KR101476049B1 - Method for restoring of puncturing date using LDPC Decoding system and apparatus thereof - Google Patents

Abstract

The present invention relates to a method for restoring punctured data by using an LDPC decoding system and an apparatus thereof. According to the present invention, provided is a method for restoring punctured data by using an LDPC decoding system, the method including: receiving a Low Density Parity Check (LDPC) decoding signal, which includes a plurality of data bits and parity bits and in which a part of the parity bits is punctured; matching an inspection node having a Log Likelihood Ratio (LLR) value among N number of detection nodes with respect to M number of bit nodes included in the LDPC decoding signal; selecting remainder inspection nodes excluding at least one first inspection node connected to the punctured parity bit among the N number of inspection nodes as effective inspection nodes and modifying the LLR value with respect to the bits connected to the effective inspection node among the each bit; and applying the modified LLR value with respect to the bits connected to the first inspection node among the each bit to restore the LLR value of the punctured parity bit. According to the method for restoring punctured data by using an LDPC decoding system and an apparatus thereof, by using a reliability of data which is not punctured in the LDPC decoding system, a decoding performance can be improved through depuncturing the punctured data.

Description

[0001] The present invention relates to a method and apparatus for restoring punctured data using an LDPC decoding system,

The present invention relates to a method and apparatus for recovering punctured data using an LDPC decoding system, and more particularly, to a method and apparatus for recovering punctured data using an LDPC decoding system. More particularly, the present invention relates to a method and apparatus for recovering punctured data using a reliability of data not punctured in an LDPC (Low Density Parity Check) To a method and apparatus for recovering punctured data using an LDPC decoding system for recovering chirped data.

Recently, low density parity check codes (LDPC codes), which improve decoding performance, are newly emerging. The LDPC is a code having a very small number of '1' in each row and column of a parity check matrix defining a code, and includes a check node, a variable node, The structure can be defined by a factor graph composed of edges.

In general, when using the IEEE 802.11n LDPC encoder / decoder, puncturing the encoded data is performed. This is a process for adjusting the LDPC-coded data to the required number of data bits constituting the OFDM symbol. However, in the normal case, the LDPC deconfiguration process of 802.11n is performed by filling the punctured data with 0 and then performing the LDPC decoding process, which causes the performance degradation of the decoding.

The technology that becomes the background of the present invention is disclosed in Korean Patent Publication No. 2007-0037249 (published Apr. 04, 2007).

An object of the present invention is to provide a method and apparatus for recovering punctured data using an LDPC decoding system capable of improving decoding performance.

The present invention relates to a method and a device for decoding data, comprising: receiving a Low Density Parity Check (LDPC) encoded signal including a plurality of data bits and a parity bit and having a parity bit punctured; Matching a check node having an LLR (Log Likelihood Ratio) value among N check nodes with respect to a check node, excluding at least one first check node to which the punctured parity bit is connected among the N check nodes, Selecting an effective check node as a valid check node, adjusting an LLR value for bits connected to the valid check node among the bits, and controlling the adjusted LLR value for the bits connected to the first check node among the bits And recovering the LLR value of the punctured parity bit by applying the punctured parity bit to the punctured parity bit. It provides a one-way.

Here, the method for recovering punctured data using the LDPC decoding system may include combining the LLR values obtained for the plurality of data bits and the parity bits and the restored LLR values for the punctured parity bits, And decoding the LDPC coded signal.

In addition, the step of adjusting the LLR value for the bits connected to the valid check node may predecode the LLR value for the bits connected to the valid check node to a value close to 1 or -1.

The step of adjusting the LLR value for the bits connected to the valid check node may include predecoding the valid check nodes in ascending order of the check node number, A first step of obtaining a sine value and a magnitude value of an initial LLR value of all bits connected to a first valid check node among the valid check nodes; A second step of obtaining a first value obtained by multiplying all the sine values of the remaining bits except for one bit and a second value which is the smallest among the magnitudes of remaining bits excluding the first bit, To the first LLR value of the first bit, and adjusting the LLR value by multiplying the first bit of the first bit by The first to third steps may be repeated to adjust the LLR values for the next bit.

