KR100860503B1 - Adaptive decision feedback equalizer using the hierarchical feedback filter and soft decision device - Google Patents

Abstract

An adaptive decision feedback equalizer is provided to prevent erroneous determination of an estimated symbol and to improve performance of a receiver by applying a soft decision device to a signal determination device using a hard decision device. An adaptive decision feedback equalizer(1) includes equalizer filters(10,30), a filter coefficient storage device, a signal output device, and a signal determination device(50). The equalizer filters prevent error propagation by removing Pre-ISI(Precursor-Inter Symbol Interference) or Post-ISI(Postcursor-Inter Symbol Interference) of an input signal having actual information based on an inputted training sequence. The filter coefficient storage device stores filter coefficients for removing the Pre-ISI or the Post-ISI measured by the equalizer filters. The signal output device multiplies the input signal with the filter coefficients stored in the filter coefficient storage device to compensate for distortion of the input signal. The signal determination device estimates accurate information about signals outputted to a soft decision device(53) and a hard decision device(51).

Description

Adaptive Decision Feedback Equalizer Using the Hierarchical Feedback Filter and Soft Decision Device

1 is a block diagram schematically illustrating a decision feedback equalizer using a hierarchical feedback filter and a soft decision device according to the present invention;

FIG. 2 is a schematic diagram illustrating a symbol feedback equalizer using a hierarchical feedback filter and a soft decision device according to the present invention. FIG.

3 illustrates a signal determination means and an adaptive equalization algorithm in a decision feedback equalizer using a hierarchical feedback filter and a soft decision device according to the present invention.

4 schematically illustrates a feedback equalization filter in a decision feedback equalizer using a hierarchical feedback filter and a soft decision device in accordance with the present invention.

5 is a flowchart schematically illustrating a method of driving a decision feedback equalizer using a hierarchical feedback filter and a soft decision device according to the present invention.

6 is a flowchart schematically illustrating a method of driving the soft decision apparatus of FIG. 5.

7 is a BER performance graph in the S-V first channel of the decision feedback equalizer using the hierarchical feedback filter and soft decision device according to the present invention and the decision feedback equalizer according to the prior art.

8 is a graph of BER performance in the S-V second channel of the crystal feedback equalizer using the hierarchical feedback filter and the soft decision device according to the present invention and the crystal feedback equalizer according to the prior art.

9 is a graph of the BER performance of the S-V third channel of the crystal feedback equalizer using the hierarchical feedback filter and the soft decision device according to the present invention and the crystal feedback equalizer according to the prior art.

10 is a graph of the BER performance of the S-V fourth channel of the decision feedback equalizer using the hierarchical feedback filter and soft decision device according to the present invention and the decision feedback equalizer according to the prior art.

      <Brief description of reference numerals for the main parts of the drawings>

1: Decision feedback equalizer using hierarchical feedback filter and soft decision device

10: feedforward equalization filter 11: delay portion

13: multiplication unit 15: adding unit

30: feedback equalization filter 30A: first hierarchical structure

31 A: sub-feedback equalization filter 33 A: multiplier

35A: summing 37A: adaptive equalization algorithm

30B: second hierarchical structure 31B: multiplier

33B: summer 40: adaptive equalization algorithm

50: signal determination means 51: hard decision device

53: soft decision device 60: adaptive equalization algorithm

The present invention relates to a crystal feedback equalizer using a hierarchical feedback filter and a soft decision device. More specifically, the feedback equalization filter uses a hierarchical structure, and by measuring errors in steps to remove the feedback. It is possible to prevent errors caused by wrong estimation symbols in the filter, and to reduce the number of taps of the feedback equalization filter to compensate for the distortion of the signal faster by using a larger sized step. By applying a soft decision device, which is an equalizer, to a signal decision means using a hard decision device, a decision feedback equalizer using a hierarchical feedback filter and a soft decision device which prevents erroneous determination of an estimated symbol and improves the receiver performance. It is about.

In general, in wireless transmission and reception of data, a transmission signal is distorted by a channel to generate inter-symbol interference (ISI), so that a receiver uses an equalizer to determine a correct signal. It is necessary to compensate for the distorted signal.

