Signal Processing 127 (2016) 282–287

Contents lists available at ScienceDirect

Signal Processing
journal homepage: www.elsevier.com/locate/sigpro

Short communication

A novel subband adaptive filter algorithm against impulsive noise

and it's performance analysis
Pengwei Wen, Sheng Zhang, Jiashu Zhang n
Sichuan Province Key Lab of Signal and Information Processing, Southwest Jiaotong University, Chengdu 610031, PR China

art ic l e i nf o a b s t r a c t

Article history: The main drawback of the sign subband adaptive filter algorithm (SSAF) is its large steady-state error. To
Received 4 December 2015 improve the performance, this paper proposes a novel normalized logarithmic subband adaptive filter
Received in revised form algorithm (NLSAF), which is derived from a new normalized logarithmic cost function. Due to possess the
4 March 2016
advantages of the normalized subband adaptive filter (NSAF) and SSAF algorithms in the proposed NLSAF
Accepted 6 March 2016
Available online 26 March 2016
algorithm, it achieves a low steady-state error and the robustness performance against impulsive noise.
Then, by using the energy conservation method, the mean-square convergence performance of the
Keywords: proposed NLSAF algorithm is presented. Simulations on system identifications demonstrate that the
logarithmic cost function proposed subband adaptive filter performs better than the conventional SSAF algorithm in an impulsive-
Subband adaptive filtering
noise scenario.
Impulsive noise
& 2016 Elsevier B.V. All rights reserved.

1. Introduction performance against impulsive noise, the SSAF was derived by

The adaptive filtering algorithms have been widely used in many
practical areas such as system identification, channel estimation, and
echo cancellation. The normalized least-mean-square (NLMS) algo-
rithm is one of the simple algorithms because it is easy to implement
and has low computational complexity [1]. However, it is well-known
that the normalized least mean square algorithm converges slowly
when the input signals are colored. In order to solve this problem, the
affine projection algorithm (APA) and its variants (e.g., see [2,3] and
the references therein) were developed to provide faster convergence
rate than the NLMS algorithm. However, the APAs require large
computational cost due to involving the matrix inversion operation in
updating the tap-weight vector.
Recent years, an attractive approach is to use the subband
adaptive filter (SAF), because it divides the colored input signal into
almost mutually exclusive multiple subband signals and each sub-
band signal is approximately white [4]. The NSAF algorithm had
been proposed and studied by Lee and Gan in [5]. Because of the
inherent decorrelating property of SAF [6], it converges faster than
the NLMS for the colored input signals. Besides, the NSAF has al-
most the same computational complexity as the NLMS, especially
for applications of long adaptive filter such as echo cancellation.
However, similarly to the conventional NLMS algorithm, the per-
formance of the NSAF algorithm will degrade when background
noise includes impulsive noise. To maintain the robustness of filter
n
Corresponding author.
E-mail addresses: 623101594@qq.com (P. Wen),
zhangsheng@my.swjtu.edu.cn (S. Zhang), jszhang@home.swjtu.edu.cn (J. Zhang).
minimizing the l1-norm of the subband a posteriori error vector of
the filter [7,8], whereas it has a large steady-state error. Then Ni
et al. proposed a variable regularization parameter SSAF (VRP-SSAF)
algorithm to further reduce the steady state error [8].
In this paper, to improve the performance of the SSAF algorithm,
we propose a new subband adaptive filter algorithm, which is de-
rived by a normalized logarithm cost function. Because it inherits
the advantages of the NSAF and SSAF algorithms, the proposed al-
gorithm has good robustness performance against impulsive noise
and small steady-state error. Then, the steady-state mean square
derivation performance of the NLSAF is analyzed in this paper,
which is based on the energy conservation relation [9], Price's
theorem [10] and some reasonable assumptions such as the in-
dependent assumption. Simulation results show that the NLSAF
algorithm achieves a lower the steady-state error compared to the
SSAF algorithm with the same speed of convergence.
The paper is organized as follows. Section 2 provides a brief
review and discussion of the SSAF algorithm. In Section 3, the
proposed NLSAF is then presented. A convergence and steady state
analysis of the proposed algorithm is carried out in Section 4. Sec-
tion 5 contains simulation results. Finally, some conclusions are
given in Section 6.
2. Review of conventional SSAF algorithm
Fig. 1 shows the structure of the subband adaptive filter [5].
Consider a desired signal

http://dx.doi.org/10.1016/j.sigpro.2016.03.006
0165-1684/& 2016 Elsevier B.V. All rights reserved.
 P. Wen et al. / Signal Processing 127 (2016) 282–287 283

d(n) = uT (n)wo + η(n) (1)