The step of reconstructing the LLR value of the punctured parity bits may be performed in a descending order of the check node numbers among the at least one check nodes connected with the punctured parity bits, A fourth step of restoring the bit nodes near the end of the parity bits and replacing the LLR values of all the bits connected to the corresponding check nodes with the adjusted LLR values according to the pre-decoding, A third value obtained by multiplying all of the sine values of the remaining bits except for the last punctured parity bit among the bits and a fourth value which is the smallest among the magnitudes of the remaining bits, And a sixth step of recovering the value obtained by multiplying the fourth value by the LLR value of the last punctured parity bit.

Here, if the fourth value is smaller than the preset threshold value in the sixth step, the LLR value of the punctured parity bit may be restored to zero.

In the restoration of the LLR value at the next check node having the check node number smaller than that of the check node, at the fourth step, the LLR value is restored at the check node among the bits connected to the next check node Bits of the punctured parity bits are replaced with LLR values previously recovered at the corresponding check nodes to obtain the sine and magnitude values, and then the last punctured parity bits The LLR value for the corresponding puncturing parity bit may be recovered by performing the fourth through sixth steps for the next puncturing parity bit having a bit node number smaller than the bit.

According to another aspect of the present invention, there is provided an LDPC encoder comprising: a signal receiver for receiving an LDPC (Low Density Parity Check) coded signal including a plurality of data bits and a parity bit and having a parity bit punctured; An LLR matching unit for matching a check node having a log likelihood ratio (LLR) value among N check nodes with respect to a bit node; and an at least one first check node connected to the punctured parity bits among the N check nodes A predecoder for selecting the remaining check nodes as valid check nodes and adjusting an LLR value for the bits connected to the valid check nodes among the bits, And an LLR restoring unit for restoring the LLR value of the punctured parity bits by applying the adjusted LLR value to the LDPC decoding unit It provides a decompression device of the punctured data using the system.

Here, the apparatus for decompressing punctured data using the LDPC decoding system may further comprise: a decoding unit configured to combine the LLR values obtained for the plurality of data bits and the parity bits and the restored LLR values for the punctured parity bits, And a decoding unit decoding the LDPC coded signal.

Also, the pre-decoding unit may predecode the LLR value for the bits connected to the valid check node to a value close to 1 or -1.

According to the method and apparatus for recovering punctured data using the LDPC decoding system according to the present invention, decoding performance is enhanced by puncturing punctured data using reliability of non-punctured data in an LDPC decoding system There is an advantage.

1 is a block diagram of an apparatus for recovering punctured data using an LDPC decoding system according to an embodiment of the present invention.
FIG. 2 is a flowchart of a method for recovering punctured data using FIG.
FIG. 3 shows an example of a matrix for explaining step S220 of FIG.
4 is an example of a connection graph between a bit node and a check node according to an embodiment of the present invention.
5 is a conceptual diagram illustrating step S240 of FIG. 2 based on FIG.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention.

The present invention relates to a method and apparatus for recovering punctured data using an LDPC decoding system, and more particularly, to a method and apparatus for recovering punctured data by using reliability of data not punctured in an LDPC (Low Density Parity Check) Discloses a technique of de-puncturing.

Generally, in the LDPC encoding process, a process of puncturing the encoded data is involved. This is to match the number of bits constituting the encoded data with the number of bits of the data constituting the OFDM symbol. The LDPC coded signal is composed of a plurality of data bits and corresponding parity bits, and the puncturing process processes some parity bits among the respective bits.

The transmitting terminal encodes the LDPC signal to transmit the signal, and the receiving terminal receives the LDPC signal and decodes (decodes) the signal. An embodiment of the present invention provides a technique for decoding and decoding punctured data included in an LDPC coded signal before receiving and decoding the LDPC coded signal at a receiving end.

