Time-domain equalizer and signal processing method thereof

Abstract

A time-domain equalizer includes a delay circuit, a weighting circuit, a controller and a summation circuit. The delay circuit receives an equalized signal and accordingly generates M delayed signals for an equalized signal. The weighting circuit applies an mth weighting of M weightings to an mth delayed signal of the M delayed signals to generate an mth weighted signal. The summation circuit sums up the M weighted signals, according to which the equalized signal is updated. The controller iteratively updates the M weightings according to a vector {right arrow over (e)}n,p=[en,p,1 . . . en,p,M], where the symbol en,p,j is defined as en,p,j=Σk(z[k]*z[k−Dp,j]*), the symbol n is an iteration index, k is a sample index, z[k] is a kth sample of the equalized signal, j is an integer index between 1 and M, and Dp,j represents a time delay amount corresponding to a jth delayed signal of the M delayed signals.

Description

This application claims the benefit of Taiwan application Ser. No. 105142447, filed Dec. 21, 2016, the subject matter of which is incorporated herein by reference.

BACKGROUND OF THE INVENTION

Field of the Invention

The invention relates in general to a time-domain equalizer, and more particularly, to a method for determining a weighting coefficient in a time-domain equalizer.

Description of the Related Art

Orthogonal frequency-division multiplexing (OFDM) technologies, featuring high spectrum utilization efficiency and simple hardware structures, are extensively applied in communication systems. An OFDM signal is consisted of multiple symbols. To prevent echo signals in a multipath from causing inter-symbol interference (ISI), a guard interval is provided at a front end of each symbol. However, in a more complex communication environment, a propagation delay may still last longer than the length of the guard interval, resulting in ISI and degraded overall system performance. Such issue cannot be resolved by frequency-domain equalizer technologies, and an additional time-domain equalizer is necessary before the frequency-domain equalizer of a receiver to eliminate or minimize the echo signals in a received signal.

Assume that an original signal transmitted from a transmitter is represented by a symbol x and a signal received at a receiver is represented by a symbol y. Without considering timing offset and frequency offset, the received signal y via multipath propagation may be represented as follows:y[k]=x[k]+Σ_p=1^pα_pe^jθp,kx[k−Δ_p]+n[k]  (1)

In equation (1), k represents a sample index, x[k] represents a k^th sample of the original signal x, y[k] represents a k^th sample of the received signal k, n[k] represents a k^th sample of a noise signal. As seen from equation (1), the received signal y is a sum of the original signal x and P echo signals. The positive integer P represents the total number of echo signals caused by a multipath transmission between the transmitter and the receiver. The receiver may obtain the value P by analyzing the received signal y. The symbols α_p, θ_p,k, and Δ_p respectively represent a magnitude amplification ratio, a phase offset and an arrival time offset of a p^th echo signal of the P echo signals (where p is an integer index between 1 and P) relative to the original signal x.

FIG. 1(A) shows a function block diagram of a time-domain equalizer. A time-domain equalizer 100 includes P similar signal generating circuits (denoted as 110₁, 110₂, . . . , and 110_P, collectively referred to as an approximate signal generating circuit 110), and a subtraction circuit 120. After the receiver has determined the value of P, the time-domain equalizer 100 is configured to include P approximate signal generating circuits 110. Each approximate signal generating circuit 110 is assigned to correspond to an echo signal in the P echo signals, and is in charge of generating an approximate signal s substantially the same to that echo signal. The subtraction circuit 120 receives the P approximate signals, which are generated from subtracting the P approximate signals from the received signal y, and the output signal z is the equalized signal. As the P approximate signals s are more similar to the respective corresponding echo signals, the equalized signal z gets closer to the original signal x.

How the approximate signal generating circuit 110_p corresponding to the p^th echo signal generates the approximate signal s is described below. FIG. 1(B) shows a function block diagram of the approximate signal generating circuit 110_p, which includes a delay circuit 111_p, a weighting circuit 112_p and a summation circuit 113_p.

The delay circuit 111_p includes M delay elements (denoted as L_p,1, L_p,2, . . . and L_p,M), each of which has an output end being a tapping node. M is a positive integer greater than 1, and m is an integer index between 1 and M. The delay circuit 111_p generates M different delayed signals for the equalized signal z by using the M delay elements. An m^th tapping node in the M tapping nodes has a time delay amount D_p,m at an input node of the delay circuit 111_p. In other words, the output end of the delay element L_p,m provides a delayed signal z[k−D_p,m].

