KR20070061244A - Apparatus and method for stable dfe using selective fbf - Google Patents

Abstract

An apparatus and a method for equalizing a determination feedback using a feedback filter selectively are provided to stably operate a determination feedback equalizer by comparing a threshold value with a mean square error of the determination feedback equalizer. An apparatus for equalizing a determination feedback using a feedback filter selectively includes an equalizing unit(41), a divergence/convergence determination unit, and a filter control unit. The equalizing unit(41) has an FFF(FeedForward Filter)(410) which compensates a distorted transmission channel by receiving a match-filtered signal, and an FBF(FeedBack Filter)(420) which reduces an ISI(Inter Symbol Interference) of the compensated transmission channel. The equalizing unit(41) operates the FFF(410) in a blind mode, and operates the FFF(410) and the FBF(420) in a determination directivity mode. The divergence/convergence determination unit determines divergence/convergence of the determination feedback equalizing apparatus based on an MSE(Mean Square Error) which is acquired through an LMS(Least Mean Square). The filter control unit controls the equalizing unit(41) to be operated in the blind mode if the divergence is determined. The filter control unit controls the equalizing unit(41) to be operated in the determination directivity mode if the convergence is determined.

Description

Crystal feedback equalizer and selective method using a feedback filter {Apparatus and Method for Stable DFE using selective FBF}

1 is a block diagram of an embodiment of a cable modem using a decision feedback equalizer using a feedback filter selectively according to the present invention;

2 is a diagram illustrating an embodiment of an impulse response of the SRRC filter of FIG. 1 according to the present invention;

3 is a flow diagram of an embodiment of a decision feedback equalization method selectively using a feedback filter according to the present invention;

4A and 4B are detailed diagrams of an embodiment of a decision feedback equalizer using the feedback filter of FIG. 1 according to the present invention;

5 is a detailed configuration diagram of an embodiment of the multi-coefficient calculation unit (MMA) of FIG. 4B according to the present invention.

* Explanation of symbols on the main parts of the drawing

410: feed forward filter 420: feedback filter

440: multi-factor calculation unit 450: slicing unit

460: mean square error calculation unit 470: divergence / convergence determination unit

471: First threshold comparison unit 472: Second threshold comparison unit

473: counter calculation unit 474: third threshold value comparison unit

480: filter control unit

The present invention relates to a decision feedback equalizer and a method for selectively using a feedback filter, and more particularly, to determine whether the equalizer diverges / converges using Mean Square Error (MSE). In case of divergence, the equalizer can be stably operated by stopping the feedback filter to operate in the blind mode, and operating the feedback filter to operate in the decision-oriented mode when converging. A decision feedback equalizer and a method for selectively using a feedback filter, which can increase the recognition rate of a received signal, and a method thereof.

In a digital communication system, due to the band-limited channel characteristics, various signals are generated as a signal transmitted from a transmitter passes through a transmission channel. Factors causing such distortion include Gaussian thermal noise, impulse noise, and additive or multiplying noise due to fading in which signal strength fluctuates in time, frequency variation, nonlinearity, and temporal dispersion. Adjacent symbols affect each other by such distortion, and interference between adjacent symbols acts as a major factor that degrades the performance of a communication system. It is the equalizer to minimize the interference between these adjacent symbols. That is, the equalizer compensates for the magnitude and delay characteristics of the signal received at the receiver, thereby improving the quality of the communication line without increasing the power of the transmitted signal or increasing the channel bandwidth.

Typical application algorithms used in the equalizer include a Least Mean Square (LMS) algorithm and a Recursive Least Square (RLS) algorithm. The LMS algorithm minimizes the mean square error (MSE) between the received signal and the determined signal. It is simpler to formulate and uses less hardware than the RLS algorithm, but the channel adaptation speed is slow. The RLS algorithm is an algorithm that minimizes the sum of squares of the weighted error signals, and it is possible to equalize the channel more effectively than the LMS algorithm by updating the filter coefficient using a cyclic method, but it has a disadvantage of complicated hardware.

In general, the equalizer is called a data aided equalizer if there is a known training symbol, or a blind equalizer if there is no known training symbol.