This is because the time-varying multipath propagation generated by a plurality of fixed obstacles, etc., is transmitted through multiple paths having different propagation times, thereby reducing the performance degradation of a high speed digital communication system in digital transmission. Since the distortion of the signal generates a bit detection error at the receiving side, the transmitting side transmits a data string previously known at the receiver for a certain period from the received signal to the channel equalization, and at the receiving side, the waveform of the distorted data pulse Compare the original waveforms to estimate the distortion of the channel.

However, since the equalizer has a feedback component, the stability of the equalizer is reduced, and as the delay time of the signal by the channel increases, the number of taps of the feedback filter used in the equalizer increases, thereby increasing the amount of determinant calculation. The increase decreases the channel adaptation speed of the equalizer, and increases the number of symbols affected by the error propagation from the decision means to the feedback filter due to a false error estimation, thereby reducing the signal correction capability of the decision feedback equalizer. There was a problem.

The present invention has been made to solve the above problems, by using a hierarchical structure of the feedback equalization filter, and by measuring the error step by step to remove it, the error due to the wrong estimation symbol in the feedback equalization filter is transmitted. Can reduce the number of taps of the feedback equalization filter, thereby compensating for the distortion of the signal faster by using a larger size step, and using a soft decision device that is a linear one tap equalizer. It is an object of the present invention to provide a decision feedback equalizer using a hierarchical feedback filter and a soft decision device that prevents erroneous determination of an estimated symbol and improves the performance of a receiver by applying the signal determination means to be used.

In order to achieve the above object, the present invention provides an equalization filter for preventing error transmission by removing the Pre-ISI or Post-ISI of the input signal containing the actual information based on the training train input; Filter coefficient storage means for storing filter coefficients for removing pre-ISI or post-ISI measured by the equalization filter; Signal output means for compensating for distortion of the input signal by multiplying the input signal by a filter coefficient stored in the filter coefficient storage means; Signal determination means for estimating accurate information about the signal output to the soft decision device and the hard decision device; It includes.

The equalization filter may include: a feedforward equalization filter for removing pre-ISI by setting filter coefficients with an adaptive equalization algorithm based on an input training sequence based on an input training sequence; A feedback equalization filter of a hierarchical structure that sets filter coefficients to remove errors caused by erroneous symbol determination and post-ISI by the adaptive equalization algorithm; Characterized in that comprises a.

Here, the feedback equalization filter may include α sub-feedback equalization filters sequentially inputting outputs of the soft decision device and sequentially delaying them; Each sub-feedback equalization filter is characterized by updating the filter coefficients through an adaptive equalization algorithm.

In this case, an error is measured in each of the sub-feedback equalization filters, and the filter coefficients are individually updated.

In the feedback equalization filter, the output of the feedforward equalization filter is sequentially input to α sub-feedback equalization filters and delayed, and a value multiplied by a filter coefficient for removing Post-ISI by the detected symbol is added up. An output first hierarchical structure; A second hierarchical structure in which Post-ISI that is not removed through an adaptive equalization algorithm is removed as an input of the output of the first hierarchical structure, and a value multiplied by a filter coefficient is added and output; Characterized in that comprises a.

In addition, the signal determining means inputted by subtracting the output value of the feedback equalization filter from the output value of the feedforward equalization filter may include: a hard decision device for compensating and outputting a signal distorted at the input; And a soft decision device having the input and a one-tap equalizer for estimating an error with a training symbol and outputting the error to the sub-feedback equalization filter. It characterized in that for outputting the determined value of the estimated symbol including.

The filter coefficients of the one-tap equalizer are updated through an adaptive equalization algorithm to the output of the soft decision device.

The input value of the signal determining means is an adaptive equalization algorithm, which updates the multiplier filter coefficients of the feedforward equalization filter and the second hierarchical multiplier filter coefficients of the feedback equalization filter through comparison with a training sequence. do.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram schematically illustrating a decision feedback equalizer using a hierarchical feedback filter and a soft decision device according to the present invention, and FIG. 2 is a decision feedback equalizer using a hierarchical feedback filter and a soft decision device according to the present invention. Figure is a schematic diagram schematically.

As shown in the figure, the decision feedback equalizer 1 using the hierarchical feedback filter and the soft decision device according to the present invention includes a feedforward equalization filter 10 and a feedback equalization filter 30. And signal determination means 50.

