The Fast Subsampled-Updating Fast Transversal Filter (FSU FTF) RLS Algorithmy

نویسندگان

  • Dirk T.M. Slock
  • Karim Maouche
  • Dirk Slock
چکیده

We present a new fast algorithm for Recursive Least-Squares (RLS) adaptive ltering that uses displacement structure and subsampled updating. The FSU FTF algorithm is based on the Fast Transversal Filter (FTF) algorithm, which exploits the shift invariance that is present in the RLS adaptation of a FIR lter. The FTF algorithm is in essence the application of a rotation matrix to a set of lters and in that respect resembles the Levinson algorithm. In the subsampled updating approach, we accumulate the rotation matrices over some time interval before applying them to the lters. It turns out that the successive rotation matrices themselves can be obtained from a Schur type algorithm which, once properly initialized, does not require inner products. The various convolutions that thus appear in the algorithm are done using the Fast Fourier Transform (FFT). For relatively long lters, the computational complexity of the new algorithm is smaller than the one of the well-known LMS algorithm, rendering it especially suitable for applications such as acoustic echo cancellation. R esum e Nous pr esentons un nouvel algorithme rapide des moindres carr es r ecursifs (MCR), bas e sur la structure de d eplacement et la mise a jour sousechantillonn ee. Le FSU FTF est d eriv e a partir de l'algorithme FTF qui exploite une certaine propri et e d'invariance sous l'op eration de d ecalage qui est inh erente au ltrage adaptatif par les MCR. L'algorithme FTF peut être vu comme l'application d'une matrice de rotation a un ensemble de ltres et s'apparente de ce point de vue a l'algorithme de Levinson. Dans une approche de mise a jour du ltre adaptatif sousechantillonn ee, nous accumulons pendant un bloc d' echantillons les matrices de rotation successives puis, nous appliquons la matrice r esultat aux ltres. Ces matrices de rotation peuvent être obtenues en utilisant une proc edure de type Schur. On evite ainsi les calculs de produits scalaires, sauf pour l'initialisation de cet algorithme FTF-Schur. Les convolutions qui apparaissent ainsi a di erents endroits dans l'algorithme sont e ectu ees a l'aide de la Transform ee de Fourier Rapide (TFR). Pour des ltres relativement longs, la complexit e du nouvel algorithme est plus faible que celle du LMS, ce qui le rend tr es adapt e pour la r esolution de probl emes tels que celui de l'annulation d' echo acoustique.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Fast subsampled-updating stabilized fast transversal filter (FSU SFTF) RLS algorithm for adaptive filtering

We present a new, doubly fast algorithm for Recursive Least-Squares (RLS) adaptive ltering that uses displacement structure and subsampled-updating. The Fast Subsampled-Updating Stabilized Fast Transversal Filter (FSU SFTF) algorithm is mathematically equivalent to the classical Fast Transversal Filter (FTF) algorithm. The FTF algorithm exploits the shift invariance that is present in the RLS a...

متن کامل

The Fast Subsampled-Updating Fast Newton Transversal Filter (FSU FNTF) for Adapting Long FIR Filters

The FNTF algorithm starts from the RLS algorithm for adapting FIR lters. The FNTF algorithm approximates the Kalman gain by replacing the sample covariance matrix inverse by a banded matrix (AR(M) assumption for the input signal). The approximate Kalman gain can still be computed using an exact recursion that involves the prediction parts of two Fast Transversal Filter (FTF) algorithms of order...

متن کامل

The Fast Subsampled-Updating Fast Newton Transversal Filter (FSU FNTF) Algorithm for Adaptive Filtering Based on a Schur Procedure and the FFT

The Fast Newton Transversal Filter (FNTF) algorithm starts from the Recursive LeastSquares algorithm for adapting an FIR lter of length N. The FNTF algorithm approximates the Kalman gain by replacing the sample covariance matrix inverse by a banded matrix of total bandwidth 2M+1 (AR(M) assumption for the input signal). In this algorithm, the approximate Kalman gain can still be computed using a...

متن کامل

The Fast Subsampled - Updating Fast A ne Projection ( FSU FAP ) Algorithm

The Fast A ne Projection (FAP) Algorithm is the fast version of the AP algorithm which is a generalization of the well-known Normalized Least-Mean-Square (NLMS) algorithm. The AP algorithm shows performances that are near to those of the Recursive Least-Squares algorithms while its computational complexity is nearly the same as the LMS algorithm one. Moreover, recent research has enlightened th...

متن کامل

Leading Element Dichotomous Coordinate Descent Exponential Recursive Least Squares Algorithm for Multichannel Active Noise Control

In this paper, a new multichannel modified filtered-x (MFX) recursive least square (RLS) algorithm for active noise control (ANC) based on leading element dichotomous co-ordinate descent (LEDCD) iterations is proposed. It is shown that the proposed algorithm has less than half of the complexity of MFX fast transversal filter (FTF) algorithm with good performance for ideal plant models and impro...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 1998