Blind equalizer algorithms include Reduced Constellation Algorithm (RCA), Constant Modulus Algorithm (CMA) and Algorithm using multiple coefficients.

First, the RCA algorithm starts the channel adaptation by reducing the constellation of the transmission signal and returns the adaptation after the channel adaptation.

Next, the CMA algorithm is a blind algorithm that draws a circle around the origin of the array and calculates the distance from the circle and applies the tap coefficient in the direction of reducing the distance. In terms of convergence speed, it shows a slow convergence speed when the eye pattern is closed and a fast convergence speed when the eye pattern is open.

Finally, the MultiModulus Algorithm (MMA) algorithm is similar to the CMA algorithm, but it adapts the tap coefficient in the direction of setting the reference value on the imaginary axis and the real axis and reducing the distance from the reference. It is a proposed algorithm suitable for quadrature modulation such as modulation, carrierless amplitude and phase modulation (CAP).

In addition, the equalizer is called a linear equalizer if there is no feedforward filter (FFF) according to the presence of a feedforward filter (FFF), and a non-linear equalization if FFF is present. It is called.

As an example, a system using a decision feedback equalizer will be described using a CATV (Cable Television) modem in a hybrid fiber coaxial (HFC) network as an example. However, it should be noted that the cable system using the crystal feedback equalizer according to the present invention is not limited to the CATV (Cable Television) modem in the HFC network.

The Data Over Cable Service Interface Specification (DOCSIS) standard, launched in 1997 by CableLabs in the United States, aims to transmit and receive broadcast and digital data using the HFC network. The DOCSIS 3.0 specification is currently in progress, and DOCSIS 3.0 requires hundreds of Mbps. For this high speed data communication, bandwidth-efficient modulation and demodulation methods such as 64QAM and 256QAM are used.

The CATV (Cable Television) modem does not use the same structure as the preamble, so the channel must be compensated with the received symbol, and the blind equalizer performs this function.

In the conventional decision feedback equalization technique, a "bridge" that operates a decision feedback equalizer after compensating nulls by a bridged tap (BT) using a null compensation filter and a null tracker in front of the equalizer at the receiving end is used. High-speed digital subscriber network modem (Domestic Patent Registration No. 10-2002-0079723, registered on Dec. 13, 2002) having an adaptive compensation filter of null by the tap (hereinafter referred to as 'first prior art') The first prior art transmits using a null tracking unit and a null compensation filter to compensate for frequency nulls caused by branching of a subscriber line in front of the equalizer in a CAP or QAM high speed digital subscriber network modem. A method of minimizing an error rate and a data receiving apparatus using the method are provided.

In the conventional blind equalization technique, since there are many signal levels in the case of higher-order QAM, a blind equalization method and apparatus in a receiver using a blind algorithm (MMA) that combines the groups into several groups (US Patent No. 1997-0666963, 1997) 12. 09 registration) (hereinafter referred to as 'second prior art'), the second prior art relates to blind equalization in a receiver of a communication equipment, which is a multi-coefficient blind equalization technique. Algorithm

The first prior art is a technique for compensating for frequency nulls to minimize transmission error rate, and the second prior art is a technique for stably performing blind equalization using a multi-coefficient algorithm. In the second prior art), there is a problem in that there is a limit in reducing the transmission error rate because both the feedforward filter and the feedback filter are operated in the blind mode (see [Table 3]).

The present invention has been proposed in order to meet the above demands. In the case of determining whether the equalizer diverges / converges using Mean Square Error (MSE), and in accordance with the determined result, the blind mode By stopping the feedback filter so as to operate the feedback filter and operating the feedback filter to operate in the decision-oriented mode when converging, the equalizer can be stably operated and the recognition rate of the received signal can be increased. It is an object of the present invention to provide a decision feedback equalizer and a method for selectively using a feedback filter.

Other objects and advantages of the present invention can be understood by the following description, and will be more clearly understood by the embodiments of the present invention. It will also be appreciated that the objects and advantages of the present invention may be realized by the means and combinations thereof indicated in the claims.

