KR100408281B1 - Signal reproducing apparatus and method for compensating for channel distortion - Google Patents

Abstract

The present invention relates to a signal reproducing apparatus and a method for compensating for channel distortion, and a signal reproducing method for compensating for channel distortion for reproducing original information from a signal passing through a predetermined channel, detects a transition occurrence portion of the channel passing signal. Doing; Calculating channel distortion occurring in the transition generation portion; Obtaining and correcting an error of a signal passing through the transition portion by reflecting the amount of distortion; And reproducing the original information by comparing the entire channel pass signal including the signal of the error-corrected transition generating portion with a predetermined threshold value.

According to the present invention, the correction by reflecting the channel distortion only in the region where the maximum error can occur is reduced the complexity according to the algorithm implementation, the hardware is simplified accordingly, the hardware is changed even if the channel model and input signal conditions change Signal processing is possible without the need for more accurate error correction.

Description

Signal reproducing apparatus and method for compensating for channel distortion

The present invention relates to a signal reproducing apparatus and method, and more particularly, to a signal reproducing apparatus that compensates for channel distortion for reproducing after tangential teal and error by finding a portion containing the most error in a signal read from a predetermined channel. And to a method.

In general, a slicer method, a partial response maximum likelihood (PRML) method, or a decision feedback equalizer (DFE) method may be used to read and reproduce a signal that has passed through a communication channel or a channel on a recording medium. have.

The slicer consists of a threshold detector and uses the analog equalizer to shape the waveform through a signal from the channel. The slicer determines the signal above the reference level as 1 and the signal below the reference level as 0.

1 illustrates a general PRML structure, in which an equalizer composed of a feed forward filter and a Least Mean Square (LMS) algorithm block 100, an error detector 110, and a Viterbi decoder 120 for adjusting coefficients of an equalizer are shown. ). The signal from the channel (hereinafter channel signal) is passed through the equalizer and LMS algorithm block 100 to produce the desired channel response. The error detector 110 detects an error from this signal. The Viterbi decoder 120 estimates the original signal before passing through the channel by finding an optimal signal path in consideration of the path for all possible errors.

2 illustrates a general DFE structure, which includes first and second FIR filters 200 and 210 and a threshold determiner 220. The first FIR filter 200 is a feed forward filter and the second FIR filter 210 is a feedback filter. The signal passing through the channel is input to the threshold determiner 220 through the first FIR filter 200. After the provisional decision is made at the threshold determiner, the provisional decision value is input to the second FIR filter 210. As the output values of the first FIR filter 200 and the second FIR filter 210 are added together, the provisional decision is gradually changed to a reliable decision by feedback.

Among the above-described conventional signal detection methods, the slicer detection method has a disadvantage in that the structure is simple but the performance is inferior. The PRML or DFE method has a high performance but a high complexity in hardware implementation. For example, considering only the multipliers of PRML and DFE, which have a high linearity, a multiplier corresponding to the number of taps is required to implement a filter. It is necessary. Therefore, in the case of the PRML or DFE method, not only the complexity is increased but also the signal processing speed is lowered, thereby making it difficult to process a fast signal.

Although there may be a method of improving these problems and detecting and correcting only the signal of the most error-producing part of the channel signal (Domestic Application No .;), this method reduces the complexity of software and hardware, but accurately corrects the error of channel distortion. There is a problem that no correction is made.

In order to reduce the complexity of hardware and software and to increase the accuracy of error correction when channel distortion occurs, a technical problem of the present invention is to detect a section having the most error among the signals to be reproduced, calculate a distortion amount of the section, The present invention provides a signal reproducing apparatus and method for compensating for channel distortion for performing error correction reflecting a distortion amount.

1 shows a general PRML structure.

2 shows a general DFE structure.

3 is a block diagram of a signal reproducing apparatus for compensating for channel distortion of the present invention.

Fig. 4 shows (a) the original input signal sequence before passing through a certain channel, (b) the form of a signal y (t) passing through a channel of a low pass filter type such as a recording medium such as an optical disk, and ( c) shows the form of a signal y (t) that has passed through a differential channel such as a recording medium such as a hard disk.

5 shows an example of (a) possible signal paths of (a) the initial signal.

6 is a flowchart of a signal reproducing method for compensating for channel distortion for explaining an operation of the present invention.