Hereinafter, embodiments of the present invention will be described in detail based on the above description.

FIG. 1 is a block diagram of an apparatus for recovering punctured data using an LDPC decoding system according to an embodiment of the present invention, and FIG. 2 is a flowchart of a method for recovering punctured data using FIG. 1 and 2, a restoration apparatus 100 according to an embodiment of the present invention includes a signal receiving unit 110, an LLR matching unit 120, a pre-decoding unit 130, an LLR restoring unit 140, (150).

First, the signal receiving unit 110 receives an LDPC coded signal from a transmitting end (not shown) (S210). Here, the LDPC encoded signal includes a plurality of data bits and parity bits as described above, and at least one parity bit among the parity bits exists in a punctured state.

Next, the LLR matching unit 120 matches the check nodes having the LLR (Log Likelihood Ratio) value among the N check nodes for the M bit nodes included in the LDPC coded signal (S220).

The step S220 is related to the fact that the structure of the LDPC is defined by a factor graph including a check node, a variable node, and an edge connecting the check node and the variable node.

FIG. 3 shows an example of a matrix for explaining step S220 of FIG. 3 is an example in which the LDPC-coded code is composed of 12 bits including 6 data bits and 6 parity bits corresponding thereto for convenience of explanation, and the present invention is not necessarily limited thereto.

FIG. 3 shows an example in which the fifth and sixth parity bits of the six parity bits are punctured, and the corresponding bit nodes P5 and P6 are indicated by shading.

In FIG. 3, since a total of 12 data are used to transmit six data, the data rate is 1/2 (= 6/12). The data rate determines the number of uses of the parity bits. In general, the lower the data rate, the greater the number of parity bits used and the better the reliability of the system.

FIG. 3 shows an N × M matrix table corresponding to M = 12 and N = 6, in which a total of 12 bit nodes B1 to B6 and P1 to P6 exist corresponding to the number of bits included in the LDPC code, A total of six check nodes CN1 to CN6 (N = 6) exist corresponding to the number of data bits.

The LLR values at check nodes corresponding to each bit node are listed in each matrix element cell in Fig. However, not all cells have an LLR value, and some have an LLR value in a null (empty) state. 3, the LLR values at the check nodes corresponding to the bit nodes P1 to P6 are omitted.

The step S220 is a process of matching check nodes having LLR values to M bit nodes. Through this, a graph connecting each check node and a bit node with an edge can be constructed. In the example of FIG. 3, the bit node B1 will be connected to the check nodes CN1, CN2 and CN5, respectively, and the bit node B2 will be connected to the check nodes CN1, CN3 and CN6, respectively.

4 is an example of a connection graph between a bit node and a check node according to an embodiment of the present invention. In the graph of FIG. 4, it is confirmed that the bit nodes P5 and P6 corresponding to the positions of the punctured parity bits among the 12 bit nodes are connected to at least one of CN4, CN5, and CN6 among the six check nodes have. In addition, bit node B1 is connected to all check nodes, and B2 is connected to CN2.

Here, it should be understood that the results of FIG. 4 are not derived from the matrix of FIG. 3, but FIGS. 3 and 4 are different examples. However, FIGS. 3 and 4 show that the bit configuration example of the LDPC is the same, and that the number of used bit nodes, the number of check nodes, and the positions of punctured bits are all the same. Hereinafter, the graph of FIG. 4 will be described as an example.

After the inter-node matching is performed using the LLR values as described above, the pre-decoding unit 130 performs a predecoding operation before the actual decoding (decoding) operation on the received LDPC encoded signal (S230).

More specifically, in step S230, the pre-decoding unit 130 extracts the first check nodes CN4, CN5, CN6 connected to the punctured parity bits among the six check nodes (N = 6) Select the remaining check nodes as valid check nodes (CN1, CN2, CN3). Thereafter, the LLR value for the bits connected to the valid check nodes (CN1, CN2, and CN3) among the 12 bits (M = 12) included in the LDPC coded signal is adjusted.

