Abstract
Polynomial nonlinear system identification suffers from numerical instability related to the ill-conditioning of the involved matrices. Orthogonal methods consist in conditioning the input signal in order to reduce the eigenvalues spread of the correlation matrix. The selection of an appropriate orthogonalization procedure, for robustness improvement, resort to signal statistics. Consequently, several orthogonal polynomial bases are proposed in the literature. Most of them use iterative processes and imply a considerable computational cost. Actually, with the growth of real-time applications, it is important to generate non-iterative orthogonal procedures, allowing optimization of algorithm-architecture adequacy. Our paper’s motivation is based on the complexity reduction aspect related to the orthogonalization process. Therefore we propose to focus on closed-form expressions of commonly used orthogonal polynomials bases namely the Shifted-Legendre orthogonal polynomials and the Hermite polynomials. A robustness study in terms of numerical stability enhancement of the two bases is carried on. Through comparative simulations results, the basis allowing the best matrix conditioning and an ease generalization for real-time applications with less restrictive hypothesis is selected in order to study the robustness of a polynomial nonlinear system identification scheme. Computer simulations are carried out to emphasize the advantages of the proposed scheme using different performance criteria for both the optimal case and the adaptive case, in terms of numerical stability and convergence rate. We propose to experiment the identification of the power amplifiers, for radio mobile applications and the loudspeakers for audio applications.
Similar content being viewed by others
References
John Mathews, V. (1995). In Orthogonalization of correlated gaussian signals for volterra system identification, IEEE signal processing letters Vol. 2.
Ogunfunmi, T., & Chang, S.L. (2001). In Second-order adaptive volterra system identification based on discrete nonlinear wiener model, IEE proceedings vision, image and siganl processing Vol. 148.
Kibangou, A.Y., Favier, G., Hassani, M.M. (2005). In Selection of generalized orthonormal bases for second order Volterra filters, Signal Processing (Vol. 85, pp. 2371–2385).
Zayani, R., Bouallegue, R., Roviras, D. (2008). Adaptive predistortions based on neural networks associated with levenberg-marquardt algorithm for satellite down links. Eurasip Journal on Wireless Communications and Networking, 2008. Article ID 132729.
Scott, I., & Mulgrew, B. (1997). Nonlinear system identification and prediction using orthonormal functions. IEEE Transactions on Signal Processing, 45(7).
Raich, R., Qian, H., Tong Zhou, G. (2004). Orthogonal polynomials for power amplifier modeling and predistorter design. IEEE Transactions on Vehicular Technology, 53(5).
Raich, R., & Tong Zhou, G. (2004). Orthogonal polynomials for complex gaussian process. IEEE Transactions on Signal Processing, 52(10).
Favier, G., & Bouilloc, T. (2010). In Identification de modèles de Volterra basée sur la décomposition PRAFAC de leurs noyaux et le filtre de Kalman étendu, irevues (Vol. 27, pp. 27–51).
Ruotsalainen, H. (2010). Investigation of orthogonal basis expansions for adaptive wiener models,Thesis of Master of Science in Technology, aaltouniversity school of science and technology.
Kay, S.M. (1993). Fundamentals of Statistical Signal Processing: Estimation Theory, Englewood Cliffs: NJ: Prentice-Hall.
Haykin, S. (1996). Adaptive Filter Theory, Third Edition: Prentice Hall, Inc. Englewood, N.J.
Mohamed Mahmoud, O.M., & Jaidane Saidane, M. (2009). On the Parameters Estimation of the Generalized Gaussian Mixture Model, EUSIPCO, Glasgow, Scotland.
Gazor, S., & Zhang, W. (2003). Speech probability distribution. IEEE Signal Processing Letter, 10(7).
Majoul, T., Raouafi, F., Jaidane, M. (2007). Turbo codes for audio watermarking: the generalized gaussian noise channel model, ICECS,Marrakech.
Saied Bouajina, S., Jaidane Saidane, M., Ghannouchi, F. (2008). In Weighted criteria for RF Power Amplifiers Identification in Wide Band Context,ICECS 2008 Vol. 2008. Malta.