Here, the decision feedback equalizer 1 using a hierarchical feedback filter and a soft decision device transmits a digital signal by transmitting the same signal through multiple paths having different propagation times. In order to eliminate inter-symbol interference (hereinafter, referred to as ISI), which is the biggest deterioration in the performance of high-speed digital communication systems, the various characteristics satisfy the ideal characteristics. In this case, the distortion generated in transmitting the signal causes a bit detection error at the receiving side, such that the entire signal cannot be restored or a completely different signal is received.

Thus, by compensating for the characteristics of the channel by processing the received signal passing through the distorted channel deviating from the ideal characteristic, it is provided to reduce the Bit Error Rate (BER) at the receiving side.

In addition, a decision feedback equalizer (DFE) 1 using the hierarchical feedback filter and the soft decision device is a data string previously known at the receiver for a predetermined period at the transmitter to perform channel equalization from the received signal. And the receiving side compares the distorted data pulse waveform through the channel and estimates the distortion level of the channel. The period of transmitting the determined data sequence is called a training mode, and the data transmitted in this period is transmitted. The row is called the training sequence.

In this case, the training sequence generally includes a pseudo random sequence, and after the training period, the determined data is used instead of the training sequence. This period is referred to as a decision decision mode (DD). do.

Therefore, in the decision feedback equalizer (DFE) 1 using the hierarchical feedback filter and the soft decision device, the training sequence or the determined data is defined as a reference signal, and the difference between the existing signal and the output is determined as an error. The signal is extracted, and the error signal updates the equalizer tap.

In addition, a decision feedback equalizer (DFE) 1 using the hierarchical feedback filter and the soft decision device 1 is a feedforward equalization filter that removes Precursor-ISI (hereinafter referred to as Pre-ISI). It has a symmetrical structure consisting of a feedback equalization filter 30 that removes 10 and Postcursor-ISI (hereinafter referred to as Post-ISI).

Here, the feedforward equalization filter 10 removes the pre-ISI through N (Z ^-1 ,... Z ^-1 ) delay units 11 which sequentially delay the input training sequence. In order to multiply the filter coefficients (C _o ,... C _{N -1} ) by the multiplication unit to the signals, and add the total N signals in the adder unit 15 to output, the interference by Pre-ISI can be eliminated. The filter coefficients C _o ,... C _{N -1} set by the feedforward equalization filter 10 during the training sequence are stored in the filter coefficient storage means (not shown).

The training sequence used to set the filter coefficient for removing the pre-ISI is preferably stored in a training sequence storage buffer (not shown) and recycled when the filter coefficient is set for removing the post-ISI.

In addition, when the data output from the feedforward equalization filter 10 equalizes the signal in the adaptive equalization algorithm 40, the data are equalized through the filtering and the filter is updated. The feedforward equalization filter 10 and the feedback equalization filter (10) are temporarily stored in the storage means (not shown) and then used for symbol equalization, and using the training sequence stored in the training sequence storage means (not shown) to measure errors. 30) the error in.

In this case, when the filter coefficients are updated by using each of the adaptive equalization algorithms, storage means or the like which are temporarily stored in order to use the updated filter coefficients in the next symbol are used.

On the other hand, the feedback equalization filter 30 is divided into two stages of a hierarchical structure in order to surely remove the post-ISI step by step. In the first hierarchical structure 30A, the soft decision apparatus 53 of the signal determination means 50 is provided. ) Outputs are input, and the input values are sequentially provided with a total of L delays, i.e., a subfeedback equalization filter 31A having an L number of taps.

Here, in the first hierarchical structure 30A, a filter coefficient for removing post-ISI due to a detected symbol is set. In the sub-feedback equalization filter 31A, a plurality of delays are sequentially delayed and α It is output to the multiplication section 33A, and each filter coefficient b ¹¹ ₁ ,... B ¹¹ _L is multiplied with the output from the sub-feedback equalization filter 31A and summed in the α adding units 35A.

The sub feedback equalization filter 31A is used to equalize the detected symbols, and the result values output from the adder 35A use α adaptive equalization algorithms 37A. The filter coefficients b ¹¹ ₁ ,... B ¹¹ _{L of} the multiplication unit 33A are updated in comparison with the training sequence.

In addition, an output value output from the first hierarchical structure 30A is input as an input value from the second hierarchical structure 30B. In the second hierarchical structure 30B, the multiplier 31B is used as the input value. Each multiplier is used to multiply by the filter coefficients b ²¹ ₁ ,... B ^{2 α} _α , add the sums through the adder 33B, and output the output value of the feedback equalization filter 30.