The check node connected to the punctured bit node performs the pre-decoding using the valid check node connected to the non-punctured bit node because the check node returning value is less in the value of the check node.

Then, the LLR restoring unit 140 corrects the adjusted LLRs (LLRs) for the bits connected to the first check nodes CN4, CN5, and CN6 among the 12 bits of the LDPC coded signal (De-puncturing) the LLR value of the punctured parity bit by applying a value to the punctured parity bit (S240).

Thereafter, the decoding unit 150 decodes the LDPC encoded signal by combining the LLR values obtained for the plurality of data bits and the parity bits, and the recovered LLR values for the punctured parity bits (S250).

Hereinafter, steps S230 through S250 will be described in more detail.

First, in step S230 of adjusting the LLR value for the bits connected to the valid check nodes CN1, CN2, and CN3, the pre-decoding unit 130 receives the bits connected to the valid check nodes CN1, CN2, Lt; RTI ID = 0.0 > 1 < / RTI > or a value close to -1. This is the process of adjusting the LLR value for the bits connected to the valid check node to be finally close to an ideal value of 1 or -1.

For example, when the initial LLR value of the bit corresponding to the bit node B1 at the beginning is 0.1 and the initial LLR value of the bit corresponding to B5 is -0.5 as shown in FIG. 3, a positive value of 0.1 in the pre-decoding process becomes 1 A negative value of -0.5 is tuned to a value closer to -1. That is, a smaller value of the LLR value is adjusted to a smaller value, and a larger value is adjusted to a larger value. This pre-decoding process enhances the reliability from the bit nodes B1 to P4.

A specific embodiment of the pre-decoding according to step S230 is as follows. For convenience of explanation, the connection graph of each node refers to the example of FIG.

The pre-decoding is performed in the order of the smallest check node number (CN1-> CN2-> CN3) among the valid check nodes (CN1, CN2 and CN3) -> B2 -> ... -> B6). Hereinafter, a predecoding process will be described with respect to the first CN1.

First, a sine value and a size value of the first LLR value of all bits connected to the first valid check node CN1 among the valid check nodes CN1, CN2, and CN3 are obtained (step 1).

In FIG. 4, it is assumed that the bit nodes connected to CN1 are B1, B5, P1 and P2, and the first LLR values of the bits corresponding to B1, B5, P1 and P2 are -0.5, 0.4, 0.2 and -0.3. In this case, the sine value and the magnitude value for the first LLR value of each bit are arranged as shown in Table 1.

B1 B5 P1 P2 LLR value -0.5 0.4 0.2 -0.3 Sine value -One One One -One Size value 0.5 0.4 0.2 0.3

From Table 1, it can be seen that the sine value for the LLR value is -1 if the LLR value is negative, and is 1 if the LLR value is positive. Also, it can be seen that the magnitude value for the LLR value corresponds to the absolute value thereof.

Thereafter, the first valid check node CN1 multiplies all the sine values of all the bits (related to B5, P1, and P2) except the first bit (B1 related) having the smallest node number among all the bits connected to the first valid check node CN1 (0.2) of the smallest value (-1 = 1 x 1 x (-1)) and the size value (0.4, 0.2, 0.3) of the remaining bits (B5, P1 and P2) (Step 2).

Then, the value (-0.2) obtained by multiplying the first value (-1) by the second value (0.2) is added to the first LLR value (-0.5) of the first bit (B1 related) The LLR value is adjusted to -0.7 (= (- 0.5) + (- 0.2)) (third step).

Next, the same process as the first step to the third step is repeated for the next bit (related to B5, P1, and P2) after the first bit (related to B1), and the LLR value In the same manner. That is, after B1, the LLR value of B5 is adjusted by performing the first to third steps for B5, and thereafter, the LLR value is adjusted by P1 and P2 in the same manner, respectively.