According to an aspect of the present invention, a feedback feedback equalization device selectively using a feedback filter includes a feedforward filter (FFF) and a compensation for compensating a distorted transmission channel by receiving a matched filtered signal. It includes a feedback filter (FBF: FeedBack Filter) for reducing the Inter Symbol Interference (ISI) of the transport channel, in the blind mode, only the feedforward filter is operated, in the decision-oriented mode the feedforward filter and the feedback filter is operated Equalization means for; Divergence / convergence determination means for judging divergence / convergence of the decision feedback equalizer using a Mean Square Error (MSE) obtained through a Least Mean Square (LMS) algorithm; And a filter control means for controlling the equalization means to operate in a blind mode if it is determined to diverge, and to control the equalization means to operate in a decision-oriented mode according to the determined divergence / convergence. .

Meanwhile, the present invention provides an equalization method applied to a decision feedback equalizer including a feed forward filter and a feedback filter, comprising: an initial blind equalization step of performing equalization by operating only the feed forward filter; An divergence / convergence determination step of determining whether divergence / convergence of the decision feedback equalization device is diverged using an average squared error value obtained through an LMS algorithm for the equalization result equalized in the initial blind equalization step; As a result of the determined divergence / convergence determination, if it is determined that divergence is determined, blind equalization that operates only the feedforward filter is performed, and feedback to the divergence / convergence determination step is performed to determine divergence / convergence with respect to the blind equalization result. Blind equalization step; And the divergence / convergence determination result, if it is determined that the convergence is performed, the feedforward filter and the feedback filter operate to perform decision-oriented equalization, and to diverge / converge the decision-oriented equalization result. A decision-oriented equalization step that feeds back to the convergence decision step.

The tap coefficients of the decision feedback equalizer are updated using the Least Mean Square (LMS) algorithm, and the equalizer is stable by checking the divergence of the equalizer using the mean square error (MSE) and the threshold of the decision feedback equalizer. In the blind mode, the FBF (Feedback Filter) is stopped, and in the decision-oriented mode, the FBF is operated and the input of the FBF can be selectively selected.

The above objects, features and advantages will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, whereby those skilled in the art may easily implement the technical idea of the present invention. There will be. In addition, in describing the present invention, when it is determined that the detailed description of the known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram of an embodiment of a cable modem using a decision feedback equalizer using a feedback filter selectively according to the present invention.

A cable modem using a decision feedback equalizer using a feedback filter according to the present invention, as shown in FIG. 1, includes a transmitter 11 and a receiver 13. Here, the transmission unit 11 is a transmission bit generation unit 111 for generating an arbitrary bit, a modulation unit (M-ary modulation) 112 for mapping the input bit string to a symbol of 64QAM or 256QAM, a symbol An upsampling unit 113 for up-sampling at a constant multiple of the speed, and an SRRC (Square Root Raised Cosine) filter 114 that is a matching filter for minimizing the effects of noise added during signal transmission.

In addition, the receiver 13 includes an SRRC filter 135 used as a matching filter, a down sampling unit 134 for down sampling at a constant multiple of the symbol rate, a blind equalizer 133 for compensating a channel, And a demodulation unit 132 for demodulating according to the data rate and a reception bit storage unit 131 for storing the received bits.

In addition, in FIG. 1, the channel 200 and the additive white gaussian noise (AWGN) for modeling the influence of the HFC network are added.

Parameters used in each component of FIG. 1 are shown in [Table 1].

Block 1 parameter M-ary modulation Bits per symbol: 6 (64QAM) UP sampling section 4 times the symbol speed SRRC Filter Length: -16T ~ 16T, Alpha = 0.2 Down sampling 2 times the symbol rate M-ary demodulation Bits per symbol: 6 (64QAM)

2 is an exemplary diagram illustrating an impulse response of the SRRC filter of FIG. 1 according to the present invention.

When the impulse response is g (t) on the time axis of the SRRC filter 135, g (t) is expressed by Equation 1 below.

3 is a flow diagram of an embodiment of a decision feedback equalization method selectively using a feedback filter in accordance with the present invention.

First, the decision feedback equalizer DFE initializes tap coefficients and step sizes mu_b and mu_dd of the feedforward filter FFF and the feedback filter FBF (301). In operation 302, the feedforward filter FFF operates in a blind mode.

