A Proportionate Affine Projection Algorithm for the Identification of Sparse Bilinear Forms

Identification of sparse impulse responses was addressed mainly in the last two decades with the development of the so-called “proportionate”-type algorithms. These algorithms are meant to exploit the sparseness of the systems that need to be identified, with the purpose of improving the convergence rate and tracking of the conventional adaptive algorithms used in this framework. Nevertheless, the system identification problem becomes more challenging when the parameter space is large. This issue can be addressed with tensor decompositions and modelling. In this paper, we aim to identify sparse bilinear forms, in which the bilinear term is defined with respect to the impulse responses of a spatiotemporal model. In this context, we derive a proportionate affine projection algorithm for the identification of such bilinear forms. Experimental results highlight the good behavior of the proposed solution.

[1]  Jacob Benesty,et al.  Sparse Adaptive Filters for Echo Cancellation , 2010, Synthesis Lectures on Speech and Audio Processing.

[2]  Jacob Benesty,et al.  On Regularization in Adaptive Filtering , 2011, IEEE Transactions on Audio, Speech, and Language Processing.

[3]  Markus Rupp,et al.  A tensor LMS algorithm , 2015, 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[4]  Jacob Benesty,et al.  Adaptive filtering for the identification of bilinear forms , 2018, Digit. Signal Process..

[5]  Kazuhiko Ozeki,et al.  An adaptive filtering algorithm using an orthogonal projection to an affine subspace and its properties , 1984 .

[6]  Jan Skoglund,et al.  Practically efficient nonlinear acoustic echo cancellers using cascaded block RLS and FLMS adaptive filters , 2017, 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[7]  Jacob Benesty,et al.  A Proportionate NLMS Algorithm for the Identification of Sparse Bilinear Forms , 2018, 2018 41st International Conference on Telecommunications and Signal Processing (TSP).

[8]  Jianming Liu,et al.  Proportionate Adaptive Filtering for Block-Sparse System Identification , 2015, IEEE/ACM Transactions on Audio, Speech, and Language Processing.

[9]  D. Harville.  Matrix Algebra From a Statistician's Perspective , 1998 .

[10]  Akihiko Sugiyama,et al.  A generalized proportionate variable step-size algorithm for fast changing acoustic environments , 2004, 2004 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[11]  Jacob Benesty,et al.  On the Identification of Bilinear Forms With the Wiener Filter , 2017, IEEE Signal Processing Letters.

[12]  M. Sondhi,et al.  On the evaluation of estimated impulse responses , 1998, IEEE Signal Processing Letters.

[13]  Donald L. Duttweiler,et al.  Proportionate normalized least-mean-squares adaptation in echo cancelers , 2000, IEEE Trans. Speech Audio Process..

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

[15]  Milos Doroslovacki,et al.  Proportionate adaptive algorithms for network echo cancellation , 2006, IEEE Transactions on Signal Processing.

[16]  Er-Wei Bai,et al.  Convergence of the iterative Hammerstein system identification algorithm , 2004, 2004 43rd IEEE Conference on Decision and Control (CDC) (IEEE Cat. No.04CH37601).