Here, a value obtained by subtracting the output of the feedback equalization filter 30 from the output value of the feedforward equalization filter 10 is input as an input value of the signal determination means 50, and each of the hard-decision device, 51) and a soft-decision device 53.

Then, the input value of the signal determination means 50 is compared with the training signal by the adaptive equalization algorithm 40 to filter coefficients C _o ,... C _{N-1 of} the feed forward equalization filter 10. And filter coefficients b ²¹ ₁ ,... B ^2α _α of the second hierarchical structure 30B of the feedback equalization filter 30, wherein the updated filter coefficients are stored in the filter coefficient storage means (not shown). Is preferably stored in.

In addition, an output value of the crystal feedback equalizer 1 using the hierarchical feedback filter and the soft decision device may be obtained through the hard decision device 51, and the feedback equalization filter 30 may be obtained through the soft decision device 53. The output value is exported to be input to the first hierarchical structure 30A.

At this time, the filter coefficient of the one-tap equalizer used as the soft decision apparatus 53 is updated by using the output value of the soft decision apparatus 53 and the adaptive equalization algorithm 60.

As described above, filter means such as the feedforward equalization filter 10 and the feedback equalization filter 30, which have the respective filter coefficient values updated during the training interval through the above process, and have completed the training sequence and are performed on the actual data. And the signal determination means 50 to compensate for the distorted signal.

That is, in the section where the training sequence section ends and the actual data symbol is received, each filter coefficient for removing pre-ISI and post-ISI stored in the received data symbol and the filter coefficient storage means (not shown) It is to compensate for the distorted signal using the value.

3 is a diagram illustrating a signal determination means and an adaptive equalization algorithm in a decision feedback equalizer using a hierarchical feedback filter and a soft decision device according to the present invention, and FIG. 4 is a hierarchical feedback filter and a soft decision device according to the present invention. The feedback equalization filter of the crystal feedback equalizer using the schematic diagram, which will be described with reference to FIG. As shown in the figure, the signal determination means 50 according to the present invention uses a soft decision apparatus 53 using a linear one tap equalizer in order to reduce the error of the determination.

)은 상기 연판정 장치(53)를 거쳐 훈련 심볼과의 오차 추정에 따른 추정값(

)이 발생되면, 이를 상기 궤환 등 화 필터(30)의 입력으로 보낸다.Here, the input value of the signal determining means 50, which is a value obtained by subtracting the output value of the feedback equalization filter 30 from the output value of the feedforward equalization filter 10 (

) Is an estimated value according to the error estimation with the training symbol via the soft decision apparatus 53 (

Is generated, it is sent to the input of the feedback equalization filter (30).

The symbol as described above passes through a feedback equalization filter 30 having a hierarchical structure divided into a first hierarchical structure 30A and a second hierarchical structure 30B, and sets filter coefficients in two steps.

In the first hierarchical structure 30A, a filter coefficient for removing post-ISI for a detected symbol is set to equalize the detected symbol input to each sub-feedback equalization filter 31A. This is output through multiplier 33A and adder 35A, which is the input value of second hierarchical structure 30B.

In addition, the second hierarchical structure 30B detects the Post-ISI not completely removed through the adaptive equalization algorithm 37A, and outputs the same.

Here, the output value of the feedforward equalization filter 10 and the output values of the first hierarchical structure 30A and the second hierarchical structure 30B of the feedback equalization filter 30 are represented by Equation 1 below.

Here, b ¹ _i (b ¹¹ ₁ ,... B ^{1 α} _α ) and b ² _i (b ²¹ ₁ ,... B ^{2 α} _α ) are the first hierarchical structure 30A and the second hierarchical structure, respectively. The filter coefficients at 30B are shown, which are the filter coefficients of the feedback equalization filter 30 in the first hierarchical structure 30A and the second hierarchical structure 30B.

B ¹ _out and B ² _out represent output values in the first hierarchical structure 30A and the second hierarchical structure 30B, and the i-th sub-feedback equalization filter 31A in the first hierarchical structure 20A. The output value becomes the input signal of the i-th multiplier 31B in the second hierarchical structure 30B via the multiplier 33A and the adder 35A.