Thereafter, the first threshold comparison unit 471 compares the mse_dd value, which is the mean square error MSE, with the first threshold value thr 1, and determines whether the mean square error value is less than the first threshold value (303). . When the mse_dd value, which is the mean square error MSE, is greater than or equal to the first threshold value in the determination result of "303", the decision feedback equalizer DFE is divergent, so that the step sizes mu_b and mu_dd are reduced (304), and the feed forward is performed. The process of initializing the tap coefficients and the step sizes mu_b and mu_dd of the filter FFF and the feedback filter FBF is repeated (301).

As a result of determination “303”, when the mse_dd value of the mean square error MSE is less than the first threshold value threshold 1, the second threshold comparison unit 472 determines that the mse_dd value of the mean square error MSE is the second threshold value. compare again to (thr 2) (305). If the mse_dd value, which is the mean square error (MSE), is greater than or equal to the second threshold value (thr 2), the counter is set to 0 (306), and the blind mode in which only the feed forward filter (FFF) is operated is operated. The process is repeated (302).

As a result of the determination of "305", the counter calculator 473 increases the counter cnt by 1 while continuously operating in the blind mode when the mse_dd value, which is the mean square error MSE, is less than the second threshold value thr 2 ( 307).

Thereafter, the third threshold comparison unit 474 determines whether the increased counter value is greater than the third threshold thr 3 (308). As a result of the determination of "308", if the increased counter value is less than or equal to the third threshold 3, the process of operating in the blind mode in which the counter value is set to 0 and only the feedforward filter FFF is operated is repeated (302).

As a result of the determination of “308”, when the increased counter value exceeds the third threshold thr 3, the feedforward filter FFF and the feedback filter FBF operate in the decision-oriented mode.

At this time, the mean square error (mse_dd) is continuously checked whether the value is greater than or less than the first and second threshold values, and whether the counter value is greater or less than the third threshold value.

Table 2 shows the parameters used in the crystal feedback equalizer according to the present invention.

parameter value Number of taps in the feed forward filter (FFF) 24 Initial value of the feed forward filter (FFF) [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0] Number of taps in the feedback filter (FBF) 6 Initial value of feedback filter (FBF) [0, 0, 0, 0, 0, 0] First threshold 1 0.5 Second threshold (thr 2) 0.01190476 Third threshold (thr 3) 64 Step size of blind mode (mu_b)

Size of decision-oriented mode (mu_dd)

4A and 4B are detailed configuration diagrams of an embodiment of the crystal feedback equalizer using the feedback filter of FIG. 1 according to the present invention. Here, A, B, C, and D of FIG. 4A and A ', B', C 'and D' of FIG. 4B are connected to A-A ', B-B', C-C 'and D-D', respectively. It is.

The decision feedback equalizer which selectively uses the feedback filter includes an equalizer 41, a divergence / convergence determination unit 42, and a filter controller 43. The equalizer 41 includes a feedforward filter 410 and a feedback filter. 420, a down sampling unit 430, and the divergence / convergence determination unit 42 includes a slicing unit 450, a subtraction unit 451, an average square error calculation unit (MSE) 460, The first threshold comparator 471, the second threshold comparator 472, the counter calculator 473, and the third threshold comparator 474 are included, and the filter controller 43 is a multi-factor generator. 440, a first multiplexer 481, a second multiplexer 482, and a third multiplexer 483. Hereinafter, each component will be described.

As shown in FIG. 4A, the feedforward filter (FFF) 410 includes a delay unit 411, a conjugate complex number 412, a multiplier 413, an adder 414, a delay unit 415, and a multiplication. Section 416.

The input x (n / K) of the feedforward filter (FFF) 410 may be a constant multiple of the symbol rate or may be a symbol rate. As an example of the present invention, K has a value of 2 and K of 2 means that a value oversampled twice the symbol rate is input and two samples per symbol are input.

The input xb (n) of the feedback filter (FBF) 420 is an output of the slicing unit 450 and is input at a symbol rate, and the feedback filter (FBF) 420 is a delay unit (shown in FIG. 4A). 421, a conjugate complex number 422, a multiplier 423, a multiplexer 424, an adder 425, a delay unit 426, and a multiplier 427.