The weighting circuit 112_p includes M multipliers (denoted as 112_p,1, 112_p,2, . . . and 112_p,M), in which the m^th multiplier 112_p,m multiplies the delayed signal z[k−D_p,m] by an m^th weighting w_p,m to generate a weighted signal. The M weighted signals are summed by the summation circuit 113_p, with a summation result represented as:s_p=Σ_m=1^Mw_p,mz[k−D_p,m]  (2)

According to the structures depicted in FIG. 1(A) and FIG. 1(B), the output signal z of the time-domain equalizer 100 may be represented as:z[k]=y[k]−Σ_p=1^pΣ_m=1^Mw_p,mz[k−D_p,m]  (3)

SUMMARY OF THE INVENTION

A time-domain and a signal processing method provided by the present invention define a new cost function, and computes on the basis of the least mean square (LMS) algorithm, so as to accordingly identify, from weighting circuits of a time-domain equalizer, a weighting w_p,m that most effectively eliminates echo signals, and further solve the issue of ISI at a receiver.

According to an embodiment of the present invention, a time-domain equalizer includes a delay circuit, a weighting circuit, a controller and a summation circuit. The delay circuit receives an equalized signal, and accordingly generates M delayed signals, where M is a positive integer greater than 1. The weighting circuit adopts M weightings, and applies an m^th weighting of the M weightings to an m^th delayed signal of the M delayed signals to generate an m^th weighted signal, where m is an integer index between 1 and M, wherein M weighted signals are generated accordingly. The summation circuit sums up the M weighted signals, according to which the equalized signal is updated. The controller iteratively updates the M weightings according to a vector {right arrow over (e)}_n,p=[e_n,p,1 . . . e_n,p,M], where the symbol e_n,p,j is defined as e_n,p,j=Σ_k(z[k]*z[k−D_p,j]*), n represents an iteration index, k represents a sample index, z[k] is a k^th sample of the equalized signal, j is an integer index between 1 and M, D_p,j represents a j^th delay amount corresponding to a j^th delayed signal of the M delayed signals.

According to another embodiment of the present invention, a signal processing method includes following steps. M delayed signals of an equalized signal are generated, where M is a positive integer greater than 1. M weightings are adopted, and an m^th weighting of the M weightings is applied to an m^th delayed signal of the M delayed signals to generate an mth weighted signal, where m is an integer index between 1 and M. The M weighted signals are summed up to update the equalized signal. In the signal processing method, the M weightings are iteratively updated according to a vector {right arrow over (e)}_n,p=[e_n,p,1 . . . e_n,p,M], where the symbol e_n,p,j is defined as e_n,p,j=Σ_k(z[k]*z[k−D_p,j]*), n represents an iteration index, k represents a sample index, z[k] is a k^th sample of the equalized signal, j is an integer index between 1 and M, D_p,j represents a j^th delay amount corresponding to a j^th delayed signal of the M delayed signals.

The above and other aspects of the invention will become better understood with regard to the following detailed description of the preferred but non-limiting embodiments. The following description is made with reference to the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1(A) (prior art) is a function block diagram of a time-domain equalizer;

FIG. 1(B) is a function block diagram of a similar signal generating circuit in a time-domain equalizer;

FIG. 2 is a function block diagram of a time-domain equalizer according to an embodiment of the present invention;

FIG. 3 and FIG. 4 are partial function block diagrams of a controller according to the present invention; and

FIG. 5 is a flowchart of a signal processing method according to an embodiment of the present invention.

It should be noted that, the drawings of the present invention include functional block diagrams of multiple functional modules related to one another. These drawings are not detailed circuit diagrams, and connection lines therein are for indicating signal flows only. The interactions between the functional elements/or processes are not necessarily achieved through direct electrical connections. Further, functions of the individual elements are not necessarily distributed as depicted in the drawings, and separate blocks are not necessarily implemented by separate electronic elements.

DETAILED DESCRIPTION OF THE INVENTION

In a time-domain equalizer and a signal processing method provided by the present invention, a new cost function is defined, and computation is performed on the basis of a least mean square (LMS) method to accordingly identify a weighting most capable of effectively eliminating echo signals for each weighting circuit in the time-domain equalizer.