In order to solve the above problems, a signal reproducing method for compensating for channel distortion for reproducing original information from a signal passing through a predetermined channel comprises: detecting a transition occurrence portion of the channel passing signal; Calculating an amount of channel distortion occurring in the transition generating portion; Obtaining and correcting an error of a signal passing through the transition portion by reflecting the channel distortion amount; And reproducing the original information by comparing the entire channel pass signal including the signal of the error-corrected transition generation part with a predetermined threshold value.

The detecting of the transition generation part may include modeling a reference signal having the same level component and a transition direction as the level component of the transition generation part.

Preferably, the channel distortion amount is calculated by adding the respective distortion amount values calculated in the j transition portions previously generated.

Preferably, the respective distortion amounts are calculated by subtracting the signal on the transition portion of the channel signal from the reference signal.

The step of obtaining and correcting an error of the transition generation part may be performed by finding a signal on a path that minimizes a difference between the reference signal and the distortion amount from among a plurality of possible path signals passing through the transition generation part. It is preferable to calculate and correct as a path signal.

The optimum error path signal is preferably corrected as being replaced by the reference signal plus the distortion amount.

In order to solve the above problems, a signal reproducing method of reproducing original information from a signal passing through a predetermined channel includes: determining a portion having a high probability of error occurrence from the signal; Setting a reference signal for the portion having a high probability of error and determining a new reference signal by adding a distortion amount to the reference signal when the distortion amount obtained in the previous step is present; Calculating a plurality of error paths from the signal of the portion having the high probability of error occurrence; Finding an error path having the smallest difference between each signal on the plurality of error paths and the reference signal, and determining the optimal error path; Correcting the determined optimal error path signal; And comparing the channel pass signal including the signal portion with the error corrected with a predetermined threshold and restoring original information from the result.

The portion where the error occurrence probability is large is preferably a portion where a transition occurs in the channel signal.

The distortion amount is preferably a value obtained by averaging each of the distortion amounts calculated in the j transition generation portions previously generated.

The reference signal is preferably a signal modeled when the original information passes through the channel without error.

In order to solve the above problems, a signal reproducing apparatus for reproducing original information from a signal passing through a predetermined channel includes: a maximum error region determination unit for determining a maximum error occurrence region in which the maximum error occurs among the signals; The reference signal modeled on the assumption that the distortion amount generated in the channel signal is calculated, a plurality of error paths are obtained on the error area determined by the maximum error area determiner, and the channel passes without distortion from among the error paths. An optimum path searcher for finding and correcting a path signal having a minimum difference from the added value by adding the distortion amount; And a signal recovery unit for comparing the channel signal reflecting the portion of the signal corrected by the optimal path search unit with a predetermined threshold value and recovering original information according to the result.

The maximum error region determiner may determine, as the maximum error region, a section in which a probability that channel input and output signals do not coincide in the predetermined channel is greater than a reference error rate determined by channel characteristics.

Preferably, the maximum error region determiner determines a section in which a level shift occurs in the channel signal as the maximum error region.

In the case of an optical disc, the maximum error area determination unit preferably sets a predetermined level to determine a predetermined section around the channel signal as a maximum error possible area from a crossing point of the channel signal for this level.

In the case of a hard disk, the maximum error area determination unit preferably sets two upper and lower predetermined levels to determine a predetermined section around the channel signal as the maximum error possible area from the intersection of the channel signal with respect to this level.

The optimal path search unit preferably finds the minimum error path and performs correction to replace the signal of the error path with a signal when the channel passes through the channel without distortion.

The distortion amount is preferably a value obtained by averaging the distortion amounts calculated from the j error transition generation portions of the channel signal.

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

3 is a block diagram of a signal reproducing apparatus for compensating for channel distortion of the present invention, and includes a maximum error region determiner 300, an optimal path search unit 310, and a signal recovery unit 320. The maximum error area determiner 300 obtains the maximum error area possible from the input signal y (t) which is the result of the first signal a _k being read through the corresponding recording medium having the channel characteristic of h (t). Decide y (t) is expressed as Equation 1 below.

Where a _k is a digital signal having components of {1, -1} or represented by L levels, k is the number of sampling times, t is the time, T is the sampling interval, and n (t) is the white Gaussian noise. Indicates. h (t) is a channel transfer function that represents the characteristics of the medium in the transmission and storage of digital signals.

Figure 4 (a) is a sequence of the original input signal (a _k ) before passing through the channel (b) is a channel of the low pass filter type, such as a recording medium, such as an optical disk, the original input signal (a _k ) Signal (y (t)) when passed, and (c) shows the signal (y (t)) when the original input signal a _k has passed through a differential channel such as a recording medium such as a hard disk. will be.