이고, 상기 수학식 1로부터 나타난 제1 계층 구조(30A) 및 제2 계층 구조(30B)의 출력값의 식으로부터 본 발명에 따른 계층적 궤환 필터 및 연판정 장치를 이용한 결정 궤환 등화기(1)의 결과적인 출력값은 하기 수학식 2와 같이 나타낼 수 있고, 하기 p는 각각의 서브 궤환 등화 필터(31A)를 나타내기 위하여 이용한다.In addition, the output value detected from the soft decision apparatus 53 additionally provided in the signal determining means 50 is

From the equations of the output values of the first hierarchical structure 30A and the second hierarchical structure 30B shown in Equation 1, the decision feedback equalizer 1 using the hierarchical feedback filter and soft decision device according to the present invention The resulting output value can be expressed as Equation 2 below, and p is used to represent each sub-feedback equalization filter 31A.

Here, the tap coefficients of the feedforward equalization filter 10 and the feedback equalization filter 30 use an adaptive equalization algorithm (LMS), and in the case of the feedback equalization filter 30, the subfeedback equalization filter 31A is generally used. An error used when updating the tap coefficient is shown, and the updated filter tap coefficient may be expressed by Equation 3 below.

Here, e (n) represents an error between the output values of the decision feedback equalizer 1 using the hierarchical feedback filter and the soft decision device in the training symbol, respectively, and e ¹ _i (n) represents the training symbol and the first hierarchical structure ( An error in each sub-feedback equalization filter 31A of 30A) is shown, and is expressed as in Equation 4 below.

FIG. 5 is a flowchart schematically illustrating a method of driving a decision feedback equalizer using a hierarchical feedback filter and a soft decision device according to the present invention, and will be described with reference to FIG. 1. As shown in the figure, the decision feedback equalizer 1 using the hierarchical feedback filter and the soft decision device according to the present invention has the above configuration and is based on a training sequence input at the beginning of operation. According to the filter coefficients Co,..., C _{N −1} of the feedforward equalization filter 10, the interference by the pre-ISI is removed (S10), and based on this, the feedforward equalization filter during the training sequence ( The filter coefficients Co,..., C _{N -1} set by 10) are stored in a filter coefficient storage buffer (not shown).

The output value output through the soft decision device 53 which is the one-tap equalizer of the feed forward equalization filter 10 is input to the feedback equalization filter 30 (S20).

In addition, the output value input to the feedback equalization filter 30 is delayed in the sub feedback filter 31A of the first hierarchical structure 30A to compare the data and the training sequence through the multiplier 33A and the adder 35A. The filter coefficients b ¹¹ ₁ ,..., B ^{1 α} ₁ composed of L pieces are updated (S30).

Here, the output value of the first hierarchical structure 30A is input as an input value of the second hierarchical structure 30B, and the output values through the multiplier 31B and the adder 33B of the second hierarchical structure 30B are It is output as the total output value of the crystal feedback equalizer 1 using the hierarchical feedback filter and the soft decision device according to the present invention (S40).

Then, a value obtained by subtracting the output value of the feedback equalization filter 30 from the output value of the feedforward equalization filter 10 is input to the input value of the signal determination means 50 (S50), and the signal determination means 50 The input value is compared with the training signal by the adaptive equalization algorithm 40 to determine the feedforward equalization filter coefficients Co, ..., C _{N -1} and α of the second hierarchical structure 30B of the feedback equalization filter 30. The filter coefficients b ²¹ ₁ ,..., B ^{2 α} _α are updated (S60).

In addition, the output from the soft decision apparatus 53 enters the input of the feedback equalization filter 30 and updates the filter coefficients of the one-tap equalizer of the soft decision apparatus 53 using the adaptive equalization algorithm 60. .

Each of the updated filter coefficients is preferably stored in a filter coefficient storage means (not shown), and operates to compensate for the distorted signal with respect to the actual data using the filter coefficient storage means (S80).

FIG. 6 is a flowchart schematically illustrating a method of driving the soft decision apparatus of FIG. 5, which will be described with reference to FIG. 1. As shown in the figure, the crystal feedback equalizer 1 using the hierarchical feedback filter and the soft decision device according to the present invention has the configuration as described above, and the step S70 is a linear decision device 53. The estimated error between the output of the one-tap equalizer and the training symbol is input to the feedback equalization filter 30 (S70-1).