When the pre-decoding is performed with respect to CN1 as described above, the LLR values for each bit (related to B1, B5, P1, and P2) connected to CN1 are all adjusted. Next, pre-decoding is sequentially performed in the same manner also for CN2 and CN3 of the following next rank. However, if the same bit as the bit used in the previous step is used for the operation, the LLR value of the corresponding bit is replaced with the LLR value adjusted in the previous step. Thus, it can be pre-decoded to a more optimal LLR value for each bit.

That is, in the pre-decoding process for the next valid check node CN2, the same bit node numbers B1, B2, and B3 as the bits pre-decoded in the last valid check node CN1 among the bits of the bit nodes connected to the next valid check node CN2, P2), the LLR value of the corresponding bit is replaced with the previously adjusted LLR value at the time of predecoding the immediately preceding validated check node CN1 (in the case of B1, it is replaced with a value of -0.7 instead of -0.5) And obtains the sine value and the magnitude value, and performs the second and third steps based on the sine value and the magnitude value to adjust the LLR value again. This also applies to CN3.

When all the pre-decoding operations from CN1 to CN3 are completed, the LLR value finally adjusted by pre-decoding is obtained for each bit corresponding to each bit node. The final adjusted LLR value is then used to recover the LLR value for the punctured parity bit.

The following describes step S240 of restoring the LLR value of the punctured parity bit. In step S240, the adjusted LLR value is applied to the bits connected to the check nodes CN4, CN5, and CN6 to which the punctured parity bits are connected, from among 12 bits of the 12 (M = 12) Then, through this, the LLR value of the punctured parity bit is recovered.

In step S240, unlike the step S230, the punctured parity bit is executed in the order of the check nodes (CN6-> CN5-> CN4) among the check nodes (CN4, CN5, CN6) In the check node (ex, CN6), restoration is first performed on the bit nodes (P6 near the end of P5 and P6) closest to the end of the punctured parity bits.

Hereinafter, the first CN 6 to be performed will be described. 5 is a conceptual diagram illustrating step S240 of FIG. 2 based on FIG. 4 and 5, the LLR value of all the bits (B1 B6, P1, and P6 related to the check node CN6) is replaced with the LLR value adjusted according to the pre-decoding (S230) (Step 4).

For convenience of explanation, it is assumed that the adjusted LLR values of B1, B6, and P1 through the pre-decoding process are -0.9, 0.8, and 0.7, respectively. In this case, each sine value is -1, 1, 1, and the size values are 0.9, 0.8, and 0.7.

Subsequently, the sign of the remaining bits (B1, B6, P1 related) except for the last punctured parity bit (related to P6) among the bits (B1, B6, P1, P6 related to the check node CN6) (-1 = (-1) x 1 x 1) obtained by multiplying all the values of the remaining bits and the smallest fourth value (0.7) of the magnitudes of the remaining bits (step 5). The principle of the fifth step is the same as the principle of the second step used in the predecoding, and thus a detailed description thereof will be omitted.

Then, the value (-0.7) obtained by multiplying the third value (-1) by the fourth value (0.7) is restored to the LLR value of the last punctured parity bit (P6 related) (Step 6). That is, the LLR value of P6 is restored to a value of -0.7 multiplied by the third value and the fourth value.

If the magnitude of the fourth value is smaller than a predetermined threshold value (ex, 0.2) in the sixth step, the LLR value of the punctured parity bit is restored to zero. That is, in the previous example, if the fourth value is found to be less than the threshold value, the LLR value of the last punctured parity bit (related to P6) is restored to zero.

After obtaining the LLR value of the parity bit corresponding to the bit node P6 using the check node CN6, the check node CN5 having the check node number smaller than that of the check node CN6 determines the LLR value The restoration process is performed in the same manner as in the fourth to sixth steps.