Since the sum of the feed forward filter (FFF) 410 and the feedback filter (FBF) 420 is output at twice the symbol rate, it is output at the symbol rate while passing through the down sampling unit 430. In this case, the K value of the down sampling unit 430 is 2 and the output y (n) is expressed by Equation 2 below.

y (n) is input to the multi-coefficient generator (MMA) 440 and the slicing unit 450.

5 is a detailed configuration diagram of an embodiment of the multi-coefficient generator (MMA) of FIG. 4B according to the present invention.

As shown in FIG. 5, the multi-coefficient generator 440 includes a splitter 51 for dividing an input complex number into a real part and an imaginary part, and a square part 52 that squares the real part and the imaginary part, respectively. 53), subtracting sections 54 and 56, and multiplication sections 55 and 57. The multi-coefficient generator 440 outputs a first error value err_b (n), which is a complex signal, and the first error value err_b (n) is expressed by Equation 3 below.

The value obtained by subtracting y (n) from the slicer output d (n) output from the slicing unit 450 becomes the second error value err_dd (n) and is expressed by Equation 4 below.

The mean square error value mse_dd (n), which is the output of the mean square error calculator 460 that receives the second error value err_dd and calculates the mean square error, is expressed by Equation 5 below.

The divergence / convergence determination unit 42 includes a slicing unit 450, a subtraction unit 451, a mean square error calculation unit (MSE) 460, a first threshold comparison unit 471, and a second unit. The threshold comparator 472 includes a counter calculator 473 and a third threshold comparator 474.

The first threshold comparator 471 compares the mean square error value calculated by the mean square error calculator 460 with the first threshold value, and determines that the average square error value is divergent when the average square error value is greater than or equal to the first threshold value. To reduce the step size and initialize the tap coefficients.

The second threshold comparator 472 determines to diverge when the mean square error value is less than the first threshold value and is greater than or equal to the second threshold value, and controls to set the counter value to zero.

The counter calculator 473 increments the counter value by 1 only when the mean square error value is less than or equal to the second threshold value.

The third threshold comparator 474 determines that the counter value is set to 0 when the increased counter value is less than the third threshold value, and diverges when the counter value is greater than or equal to the third threshold value.

That is, the divergence / convergence determining unit 42 multiplexes the first error value err_b in the blind mode and the second error value err_dd in the decision-oriented mode by comparing the mean square error value with the first or second threshold value. Generates a multiplex signal DFE_b. The multiplexed signal DFE_b is input to the filter controller 43 to control the equalizer 41.

The filter control unit 43 includes a multi-coefficient generator 440, a first multiplexer 481, a second multiplexer 482, and a third multiplexer 483. The filter control unit 43 stops the feedback filter to operate in the blind mode when it is determined to diverge from the divergence / convergence determination unit 42, and to operate to the decision-oriented mode when it is determined to converge. Perform a function to operate.

The first multiplexer 481 inputs 0 to the feedback filter 420 in order to operate in a blind mode when it is determined to be divergent according to the determination result of the divergence / convergence determination unit 42, and determines the convergence direction mode when it is determined to be convergent. The slicer output d (n) is input to the feedback filter 420 in order to operate as.

The second multiplexer 482 outputs a first error value err_b (n) to operate in a blind mode when it is determined to be divergence according to the determination result of the divergence / convergence determination unit 42. The second error value err_dd (n) is output.

The third multiplexer 483 outputs the first step size mu_b if it is determined to be divergent based on the determination result of the divergence / convergence determination unit 42, and outputs the second step size mu_dd if it is determined to be convergent. do.

The multiplier 484 multiplies the second step size output from the third multiplexer 483 and the output value of the second multiplexer 482. The multiplied value is input to the feedforward filter 410 and the feedback filter 420 as coefficients.