Then, in order to remove the Post-ISI by the detected symbol, the filter coefficients b ¹¹ ₁ ,..., B ^{1 α} ₁ composed of L number of taps are set (S70-2). Equalization is performed based on a symbol entering the sub-feedback equalization filter 31A of the first hierarchical structure 30A (S70-3).

Further, Post-ISI which is not removed by each adaptive equalization algorithm 37A is detected and output as the total output of the feedback equalization filter 30 (S70-4).

7 is a graph of BER performance in the S-V first channel of the decision feedback equalizer using the hierarchical feedback filter and the soft decision device according to the present invention and the decision feedback equalizer according to the prior art. As shown in the figure, the performance of the decision feedback equalizer using the hierarchical feedback filter and the soft decision device according to the present invention shows the time delay and the signal distortion according to the time delay. It is a graph compared with (BER: Bit Error Rate).

Here, when the delay time of the first channel is 21 symbols, the decision feedback equalizer using the decision feedback equalizer according to the prior art and the hierarchical feedback filter and the soft decision device according to the present invention to compensate for the distortion of the signal accordingly. It shows the performance of using 8 taps and 16 of the Feedforward Equalize Filter and Feedback Equalize Filter.

In this case, regardless of the number of taps of the filter used, it can be seen that the performance of the decision feedback equalizer using the hierarchical feedback filter and the soft decision device according to the present invention is excellent, and the number of taps used is 8 days smaller than the symbol delay time of the channel. In case it has better performance.

8 is a graph of BER performance in the S-V second channel of the decision feedback equalizer using the hierarchical feedback filter and the soft decision device according to the present invention and the decision feedback equalizer according to the prior art. As shown in the figure, the performance of the decision feedback equalizer using the hierarchical feedback filter and the soft decision device according to the present invention shows the time delay and the signal distortion according to the time delay. It is a graph compared with (BER: Bit Error Rate).

Here, when the delay time of the second channel is 34 symbols, the decision feedback equalizer using the decision feedback equalizer according to the prior art and the hierarchical feedback filter and the soft decision device according to the present invention to compensate for the distortion of the signal accordingly. This shows the performance when 16 feeds and 32 feedforward equalization filters and feedback equalization filters are used.

In this case, it can be seen that the performance of the crystal feedback equalizer using the hierarchical feedback filter and the soft decision device according to the present invention is excellent regardless of the number of taps of the used filter.

9 is a BER performance graph of the S-V third channel of the decision feedback equalizer using the hierarchical feedback filter and the soft decision device according to the present invention and the decision feedback equalizer according to the prior art. As shown in the figure, the performance of the decision feedback equalizer using the hierarchical feedback filter and the soft decision device according to the present invention shows the time delay and the signal distortion according to the time delay. It is a graph compared with (BER: Bit Error Rate).

Here, when the delay time of the third channel is 47 symbols, the decision feedback equalizer using the decision feedback equalizer according to the prior art and the hierarchical feedback filter and the soft decision device according to the present invention to compensate for the distortion of the signal according to the third channel. It shows the performance of using 32 taps and 48 equalizing feedforward equalization filter and feedback equalization filter.

In this case, regardless of the number of taps of the filter used, it can be seen that the performance of the decision feedback equalizer using the hierarchical feedback filter and the soft decision device according to the present invention is excellent, and the number of taps used is 32 days smaller than the symbol delay time of the channel. In case it has better performance.

10 is a BER performance graph of the S-V fourth channel of the decision feedback equalizer using the hierarchical feedback filter and the soft decision device according to the present invention and the decision feedback equalizer according to the prior art. As shown in the figure, the performance of the decision feedback equalizer using the hierarchical feedback filter and the soft decision device according to the present invention shows the time delay and the signal distortion according to the time delay. It is a graph compared with (BER: Bit Error Rate).

Here, when the delay time of the fourth channel is 76 symbols, the decision feedback equalizer using the decision feedback equalizer according to the prior art and the hierarchical feedback filter and the soft decision device according to the present invention to compensate for the distortion of the signal according to the fourth channel. It shows the performance when using 32 taps, 47 using, and 64 using the Feedforward Equalize Filter and Feedback Equalize Filter.