The concept of the present invention applied to the time-domain equalizer 100 shown in FIG. 1(A) and FIG. 1(B) is first discussed below. As previously stated, the output signal z of the time-domain equalizer 100 may be represented as:z[k]=y[k]−Σ_p=1^pΣ_m=1^Mw_p,mz[k−D_p,m]  (3)

One design target of the time-domain equalizer 100 of the present invention is to select appropriate time delay amount D_p,m and weighting w_p,m for the approximate signal generating circuit 110_p, such that an approximate signal s_p can be as approximate to the p^th echo signal as possible, so as to have the equalized signal z generated by the time-domain equalizer 100 in equation (3) be close to the original signal x, i.e., the influence of echo signals is eliminated.

A method adopted for determining the time delay amount D_p,m is described below. In practice, through fast Fourier transform (FFT), a receiver where the time-domain equalizer is located can estimated an approximate value (represented by D_p) of an actual arrival time delay Δ_p; however, it is difficult to calculate an accurate difference between the actual arrival time delay Δ_p and the approximate delay D_p. In the embodiment of the present invention, after learning the approximate delay D_p (a calculating method thereof being prior art), the time delay amount of each of the M delay elements included in the delay circuit 111_p can be configured according to the approximate delay D_p. For example, a circuit designer may select a reasonable lower difference threshold τ_min and a reasonable upper difference threshold τ_max based on experience. More specifically, the arrival time delay Δ_p is assumed to fall within the range from (D_p−τ_min) to (D_p+τ_max), and the time delay amounts, provided by the M delay elements, may be selected from this range. Taking M equal to 8 for example, if the time delay amount D_p,m provided by the delay element L_p,m is presented as (D_p+d_p,m), the time delay (D_p+d_p,1) provided by the delay element L_p,1 may be set as (D_p−τ_min), the time delay amount (D_p+d_p,8), provided by the delay element L_p,8 may be set as (D_p+τ_max), and the other six time delay amounts (D_p+d_p,2) to (D_p+d_p,7) may be set as equidistantly interpolated results between (D_p−τ_min) and (D_p+τ_max).

The method for determining the weighting w_p,m provided by the present invention is described below. A cost function is defined for p^th echo signal in P echo signals as below:e_n,p²=Σ_j=1^Me_n,p,j²  (4)

In equation (4), n represents an iteration index, j is an integer index between 1 and M, the e_n,p,j represents a correlation factor and is defined as:e_n,p,j=Σ_k(z[k]*z[k−D_p,j]*)  (5)

In equation (5), the delay amount D_p,j may also be rewritten as (D_p+d_p,j), and the symbol D_p is the foregoing approximate delay.

The effect of equation (5) is calculating the correlation between the equalized signal z and its delayed signal, and summing the accumulated correlation results (e.g., summing 5000 consecutive correlation results corresponding to the sample index k). Ideally, the time-domain equalizer substantially eliminates the echo signals in the received signal y, i.e., causing the output signal z of the equalizer 100 to be substantially the same to the original signal x. Theoretically, if beginning and end sections of the original signals x are not temporally correlated, the correlation of the original signal and its delayed signal approximates zero. Accordingly, the correlation between the ideal equalized signal z and its delayed signal also approximates zero. In conclusion, the more ideal a time-equalizer is, the calculation result of equation (5) approaches more to zero, also the smaller the calculation result of equation (4) becomes. The time-domain equalizer of the present invention is designed to iterate by using the LMS algorithm to identify the weighting w_p,m that minimizes the cost function e_n,p². The calculation and implementation details are given below.

First of all, equation (1) is substituted into equation (3), and the symbol of the weighting w_p,m is replaced by w_n,p,m (having denoted the iteration index n), and equation (3) is re-written as below:z[l]=x[k]+Σ_p=1^Pα_pe^jθp,kx[k−Δ_p]+n[k]−Σ_p=1^pΣ_m=1^Mw_n,p,mz[k−D_p,m]  (6)

In equation (6), the delay amount D_p,m can also be re-written as (D_p+d_p,m), and the symbol (D_p+d_p,m) is the foregoing approximate delay D_p.

