A unified framework for adaptive filter algorithms with variable step-size☆
Introduction
Adaptive filtering has been, and still is, an area of active research that plays an important role in an ever increasing number of applications such as the noise cancellation, channel estimation, the channel equalization and the acoustic echo cancellation [1], [2]. The least mean square (LMS) and its normalized version (NLMS) are the workhorses of adaptive filtering. In the presence of colored input signals, the LMS and NLMS algorithms have extremely slow convergence rates. To solve this problem a number of adaptive filtering structures based on affine subspace projections [3], [4], [5], data reusing adaptive algorithms [6], [7], [8], Block adaptive filters [2] and multirate techniques have been proposed in the literature [9], [10], [11]. In all these algorithms the selected fixed step-size can change the convergence rate and the steady-state mean square error. It is well known that the steady-state mean square error (MSE) decreases, when the step-size decreases while the convergence speed increases when the step-size increases. By optimally selecting the step-size, during the adaptation, we can obtain the both fast convergence rate and low steady-state mean square error. Important examples of the two new variable step-size (VSS) versions of the NLMS and the APA algorithm (APA) can be found in [12].
In [13], the generic adaptive filter based on the weighted, estimated Wiener–Hopf equation is proposed and then we used this framework to study the transient analysis of different adaptive filter algorithms covered by our generic adaptive filter [14]. The affine projection algorithms (APA), the transform domain adaptive filters (TDAF) [15] and the Pradhan Reddy Subband Adaptive Filters (PRSAF) [16] are the particular examples that can be covered with this framework. This framework obscures the relations, commonalities and differences, between the numerous adaptive algorithms available today [17].
Our objectives in this paper are
- •
Extend the generic adaptive filter of [14] to cover the block and data reusing adaptive filter algorithms.
- •
Develop the generic variable step-size version of the generic adaptive filter. The VSSNLMS and VSSAPA of [12] can be easily derived from this generic variable step-size adaptive filter.
- •
Based on the generic VSS adaptive filter, we introduce the two new VSS adaptive filter algorithms named the variable step-size block normalized least mean square (VSSBNLMS) and the variable step-size NDRLMS (VSSNDRLMS), which are characterized by the fast convergence speed, and reduced steady-state mean square error when compared to the ordinary BNLMS and NDRLMS adaptive filter algorithms.
We have organized our paper as follows: In the following section the generic framework with extension of more compatibility, will be briefly reviewed. In this section we will show that the Block and data reusing adaptive filter algorithms can be derived with the extension of this generic adaptive filter. In the next section, we present the generic variable step-size update equation forming the basis of our development of the VSSBNLMS and VSSNDRLMS adaptive filter algorithms. Subsequently, we derive the VSSBNLMS and VSSNDRLMS adaptive algorithms. Finally, before concluding the paper, we demonstrate the advantages of the proposed algorithms by presenting several experimental results.
Throughout the paper, the following notations are adopted:
- ∥ · ∥
Euclidean norm of a vector.
- ∥ · ∥2
squared Euclidean norm of a vector.
Σ-weighted Euclidean norm of a column vector t defined as tTΣt.
- Tr(·)
trace of a matrix.
- (·)T
transpose of a vector or a matrix.
- diag{ … }
diagonal matrix of its entries { … }.
- A ⊗ B
Kronecher product of matrices A and B.
- vec(T)
creating an M2 × 1 column vector t through stacking the columns of the M × M matrix T.
- vec(t)
creating an M × M matrix T form the M2 × 1 column vector t.
- λmax
the largest eigenvalue of a matrix.
the set of positive real numbers.
- E{ · }
expectation operator.
Section snippets
The generic update equation
The generic filter vector update equation at the center of our analysis can be stated as [13], [14], [17]
In Fig. 1 we show the prototypical adaptive filter setup. When the input signal, x(n), and the desired signal, d(n), are jointly stationary, the optimum filter minimizing E{e2(n)} is the Wiener filter. This filter is found as the solution of the Wiener–Hopf equation given bywhere ht is the M × 1 vector of filter coefficients constituting what we refer as
The generic variable step-size update equation
We now proceed by determining the optimum step-size, μo(n), instead of using μ in the VSS version of Eq. (12). The latter equation can be stated in terms of weight error vector, ϵ(n) = ht − h(n), where, as before, ht is the unknown true filter vector, asBy taking the squared Euclidean norm and expectation from both sides of Eq. (13),where B(n) = X(n)W(n), Eq. (14) can be represented in the form of
Derivation of the variable step-size adaptive filter algorithms
We now in the position to develop the two new variable step-size adaptive filter based on the generic variable step-size update equation. These algorithms are variable step-size version of BNLMS and NDRLMS algorithms presented in Table 1.
Experimental results
We justified the theoretical results presented in this paper by several computer simulations in a channel estimation setup. The unknown channel has eight taps and selected as a random length eight vector. Two different types of signals, Gaussian and uniformly distributed signals, are used in forming the input signal, x(n):which is a first order autoregressive (AR) process with a pole at ρ. For the Gaussian case, w(n) is a white, zero-mean, Gaussian random sequence having unit
Conclusions
In this paper we extended the generic adaptive filter to cover block normalized LMS (BNLMS) and normalized data reusing LMS (NDRLMS) adaptive filter algorithms. Following this we developed a generic variable step-size adaptive filter algorithm. This generic VSS adaptive filter can cover VSSNLMS and VSSAPA adaptive filter algorithms. Having applied the generic variable step-size update equation, the VSSNDRLMS and the VSSBNLMS adaptive filter algorithms were developed. These algorithms exhibit
Acknowledgment
We would like to express our thanks to ITRC for their financial support of this project (TMU 85-12-85).
