Adaptive Nonlinear System Identification in the Short-Time Fourier Transform Domain

In this paper, we introduce an adaptive algorithm for nonlinear system identification in the short-time Fourier transform (STFT) domain. The adaptive scheme consists of a parallel combination of a linear component, represented by crossband filters between subbands, and a quadratic component, which is modeled by multiplicative cross-terms. We adaptively update the model parameters using the least-mean-square (LMS) algorithm, and derive explicit expressions for the transient and steady-state mean-square error (MSE) in frequency bins for white Gaussian inputs. We show that estimation of the nonlinear component improves the MSE performance only when the power ratio of nonlinear to linear components is relatively high. Furthermore, as the number of crossband filters increases, a lower steady-state MSE may be obtained at the expense of slower convergence. Experimental results support the theoretical derivations.

[1]  Walter Kellermann,et al.  Adaptation of a memoryless preprocessor for nonlinear acoustic echo cancelling , 2000, Signal Process..

[2]  D. Westwick,et al.  Separable Least Squares Identification of Nonlinear Hammerstein Models: Application to Stretch Reflex Dynamics , 2001, Annals of Biomedical Engineering.

[3]  Martin Vetterli,et al.  Adaptive filtering in subbands with critical sampling: analysis, experiments, and application to acoustic echo cancellation , 1992, IEEE Trans. Signal Process..

[4]  Brett Ninness,et al.  Quantifying the accuracy of Hammerstein model estimation , 2000, Autom..

[5]  K. Senne,et al.  Performance advantage of complex LMS for controlling narrow-band adaptive arrays , 1981 .

[6]  Walter Kellermann,et al.  Analysis and design of multirate systems for cancellation of acoustical echoes , 1988, ICASSP-88., International Conference on Acoustics, Speech, and Signal Processing.

[7]  Yu Gong,et al.  An LMS style variable tap-length algorithm for structure adaptation , 2005, IEEE Transactions on Signal Processing.

[8]  Graham C. Goodwin,et al.  Estimated Transfer Functions with Application to Model Order Selection , 1992 .

[9]  Taiho Koh,et al.  Second-order Volterra filtering and its application to nonlinear system identification , 1985, IEEE Trans. Acoust. Speech Signal Process..

[10]  Yuantao Gu,et al.  Convergence analysis of a deficient-length LMS filter and optimal-length sequence to model exponential decay impulse response , 2003, IEEE Signal Processing Letters.

[11]  K. Senne,et al.  Performance advantage of complex LMS for controlling narrow-band adaptive arrays , 1981 .

[12]  Robert D. Nowak,et al.  Penalized least squares estimation of Volterra filters and higher order statistics , 1998, IEEE Trans. Signal Process..

[13]  Edward J. Powers,et al.  Batch and adaptive Volterra filtering of cubically nonlinear systems with a Gaussian input , 1993, 1993 IEEE International Symposium on Circuits and Systems.

[14]  D. Brillinger.  Time series - data analysis and theory , 1981, Classics in applied mathematics.

[15]  V. U. Reddy,et al.  A new approach to subband adaptive filtering , 1999, IEEE Trans. Signal Process..

[16]  Walter A. Frank.  An efficient approximation to the quadratic Volterra filter and its application in real-time loudspeaker linearization , 1995, Signal Process..

[17]  Jason Wexler,et al.  Discrete Gabor expansions , 1990, Signal Process..

[18]  Robert Bregovic,et al.  Multirate Systems and Filter Banks , 2002 .

[19]  Israel Cohen,et al.  On Multiplicative Transfer Function Approximation in the Short-Time Fourier Transform Domain , 2007, IEEE Signal Processing Letters.

[20]  Sergio Benedetto,et al.  Nonlinear Equalization of Digital Satellite Channels , 1982, IEEE J. Sel. Areas Commun..

[21]  Walter Kellermann,et al.  Orthogonalized power filters for nonlinear acoustic echo cancellation , 2006, Signal Process..

[22]  Lennart Ljung,et al.  System Identification: Theory for the User , 1987 .

[23]  David R. Brillinger,et al.  Time Series: Data Analysis and Theory. , 1982 .

[24]  Michael T. Orchard,et al.  Adaptive filtering using filter banks , 1996 .

[25]  Allen Gersho,et al.  Adaptive Cancellation of Nonlinear Intersymbol Interference for Voiceband Data Transmission , 1984, IEEE Journal on Selected Areas in Communications.

[26]  V. John Mathews,et al.  Parallel-cascade realizations and approximations of truncated Volterra systems , 1998, IEEE Trans. Signal Process..

[27]  Rudolf Rabenstein,et al.  Nonlinear acoustic echo cancellation with 2nd order adaptive Volterra filters , 1999, 1999 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings. ICASSP99 (Cat. No.99CH36258).