The maximum error region determiner 300 detects a signal (y ₁ (t)) of the section by finding the maximum error occurrence section (region) where the error occurs most frequently from y (t), as shown in Equation (2). .

Where σ is the probability that an error will occur at the saturation level of the detected signal, and p _th (y _k | a _k ) is the probability that the channel output signal y is equal to the stream a _k in which the channel output signal y was transmitted (or recorded). In other words, y ₁ (t) means a signal in which the probability that the input signal and the output signal do not coincide on the channel is greater than σ determined by the channel characteristic. This signal can be detected by a threshold method. That is, in the channel form as shown in FIG. 4 (b), an intermediate value is set to measure a crossing point for this level, and a predetermined section around the signal is detected as the maximum error possible region, and the signal spans the region. Is detected as y ₁ (k). As it can be estimated from the figure, the point where y (k) intersects a predetermined level will be a section in which a signal transition occurs, so a predetermined section around the intersection will be the maximum error possible region. In the channel form as shown in (c) of FIG. 4, the two upper and lower levels are set to measure intersection points of the levels, and a predetermined section of the surrounding area is detected as the maximum error possible region, and the signal over the region is y _It detects as ₁ (k). Here again, the point where y (k) intersects for the two levels set will be a section in which the signal transition occurs, so that a predetermined section around the intersection will be the maximum error probable area. According to channel characteristics other than those of Figs. 4B and 4C, for two or more possible channel pass signal levels, the same number of levels as the corresponding signal levels can be set to detect an error region.

The optimal path search unit 310 of FIG. 3 first calculates a reference value constituting the transition portion when the channel passes through the channel without error with respect to the region determined by the maximum error occurrence region determiner 300, that is, the transition portion. Model as shown in 3.

b _n are possible signal streams consisting of a sequence in which a level transition occurs with the original recording signal a _k as an element, an example of which is shown in Table 1.

n ... b _{n, -k} ... b _{n, -1} b _{n, 0} b _{n, 1} b _{n, 2} ... b _{n, k} ... One ...One One One One One -One -One -One One One... 2 ...One One One One One -One -One -One -One One... : : n ...One One One One One -One -One -One -One -One... n + 1 ...One One -One -One -One One One One One One... : : N ...-One -One -One -One -One One One One One One...

b _n has a sequence reflecting the minimum run length limitation d applied to the recording signal a _k . Table 1 is for the case where d = 2. The reference value represented by Equation 3 is expressed as a value convolution of b _n and -b _n with the channel transfer function h (k), respectively, according to the transition direction. Refers to a model of an ideal channel pass signal in which any possible signal with a containing sequence has passed through the channel without noise. The optimal path search unit 310 finds a plurality of possible error paths with respect to the signal y ₁ (k) of the section detected by the maximum error occurrence area determiner 300. For example, if the reference signal is {x ₁ , x ₂ , x ₃ , x ₄ , x ₅ , x ₆ } in sequence, then the corresponding y ₁ (k) is in sequence {c ₁ , c ₂ , c ₃ , c ₄ , c ₅ , c ₆ }, and other possible error paths that take into account the occurrence of an error are {c ₂ , c ₃ , c ₄ , c ₅ , c ₆ , c for the same sampling time point. ₇ }, and may also be assumed to have a sequence of {c ₀ , c ₁ , c ₂ , c ₃ , c ₄ , c ₅ } for the same sampling time point. 5 shows an example of (a) (b) error probable paths of the original signal. The optimum path search unit 310 calculates the distance (error) between the signal on each error path and the reference signal from y ₁ (k) for each of the plurality of error paths as shown in Equation 4 below.

Where m is the index of the possible error paths obtained above, 1, ..., M, and _n of x _n is the index corresponding to one of the reference signals. However, it is not necessary to calculate the above distances for all reference signals and for all error paths, and it is preferable to apply only the above-described equations by selecting some of their respective cases. According to the calculation as in Equation 4, the signal path in the case where D is the smallest is selected as the optimum error path. Once the optimal error path is selected, the absolute value of the difference between the signal on the optimal error path and the reference signal is obtained. For each successive transition region of a predetermined number of times, the value obtained by accumulating and averaging the absolute values of the difference between the above-described reference signal and the signal on the optimum error path is determined as the amount of distortion. After the amount of distortion is determined, a value obtained by adding the amount of distortion to the above-described reference signal is defined as a new reference signal, and the new reference signal is applied when detecting an optimal error path in a subsequent error occurrence area, that is, a transition section. The signal on the detected optimal error path is subjected to error correction according to a predetermined algorithm. As an example of error correction, a method of replacing the signal on the optimal error path with a reference signal may be used.