Mohammad Shams Esfand Abadi was born in Tehran, Iran, on September 18, 1978. He received the B.S. degree in Electrical Engineering from Mazandaran University, Mazandaran, Iran and the M.S. degree in Electrical Engineering from Tarbiat Modarres University, Tehran, Iran in 2000 and 2002, respectively, and the Ph.D. degree in Biomedical Engineering from Tarbiat Modarres University, Tehran, Iran in 2007. Since 2004 he has been with the Department of Electrical Engineering, Shahid Rajaee University,
References (18)
- et al.
Adaptive signal processing
(1985) Adaptive filter theory
(2002)- et al.
An adaptive filtering algorithm using an orthogonal projection to an affine subspace and its properties
Electron Commun Jpn
(1984) - et al.
Mean-square performance of a family of affine projection algorithms
IEEE Trans Signal Process
(2004) Fundamentals of adaptive filtering
(2003)- Roy S, Shynk JJ. Analysis of data-reusing LMS algorithm, Proc Midwest Symp Circuits Syst...
- et al.
Convergence analysis of the binormalized data-reusing LMS algorithm
IEEE Trans Signal Process
(2000) - et al.
Mean-square performance of data-reusing adaptive algorithms
IEEE Signal Process Lett
(2005) - et al.
A new approach to subband adaptive filtering
IEEE Trans Signal Process
(1999)
Cited by (12)
Improved subband-forward algorithm for acoustic noise reduction and speech quality enhancement
2016, Applied Soft Computing JournalCitation Excerpt :Recently, several works have been conducted to increase convergence rates and reducing the steady-state estimation error (low misadjustment) of these different adaptive filtering algorithms in acoustic noise reduction and speech enhancement applications [23–38,1,39]. Furthermore, in the same direction, many variable step-size adaptive filtering algorithms for speech enhancement application have been also proposed [23–38,1,39]. For example, in references [23–26], variable step size LMS (VSS-LMS) algorithms were proposed to improve the performance of the fixed step size LMS algorithm by using particular recursive formulas.
A new method for removal of powerline interference in ECG and EEG recordings
2015, Computers and Electrical EngineeringCitation Excerpt :Therefore, although this solution is easy to be implemented at low cost, it produces undesirable signal distortion and cannot eliminate completely the interference [12]. Another method that is potentially capable of reducing the 50/60-Hz ECG interference is an adaptive filter (AF) [13–15]. This system is based on an adaptive filter that has the capability of automatically adjusting its parameters to obtain the best performance under a variable interference [1,2].
Two-channel variable-step-size forward-and-backward adaptive algorithms for acoustic noise reduction and speech enhancement
2015, Signal ProcessingCitation Excerpt :For example, in NR-SE applications, there is an interest to control the step-size values to get more robustness against non-stationary of the inputs and also to achieve enough noise reduction and less distortion of the output signals. Recently, many variable-step-size (VSS) adaptive algorithms have been proposed [25–34]. In [27], VSS LMS (VSS-LMS) algorithms are proposed to improve performance over the FSS algorithm.
A new two-microphone Gauss-Seidel pseudo affine projection algorithm for speech quality enhancement
2017, International Journal of Adaptive Control and Signal ProcessingState Space Least Mean Fourth Algorithm for Dynamic State Estimation in Power Systems
2016, Arabian Journal for Science and Engineering
Mohammad Shams Esfand Abadi was born in Tehran, Iran, on September 18, 1978. He received the B.S. degree in Electrical Engineering from Mazandaran University, Mazandaran, Iran and the M.S. degree in Electrical Engineering from Tarbiat Modarres University, Tehran, Iran in 2000 and 2002, respectively, and the Ph.D. degree in Biomedical Engineering from Tarbiat Modarres University, Tehran, Iran in 2007. Since 2004 he has been with the Department of Electrical Engineering, Shahid Rajaee University, Tehran, Iran. During the fall of 2003 and again in the spring of 2005, he has a visiting scholar with the Signal Processing Group at the University of Stavanger, Norway. His research interests include digital filter theory and adaptive signal processing algorithms.
Ali Mahlooji Far was born in Qom, Iran, on July 1, 1965. He received the B.S. degree in Electrical Engineering from Tehran University, Tehran, Iran and the M.S. degree in Electrical Engineering from Sharif University, Tehran, Iran in 1988 and 1991, respectively, and the Ph.D. degree in Biomedical Engineering from the University of UMIST, Manchester, UK. Since 1997 he has been an Associate Professor with the Department of Electrical Engineering, Tarbiat Modarres University, Tehran, Iran. His research interests include digital signal processing, medical imaging and image analysis.
- ☆
This work was supported by the Iran Telecommunication Research Center (ITRC), Tehran, Iran (TMU 85-12-85).