N−1
A(k ) = ∑ uiT (k)ui(k).
T
where u(n) ¼[u(n),u(n  1),…,u(n M þ1)] , wo and η(n) denote the i=0 (7)
input vector, the tap-weight vector of unknown channel with the
length M and the background noise, respectively. The background
noise η(n) includes the white Gaussian background noise υ(n) and
impulsive noise ς(n), which is independent of u(n). The back- 3. The normalized logarithmic subband adaptive filter (NLSAF)
ground measurement noise υ(n) is a zero-mean Gaussian process algorithm
with variance συ2. In Fig. 1, the input signal u(n) and desired signal d
(n) are decomposed into ui(n)and di(n) by the analysis filters Hi(z) 3.1. The proposed algorithm
( i = 0, 1... N − 1), and the subband output signals yi(n) are ob-
^ (k ). N is the Although the SSAF is robustness against impulsive noise, it has
tained from ui(n) filtered by the adaptive filter w
a large steady-state error. To improve the performance of the SSAF,
number of subbands. Note that we use the variable n to index the
the new cost function combining the NSAF and SSAF algorithms is
original sequences, and k to index the decimated sequences. Then,
proposed in this section.
the signals di(n) and yi(n) are decimated to di,D(k) and yi,D(k) re-
To derive the proposed algorithm, the novel cost function using
spectively. It is easy to note that
the logarithm function in [11] is firstly defined by
di, D(k ) = uiT (k )wo + ηi(k ) (2) N−1
J (k ) = ∑ J ( ei, D(k))
and i=0 (8)
yi, D (k ) = ^ (k )
uiT (k )w where
(3)

where ui(k)¼[ui(kN),ui(kN  1),…,ui(kN  M þ1)]T and ηi(k ) is a ith ei, D(k ) ⎛ ei, D(k ) ⎞
1 ⎜ ⎟
subband noise. The ith subband output error is defined as (
J ei, D(k ) =) A(k )
− ln⎜ 1 + α
α ⎝ A(k )
⎟
⎠ (9)
ei, D(k ) = di, D(k ) − yi, D (k ). (4)
and α > 0 is the design parameter. When ei, D(k ) / A(k ) < 1, we
The original SSAF algorithm can be derived by minimizing the expand (9) with a Taylor series for small perturbations of the error
following cost function:
⎛ ⎛ ⎞2 ⎞
ei, D(k ) 1 ⎜ ei, D(k ) α 2 ⎜ ei, D(k )
eD(k ) (
J ei, D(k ) =) − α − ⎟ + …⎟
1 α⎜⎜ 2 ⎜⎝ A(k ) ⎟ ⎟⎟
J (k ) = A(k ) A(k ) ⎠
N−1 ⎝ ⎠
∑i = 0 uiT (k )ui (k ) (5)
⎛ ⎞2 3 ⎛ e (k ) ⎞3
α ⎜ ei, D(k ) ⎟ − α ⎜ i, D ⎟ + ….
where eD(k)¼ [e0,D(k),e1,D(k),…eN  1(k)]T. =
2 ⎜⎝ A(k ) ⎟
⎠ 2 ⎜⎝ A(k ) ⎟
⎠ (10)
Applying the gradient descent method, the conventional SSAF
algorithm for updating the tap-weight vector is expressed as
We note that the expand cost function includes the higher-
^ ( k + 1) = w
^ (k ) + μ U (k )sgn⎡⎣ eD(k )⎤⎦ order measures of the error. Thus, the proposed algorithm can
w
N−1 achieve smaller steady-state mean square errors through the use
∑i = 0 uiT (k )ui (k ) (6) of the higher-order statistics for small perturbations of the error.
where μ is the step-size, U (k ) = [u0(k ) , u1(k ) , ... , uN − 1(k )]. In the Taking the derivative of (9) with respect to the tap-weight vector
following, we define ^ (k )
w

Fig. 1. structure of SSAF.