At this time, if the punctured parity bit is the same as the recovered bit (related to P6) of the LLR value in the corresponding check node CN6 among the bits connected to the next check node CN5 (see FIG. 5 (P6 is connected to CN5 in addition to CN6), and the LLR value of the corresponding punctured parity bit (P6 related) is replaced with the LLR value previously restored at the check node CN6, The same procedure as in step 4 is used. Then, the fifth and sixth steps are performed for the next puncturing parity bit (P5 related) to recover the LLR value for the corresponding puncturing parity bit (P5 related). This process can also be performed for CN4. This is because P5 is connected to CN4, and the LLR value for P5 can be adjusted again by performing the above process for P5.

After the LLR value for the punctured parity bits is recovered, the LLR value adjusted according to the pre-decoding is not used at the decoding of the LDPC coded signal (S250), and all the bits excluding the punctured parity bits are used And decodes the LDPC coded signal using the recovered LLR value for the punctured parody bits.

According to the embodiment of the present invention as described above, the punctured data is filled with non-zero values using highly reliable data that is not punctured in the process of pre-LDPC decoding, and a random value (A value of 0) can be reduced. According to the present invention, the decoding performance can be improved by depuncturing the punctured data using the reliability of the non-punctured data in the LDPC decoding system.

While the present invention has been described with reference to exemplary embodiments, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims. Accordingly, the true scope of the present invention should be determined by the technical idea of the appended claims.

100: device for restoring punctured data
110: signal receiving unit 120: LLR matching unit
130: Pre-decoding unit 140: LLR restoring unit
150:

Claims (13)

