. Gas Turbines Power, 137, 1521-1533, 2015.
• [28] C.L. Chen, C.Y. Chiu, "A fuzzy neural approach to design of a Wiener printer model incorporated into model-based digital half toning", Applied Soft Computing, 12, 1288-1302, 2012.
• [29] L. Yong, T. Ying-Gan, "Chaotic system identification based on a fuzzy Wiener model with particle swarm optimization", Chinese Physics Letters, 27, 1-4, 2010.
• [30] R. Sujendran, M. Arunachalam, "Hybrid fuzzy adaptive Wiener filtering with optimization for intrusion detection", ETRI Journal, 37, 1-10, 2015.
• [31] J. Wang, Q. Zhang, L. Ljung, "Revisiting the two-stage algorithm for hammerstein system identification", Chinese Control Conf. (CDC), Shanghai, 3620-3625, 2009.
• [32] H.N. Al-Duwaish, "Identification of Wiener model using genetic algorithms", IEEE GCC Conf. & Exhib., Kuwait City,1-4, 2009.
• [33] F. Sbeity, J.M. Girault, S. Ménigot, J. Charara, "Sub and ultra harmonic extraction using several hammerstein models", Int. Conf. Comp. Syst. (ICCS), Morocco, 1-5, 2012.
• [34] N. Wiener, Nonlinear Problems in Random Theory, Wiley, New York, 1958.
• [35] M. Schetzen, The Volterra and Wiener Theories of Nonlinear Systems, Krieger, Malabar, 1980.
• [36] P. Celka, N.J. Bershad, J.M. Vesin, "Fluctuation analysis of stochastic gradient identification of polynomial Wiener systems", IEEE Transactions on Signal Proc., 48, 1820-1825, 2000.
• [37] S. Ozer, H. Zorlu, "Neuro-Fuzzy soft-switching hybrid filter for impulsive noisy environments", Turk. J. Elec. Eng. & Comp., 19, 73-85, 2011.
• [38] A. Basturk, M. E. Yuksel, "Neuro-Fuzzy soft switching hybrid filter for impulse noise removal from digital images", Proc. of the IEEE Signal Proces. Com. Ap. Conf. (SIU), Kayseri, 13-16, 2005.
• [39] M.A. Soyturk, A. Basturk, M. E. Yuksel, "A novel fuzzy filter for speckle noise removal", Turk. J. Elec. Eng. & Comp., 22, 1367-1381, 2014.
• [40] Z. Wang, Y. Shen, Z. Ji, F. Ding, "Filtering based recursive least squares algorithm for Hammerstein FIR-MA systems", Nonlinear Dynamic, 73, 1045-1054, 2013.
• [41] S. Mete, S. Ozer, H. Zorlu, "System identification application using hammerstein model", Sadhana-Academy Proceedings in Engineering Sciences, 41, 597-605, 2016.
• [42] S.J. Nanda, G. Panda, B. Majhi, "Development of immunized PSO algorithm and its application to Hammerstein model identification", IEEE Congress on Evoluti. Comp., Trondheim, 3080-3086, 2009.
• [43] M. H. Calp, “İşletmeler için Personel Yemek Talep Miktarının Yapay Sinir Ağları Kullanılarak Tahmin Edilmesi”, Politeknik Dergisi, 2019. DOI: 10.2339/politeknik.444380. (Basımda)
• [44] H. Zorlu, S. Mete, Ş. Özer, “System identification using hammerstein model optimized with artificial bee colony algorithm”, Omer Halisdemir University Journal of Engineering Sciences, 7(1), 83-98, 2018.
• [45] S. Mete, S. Ozer, H. Zorlu, "System identification using Hammerstein model optimized with differential evolution algorithm", International Journal of Electronics and Communications (AEU), 70, 1667-1675, 2016.
• [46] Ş. Özer, H. Zorlu, S. Mete, “A comparison study of system identification using hammerstein model”, IEEE 2015 11th International Conference on Innovations in Information Technology (IIT'15), Dubai, 367-372, 2015.
• [47] J. Jeraj, V.J. Mathews, "Stochastic mean-square performance analysis of an adaptive hammerstein filter", IEEE Transaction on Signal Proces., 54, 2168-2177, 2006.
• [48] H. Zorlu, S. Ozer, "Identification of nonlinear volterra systems using differential evolution algorithm", National Conf. on Electrical, Electronics and Computer Engineering, Bursa, 630-633, 2010.
• [49] S. Mete, Ş. Özer, H. Zorlu, “System identification using hammerstein model”, 22nd IEEE Signal Processing and Communications Applications Conference, Trabzon, 1303-1306, 2014.