When the multiplexed signal DFE_b is 0, the multiplexer 424 has a value of 0, and the first multiplexer 481 inputs a value of xb (n), that is, 0 to the feedback filter 420. In addition, the second multiplexer 482 causes the error value err (n) to be the first error value err_b (n) of the decision-oriented mode, and the third multiplexer 483 has a step size mu. The first step size mu_b of the blind mode is obtained. The decision feedback equalizer DFE operates in the blind mode, and the output value of the feedback filter FBF becomes 0, so that only the feed forward filter FFF operates. When DFE_b is 1, the blind equalizer operates in the decision directing mode, and the first multiplexer 481 to the third multiplexer 483 and the multiplexer 424 output a value input as 1.

[Table 3] below shows the case in which the output feedback DFE_b of the divergence / convergence determination unit 42 is forcibly set to 1 when the decision feedback equalizer DFE operates in the blind mode (first case). It is a result of simulating a bit error rate (BER) when the blind equalization device according to the present invention is operated (second case).

Bit Energy-to-Noise Ratio (Eb / N0) 14 dB Modulation 64QAM BER of the first case 3.2895e-3 BER of the second case 2.2284e-3

In Table 3, the BER in the case of using the crystal feedback equalizer according to the present invention (second case) was 2.2284e-3, and the BER in the forced setting (first case) was 3.2895e-3. Therefore, the method using the crystal feedback equalizer turns out to have a low error rate.

The present invention described above is capable of various substitutions, modifications, and changes without departing from the technical spirit of the present invention for those skilled in the art to which the present invention pertains. It is not limited by the drawings.

As described above, the present invention can operate stably with different inputs in the blind mode and the decision-oriented mode, and can increase the recognition rate of the received signal.

That is, according to the present invention, tap coefficients of the decision feedback equalizer are updated through a Least Mean Square (LMS) algorithm, and the mean square error (MSE) of the decision feedback equalizer is compared with a threshold to determine the equalizer. By checking for divergence, the decision feedback equalizer (DFE) is operated stably, the feedback filter (FBF) is temporarily stopped when the decision feedback equalizer (DFE) operates in the blind mode, and operated again in the decision-oriented mode. By doing so, there is an effect that can improve the performance of the crystal feedback equalizer (DFE).

Claims (9)

In the decision feedback equalizer which selectively uses a feedback filter, A feed forward filter (FFF) for compensating for the distorted transmission channel by receiving the matched filtered signal and a feedback filter (FBF) for reducing inter symbol interference (ISI) of the compensated transmission channel Equalization means for operating only the feedforward filter in a blind mode and operating the feedforward filter and the feedback filter in a decision-oriented mode; Divergence / convergence determination means for judging divergence / convergence of the decision feedback equalizer using a Mean Square Error (MSE) obtained through a Least Mean Square (LMS) algorithm; And Filter control means for controlling the equalization means to operate in a blind mode if it is determined that divergence is determined according to the determined divergence / convergence. A decision feedback equalization device for selectively using a feedback filter comprising a. The method of claim 1, The equalization means, A feedforward filter for receiving a matched filtered signal and filtering to compensate for the distorted transmission channel by using different error values and the input matched filtering signal according to the blind mode or the decision directing mode; A feedback filter for receiving a slicer output signal and filtering to reduce ISI of the compensated transmission channel in the decision-oriented mode; Adding means for adding a filtering signal of the feedforward filter and a filtering signal of the feedback filter; And Down-sampling means for down-sampling the added signal at a symbol rate to output an equalized signal A decision feedback equalization device for selectively using a feedback filter comprising a. The method according to claim 1 or 2, The divergence / convergence determination means, Slicing means for receiving an output signal from the down sampling means to generate the slicer output signal and subtracting the generated slicer output signal from the output signal to calculate a second error value; Mean square error calculation means for calculating a mean square error value according to the LMS algorithm using the second error value; First threshold comparison means for controlling to diverge if the mean squared error value is equal to or greater than the first threshold value and to reduce the step size and initialize the tap coefficient; Second threshold comparison means for determining a divergence if the mean square error value is less than the first threshold but greater than a second threshold, and sets a counter value to zero; Counter calculation means for incrementing the counter value by one only when the mean square error value is less than or equal to the second threshold value; And Third threshold comparison means for determining if the increased counter value is less than a third threshold value and restoring the counter value to zero and diverging; if it is greater than or equal to the third threshold value, third threshold value comparing means A decision feedback equalization device for selectively using a feedback filter comprising a. The method of claim 3, wherein The filter control means, Multi-coefficient generating means for receiving the output signal and generating a first error value using a multi-coefficient algorithm; If it is determined that the divergence / convergence determination means is divergent, the equalizer means inputs 0 to the feedback filter to operate in a blind mode, and if it is determined to converge, output the slicer to the feedback filter to operate the equalization means in decision-directed mode. First multiplexing means for inputting; Second multiplexing means for outputting the first error value to operate in a blind mode if it is determined to diverge and for outputting the second error value if determined to be convergent; Third multiplexing means for outputting a first step size if the divergence is determined, and outputting a second step size if the convergence is determined; And Multiplication means for multiplying a second step size output from said third multiplexing means with an output value of said second multiplexing means; A decision feedback equalization device for selectively using a feedback filter comprising a. The method of claim 4, wherein The multi-factor generating means, A decision feedback equalization device selectively using a feedback filter, wherein the first error value is generated using Equation 1 below. [Equation 1]