The signal recovery unit 320 compares the y (k) signal with a predetermined threshold by a simple threshold determination method on the channel signal y (k) including the signal portion corrected by the optimal path search unit 310. As a result, the original information is recovered.

6 is a flowchart of a signal reproducing method for explaining the operation of the present invention. First, the signal reproducing method of the present invention first determines the maximum error occurrence region from the signal y (k) that has passed through the channel (step 600). . In general, the most common error in the channel pass signal is the level transition. In order to detect a level transition, it is checked whether a product of two consecutive sampling time points y (n) and y (n-1) is less than or equal to zero (step 602). If less than or equal to 0, it is determined that a level transition has occurred between the two signals, and the number of transitions generated is checked (step 604). This is to collect distortion amounts included in each transition signal at a certain number of times of transition times. If the number of times is less than or equal to j, a predetermined reference signal for the level shift generation part is modeled (step 606). The reference signal is modeled as a result of the possible signal streams b _n passing through the channel without error, consisting of a sequence in which the level transition occurs with the original recording signal a _k as an element before passing through the channel. b _k is a signal stream having the same signal component of the original signal a _k , including the level transition portion, and reflecting the run length constraint of a _k . The calculation of the reference signal x (k) modeling the ideal result when b _k passes through the channel having the channel transfer function h (t) without noise is performed as in Equation 3. If the reference signal is determined in one direction of the level transition, for example, in the level transition direction when y (n) is greater than 0 and y (n-1) is less than 0, the reference signal is determined when the level transition occurs in the opposite direction. Must be reversed. Once the reference signal is modeled, the next transition takes a reference signal in which only the sign of b _k is changed to + or-in consideration of the transition direction. If the number of transitions is greater than j, a value obtained by adding the final distortion amount obtained in the previous step to the reference signal obtained in step 606 is set as the reference signal (step 608). If the reference signal is set, a plurality of possible error paths are found for the signal y ₁ (k) of the maximum error occurrence area (step 601). This is the same as the error path extraction method described in FIG. The difference between the signal on each error path and the reference signal corresponding to the level transition direction of the error path is obtained, and an error path having the smallest absolute value is detected as the optimum error path (step 612). If the number of transitions is less than j, the absolute value of the difference between the detected optimal error path signal and the reference signal is stored as a distortion amount (step 614). If the number of transition occurrences is the j-th, the final distortion amount obtained by adding the distortion amounts stored in step 614 and the current distortion amounts and obtaining the average value is calculated (step 616). When the final distortion amount is obtained, the reference signal for the next transition generation portion uses a value obtained by adding the final distortion amount to the reference signal as a new reference signal as in step 608. The signal on the optimal error path is subjected to an error correction step by a predetermined error correction algorithm (step 618). An example of error correction is a correction in which the reference signal is replaced with a signal on an optimal error path. The corrected optimal error path signal is restored to the original signal by a simple comparison with a threshold value (step 620).

FIG. 7 shows channel characteristics of one channel in which the bit error rate (BER) for the sampling point along the horizontal axis shows an impulse response characteristic.

FIG. 8 illustrates the method according to the present invention to which the values of m = 3, n = 1, x = {0.6, 0.2, -0.2, -0.6} for the channel shown in FIG. The graph of BER is shown. It can be seen that the higher the signal-to-noise ratio (SNR), the lower the bit error rate when the method of the present invention is applied.

9 is a graph showing BER according to the amount of distortion when channel distortion correction is performed and when channel distortion correction is not performed. Here, it can be seen that the bit error rate is reduced when the channel distortion correction is performed than when the channel distortion correction is not performed.

10 is a graph showing the BER according to the change of the distortion amount in each case using the method of the present invention and the adaptive PRML method, and it can be seen that the bit error rate is generally low when the method of the present invention is used.

As such, the error correction algorithm is applied only to the maximum error occurrence portion of the channel signal, and at the same time, the error correction for the channel distortion can be performed at the same time, so that more accurate error correction can be achieved while reducing the complexity and time according to the algorithm implementation. . In the apparatus and method of the present invention, even when the channel model and the minimum run length condition of the input signal and the like change, only by resetting the value of the reference signal x (k), it is possible to process the signal in accordance with the change.