284 P. Wen et al. / Signal Processing 127 (2016) 282–287

Table 1 0.3
Summary of the computational complexity.
0.2
Algorithm Multiplication Addition
0.1
NLMS 3Mþ 1 3M
SSAF Mþ M/N þ 3NLþ 2N 2MNþ 1 0
VRP-SSAF Mþ M/N þ 3NLþ (2þ 2N)N 3MN  N  M þ 2
NLSAF 2Mþ M/N þ 3NLþ 2N þ 2 2MNþ 2 -0.1

-0.2

-0.3
(
αf ei, D(k ) )
( ) (
∇w J ei, D(k ) = ∇w f ei, D(k ) ) 1 + αf -0.4
( ei,D(k)) (11)
-0.5
where
-0.6
ei, D(k ) 0 200 400 600 800 1000
( )
f ei, D(k ) =
A(k )
. Samples
(12)
Fig. 2. Measured impulse response of the acoustic echo path.
Because the new cost function (9) is a convex function, it has no
local minima. Then using the gradient descent method, the NLSAF
algorithm updates the weight vector as
N−1
^ ( k + 1) = w
w ^ (k ) + μ ∑ ui(k)
(
∂f ei, D(k ) ) ( )
αf ei, D(k )
i=0 (
∂ei, D(k ) 1 + αf ei, D(k ) )
αui(k)ei, D(k)
N−1
^ (k ) + μ A(k)
=w ∑ ei, D(k)
i=0 1+α
A(k)
N−1
^ (k ) + μ αui (k )ei, D(k )
=w ∑
i=0 A(k ) + α A(k ) ei, D(k ) (13)

when the error signal is sufficiently small such that

A(k ) + α A(k ) ei, D(k ) ≈ A(k ), the algorithm (13) resembles the
NSAF scheme. Otherwise, if the error signal is enough large so that
A(k ) + α A(k ) ei, D(k ) ≈ α A(k ) ei, D(k ) in (13), the proposed algo-
rithm behaves like the SSAF scheme. Hence, the NLSAF algorithm
intrinsically combines the NSAF and SSAF algorithms to improve
Fig. 3. MSDs comparison of the proposed algorithm with different κ , SNR¼ 30 dB
the performance, which provides smaller steady-state mean ( μ = 0.005, α = 1500, p = 0.001).
square error for small perturbations and stability for large
perturbations.

3.2. Complexity analysis

The computational complexity of the proposed algorithm is com-

pared with that of conventional algorithms in terms of the total
number of additions and multiplications. With the length L of the
analysis filters, the number of subbands N and the filter length M, the
computational requirements for these adaptive filters with corre-
sponding adaptive algorithm are summarized in Table 1. Compared
with the SSAF algorithm, the proposed (NLSAF) algorithm only has
slightly more computations. The NLSAF algorithm behaves better
performance than other algorithms, shown in Section 5.

4. Performance analysis of the NLSAF algorithm

In this section, the convergence and steady state analysis of the

proposed algorithm will be presented by using the energy con- Fig. 4. MSDs comparison of the proposed algorithm with different numbers of
servation method [9]. To facilitate the analysis, we use the fol- subbands, SNR ¼30 dB ( μ = 0.01, α = 2000, p = 0.001).

lowing assumptions.

Assumption 1. The impulsive noise ς(n) is usually modeled as a

Bernoulli–Gaussian (BG) process, i.e., ς (n) = ω(n)ε(n) [11–16], where Gaussian with zero-mean and variance σε2 = κσυ2, κ > > 1. Here, p
ω(n) is a Bernoulli process with the probability density function denotes the probability of the occurrence of the impulsive noise.
described by P (ω(n) = 1) = p and P (ω(n) = 0) = 1 − p, and ε(n) is Accordingly, the additive noise,η(n) = ς (n) + υ(n), is a contaminated
 P. Wen et al. / Signal Processing 127 (2016) 282–287 285

Fig. 5. MSDs comparison of the proposed algorithm for different α with impulsive
measurement noise, SNR ¼30 dB ( p = 0.001).