In this case, regardless of the number of taps of the filter used, it can be seen that the performance of the decision feedback equalizer using the hierarchical feedback filter and the soft decision device according to the present invention is excellent, which can increase the accuracy of the estimation even in the same equalization range. Means.

Lastly, in order to update the filter tap coefficients, the error between each filter output value and the training symbol is measured and used. The structure of the feedback equalization filter is hierarchically filtered to remove the error step by step. It is possible to attenuate the occurrence of error propagation in the feedback equalization filter, and by reducing the number of taps of the feedback equalization filter, a larger step size can be applied to obtain a faster MSE to reduce an error.

Although the preferred embodiments of the present invention have been described above by way of example, the scope of the present invention is not limited to such specific embodiments, and those skilled in the art are appropriate within the scope described in the claims of the present invention. It will be possible to change.

As described above, the present invention having the configuration as described above uses a hierarchical structure of a feedback equalization filter, measures errors in steps, and removes them, thereby transmitting an error due to an incorrect estimation symbol in the feedback equalization filter. Can reduce the number of taps of the feedback equalization filter, thereby compensating for the distortion of the signal faster by using a larger size step, and using a soft decision device that is a linear one tap equalizer. By applying to the signal determination means to be used, it is possible to prevent an erroneous determination of the estimated symbol and to improve the performance of the receiver.

Claims (8)

An equalization filter that prevents error transmission by removing Precursor-Inter Symbol Interference (Pre-ISI) or Postcursor-Inter Symbol Interference (Post-ISI) of an input signal containing actual information based on the training sequence input; Filter coefficient storage means for storing filter coefficients for removing pre-ISI or post-ISI measured by the equalization filter; Signal output means for compensating for distortion of the input signal by multiplying the input signal by a filter coefficient stored in the filter coefficient storage means; Signal determination means for estimating accurate information about the signal output to the soft decision device and the hard decision device; Decision feedback equalizer using a hierarchical feedback filter and soft decision device comprising a. The method of claim 1, The equalization filter A feedforward equalization filter for removing pre-ISI by setting filter coefficients with an adaptive equalization algorithm on an input signal containing actual information based on the input training sequence; A feedback equalization filter of a hierarchical structure that sets filter coefficients to remove errors caused by erroneous symbol determination and post-ISI by the adaptive equalization algorithm; Crystal feedback equalizer using a hierarchical feedback filter and soft decision apparatus comprising a. The method of claim 2, The feedback equalization filter may include one or more sub-feedback equalization filters sequentially receiving outputs of the soft decision device and sequentially delaying them; Each sub-feedback equalization filter is a decision feedback equalizer using a hierarchical feedback filter and a soft decision device, characterized in that for updating the filter coefficients through an adaptive equalization algorithm. The method of claim 3, The decision feedback equalizer using the hierarchical feedback filter and the soft decision device, characterized in that the error is measured in each of the sub-feedback equalization filters and the filter coefficients are individually updated. The method of claim 2, The feedback equalization filter A first hierarchical structure in which an output of the feedforward equalization filter is sequentially input to one or more sub-feedback equalization filters and delayed, and a value multiplied by a filter coefficient for removing Post-ISI by a detected symbol is added and output; A second hierarchical structure in which Post-ISI that is not removed through an adaptive equalization algorithm is removed as an input of the output of the first hierarchical structure, and a value multiplied by a filter coefficient is added and output; Crystal feedback equalizer using a hierarchical feedback filter and soft decision apparatus comprising a. The method of claim 1, The signal determining means input by subtracting the output value of the feedback equalization filter from the output value of the feedforward equalization filter A hard decision device for compensating and outputting the distorted signal at the input; And A soft decision device having the input and a one tap equalizer for estimating an error with a training symbol and outputting the error to the sub-feedback equalization filter; A decision feedback equalizer using a hierarchical feedback filter and a soft decision device, characterized in that for outputting the decision value of the estimated symbol including. The method of claim 6, And a hierarchical feedback filter and a decision feedback equalizer using the soft decision device, wherein the filter coefficients of the one-tap equalizer are updated through an adaptive equalization algorithm to the output of the soft decision device. The method of claim 1, The input value of the signal determining means is an adaptive equalization algorithm, which updates the multiplier filter coefficients of the feedforward equalization filter and the second hierarchical multiplier filter coefficients of the feedback equalization filter through comparison with a training sequence. Feedback equalizer using red feedback filter and soft decision device.

Images