[28]  Karen O. Egiazarian,et al.  A new variable length LMS algorithm: theoretical analysis and implementations , 2002, 9th International Conference on Electronics, Circuits and Systems.

[29]  Israel Cohen,et al.  Adaptive System Identification in the Short-Time Fourier Transform Domain Using Cross-Multiplicative Transfer Function Approximation , 2008, IEEE Transactions on Audio, Speech, and Language Processing.

[30]  Khaled A. Mayyas,et al.  Performance analysis of the deficient length LMS adaptive algorithm , 2005, IEEE Transactions on Signal Processing.

[31]  Rik Pintelon,et al.  Modified AIC and MDL model selection criteria for short data records , 2004, IEEE Transactions on Instrumentation and Measurement.

[32]  J. Rissanen,et al.  Modeling By Shortest Data Description* , 1978, Autom..

[33]  H. Akaike.  A new look at the statistical model identification , 1974 .

[34]  Demetrios G. Lainiotis,et al.  A partitioned adaptive approach to nonlinear channel equalization , 1998, IEEE Trans. Commun..

[35]  Minyue Fu,et al.  A blind approach to Hammerstein model identification , 2002, IEEE Trans. Signal Process..

[36]  Giovanni L. Sicuranza,et al.  Simplified volterra filters for acoustic echo cancellation in GSM receivers , 2000, 2000 10th European Signal Processing Conference.

[37]  G. Schwarz.  Estimating the Dimension of a Model , 1978 .

[38]  Torbjörn Wigren,et al.  Quantifying model errors caused by nonlinear undermodeling in linear system identification , 1996 .

[39]  George-Othon Glentis,et al.  Efficient algorithms for Volterra system identification , 1999, IEEE Trans. Signal Process..

[40]  Edward J. Powers,et al.  On the identification of a third-order Volterra nonlinear system using a frequency-domain block RLS adaptive algorithm , 1990, International Conference on Acoustics, Speech, and Signal Processing.

[41]  Giovanni Ramponi,et al.  Quadratic digital filters for image processing , 1988, IEEE Trans. Acoust. Speech Signal Process..

[42]  Jae S. Lim,et al.  Signal estimation from modified short-time Fourier transform , 1983, ICASSP.

[43]  Yves Rolain,et al.  Identification of linear systems with nonlinear distortions , 2003, Autom..

[44]  W. Rugh.  Nonlinear System Theory: The Volterra / Wiener Approach , 1981 .

[45]  Wei-Ping Zhu,et al.  Compensation of Loudspeaker Nonlinearity in Acoustic Echo Cancellation Using Raised-Cosine Function , 2006, IEEE Transactions on Circuits and Systems II: Express Briefs.

[46]  Y. Selen,et al.  Model-order selection: a review of information criterion rules , 2004, IEEE Signal Processing Magazine.

[47]  M. Portnoff.  Time-frequency representation of digital signals and systems based on short-time Fourier analysis , 1980 .

[48]  Yekutiel Avargel.  PERFORMANCE ANALYSIS OF CROSS-BAND ADAPTATION FOR SUBBAND ACOUSTIC ECHO CANCELLATION , 2006 .

[49]  Israel Cohen,et al.  Modeling and Identification of Nonlinear Systems in the Short-Time Fourier Transform Domain , 2010, IEEE Transactions on Signal Processing.

[50]  Huijuan Cui,et al.  Convergence analysis of a deficient-length LMS filter and optimal-length sequence to model exponential decay impulse response , 2003, IEEE Signal Process. Lett..

[51]  S. Haykin,et al.  Adaptive Filter Theory , 1986 .

[52]  Yariv Ephraim,et al.  Recent Advancements in Speech Enhancement , 2004 .

[53]  Israel Cohen,et al.  Representation and Identification of Nonlinear Systems in the Short-Time Fourier Transform Domain , 2010 .

[54]  Huijuan Cui,et al.  LMS algorithm with gradient descent filter length , 2004, IEEE Signal Processing Letters.

[55]  Ephraim.  Speech enhancement using a minimum mean square error short-time spectral amplitude estimator , 1984 .

[56]  Gerhard Schmidt,et al.  Acoustic echo control. An application of very-high-order adaptive filters , 1999, IEEE Signal Process. Mag..

[57]  Edward J. Powers,et al.  A digital method of modeling quadratically nonlinear systems with a general random input , 1988, IEEE Trans. Acoust. Speech Signal Process..

[58]  D.G.M. Cruickshank,et al.  Linear equalisers, with dynamic and automatic length selection , 2001 .

[59]  Régine Le Bouquin-Jeannès,et al.  Nonlinear acoustic echo cancellation based on Volterra filters , 2003, IEEE Trans. Speech Audio Process..

[60]  Israel Cohen,et al.  System Identification in the Short-Time Fourier Transform Domain With Crossband Filtering , 2007, IEEE Transactions on Audio, Speech, and Language Processing.