Fig. 7. MSDs comparison of the proposed algorithm with impulsive measurement

noise ( p = 0.001) SNR¼ 30 dB. (a) M¼ 512.(b) M ¼ 32.

Gaussian with zero-mean and variance ση2 = ση2/N .

i

^ (k ), the subband input

Assumption 3. The tap-weight vector w
vector ui(k ), and the subband noise ηi(k ) are statistically
independent.

Assumption 4. Owing to the inherent decorrelating property of SAF,

we can assume that each subband input signal is close to a white
signal. That is E{ui(k )uiT (k )} ≈ σu2i(k )IM and E{uiT (k )ui(k )} ≈ Mσu2i(k ),
where IM is an M × M identity matrix.

4.1. Convergence analysis

The convergence of the proposed algorithm is analyzed based

Fig. 6. MSDs comparison of the NLMS, SSAF, VRP-SSAF and proposed algorithm
with different SNR in the impulsive measurement noise ( p = 0.001) (a) SNR¼20 dB, on the mean square deviation (MSD). Defining the tap-weight
(b) SNR¼ 30 dB. ^ (k ), and inserting it in (13), we obtain
error vector v(k ) = w (k ) − w
o

Gaussian process with zero-mean and variance N−1

αui (k )ei, D(k )
v( k + 1) = v(k ) − μ ∑
i=0 A(k ) + α A(k ) ei, D(k ) (15)
ση2 = συ2 + pσN2 = ( 1 + pκ )συ2 (14)
By taking the squared Euclidean norm of both sides of (15), we
Assumption 2. The subband noises ηi(k ) are contaminated have
286 P. Wen et al. / Signal Processing 127 (2016) 282–287

N−1
vT (k )ui (k )ei, D(k ) 4.2. Steady-state analysis
‖v( k + 1)‖2 = ‖v(k )‖2 − 2μα ∑
i=0 A(k ) + α A(k ) ei, D(k )
In this section, the mean square derivation of the proposed
N−1
2 2 ui (k )T ui (k )ei2, D(k ) algorithm at steady state will be performed. When an adaptive
+μ α ∑ 2 filter operates in steady state, the MSD satisfies c (k + 1) = c (k ), as
i=0 ( A(k) + α A(k ) ei, D(k ) ) (16) k → ∞, we will get an equation on the steady-state MSD from (22)
i.e., c (∞), as shown in (24)
Taking the mathematical expectation of (16), we get
N −1 σ u2 (∞) ⎛ ei2, D (∞) ⎞
N−1 ⎛ ⎞ 2μα ∑ i
E⎜ ⎟c (∞)
ea, i (k )ei, D(k ) 2 2 ⎜ ⎟
E⎜⎜ ⎟ ( ) ( ) ( )
pκ + 1 σ υ /N + σ u (∞)c (∞) ⎝ E A(k ) + α E A(k ) ei, D (∞) ⎠
c ( k + 1) = c (k ) − 2μα ∑ ⎟
i =0 i
i=0 ⎝ A(k ) + α A(k ) ei, D(k ) ⎠ ⎛ ⎞
N −1 ⎜ ei2, D (∞) ⎟
N−1
⎛ ⎞ =μ 2α 2M ∑ σ u2 (∞)E⎜ ⎟
⎜ ui (k )T ui (k )ei2, D(k ) ⎟ i =0
i ⎜⎜
(
2⎟
E ( A(k )) + α E ( A(k )) ei, D (∞) ⎟⎠ )
+ μ2 α 2 ∑ E⎜ 2⎟
⎝ (24)
i = 0 ⎜ A(k ) + α A(k ) e (k )
( )⎟
⎝ i, D ⎠ (17) As can be seen, it is difficult to find a closed-form solution of
2 c (∞) from (24), because it is a transcendental equation. However,
where c (k ) = E[‖v(k )‖ ] represents the MSD at kth iteration. De-
we can use some numerical computation methods to obtain its
noting the a priori subband error ea, i = vT (k )ui(k ).
numerical solution, such as bisection method [18]. Besides, when
Assuming that ea, i(k ) is a zero-mean Gaussian process for suf- step-size μ is very small, we can use an assumption to solve (24),
ficiently long filters. To simplify the evaluation of the above ex- i.e., c (∞)σu2i(∞) < < συ2/N [12]. In this case, the steady-state MSD of
pectation, we use the Price's theorem and Assumption 1 in (17) the proposed algorithm is approximated as
and have the following approximation.
⎛ ⎞
ei2, D(∞)
⎛ ea, i (k )ei, D (k )
⎞
(
E ea, i1(k )e i1, D (k ) ⎜
⎛
) e i2, D (k ) ⎞
σu2i(∞)E⎜ ⎟
E⎜⎜ ⎟≈p E⎜ 1 ⎟ ⎜ 2⎟
⎟ ⎟ μαM
N−1
⎝( E ( A( k ) ) + α E ( A ( k )) e (∞) )⎠
⎝ A(k ) + α A(k ) ei, D (k ) ⎠ (
E e i2, D (k )
1 ) ⎝ A(k ) + α A(k ) e i1, D (k ) ⎠ c (∞) = ∑ ⎛
i, D