Equation (6) is substituted into the expression z[k]*z[k−D_p,j]*:z[k]*z[k−D_p,j]*=(x[k]+Σ_p=1^Pα_pe^jθp,kx[k−Δ_p])*z[k−D_p,j]*+n[k]z[k−D_p,j]*−Σ_p=1^PΣ_m=1^Mw_n,p,mz[k−D_p,m]*z[k−D_p,j]*=(x[k]+Σ_p=1^Pα_pe^jθp,kx[k−Δ_p])*(x[k−D_p,j]+Σ_p=1^Pα_pe^jθp,kx[k−(D_p,j+Δ_p)]+n[k−D_p,j]−Σ_p=1^PΣ_m=1^Mw_n,p,mz[k−(D_p,m+D_p,j)])*+n[n]z[k−D_p,j]*−Σ_p=1^PΣ_m=1^Mw_n,p,mz[k−D_p,m]*z[k−D_P,j]*  (7)

It is assumed that the beginning and end sections of the original signal x are not temporarily correlated and have an average value of zero. Further, it is also assumed that the average value of noise signal is zero. If equation (7) is substituted into equation (5), after accumulating for a period of time, the accumulation values of some terms in equation (7) approximate zero, such that the calculation result of equation (5) can be simplified to approximate:e_n,p,j^≈Σ_k(α_pe^jθp,kx[k−Δ_p]x[k−D_p,j]*−Σ_m=1^Mw_n,p,mz[k−D_p,m]*z[k−D_p,j]*)  (8)

According to the concept of the LMS algorithm, partial differentiation is performed on equation (4) by using the weightings w_n,p,m as partial derivatives to obtain the weighting w_n,p,m that minimizes the cost function e_n,p². By introducing equation (8) during the partial differentiation, it is obtained that:

$\begin{array}{cc}\begin{array}{c}\frac{\partial e_{n,p}^2}{\partial w_{n,p,m}}=\sum _{j=1}^M\frac{\partial e_{n,p,j}^2}{\partial w_{n,p,m}}=\sum _{j=1}^{M}2e_{n,p,j}\frac{\partial e_{n,p,j}}{\partial w_{n,p,m}}\\ =\sum _{j=1}^M-2e_{n,p,j}*\left({\Sigma }_{k}z\left[k-D_{p,m}\right]*{z\left[k-D_{p,j}\right]}^{*}\right)\\ =-2{\stackrel{\rightharpoonup }{e}}_{n,p}*{\stackrel{\rightharpoonup }{z}}_{n,p,m}^T\end{array}& \left(9\right)\end{array}$

In the above equation, the vector {right arrow over (e)}_n,p=[e_n,p,1 . . . e_n,p,M] (refer to equation (5) for its definition), and the vector {right arrow over (z)}_n,p,m is:[Σ_kz[k−D_p,m]*z[k−D_p,1]* . . . Σ_kz[k−D_p,m]*z[k−D_p,M]*]  (10)

According to the deduced results of the LMS algorithm and equation (9), the weighting w_n,p can be updated by using the following equation, thus gradually minimizing the cost function e_n,p²:

$\begin{array}{cc}{\stackrel{\rightharpoonup }{w}}_{n+1,p}=\left[\begin{array}{c}{w}_{n+1,p,1}\\ ⋮\\ w_{n+1,p,M}\end{array}\right]={\stackrel{\rightharpoonup }{w}}_{n,p}-\mu \nabla e_{n,p}^2=\left[\begin{array}{c}{w}_{n,p,1}\\ ⋮\\ w_{n,p,M}\end{array}\right]+2\mu \left[\begin{array}{c}{\stackrel{\rightharpoonup }{e}}_{n,p}{\stackrel{\rightharpoonup }{z}}_{n,p,1}^T\\ ⋮\\ {\stackrel{\rightharpoonup }{e}}_{n,p}{\stackrel{\rightharpoonup }{z}}_{n,p,M}^T\end{array}\right]& \left(11\right)\end{array}$

In equation (11), μ represents an adjustable parameter in the LMS algorithm, is associated with the speed of iteration, and can be determined by a circuit designer instead of being limited to a predetermined value.

As shown in FIG. 2, if the concept of the present invention is applied to the time-domain equalizer in FIG. 1(A) and FIG. 1(B), the time-domain equalizer 100 may include a controller 130, which determines the vector {right arrow over (e)}_n,p=[e_n,p,1 . . . e_n,p,M] in the cost function e_n,p², and accordingly iteratively updates the weighting w_n,p used in the weighting circuits 112 in the similar signal generating circuit 110.