Receiving a Low Density Parity Check (LDPC) encoded signal including a plurality of data bits and a parity bit and puncturing some parity bits;
Based on whether there is a log likelihood ratio (LLR) value at N check nodes corresponding to M bit nodes included in the LDPC coded signal, each of the bit nodes is divided into N check nodes Matching the selected check node with the LLR value;
Selecting a check node other than at least one first check node to which the punctured parity bit is connected among the N check nodes as an effective check node and selecting an LLR value for the bits connected to the valid check node among the bits ;
Applying the adjusted LLR value to bits connected to the first check node among the bits to recover an LLR value of the punctured parity bit; And
And decoding the LDPC coded signal by combining the plurality of data bits and an initially obtained LLR value for the parity bit with the recovered LLR value for the punctured parity bit,
Wherein adjusting the LLR value for the bits connected to the valid check node comprises:
Predicting an LLR value for bits connected to the valid check node to a value close to 1 or -1, and if the LLR value for the bits connected to the valid check node is a positive value, 1, and if the LLR value is a negative value, predecoding the LLR value to a value between -1 and -1. delete delete The method according to claim 1,
Wherein adjusting the LLR value for the bits connected to the valid check node comprises:
The pre-decoding is performed in descending order of the check node number among the valid check nodes, and in the same effective check node,
A first step of obtaining a sine value and a magnitude value of an initial LLR value of all bits connected to a first valid check node among the valid check nodes;
A second value obtained by multiplying all of the sine values of the remaining bits except for the first bit among all the bits connected to the first valid check node and a second smallest value among the size values of the remaining bits excluding the first bit, A second step; And
And a third step of adding the value obtained by multiplying the first value by the second value to the first LLR value of the first bit to adjust the LLR value,
And the LLR values for the next bits are adjusted by repeating the first to third steps for the next bit after the first bit, respectively. The method of claim 4,
Wherein restoring the LLR value of the punctured parity bit comprises:
Wherein a check node number of at least one check node to which the punctured parity bits are connected is performed in a descending order of a check node number and a bit node near the end of the punctured parity bit is first restored in the check node,
A fourth step of replacing the LLR value of all bits connected to the corresponding check node with the LLR value adjusted according to the pre-decoding;
A third value obtained by multiplying all the sine values of the LLR values of the remaining bits except the last punctured parity bit among the bits connected to the corresponding check node and a fourth value of the smallest size value of the remaining bits Step 5; And
And a sixth step of restoring the value obtained by multiplying the third value by the fourth value to the LLR value of the last punctured parity bit,
In the restoration of the LLR value at the next check node having a check node number smaller than the check node,
If there is a punctured parity bit equal to the recovered bit of the LLR value at the corresponding check node among the bits connected to the next check node in the fourth step, And the sine and magnitude values are obtained by replacing the LLR values with the restored LLR values,
And performing an LLR value for the puncturing parity bits by performing the fourth through sixth steps for the next puncturing parity bit having a bit node number smaller than the last punctured parity bit, A method for restoring punctured data. The method of claim 5,
And restoring the LLR value of the punctured parity bits to 0 when the fourth value is smaller than a preset threshold value in the sixth step. delete A signal receiving unit receiving a low density parity check (LDPC) encoded signal including a plurality of data bits and parity bits and puncturing some parity bits;
Based on whether there is a log likelihood ratio (LLR) value at N check nodes corresponding to M bit nodes included in the LDPC coded signal, each of the bit nodes is divided into N check nodes An LLR matching unit for matching with some check nodes in which the LLR value exists;
Selecting a check node other than at least one first check node to which the punctured parity bit is connected among the N check nodes as an effective check node and selecting an LLR value for the bits connected to the valid check node among the bits A pre-
An LLR restoring unit for restoring an LLR value of the punctured parity bit by applying the adjusted LLR value to bits connected to the first check node among the bits; And
And a decoding unit for decoding the LDPC coded signal by combining the plurality of data bits and an initially obtained LLR value for the parity bit and the recovered LLR value for the punctured parity bit,
The pre-
Predicting an LLR value for bits connected to the valid check node to a value close to 1 or -1, and if the LLR value for the bits connected to the valid check node is a positive value, 1, and if the LLR value is a negative value, predecoding the LLR value to a value between -1 and -1. delete delete The method of claim 8,
The pre-
The pre-decoding is performed in descending order of the check node number among the valid check nodes, and in the same effective check node,
A first step of obtaining a sine value and a size value of the first LLR value of all bits connected to the first valid check node among the valid check nodes,
A second value obtained by multiplying all of the sine values of the remaining bits except for the first bit among all the bits connected to the first valid check node and a second smallest value among the size values of the remaining bits excluding the first bit, The second step, and
Performing a third step of adjusting the LLR value by adding a value obtained by multiplying the first value by a second value to the first LLR value of the first bit,
And an LLR value for the next bit is adjusted by repeating the first to third steps for the next bit after the first bit, respectively. The method of claim 11,
Wherein the LLR restoring unit comprises:
Wherein a check node number of at least one check node to which the punctured parity bits are connected is performed in a descending order of a check node number and a bit node near the end of the punctured parity bit is first recovered in the check node,
A fourth step of replacing the LLR value of all bits connected to the corresponding check node with the LLR value adjusted according to the pre-decoding;
A third value obtained by multiplying all the sine values of the LLR values of the remaining bits except the last punctured parity bit among the bits connected to the corresponding check node and a fourth value of the smallest size value of the remaining bits Step 5; And
And a sixth step of restoring a value obtained by multiplying the third value by the fourth value to an LLR value of the last punctured parity bit
In the restoration of the LLR value at the next check node having a check node number smaller than the check node,
If there is a punctured parity bit equal to the recovered bit of the LLR value at the corresponding check node among the bits connected to the next check node in the fourth step, And the sine and magnitude values are obtained by replacing the LLR values with the restored LLR values,
An LDPC decoding system for performing the fourth through sixth steps on the next puncturing parity bits having a bit node number smaller than the last punctured parity bit to recover an LLR value for the corresponding puncturing parity bit, A device for restoring punctured data. The method of claim 12,
Wherein the LLR restoring unit comprises:
And restores the LLR value of the corresponding punctured parity bit to 0 when the fourth value is smaller than a preset threshold value in the sixth step.

Images