(here,

Is the first error value,

Indicates output signal.) The method of claim 5, The mean square error calculation means, A decision feedback equalization device selectively using a feedback filter, wherein the mean square error value is calculated using Equations 2 and 3 below. [Equation 2]

(here,

Is the second error value,

Silver slicer output,

Indicates output signal.) [Equation 3]

(here,

Is the mean squared error value,

Represents the second error value.) An equalization method applied to a decision feedback equalizer including a feedforward filter and a feedback filter, An initial blind equalization step of performing equalization by operating only the feedforward filter initially; An divergence / convergence determination step of determining whether divergence / convergence of the decision feedback equalization device is diverged using an average squared error value obtained through an LMS algorithm for the equalization result equalized in the initial blind equalization step; As a result of the determined divergence / convergence determination, if it is determined that divergence is determined, blind equalization that operates only the feedforward filter is performed, and feedback to the divergence / convergence determination step is performed to determine divergence / convergence with respect to the blind equalization result. Blind equalization step; And As a result of the determined divergence / convergence determination, if it is determined that convergence is performed, the feedforward filter and the feedback filter operate to perform decision-oriented equalization, and to determine whether divergence / convergence is determined for divergence / convergence with respect to the decision-oriented equalization result. Decision-Oriented Equalization Step Feedback to Convergence Decision Equalization method applied to the decision feedback equalizer for selectively using a feedback filter comprising a. The method of claim 7, wherein The divergence / convergence determination step, A first error value is calculated from the output signal of the feedforward filter using a multi-coefficient algorithm, the output signal of the feedforward filter is sliced to generate a slicer output signal, and the second error is generated using the generated slice output signal. An error / average square error calculation step of calculating a value and calculating a mean square error value according to an LMS algorithm using the calculated second error value; A first threshold value comparing step of determining a divergence if the mean squared error value is equal to or greater than the first threshold value and reducing a step size and initializing a tap coefficient; A second threshold value comparing step of setting a counter value to 0 when the average square error value is less than the first threshold value and is equal to or greater than a second threshold value; A counter calculation step of incrementing a counter value by one only when the mean square error value is less than or equal to the second threshold value; And A third threshold comparison step of determining that the counter value is set to 0 when the increased counter value is less than a third threshold value, and diverging when the counter value is greater than or equal to the third threshold value; Equalization method applied to the decision feedback equalizer for selectively using a feedback filter comprising a. The method according to claim 7 or 8, The filter control step, In the divergence / convergence determination step, if the divergence is determined to be controlled, the feedback filter is inputted to the feedback filter to operate in the blind mode, and if the convergence is determined, the slicer output is input to the feedback filter to operate in the decision-oriented mode. An input control step; And If it is determined that the divergence is determined, the decision feedback equalizer is controlled to operate in the blind mode by using the first error value and the first step size according to a multi-coefficient algorithm, and when the convergence is determined, the second error value and An error / step size control step of controlling the decision feedback equalizer to operate in the decision-directed mode using two step sizes Equalization method applied to the decision feedback equalizer for selectively using a feedback filter comprising a.

Images