Kang, D.J., Hwang, D.O., Nam, S.W., Powers, E.J. (2002). Adaptive volterra filtering usign M-band wavelet transform and Gram-Schmidt orthogonalisation. IEE Electronics Letters, 38(6).
Rocher, R., Menard, D., Sentieys, O., Scalart, P. (2007). In Analytical accuracy evaluation of Fixed-Point Systems, European Signal Processing Conference, EUSIPCO, (pp. 2316–2320). Poznan.
Giueseppe Thadeu Freitas de Abreu (2005). Closed-form correlation functions of generalized hermite wavelets. IEEE Transactions on Signal Processing, 53(6).
Mezghani- Marrakchi, I., Mahé, G., Djaziri-Larbi, S., Jaidane, M., Turki-Hadj Alouane, M. (2014). Nonlinear audio systems identification through audio input Gaussianization. IEEE/ACM Transactions on Audio, Speech and Language Processing, 22(1), 41–53.
Ding, L., & Zhou, G.T. (2004). Effects of even-order nonlinear terms on power amplifier modeling and predistortion linearization. IEEE Transactions on Vehicular Technology, 53, 156–162.
Chen, N., Zhou, G.T., Qian, H. (2007). Power efficiency improvements through peak-to-average power ratio reduction and power amplifier linearization. EURASIP Journal on Advances in Signal Processing, 2007, 7. doi:10.1155/2007/20463. Article ID 20463.
Klippel, W. (1996). Modeling the nonlinearities in horn loudspeakers. Journal of the Audio Engineering Society, 44(6), 470–480.
Tronchin, L. (2012). The emulation of nonlinear time-invariant audio systems with memory by means of Volterra series. Journal of the Audio Engineering Society, 60(12), 984–996.
Kallel Smaoui, M., Ben Jemaa, Y., Jaidane, M. (2010). A nonlinear adaptive filtering scheme based on a non-iterative orthogonalization procedure, ISIVC,Marrakech,Maroc.
Hammi, O., Boumaiza, S., Ghannouchi, F.M. (2007). On the robustness of digital predistortion function synthesis and average power tracking for highly non linear power amplifiers. IEEE Transactions Micro. Theory and Tech., 55, 1362–1389.
Bouajina-Saied, S., Hammi, O., Jaidane, M., Ghannouchi, F. (2010). Experimental approach for robust identification of RF power amplifier behavioral models using polynomial structures, IET Microwaves. Antennas and Propagation, 4(11), 1818–1828.
Saleh, A. (1981). Frequency-independent and frequency-dependent nonlinear models of TWT amplifiers. IEEE Transactions on Communications, 29(11), 1715–1720.
Fu, J., & Zhu, W.P. (2008). A nonlinear acoustic echo canceller using sigmoid transform in conjunction with RLS algorithm. IEEE Transactions on Circuits ans Systems-II: Express Briefs, 55(10).
Klippel, W. (2006). Tutorial: loudspeaker nonlinearities- causes, parameters, symptoms. Journal Audio Engineering Society, 54(10).
Lascaux, P., & Theodor, R. (1986). Analyse numérique matricielle adaptée à l’art de l’ingénieur, Vol.1. Masson.
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
For a matrix A, the condition number is defined as:
where λ m a x and λ m i n are respectively the maximum and the minimum eigenvalue of the matrix.
A matrix is said to be ill-conditioned if the condition number is much larger than 1, which induces difficulties to inverse correctly the matrix and hence to estimate accurately the NL system coefficients set with relation A x = b.
In fact, if we introduce perturbation on system:
Δ x is the error caused by perturbations Δ A and Δ b.
In [30], it is demonstrated that the relative error estimation of x o p t noted \(\delta (\mathbf {x})\;=\;\frac {\|\mathbf {\Delta x}\|}{\|\mathbf {x}\|}\) is quantified as:
Rights and permissions
About this article
Cite this article
Smaoui, M.K., Jemâa, Y.B. & Jaidane, M. Selection of a Closed-Form Expression Polynomial Orthogonal Basis for Robust Nonlinear System Identification. J Sign Process Syst 81, 285–303 (2015). https://doi.org/10.1007/s11265-014-0948-2
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11265-014-0948-2