According to the present invention, the correction by reflecting the channel distortion only in the region where the maximum error can occur is reduced the complexity according to the algorithm implementation, the hardware is simplified accordingly, the hardware is changed even if the channel model and input signal conditions change Signal processing is possible without the need for more accurate error correction.

Claims (17)

A signal reproduction method for compensating for channel distortion for reproducing original information from a signal passing through a predetermined channel, Detecting a transition occurrence portion of the channel pass signal; Calculating an amount of distortion occurring in the transition generating portion; Obtaining and correcting an error of a signal passing through the transition portion by reflecting the amount of distortion; And And reproducing original information by comparing an entire channel pass signal including a signal of the error-corrected transition generating portion with a predetermined threshold value. The method of claim 1, wherein the detecting of the transition part comprises: And modeling a reference signal having the same level component and a transition direction as the level component of the transition generating portion. The method of claim 1, wherein the calculation of the distortion amount, A method of compensating for channel distortion, characterized in that the average value is calculated by adding up the respective distortion amounts calculated from the j transition portions previously generated. The method of claim 2 or 3, wherein each of the distortion amount, And subtracting the signal on the transition portion of the channel signal from the reference signal. The method of claim 4, wherein obtaining and correcting an error of the transition occurrence portion comprises: Channel distortion, characterized in that to find the signal on the path of the plurality of possible paths passing through the transition portion is the minimum difference between the reference signal and the amount of distortion is calculated as the optimal path signal and corrected Signal reproduction method to compensate. 6. The method of claim 5, wherein the optimal error path signal is corrected as being replaced by the reference signal plus the distortion amount. A signal reproduction method for compensating for channel distortion for reproducing original information from a signal passing through a predetermined channel, Determining a portion having a high probability of error occurrence from the signal; Setting a reference signal for the portion having a high probability of error and determining a new reference signal by adding a distortion amount to the reference signal when the distortion amount obtained in the previous step is present; Calculating a plurality of error paths from the signal of the portion having the high probability of error occurrence; Finding an error path having the smallest difference between each signal on the plurality of error paths and the reference signal, and determining the optimal error path; Correcting the determined optimal error path signal; And And comparing the channel pass signal including the error-corrected signal portion with a predetermined threshold and restoring original information from the result. The method of claim 7, wherein And a portion having a high probability of error occurrence is a portion where a transition occurs in the channel signal. The method of claim 8, wherein the distortion amount, A method for compensating for channel distortion, characterized in that the sum of the distortion amounts calculated in the previously generated j transition generation parts is averaged. The method of claim 7, wherein the reference signal, The signal reproduction method for compensating for channel distortion, characterized in that the signal is modeled when the original information passes through the channel without error. A signal reproducing apparatus for compensating for channel distortion for reproducing original information from a signal passing through a predetermined channel, A maximum error area determiner configured to determine a maximum error occurrence area in which the maximum error occurs among the signals; The reference signal modeled on the assumption that the distortion amount generated in the channel signal is calculated, a plurality of error paths are obtained on the error area determined by the maximum error area determiner, and the channel passes without distortion from among the error paths. An optimum path searcher for finding and correcting a path signal having a minimum difference from the added value by adding the distortion amount; And And a signal recovery unit for comparing the channel signal reflecting the portion of the signal corrected by the optimum path searcher with a predetermined threshold and recovering original information according to the result. . The method of claim 11, wherein the maximum error area determiner, And determining a section in which the probability that channel input and output signals do not coincide in the predetermined channel is greater than a reference error rate determined by channel characteristics as a maximum error region. The method of claim 11, wherein the maximum error area determiner, And a section in which a level shift occurs in the channel signal as a maximum error region. The method of claim 11, wherein the maximum error area determiner, In the case of an optical disc, a signal reproduction device for compensating for channel distortion, characterized in that a predetermined level is set to determine a predetermined section around the channel as a maximum error possible area from a crossing point of the channel signal for this level. The method of claim 11, wherein the maximum error area determiner, In the case of a hard disk, a signal reproduction apparatus for compensating for channel distortion, characterized in that a predetermined level is set at two upper and lower levels to determine a predetermined section around the channel signal as a maximum error possible area from the intersection of the channel signal for this level. The method of claim 11, wherein the optimum path search unit, And correcting for finding the minimum error path and replacing the signal in the error path with a signal when the channel passes through the channel without distortion. The method of claim 13, The distortion amount is a signal reproduction method for compensating for channel distortion, characterized in that the average of the sum of the respective distortion amounts calculated in the j error transition generation portion of the channel signal.