In an embodiment, the controller 130 updates the weighting w_n,p according to equation (11). More specifically, the controller 130 may first calculate the vector {right arrow over (e)}_n,p and the vector {right arrow over (z)}_n,p,m, and then multiply the product of the vector {right arrow over (e)}_n,p and the vector {right arrow over (z)}_n,p,m^T by 2 μ to obtain the last term in equation (11). By adding this term with the current weighting {right arrow over (w)}_n,p, an updated weighting {right arrow over (w)}_n+1,p is generated. FIG. 3 shows a partial function block diagram of the controller 130 of this embodiment, illustrating how calculation is performed by using the signal z[k] and the delayed signal z[k−D_p,j] generated by the delay circuit 111_p to update the weighting w_n,p,1 of the first multiplier (m=1) in the weighting circuit 112_p to w_n+1,p,1. A calculation circuit 131 performs calculation corresponding to equation (5) to determine the vector {right arrow over (e)}_n,p. A calculation circuit 132 performs calculation corresponding to equation (10) to determine the vector {right arrow over (z)}_n,p,m^T. A multiplier 133 multiplies output signals of the calculation circuits 131 and 132. A multiplier 134 multiples an output signal of the multiplier 133 by 2 μ. An adder 135 adds an output signal of the multiplier 134 and the weighting {right arrow over (w)}_n,p. In practice, the calculation circuits 131 and 132 may receive the delayed signal z[k−D_p,j], generated by the delay circuit 111_p, as its input signal.

One person can understand that, without departing from the spirit of the present invention, there are numerous other circuit configurations for implementing the controller 130. In practice, the controller may be realized by using various control and processing platforms, including fixed and programmable logic circuits, e.g., programmable logic gate arrays, application-specific integrated circuits, microcontrollers, microprocessors, and digital signal processors. Further, the controller 130 may be designed to complete its task through executing a processor command stored in a memory. It should be noted that, processing details of delay, vector transpose, addition, multiplication and summation are generally known to one person skilled in the art, and shall be omitted herein.

In addition to the typical LMS algorithm, the time-domain equalizer of the present invention may also determine the weighting w_n,p,m capable of minimizing the cost function e_n,p² by variations of the time-domain equalizer. Taking the improved proportionate normalized least mean square (IPNLMS) algorithm for example, the controller 130 may determine the M parameters [k_n,p,1 . . . k_n,p,M] for adaptively adjusting the vector {right arrow over (e)}_n,p and the vector {right arrow over (z)}_n,p,m^T further according to the weighting [w_n,p,1 . . . w_n,p,M]; that is, iteratively updating the weighting w_n,p according to an equation below:

$\begin{array}{cc}{\stackrel{\rightharpoonup }{w}}_{n+1,p}=\left[\begin{array}{c}{w}_{n+1,p,1}\\ ⋮\\ w_{n+1,p,M}\end{array}\right]=\left[\begin{array}{c}{w}_{n,p,1}\\ ⋮\\ w_{n,p,M}\end{array}\right]+\mu \left[\begin{array}{c}{k}_{n,p,1}{\stackrel{\rightharpoonup }{e}}_{n,p}{\stackrel{\rightharpoonup }{z}}_{n,p,1}^T\\ ⋮\\ k_{n,p,M}{\stackrel{\rightharpoonup }{e}}_{n,p}{\stackrel{\rightharpoonup }{z}}_{n,p,M}^T\end{array}\right]& \left(12\right)\end{array}$

The M parameters [k_n,p,1 . . . k_n,p,M] are generated according to an equation below:

$\begin{array}{cc}{\alpha }_{n,p,l}=\{\begin{array}{c}-0.5,w_{n,p,l}>0.1*\max \left(w_{n,p,l}\right)\\ +0.5,w_{n,p,l}\le 0.1*\max \left(w_{n,p,l}\right)\end{array}& \left(13\right)\\ k_{n,p,l}=\frac{1-a_{n,p.l}}{2M}+\left(1+{\alpha }_{n,p,l}\right)\frac{w_{n,p,l}}{2w_{n,p,l}+\varepsilon }& \left(14\right)\end{array}$

In equation (14), ϵ is an adjustable parameter in the IPNLMS, and I is an integer index between 1 and M.

$\begin{array}{cc}w_{n,p,l}=\sum _{l=1}^Mw_{n,p,l}& \left(15\right)\end{array}$

As shown in FIG. 4, the controller 130 further includes an adaptive adjusting circuit 136 and a multiplier 137. The adaptive adjusting circuit 136 generates the above parameters [k_n,p,1 . . . k_n,p,M]. The multiplier 137 multiplies the parameter k_n,p,1 generated by the adaptive adjusting circuit 136 by the input signal of the multiplier 133, so as to adaptively adjust the vector {right arrow over (e)}_n,p and the vector {right arrow over (z)}_n,p,m^T.