⎞
2 σu2 (∞) ei2, D(∞)
E ea, i2(k )e i2, D (k ) ⎛⎜
( ) e i2 , D (k ) ⎞
⎟
i=0 i
E⎜ ⎟
2 2
+ ( 1 − p) E⎜ ⎟ ( pκ + 1)συ / N ⎝ E( A(k)) + α E( A(k)) ei, D(∞) ⎠ (25)
(
E e i2 , D (k )
2 )
⎝ A(k ) + α A(k ) e i2, D (k ) ⎠ (18)
The relation (25) reveals that the steady-state MSD of the
Likewise, assuming ‖ui(k )‖2 is uncorrelated with f (ei, D(k )) for NLSAF algorithm will increase as the step-size increases.
sufficiently long filters [17], we have
⎛ ⎞ ⎛ ⎞
⎜ u i(k )T u i(k )ei2, D(k ) ⎟ ⎡ 2⎤
⎜ ei2, D(k ) ⎟ 5. Simulation results
E⎜ 2
⎟ ≈ E⎣ ‖u i(k )‖ ⎦E⎜ 2
⎟
⎜ A(k ) + α A(k ) e (k ) ⎟ ⎜ A(k ) + α A(k ) e (k ) ⎟
⎝ ( i, D ) ⎠ ⎝ ( i, D ) ⎠ (19) In this section, the performance of the proposed algorithm is
evaluated through Monte Carlo (MC) simulations in the system
where
identification. The cosine modulated filter bank is used in all the
ei2, D(k ) subband filter algorithms. We use the measured impulse response
(
f ei, D(k ) =) 2 of the acoustic echo path as the unknown system, which is de-
( A(k) + α A(k ) ei, D(k ) ) (20) picted in Fig. 2. In the exact-modeling scenario, the echo path is
2
truncated to the first 32 or 512 tap weights. The number of sub-
Assuming that the fluctuation of ‖ui(k )‖ from one iteration to bands N is 4 for prototype filter length L¼32. The colored input
the next is small enough [4] for sufficiently long filters, we can use signal is obtained by filtering a zero-mean white Gaussian random
the approximation sequence through a first-order system ϕ(z ) = 1/(1 − 0.9z −1). All the
simulation results are obtained by ensemble averaging over 50
A(k ) ≈ E( A(k )) (21)
independent trials. In all simulations, the measurement noise υ(n)
Substituting (18)–(21) into (17) yields is added to y(n) = woT (n)u(n) with 20 dB or 30 dB signal-to-noise
ratios (SNR).
N −1 σ u2 c (k ) ⎛ ei2, D (k ) ⎞
In Fig. 3, we selected different values for the parameter
c ( k + 1) = c (k ) − 2μα ∑ i
E⎜ ⎟
2 2 ⎜ ⎟ κ (κ = 1, 10, 100, 500 and 1000). The adaptive filter tap-length is
i =0 ( p κ + 1 )υ
σ / N + σ uic ( k ) ⎝ (
E A ( k ) ) + α E ( A(k )) ei, D (k ) ⎠
⎛ ⎞ 512. As can be seen, the proposed algorithm for different κ has the
N −1 ⎜ ei2, D (k ) ⎟
+ μ 2α 2M ∑ σ u2 E⎜ ⎟ same initial convergence speed. By decreasing this parameter the
i ⎜
⎜ E ( A(k )) + α E ( A(k )) e (k ) ⎟⎟
2
i =0
⎝ ( i, D ⎠) (22) steady-state MSD reduces. However, the steady-state MSDs for
different κ only exist a small deviation. The impact of the para-
For the algorithm to be stable, the mean-square deviation must meter can be ignored. In the following simulations, the value of κ
decrease iteratively, implying that c (k + 1) − c (k ) < 0. Thus the is chosen as 1000.
step-size has to fulfill the condition Fig. 4 shows the performance of the proposed algorithm using
different number of subbands. The adaptive filter tap-length is 512.
⎧ ⎫
⎪ N−1 Similar to the SSAF algorithm, with the number of subbands increas-
c (k)σu2 ⎛ ei2, D(k) ⎞⎪
⎪ ∑i = 0 i
E⎜ ⎟⎪ ing, the convergence speed of the proposed algorithm becomes slow
⎪ ( pκ + 1)συ2 / N + σu2ic (k) ⎝ E( A(k)) + α E( A(k)) ei, D(k) ⎠ ⎪ and the steady-state MSD reduces. Simultaneously, with the number
0 < μ < 2⎨ ⎬
⎪ ⎛ ⎞ ⎪ of subbands increasing, the computational complexity also increases.
N−1 2 ⎜ ei2, D(k)
⎪ αM ∑i = 0 σuiE ⎟ ⎪
⎜ 2⎟ Thus, a suitable number of subbands are determined as a trade-off
⎪ ⎝ ( E( A(k)) + α E( A(k)) ei, D(k) ) ⎠ ⎪
⎩ ⎭ (23) between convergence speed and steady-state MSD.
 P. Wen et al. / Signal Processing 127 (2016) 282–287
In Fig. 5, we selected different values for the parameter
α (α = 100, 500, 1000, 1500 and 2000). The length of the adaptive
filter is 512. As can be seen, with a small parameter α , the pro-
posed algorithm yields a reduced steady-state MSD at the expense
of a poor convergence speed. Therefore, the parameter α is de-
termined as a trade-off between convergence speed and steady-
state misalignment. To obtain a small misalignment and fast
convergence rate, the value of α is chosen as 1500 in the following
simulations.
Fig. 6 compares the MSDs of the NLMS, conventional SSAF, VRP-
SSAF and the proposed algorithm with the fixed step-size and
different SNRs. The length of the adaptive filter is 512. As can be
seen, the proposed algorithm has the robust performance against
impulsive noise. And it achieves a faster convergence rate and a
smaller steady-state error, compared to the conventional SSAF and
VRP-SSAF.
Fig. 7 shows that the steady-state MSD curve of theoretical and
simulation results for different M, where the dotted lines represent
the theoretical steady-state MSDs. As can be seen, the proposed
theoretical results calculated by (25) are in good agreement with
simulation results. Therefore, it is concluded that (25) is an ap-
proximate solution of the transcendental Eq. (24).
6. Conclusion
In this paper, we propose a novel normalized logarithmic cost
function. Using the general steepest descent method, the NLSAF
algorithm is derived from the new cost function. With the benefit
of the robust logarithmic cost function, the proposed NLSAF al-
gorithm is able to suppress the effect of impulsive noise on the
filter weight update. Simultaneously, the NLSAF can obtain a low
steady-state error. The complexity, convergence and steady state
analysis of the NLSAF algorithm was carried out based on the
energy conservation method. Compared to the SSAF and VRP-SSAF
algorithm, simulations on system identification show that the
proposed algorithm obtains 3 dB attenuation in term of the stea-
dy-state MSD than the conventional SSAF at the same convergence
rate. With different tap lengths and step sizes, the theoretical re-
sults of NLSAF have supported our theoretical analysis on the
steady-state MSD.
Appendix A
In this appendix, we give a detailed derivation about (17).
ei1, D(k ) = ea, i (k ) + ςi(k )
(A1)
ei2, D(k ) = ea, i (k ) + υi (k )
(A2)
287
E⎡⎣ ei21, D(k )⎤⎦ = E⎡⎣ ea2, i (k )⎤⎦ + ( κ + 1)συ2/N (A3)
E⎡⎣ ei22, D(k )⎤⎦ = E⎡⎣ ea2, i (k )⎤⎦ + συ2/N (A4)
E⎡⎣ ei2, D(k )⎤⎦ = pE⎡⎣ ei21, D(k )⎤⎦ + ( 1 − p)E⎡⎣ ei22, D(k )⎤⎦
=E⎡⎣ ea2, i (k )⎤⎦ + ( pκ + 1)συ2/N (A5)
According to the definitions of the previous MSD(k ) and ea, i(k ),
we have the following relation for each subband
E⎡⎣ ea2, i (k )⎤⎦ = σu2i(k )c (k ) (A6)
Hence, combining (A3)–(A6), (17) can be derived.
References
[1] B. Widrow, S.D. Sterns, Adaptive Signal Processing, Prentice-Hall, Englewood
Cliffs, NJ, 1985.
[2] S. Zhang, J. Zhang, Modified variable step-size affine projection sign algorithm,
Electron. Lett. 49 (20) (2013) 1264–1265.
[3] T. Shao, Y.R. Zheng, J. Benesty, An affine projection sign algorithm robust
against impulsive interference, IEEE Signal Process. Lett. 17 (4) (2010)
327–330.
[4] K.A. Lee, W.S. Gan, S.M. Kuo, Subband adaptive Filtering: Theory and Im-
plementation, Wiley, Hoboken, NJ, 2009.
[5] K.A. Lee, W.S. Gan, Improving convergence of the NLMS algorithm using
constrained subband updates, IEEE Signal Process. Lett. 11 (9) (2004) 736–739.
[6] K.A. Lee, W.S. Gan, Inherent decorrelating and least perturbation properties of
the normalized subband adaptive filter, IEEE Trans. Signal Process. 54 (11)
(2006) 4475–4480.
[7] J.H. Kim, J.H. Chang, S.W. Nam, Sign subband adaptive filter with l1-norm
minimisation-based variable step-size, Electron. Lett. 49 (21) (2013)
1325–1326.
[8] J. Ni, F. Li, Variable regularisation parameter sign subband adaptive filter,
Electron. Lett. 46 (24) (2010) 1605–1607.
[9] A.H. Sayed, Fundamentals of Adaptive Filtering, Wiley–Interscience, New York,
2003.
[10] S. Zhang, J. Zhang, Transient analysis of zero attracting NLMS algorithms
without Gaussian input signal, Signal Process. 97 (2013) 100–109.
[11] M.O. Sayin, N.D. Vanli, S.S. Kozat, A novel family of adaptive filtering algo-
rithms based on the logarithmic cost, IEEE Trans. Signal Process. 62 (17) (2014)
4411–4424.
[12] Y.P. Li, T.S. Lee, B.F. Wu, A variable step-size sign algorithm for channel esti-
mation, Signal Process. 102 (2014) 304–312.
[13] S. Zhang, J. Zhang, Enhancing the tracking capability of recursive least p-norm
algorithm via adaptive gain factor, Digit. Signal Process. 30 (2014) 67–73.
[14] J. Shin, J. Yoo, P. Park, Variable step-size sign subband adaptive filter, Signal
Process. Lett. 20 (2) (2013) 173–176.
[15] J.W. Yoo, J.W. Shin, P.G. Park, A band-dependent variable step-size sign sub-
band adaptive filter, Signal Process. 104 (2014) 407–411.
[16] J. Ni, X. Chen, J. Yang, Two variants of the sign subband adaptive filter with
improved convergence rate, Signal Process. 96 (2014) 325–331.
[17] T.Y. Al-Naffouri, A.H. Sayed, Transient analysis of adaptive filters with error
nonlinearities, IEEE Trans. Signal Process 51 (2003) 653–663.
[18] J.F. Traub, Iterative Methods for the Solution of Equations, Am. Math. Soc.
(1982).