FIG. 5 shows a flowchart of a signal processing method according to another embodiment of the present invention. In step S501, an equalized signal is received, and M delayed signals are generated accordingly, where M is a positive integer greater than 1. In step S502, M weightings are adopted, and an m^th weighting of the M weightings is applied to an m^th delayed signal of the M delayed signals to generate an m^th weighted signal, where m is a positive integer between 1 and M. In step S503, the M weighted signals are summed up to update the equalized signal. In step S504, the M weightings used in the step S502 are iteratively updated according to the vector e_n,p=[e_n,p,1 . . . e_n,p,M], where the symbol e_n,p,j is defined as:e_n,p,j=Σ_k(z[k]*z[k−D_p,j]*)  (16)

In equation (16), n represents an iteration index, k represents a sample index, z[k] is a k^th sample of the equalized signal, j is an integer index between 1 and M, and D_p,j represents a j^th time delay amount corresponding to a j^th delayed signal of the M delayed signals.

One person skilled in the art can understand that the operation variations applied to the time-domain equalizer 100 are applicable to the signal processing method in FIG. 5, and shall be omitted herein.

It should be noted that, the mathematical expressions in the disclosure are for explaining principles and logics associated with the embodiments of the present invention. Unless otherwise specified, these mathematical expressions do not form limitations on the present invention. One person skilled in the art can understand that, there are numerous technologies for realizing the physical forms corresponding to these mathematical expressions.

While the invention has been described by way of example and in terms of the preferred embodiments, it is to be understood that the invention is not limited thereto. On the contrary, it is intended to cover various modifications and similar arrangements and procedures, and the scope of the appended claims therefore should be accorded the broadest interpretation so as to encompass all such modifications and similar arrangements and procedures.

Claims

1. A time-domain equalizer, comprising: a delay circuit, receiving an equalized signal, and accordingly generating M delayed signals, where M is a positive integer greater than 1;a weighting circuit, adopting M weightings, and applying an mth weighting of the M weightings on an mth delayed signal of the M delay signals to generate an mth weighted signal, where m is an integer index between 1 and M, wherein M weighted signals are generated accordingly;a controller, iteratively updating the M weightings according to a vector {right arrow over (e)}n,p=[en,p,1 . . . en,p,M], where en,p,j is defined as en,p,j=Σk(z[k]*z[k−Dp,j]*), n represents an iteration index, k represents a sample index, z[k] is a kth sample of the equalized signal, p corresponds to a pth echo signal, j is an integer index between 1 and M, Dp,j represents a jth time delay amount corresponding to a jth delayed signal of the M delayed signals, and z[k−Dp,j]* represents z[k] delayed by delay Dp,j; andsummation circuit, summing up the M weighted signals, according to which the equalized signal is updated.

2. The time-domain equalizer according to claim 1, wherein the controller updates the M weightings adopted by the weighting circuit according to an equation:

3. The time-domain equalizer according to claim 1, wherein: the controller further comprises an adaptive adjusting circuit that determines M parameters [kn,p,1 . . . kn,p,M] according to the M weightings [wn,p,1 . . . wn,p,M]; andthe controller updates the M weightings according to an equation:

4. A signal processing method, comprising: receiving an equalized signal, and accordingly generating M delayed signals, where M is a positive integer;adopting M weightings, and applying an mth weighting of the M weightings to an mth delayed signal of the M delay signals to generate an mth weighted signal, where m is an integer index between 1 and M;iteratively updating the M weightings according to a vector {right arrow over (e)}n,p=[en,p,1 . . . en,p,M]. where en,p,j is defined as en,p,j=Σk(z[k]*z[k−Dp,j]*), n represents an iteration index, k represents a sample index, z[k] is a kth sample of the equalized signal, p corresponds to a pth echo signal, j is an integer index between 1 and M, Dp,j represents a jth time delay amount corresponding to a jth delayed signal of the M delayed signals, and z[k−Dp,j]* represents z[k] delayed by delay Dp,j; andsumming up the M weighted signals to update the equalized signal.

5. The signal processing method according to claim 4, wherein the step of iteratively updating the M weightings comprising updating the M weightings according to an equation:

6. The signal processing method according to claim 4, further comprising: determining the M parameters [kn,p,1 . . . kn,p,M] according to the M weightings [wn,p,1 . . . wn,p,M];and the step of iteratively updating the M weightings comprises updating the M weightings according to an equation: