KR20130044255A - Apparatus and method for channel encoding and decoding in communication system using low-density parity-check codes - Google Patents

Abstract

PURPOSE: A channel encoding/decoding method and a device thereof in a communication system using a LDPC code are provided to generate a LDPC code having different code word length using information of a given parity check matrix. CONSTITUTION: A demodulator(1230) demodulates a transmitted signal from a transmitter. A punch pattern determiner or estimator(1220) determines location information of a punched parity bit. An LDPC decoder(1240) decodes data using the location information of the punched parity bit. The location information of the punched parity bit determines the number of the punched parity bits and determines the number of punched parity bit groups based on the number of the punched parity bits. The location information of the punched parity bit is determined by an order of a predetermined punched parity bit groups. [Reference numerals] (1210) Control unit; (1220) Reduced/punched pattern determiner or estimator; (1230) Demodulator; (1240) LDPC decoder

Description

{APPARATUS AND METHOD FOR CHANNEL ENCODING AND DECODING IN COMMUNICATION SYSTEM USING LOW-DENSITY PARITY-CHECK CODES}

The present invention relates to a communication system using a low-density parity-check (LDPC) code, and more particularly, to a channel encoding / decoding method and apparatus for generating a specific type of LDPC code. It is about.

In a wireless communication system, the performance of a link is significantly degraded due to various noises, fading and inter-symbol interference (ISI) of a channel. Therefore, it is essential to develop overcoming techniques for noise, fading and ISI in order to realize high-speed digital communication systems requiring high data throughput and reliability, such as next generation mobile communication, digital broadcasting, and portable Internet. In recent years, error-correcting codes have been actively studied as methods for efficiently restoring information distortion and improving communication reliability.

The LDPC codes are typically represented using graphical representations, and many features can be analyzed through graph theory, algebra, and probability-based methods. Generally, a graph model of a channel code is useful not only for describing codes but also to correspond information of encoded bits to a vertex in a graph and to correspond each bit relation to edges in a graph , It is possible to derive a natural decryption algorithm because each vertex can be regarded as a communication network for exchanging messages determined by each line. For example, the decoding algorithm derived from trellis, which is a kind of graph, includes the well-known Viterbi algorithm and BCJR (Bahl, Cocke, Jelinek and Raviv) algorithm.

The LDPC code is generally defined as a parity-check matrix and can be expressed using a bipartite graph collectively referred to as a Tanner graph. The bipartite graph indicates that the vertices constituting the graph are divided into two different types. In the case of the LDPC code, a binary graph composed of a variable node and a vertex called a check node Is expressed. The variable node corresponds one-to-one to the encoded bit.

Figure 1 shows an example of the parity check matrix H ₁ of the LDPC code consisting of four rows and eight columns. Referring to FIG. 1, since there are 8 columns, it means an LDPC code that generates a codeword having a length of 8, and each column corresponds to an encoded 8 bit.

FIG. 2 is a diagram illustrating a Tanner graph corresponding to H ₁ of FIG. 1.

Referring to FIG. 2, the Tanner graph of the LDPC code includes eight variable nodes x ₁ (202), x ₂ (204), x ₃ (206), x ₄ (208), x ₅ (210), x ₆ (212), x ₇ (214), x ₈ (216) and four check nodes (218, 220, 222, 224). Here, the i th column and the j th row of the parity check matrix H ₁ of the LDPC code correspond to the variable nodes x _i and j th check nodes, respectively. In addition, a value of 1, i.e., a non-zero value of a point where the i-th column and the j-th row of the parity check matrix H ₁ of the LDPC code intersect, refers to the variable node x _i on the Tanner graph as shown in FIG. An edge exists between and the j th test node.

In the Tanner graph of the LDPC code, the degree of the variable node and the check node means the number of line segments connected to each node, which is determined in the column or row corresponding to the node in the parity check matrix of the LDPC code. It is equal to the number of nonzero entries. For example, in FIG. 2, the variable nodes x ₁ 202, x ₂ 204, x ₃ 206, x ₄ 208, x ₅ 210, x ₆ 212, and x ₇ ( 214, and the order of x ₈ (216) is 4, 3, 3, 3, 2, 2, 2, 2, respectively, in order, and the order of test nodes 218, 220, 222, and 224, respectively, in order. 6, 5, 5, 5. In addition, the number of non-zero elements in each column of the parity check matrix H ₁ of FIG. 1 corresponding to the variable nodes of FIG. 2 is equal to the above-described orders 4, 3, 3, 3, 2, 2, and 2. , The same as 2, and the number of nonzero elements in each row of the parity check matrix H ₁ of FIG. 1 corresponding to the check nodes of FIG. 2 is the above-described orders 6, 5, 5, 5 In order.

In order to express the degree distribution of the nodes of the LDPC code, the ratio of the number of variable nodes with order _i to the total number of variable nodes is f _i , and the number of check nodes with order j and the total number of check nodes is f _i . Let g _{j be} the ratio with the number. For example, for the LDPC code to the Figure correspond to the 1 and 2 is _{f 2 = 4/8, f} 3 = 3/8, f 4 = 1/8, i ≠ 2, 3, about 4 f _i = 0, g ₅ = 3/4, g ₆ = 1/4, and g _j = 0 for j ≠ 5,6. When the length of the LDPC code is N, that is, the number of columns is N, and the number of rows is N / 2, the density of nonzero elements in the parity check matrix having the above degree distribution is expressed by Equation 1 below. Is calculated as

When N increases in Equation 1, the density of 1 in the parity check matrix continues to decrease. In general, since the density of non-zero elements is inversely proportional to the code length N, the non-zero elements have a very low density. The term low-density in the name of the LDPC code is derived for this reason.

FIG. 3 schematically illustrates an LDPC code adopted as a standard technology in Digital Video Broadcasting-Satelliete transmission 2nd generation (DVB-S2), which is one of the European digital broadcasting standards.

은 LDPC 부호어 길이고,

은 정보어의 길이이고,

은 패리티 길이를 의미한다. 그리고,

이 성립하도록 정수

과

를 결정한다. 이때,

도 정수가 되도록 한다. 편의상 도 3의 패리티 검사 행렬을 제 1 패리티 검사 행렬 H₁이라 한다. Referring to Figure 3,

Is the LDPC codeword length,

Is the length of the information word,

Means parity length. And,

To establish this constant

and

. At this time,

Is also an integer. For convenience, the parity check matrix of FIG. 3 is referred to as a first parity check matrix H ₁ .

번째 열(column)부터

번째 열까지의 구조는 이중 대각(dual diagonal) 형태이다. 따라서, 패리티 부분에 해당하는 열의 차수(degree) 분포는 그 값이 '1'인 마지막 열을 제외하고 모두 '2'를 가진다.Referring to FIG. 3, a part corresponding to a parity part in the parity check matrix, that is,

From the first column

The structure up to the second column is a dual diagonal form. Therefore, the degree distribution of the column corresponding to the parity part has '2' except for the last column in which the value is '1'.

번째 열까지의 구조를 이루는 규칙은 다음과 같다. In the parity check matrix, the portion corresponding to the information word portion, i.e.,

The rules for the structure up to the first column are as follows.

개의 열을

개의 열들로 구성된 복수 개의 그룹으로 그룹화(grouping)하여, 총

개의 열 그룹(column group)을 생성한다. 각 열 그룹에 속해있는 각각의 열을 구성하는 방법은 하기 규칙 2에 따른다. Rule 1: Corresponding Information Words in the Parity Check Matrix

Columns

Grouping into multiple groups of columns

≪ / RTI > column groups. How to configure each column belonging to each column group follows the rule 2.

번째

열 그룹의 각 0 번째 열에서의 1의 위치를 결정한다. 여기서, 각

번째 열 그룹의 0 번째 열의 차수를

라 할 때, 각 1이 있는 행의 위치를

이라 가정하면,

번째 열 그룹 내의

번째 열에서 1이 있는 행의 위치

는 하기 <수학식 2>와 같이 정의된다. Rule 2: First

th

Determine the position of 1 in each 0th column of the row group. Here,

Order of the 0th column of the 1st column group

Is the position of the row

Assuming that

Within the first column group

The position of the row with 1 in the third column

Is defined as in Equation 2 below.

번째

열 그룹 내에 속하는 열들의 차수는 모두

로 일정함을 알 수 있다. According to the rule above

th

The order of the columns belonging to the column group is

It can be seen that the constant.

일때, 3개의 열 그룹의 0 번째 열에 대한 1을 가지는 행의 위치 정보에 대한 3개의 시퀀스는 다음과 같이 나타낼 수 있다. 여기서, 상기 시퀀스는 무게 1 위치 시퀀스(weight-1 position sequence)라고 한다.As a specific example

In this case, three sequences of position information of a row having 1 for the 0th column of the 3 column group may be expressed as follows. Here, the sequence is called a weight-1 position sequence.

The weight 1 position sequence of the row with the 0 th 1 of each column group may indicate only the sequence where the weight 1 is positioned for each column group as follows.

0 1 2

0 11 13

0 10 14

번째 열의

번째 무게 1 위치 시퀀스는

번째 열 그룹에서 1을 가지는 행의 위치 정보를 순차적으로 나타낸 것이다. That is,

Of the first column

The first weight 1 position sequence

Position information of a row having a 1 in the first column group is sequentially displayed.

If the parity check matrix is configured using the information corresponding to the specific example and <Rule 1> and <Rule 2>, an LDPC code having the same concept as the DVB-S2 LDPC code shown in FIG. 4 may be generated.

It is known that the DVB-S2 LDPC code designed through <Rule 1> and <Rule 2> can be efficiently encoded using a structural form. Each step of the LDPC encoding process using the parity check matrix of the DVB-S2 will be described with the following example.

,

,

,

를 특징으로 하는 DVB-S2 LDPC 부호를 이용하는 부호화 과정을 설명하였다. 또한 설명의 편의를 위해 길이가

인 정보어 비트들을

로 나타내고, 길이가

인 패리티 비트들을

로 나타낸다. As a specific example below

,

,

,

The encoding process using the DVB-S2 LDPC code, which has been described above, has been described. Also for convenience of description the length

Information bits

And the length is

Parity bits

Respectively.

. Step 1 : The LDPC encoder initializes parity bits as follows.

.

Step 2 : The LDPC coder reads the information of the row 1 in the column group from the 0 th weight 1 position sequence of the sequences representing the stored parity check matrix.

0 2084 1613 1548 1286 1460 3196 4297 2481 3369 3451 4620 2622

를 이용하여 하기의 <수학식 3>과 같이 특정 패리티 비트

들을 업데이트한다. 여기서,

는 각각의

값을 의미한다. The called information and the first information word bit

As shown in Equation (3) below,

Lt; / RTI &gt; here,

Respectively,

Lt; / RTI >

는

로 표기하기도 하며,

는 이진(binary) 덧셈을 의미한다. In Equation (3)

The

Sometimes referred to as

Means binary addition.

이후의 다음 359개의 정보어 비트

,

에 대해서 먼저 하기의 <수학식 4>에 대한 값을 구한다. Step 3 :

The following 359 information bits

,

First, the value for Equation 4 below is obtained.

는 각각의

값을 의미한다. 상기 <수학식 4>는 <수학식 2>와 동일한 개념의 수학식임에 유의한다. In Equation (4)

Respectively,

Lt; / RTI &gt; It is noted that Equation 4 is a mathematical equation having the same concept as Equation 2.

에 대해서

을 업데이트한다. 예를 들어

, 즉,

에 대해서 하기의 <수학식 5>와 같이

들을 업데이트한다. Next, a similar operation to <Equation 3> is performed using the value obtained in Equation 4. In other words,

about

Update E.g

, In other words,

As shown below in Equation (5)

Lt; / RTI &gt;

임에 유의한다. 위와 같은 과정을

에 대해서 마찬가지로 진행한다. In the case of Equation 5

. Same process as above

Proceed similarly for.

에 대해서 첫 번째 무게 1 위치 시퀀스인

의 정보를 호출하고, 특정

을 업데이트한다. 여기서,

는 을 의미한다.

이후의 다음 359개의 정보어 비트

에 대해서 <수학식 4>를 유사하게 적용하여

를 업데이트한다. Step 4 : 361th information word bit as in Step 2 above

For the first weight 1 position sequence

Call information, and specific

Update here,

The .

The following 359 information bits

Equation 4 is similarly applied to

Lt; / RTI &gt;

Step 5 : Repeat the steps 2, 3 and 4 for all 360 groups of information word bits.

Step 6: Finally, the parity bit is determined through Equation 6.

들이 LDPC 부호화가 완료된 패리티 비트들이다. In Equation (6)

These are parity bits for which LDPC encoding is completed.

As described above, in DVB-S2, LDPC encoding is performed through the steps 1 to 6.

In order to apply the LDPC code to the actual communication system, it must be designed to meet the data rate required in the communication system. Especially, the communication system supporting various broadcasting services as well as the adaptive communication system applying the Hybrid Automatic Retransmission Request (HARQ) method and the Adaptive Modulation and Coding (AMC) method can meet the requirements of the system. Accordingly, LDPC codes having various codeword lengths are required to support various data transmission amounts.

However, as described above, the LDPC code used in the DVB-S2 system has only two types of codeword lengths due to its limited use, and requires an independent parity check matrix. For this reason, there is a need for a method that supports various codeword lengths to increase the scalability and flexibility of the system. In particular, in the DVB-S2 system, data transmission of hundreds to thousands of bits is required to transmit signaling information. Since only two DVB-S2 LDPC codes are 16200 and 64800, support for various codeword lengths is essential. .

In addition, storing an independent parity check matrix for each codeword length of an LDPC code reduces memory efficiency. Therefore, it is not necessary to design a new parity check matrix, and to efficiently convert various codeword lengths from a given parity check matrix. Support is needed.

The present invention for solving the problems of the prior art as described above is a communication using a low-density parity check code for generating LDPC code having a different codeword length by using puncturing or shortening from a given LDPC code The system provides a channel coding / decoding method.

In addition, the present invention provides a channel encoding / decoding method and apparatus in a communication system using a low density parity check code that ensures optimal performance in consideration of the DVB-S2 structure.

In accordance with another aspect of the present invention, a channel decoding method includes: determining a number of parity bits to be punctured in a channel encoding method in a system using a low density parity check code; Determining the number of parity bit groups to be punctured based on the number of parity bits to be punctured; And puncturing the parity bits based on the determined number of parity bit groups to be punctured and a predetermined order of the parity bit groups to be punctured, wherein the parity bit group includes a plurality of parity bits in one parity bit group. It is configured to have a constant interval.

According to an embodiment of the present invention, in a channel encoding apparatus in a system using a low density parity check code, the channel decoding apparatus determines the number of parity bits to be punctured and performs puncturing based on the number of parity bits to be punctured. Determining a number of parity bit groups and puncturing a parity bit on the basis of the determined number of parity bit groups to be punctured and a predetermined order of parity bit groups to be punctured, wherein the parity bit group is one; The plurality of parity bits in the parity bit group of are configured to have a constant interval.

The present invention can generate a separate LDPC code having a different codeword length by using information of a given parity check matrix in a communication system using an LDPC code.

The present invention also applies perforation to optimize the performance of the DVB-S2 LDPC code.

1 is a diagram showing an example of a parity check matrix of an LDPC code having a length of 8;
2 is a Tanner graph of an example of a parity check matrix of an LDPC code having a length of 8;
3 is a schematic structural diagram of a DVB-S2 LDPC code;
4 is an example of a parity check matrix of an LDPC code of DVB-S2 type;
5 is a block diagram of a transceiver of a communication system using an LDPC code;
FIG. 6 is a diagram illustrating a first example of applying irregular puncturing to the LDPC code of FIG. 4; FIG.
FIG. 7 illustrates a second example in which regular puncturing is applied to the LDPC code of FIG. 4;
FIG. 8 is a diagram illustrating a third example in which regular puncturing is applied to the LDPC code of FIG. 4. FIG.
9 is a flowchart for generating an LDPC code having a different codeword length from a parity check matrix of an stored LDPC code according to an embodiment of the present invention;
10 is a flowchart illustrating an LDPC decoding method in a receiving apparatus when a puncturing pattern is applied according to an embodiment of the present invention;
11 is a block diagram of a transmitter using an LDPC code to which puncturing and shortening proposed in the present invention are applied.
12 is a block diagram of a receiver using an LDPC code to which puncturing and shortening proposed in the present invention are applied.

Preferred embodiments according to the present invention will be described in detail with reference to the accompanying drawings. It should be noted that in the following description, only parts necessary for understanding the operation according to the present invention will be described, and descriptions of other parts will be omitted so as not to obscure the subject matter of the present invention.

An embodiment of the present invention proposes a method and apparatus for supporting an LDPC code having various codeword lengths by using a parity check matrix of a structured LDPC code of a specific type. In addition, an embodiment of the present invention proposes an apparatus and a control method supporting various codeword lengths in a communication system using a specific type of LDPC code. In particular, an embodiment of the present invention proposes a method and apparatus for generating a smaller LDPC code using a parity check matrix of a given LDPC code.

5 is a block diagram of a transceiver of a communication system using an LDPC code.

는 수신기(530)로 전송되기 전에 송신기(510)의 LDPC 부호화기(encoder)(511)로 입력된다. 그러면 상기 LDPC 부호화기(encoder)(511)는 입력된 메시지

를 부호화하여 출력한 부호화 신호를 변조기(Modulator)(513)로 전송한다. 상기 변조기(513)는 상기 부호화된 신호를 변조한 후 무선 채널(520)을 통해 수신기(530)로 전송한다. 그러면, 수신기(530)의 복조기(Demodulator)(531)는 수신된 신호 r을 복조한 후, 복조된 신호 x를 LDPC 복호기(Decoder)(533)로 출력한다. 상기 LDPC 복호기(533)는 무선 채널(520)을 통해 받은 데이터를 통해 메시지의 추정치(estimatation value)

를 추정해낸다. 5, the message

Is input to the LDPC encoder 511 of the transmitter 510 before being transmitted to the receiver 530. The LDPC encoder 511 then inputs an input message.

Transmits the encoded signal output by encoding the to the modulator (513). The modulator 513 modulates the encoded signal and transmits the modulated signal to the receiver 530 through the wireless channel 520. The demodulator 531 of the receiver 530 demodulates the received signal r and then outputs the demodulated signal x to the LDPC decoder 533. The LDPC decoder 533 may estimate a message through data received through the wireless channel 520.

.

The LDPC encoder 511 generates a parity check matrix according to a codeword length required by the communication system in a preset manner. In particular, in the embodiment of the present invention, the LDPC encoder 511 may support various codeword lengths without the need for additional additional storage information by using the LDPC code.

In the embodiment of the present invention, a method of supporting various codeword lengths uses a method called puncturing or shortening. The puncturing method refers to a method of generating LDPC codewords by performing LDPC encoding from a given specific parity check matrix and then substantially not transmitting a specific part of the LDPC codeword. Therefore, at the receiving end, the untransmitted portion is determined to be an erasure.

In order to understand the puncturing method, the parity check matrix of the DVB-S2 LDPC code of FIGS. 3 and 4 will be described in detail.

이고, 앞부분은 길이가

인 정보어 비트들

이 대응되고, 뒷부분은 길이가

인 패리티 비트들

이 대응된다. The parity check matrix of the DVB-S2 LDPC code of FIG.

The front part is

Information word bits

This corresponds, and the back part has a length

In parity bits

This corresponds.

인 이미 생성된 LDPC 부호어 중에서

개의 약속된 위치에 있는 비트들을 단지 전송하지 않음으로써 길이가

인 LDPC 부호어를 전송하는 것과 동일한 효과를 얻는다. 패리티 검사 행렬에서 천공된 비트들에 해당하는 열들은 복호 과정에서 모두 그대로 사용되므로 단축법과는 차이가 있다. The above puncturing method is generally applicable to both information word bits and parity bits. In addition, although the puncturing and shortening methods have in common that the codeword length of a code is shortened, unlike the shortening method, the puncturing method is not a concept of limiting a value of a specific bit. The puncturing method is a method of treating the receiver with an erasure by not transmitting a specific information word bit or a specific part of the generated parity bit. In other words, the length

Of already generated LDPC codewords

By not transmitting the bits at the two promised positions,

The same effect as transmitting an LDPC codeword is obtained. The columns corresponding to bits punctured in the parity check matrix are used as they are in the decoding process, which is different from the shorthand method.

Since the position information on the punctured bits can be shared or estimated by the transmitter and the receiver in the same time as the system is set, the receiver performs the decoding only by processing the punctured bits.

이고, 정보어의 길이는 변함없이

이므로 부호율이

이 되어 처음 주어진 부호율

보다 항상 크게 된다. The puncturing method uses the codeword length

The length of the information word

Since the code rate is

Is the first given code rate

Always bigger.

The following describes a puncturing method and a shortening method suitable for DVB-S2 LDPC codes. The DVB-S2 LDPC code is a kind of LDPC code having a specific structure as mentioned in the background art. Therefore, shortening and puncturing can be applied more efficiently than in the case of a general LDPC code.

,

인 DVB-S2 LDPC 부호로부터 단축법과 천공법을 통하여 우리가 최종적으로 얻고자 하는 LDPC 부호의 부호어 길이와 정보어 길이를 각각

,

이라 하자. 만일 우리가

,

라고 정의하면, DVB-S2 LDPC 부호의 패리티 검사행렬에서

비트만큼 단축을 취하고,

비트만큼 천공을 취하면 부호어 길이와 정보어 길이를 각각

,

인 상기 LDPC 부호를 생성할 수 있다. 이렇게 생성된 상기 LDPC 부호는

또는

일 때, 부호율이

가 되어 일반적으로 DVB-S2 LDPC 부호의 부호율

와는 다르게 되므로 대수적 특성이 변하게 된다. 여기서

인 경우에는 단축이나 천공을 모두 적용하지 않거나 또는 단축만 취한 경우에 해당된다. For convenience of explanation, the codeword length and information word length

,

The codeword length and information word length of the LDPC code that we finally obtain from the DVB-S2 LDPC code

,

. If we

,

Is defined in the parity check matrix of the DVB-S2 LDPC code.

Taking the bit as short as possible,

When puncturing bit by bit, the codeword length and information word length

,

The LDPC code may be generated. The generated LDPC code is

or

, The code rate is

Code rate of DVB-S2 LDPC code

And the algebraic characteristic is changed. here

If is not applied to both shortening or drilling or if only shortening is taken.

이며, 3개의 열 그룹의 0 번째 열에 대한 무게 1 위치 시퀀스가 아래와 같음에 유의한다.The characteristics of applying parity puncturing to the DVB-S2 LDPC code will be described in detail with reference to FIG. 4. 4 is

Note that the weight 1 position sequence for the 0th column of the 3 column group is as follows.

0 1 2

0 11 13

0 10 14

번째 열의

번째 무게 1 위치 시퀀스는

번째 열 그룹에서 1을 가지는 행의 위치 정보를 순차적으로 나타낸 것이다. remind

Of the first column

The first weight 1 position sequence

Position information of a row having a 1 in the first column group is sequentially displayed.

FIG. 6 is a diagram illustrating an example in which irregular puncturing is applied to the LDPC code of FIG. 4. Since the parity bits punctured in FIG. 6 are lost by the decoder, the performance improvement effect is not significantly improved in the LDPC decoding process as compared to the other bits that are not lost, thereby reducing reliability. As a result, other bits directly connected to the punctured parity bits, which are less reliable, are also adversely affected by the performance improvement effect during the decoding process. This effect is amplified as the number of edges directly connected to the punched bits on the Tanner graph.

For example, in FIG. 6, the bits of the 0 th information word corresponding to the 0 th column are directly connected to the punctured parity bit twice, and the third information word bit corresponding to the third column is directly connected to the punctured parity bit 1 times. The eighth information word bit corresponding to the eighth column is connected to the punctured parity bit three times. In this case, the order in which the performance improvement effect is excellent in the decoding process is the 3, 0, 8th order. That is, when the degree of varieties of information nodes is the same, the performance improvement effect is lower as the number of punctured bits increases.

Referring to FIG. 6, it can be seen that the number of perforated parities directly connected to each information word is irregular by an irregular puncturing pattern. Therefore, the reliability of each information word bit is also likely to be irregular. In other words, some information word bits are decoded more than necessary, but some information word bits can suffer severe performance degradation. Such irregular puncturing patterns may cause irregularity of serious reliability of the information bits in the decoding process.

FIG. 7 is a diagram illustrating a second example in which regular puncturing is applied to the LDPC code of FIG. 4. That is, Figure 7 is an example of applying a relatively regular puncturing pattern of a particular type.

Even if a relatively regular puncturing pattern is applied, the connection state with the information word bit may be non-uniform according to the puncturing pattern. In the case of FIG. 7, it may be more non-uniform than the irregular puncturing pattern of FIG. 6.

As described above, in the case of the LDPC code having a parity check matrix having a specific structure such as the DVB-S2 LDPC code, the connection state between the punctured parity bit and the information word bit may vary greatly according to the puncturing pattern.

The present invention proposes a puncturing pattern that provides stable decoding performance by maximally suppressing nonuniformity of reliability of information word bits during decoding process using the structural characteristics of the DVB-S2 LDPC code.

In the example of FIG. 8, since one of the configuration variables is q = 3 in the parity check matrix of FIG. 4, a puncturing pattern in which the spacing of the punctured parity bits is kept at 3 is applied. As shown in FIG. 8, each information word bit is equally connected to two punctured bits.

값에 따라 설정하면 천공된 비트와 정보어 비트 사이의 비균일성이 크게 감소 하는 이유는 DVB-S2 LDPC 부호의 구조에서부터 찾을 수 있다. 이를 설명하기 위하여 상기 도 3을 살펴본다.The interval of the perforated parity bits

The reason for the large decrease in nonuniformity between the punctured bits and the information word bits can be found in the structure of the DVB-S2 LDPC code. In order to explain this, look at Figure 3 above.

에 대해

의 배수만큼 차이가 난다. 여기서

은 LDPC 부호어 길이고,

은 정보어의 길이를 의미한다. 더 구체적으로 말하면, 특정 열그룹 내에서 연속적인 두 개의 열에서 1이 위치한 행의 인덱스들은 정확하게 모듈로(modulo)

에 대해

만큼 차이가 난다. Referring to <Rule 1>, <Rule 2>, and FIG. 3, each column group determines the position of '1' of the first column within the corresponding column group. In this case, the index of the row where 1 is located in the remaining columns is exactly modulo with the index of the row where 1 is located in the first column.

About

It differs by a multiple of. here

Is the LDPC codeword length,

Is the length of the information word. More specifically, the indices of the rows where 1 is located in two consecutive columns within a particular column group are precisely modulo.

About

As far as the difference is.

번째 패리티 비트는

번째 행에 위치한 1과 대응된다. Another feature of the DVB-S2 LDPC code is the partial matrix corresponding to parity in the parity check matrix. Referring to FIG. 3, the parity portion is a lower triangular matrix structure having all 1s in a diagonal portion.

Parity bit

Matches 1 in the first row.

간격만큼 패리티 천공이 반복되면, 특정 열그룹 내에서 천공된 패리티 비트와 연결되는 정보어 비트의 선분 수는 최대한 균일하게 된다. 예를 들어

에 대해서

번째 패리티 비트가 천공되고,

에 대해서

번째 패리티 비트가 반복적으로 천공되었다고 가정하자. 그럼 정보어 비트가

번째 패리티 비트와 연결되어 있다는 것은 해당 정보어 비트에 대응되는 열의

번째 행에 1이 있음을 의미한다. 따라서 <규칙 1>과 <규칙 2>에 따라 상기 정보어 비트와 같은 열그룹에 있는 열 중에서 상기 정보어 비트로부터

만큼 떨어져 있는 정보어 비트에 대응되는 열에는

번째 행에 1이 있음을 알 수 있다. 따라서, 천공된

번째 비트와 연결된다. From the structural characteristics of this DVB-S2 LDPC code, a certain parity is precisely punctured.

When the parity puncture is repeated by the interval, the number of segments of the information word bits connected to the parity bits punctured in the specific column group is as uniform as possible. E.g

about

The first parity bit is punctured,

about

Assume that the first parity bit has been repeatedly punctured. Then the information bit

Connected with the first parity bit means that the corresponding column of the information word bit

Means that there is 1 in the first row. Therefore, according to <Rule 1> and <Rule 2>, from the information word bit among the columns in the same column group as the information word bit

The columns corresponding to the information word bits that are separated by

Notice that there is 1 in the first row. Thus, perforated

Connected to the first bit.

은 LDPC 부호어 길이를 의미하며, 각 열그룹은

개의 열들로 이루어져 있으며

개의 패리티 비트를 천공하는 경우라고 가정하자. 또한 하기의 천공 과정은 도 9의 흐름도에 나타내었다.Since the DVB-S2 LDPC code has the same degree of variable node corresponding to all information words in one column group and one or less 1s are distributed in one row, the puncturing pattern is applied to one column group. Corresponding information word bits are concatenated with the same number of punctured bits. Therefore, the connection between the punctured bits and the information word bits has a regular advantage, and thus stable decoding can be expected in the decoding process. The general procedure for the application of the above-described puncturing method can be summarized as follows. For convenience of explanation

Means LDPC codeword length, and each column group

Consists of four rows

Assume a case of puncturing two parity bits. In addition, the following drilling process is shown in the flowchart of FIG. 9.

9 is a flowchart for generating an LDPC code having a different codeword length from a parity check matrix of a stored LDPC code according to an embodiment of the present invention.

Drilling Step 1 : The transmitting device generates an existing DVB-S2 LDPC codeword, which is shortened or shortened in step 901.

를 결정하고

를 구한다. 여기서

는

보다 같거나 작은 최대 정수를 의미한다. Drilling Step 2 : Number of Lengths to be Punched in Step 903 and Step 905 in the Transmission Device (Length)

And

. here

The

Means a maximum integer greater than or equal to

,

에 대해서 천공될 패리티 비트들

,

, ...,

을 결정한다.

에 대해서

의 값들은 성능을 고려하여 사전에 미리 결정되어 있었다고 가정한다. (여기서

인 관계가 있음에 유의한다.) Drilling Step 3 : In the transmitting device, in step 907

,

Parity bits to be punctured for

,

, ...,

.

about

The values of are assumed to be predetermined in consideration of performance. (here

Note that there is a relationship.)

,

에 대해서 패리티 비트

에 대해서 모두 천공을 적용한다. 여기서 상수 B는 사전이 미리 설정된 0이 아닌 정수이다. Drilling Step 4 : In the transmitting device, in step 907

,

About the parity bit

Apply all perforations. Where the constant B is a non-zero integer preset.

에 대해서 패리티 비트

를 추가적으로 천공한다.(907 단계)
Drilling Step 5 :

About the parity bit

Drill additionally (step 907).

In step 909, the transmitter transmits bits other than the punctured bits.

개의 패리티가 천공되고, 천공 단계 5에 의해서

개의 패리티가 천공되어 총

개의 패리티가 천공되었음을 알 수 있다. 위와 같이 천공되어 전송된 DVB-S2 LDPC 부호어는 수신 장치에서 수신된 신호로부터 도 10을 참고하여 아래에서 상세히 설명될 복호 과정을 거쳐 원래의 신호로 복원한다.Looking at the drilling process, in the drilling step 3 and the drilling step 4

Parity is punctured, and by puncturing step 5

Parity perforated

It can be seen that the parity has been punctured. The above-described punctured and transmitted DVB-S2 LDPC codeword is restored to the original signal through a decoding process to be described in detail below with reference to FIG.

,

,

,

의 특성을 가지는 부호이다.
In order to more easily understand the process from the drilling step 3 to the drilling step 5 in the drilling process, the following specific example will be described. The DVB-S2 LDPC code used here is

,

,

,

This code has the characteristic of.

Example of puncturing step 1 : The transmitting device generates an existing DVB-S2 LDPC codeword that is shortened or not shortened.

를 결정하고

를 구한다. 여기서

는

보다 같거나 작은 최대 정수를 의미한다. Example of drilling step 2 : length to take drilling in the transmitting device

And

. here

The

Means a maximum integer greater than or equal to

,

에 대해서 천공될 패리티 비트들

,

, ...,

을 결정한다.

에 대해서

의 값들은 천공된 패리티 비트와 정보어 비트의 연결 관계 및 밀도 진화(density evolution) 분석 방법을 이용하여 점근적인(asymptotic) 성능이 우수한 경우를 고려하여 다음과 같이 선택하였다. Example of drilling step 3 : In the transmitting device

,

Parity bits to be punctured for

,

, ...,

.

about

The values of are selected as follows considering the case of excellent asymptotic performance using the connection relationship between the punctured parity bit and the information word bit and the density evolution analysis method.

27, 13, 29, 32, 5, 0, 11, 21, 33, 20, 25, 28, 18, 35, 8, 3, 9, 31, 22, 24, 7, 14, 17, 4, 2, 26, 16, 34, 19, 10, 12, 23, 1, 6, 30, 15

에 대해서

번째 무게 1 위치 시퀀스는

값에 대응된다.
In the sequence

about

The first weight 1 position sequence

Corresponds to the value.

,

일 경우 패리티 비트

에 대해서 모두 천공을 적용한다. (여기서 B의 값은 1로 설정했다.) Example of drilling step 4 : In the transmitting device

,

Parity bit if

Apply all perforations. (Here we set the value of B to 1.)

에 대해서 패리티 비트

를 추가적으로 천공한다. Example of drilling step 5 : In the transmitting device

About the parity bit

Drill additionally.

와

의 값들을 정의해 주는 수열 정보와

값을 알면 천공 패턴이 정확하게 정의됨을 알 수 있다. Looking at the example of the drilling step 1 to the drilling step 5, the number of bits to be drilled

Wow

Sequence information that defines the values of

Knowing the value, we can see that the perforation pattern is accurately defined.

,

,

, ..., )와 같이 표기할 때, 상기 천공 단계의 예를 다음 <표 1>과 같이 더욱 간단하게 나타낼 수 있다. Parity bits of the DVB-S2 LDPC code applied to the example of puncturing step 5 from the example of puncturing step 1 are

,

,

, ..., In the following notation, an example of the perforation step may be more simply shown in Table 1 below.

,

,

,

의 특성을 가지는 DVB-S2 LDPC 부호에 대해서 <표 2>와 같은 천공 패턴을 구할 수 있다. As another embodiment of the drilling step

,

,

,

The puncturing pattern shown in Table 2 can be obtained for the DVB-S2 LDPC code having the characteristic of.

As described above, in the exemplary embodiment of the present invention, unlike the conventional puncturing method or the simple uniform puncturing method that is commonly used for puncturing the DVB-S2 LDPC code, the DVB-S2 is used by using the structural characteristics of the DVB-S2 LDPC code. An efficient puncturing method can be applied to stabilize the performance of the LDPC code.

Referring to the method of determining the order of bits punctured in step 3 of the puncturing process of the DVB-S2 LDPC code, the order of the punctured bits is determined through the density evolution analysis method and the cycle analysis method on the Tanner graph.

Since the puncturing method not only changes the length of the LDPC codeword, but also shortens the codeword length without changing the length of the information word, the puncturing method increases the code rate. Therefore, in order to obtain the code rate and codeword length required in the system, it is easy to apply the shortening method as well as the puncturing method.

,

인 LDPC 부호로부터 단축법과 천공법을 통하여 우리가 최종적으로 얻고자 하는 LDPC 부호의 부호어 길이와 정보어 길이를 각각

,

라 할 때,

,

라고 정의하면, LDPC 부호의 패리티 검사행렬에서

비트만큼 단축을 취하고,

비트만큼 천공을 취하면 부호어 길이와 정보어 길이를 각각

,

인 상기 LDPC 부호를 생성할 수 있다. 이렇게 생성된 상기 LDPC 부호는

또는

일 때, 부호율이

가 되므로

와

을 고려하여 천공 및 단축 길이를 설정하면 된다. As described above, the given codeword length and information word length

,

The codeword length and information word length of the LDPC code that we finally want to obtain from the LDPC code

,

When we say

,

Is defined as the parity check matrix of the LDPC code.

Taking the bit as short as possible,

When puncturing bit by bit, the codeword length and information word length

,

The LDPC code may be generated. The generated LDPC code is

or

, The code rate is

Becomes

Wow

In consideration of this, the perforation and shortening length may be set.

10 is a flowchart illustrating a receiving method in a receiving apparatus according to an embodiment of the present invention.

Referring to FIG. 10, the receiving apparatus determines or estimates a puncturing / shortening pattern from the signal received in step 1001. Thereafter, in step 1003, the receiving apparatus determines whether there is a bit punctured or shortened.

If the punctured or shortened bit does not exist, the receiving device performs decoding in step 1009. However, if there is a shortened or punctured bit, the receiving device transmits the puncturing / shortening pattern to the LDPC decoder 1260 in step 1005.

The LDPC decoder 1260 sets the punctured bits to erasure in step 1007, sets the probability that the shortened bits have a value of 0 is 1, and then performs decoding in step 1009.

An example is shown in FIG. 11 to show in more detail a transmission apparatus for realizing a puncturing process for a DVB-S2 LDPC code. 11 is a block diagram of a transmitter using a punctured / shortened LDPC code according to an embodiment of the present invention.

The transmitting apparatus includes a controller 1110, a shortened pattern applying unit 1120, an LDPC code parity check matrix extractor 1140, and an LDPC encoder 1160.

The LDPC code parity check matrix extractor 1140 extracts the shortened LDPC code parity check matrix. The LDPC code parity check matrix may be extracted using a memory, may be given in a transmitting device, or may be generated in a transmitting device.

The controller 1110 controls the shortened pattern applying unit 1120 to determine a shortened pattern according to the length of the information word, and the shortened pattern applying unit 1120 has a bit having a zero value at a position corresponding to the shortened bit. Or inserts a column corresponding to a shortened bit from a parity check matrix of a given LDPC code. In the method of determining the shortened pattern, a shortened pattern stored using a memory is used, a shortened pattern is generated using a sequence generator (not shown), or the density is evolved with respect to a parity check matrix and a given information word length. It can also be obtained using an analysis algorithm.

The LDPC encoder 1160 performs encoding based on the LDPC code shortened by the controller 1110 and the shortened pattern applying unit 1120.

The controller 1110 controls the puncturing pattern applying unit 1180. The puncturing pattern applying unit 1180 determines the number of parity bits to be punctured, divides the parity bits at regular intervals, determines the number of puncturing bits to be punctured within the predetermined interval, and within the predetermined interval. The position of the puncturing parity bit corresponding to the determined number of puncturing bits to be punctured is determined, and puncturing is performed by repeating the puncturing parity bits corresponding to the determined position at the predetermined intervals.

12 is a block diagram of a receiving apparatus according to an embodiment of the present invention.

FIG. 12 shows an example of a receiving apparatus for receiving a signal transmitted in a communication system using the punctured or shortened DVB-S2 LDPC code and restoring data desired by a user from the received signal.

Referring to FIG. 12, the receiving apparatus includes a controller 1210, a shortening, a puncturing pattern determination or estimator 1220, a demodulator 1230, and an LDPC decoder 1240.

The demodulator 1230 receives and demodulates the shortened LDPC code, and transmits the demodulated signal to the shortened pattern determination or estimator 1220 and the LDPC decoder 1240.

The shortening, puncturing pattern determination or estimator 1220 estimates or determines the information on the puncturing or shortening pattern of the LDPC code from the demodulated signal under the control of the control unit 1210, and thus positions the punctured and shortened bits. Information is transmitted to the LDPC decoder 1240. In the method of determining or estimating the puncturing and shortening pattern by the shortening and puncturing pattern determination or estimating unit 1220, puncturing and shortening using a puncturing and shortening pattern stored using a memory or a generation method that is implemented in advance. A pattern may be generated or obtained using a parity check matrix and a density evolution analysis algorithm for a given information word length. In addition, the bits punctured in the LDPC decoder 1240 are processed to be lost to perform decoding.

In addition, the shortening, puncturing pattern determination or estimator 1220 may first perform the pattern determination or estimation of the shortening in the receiving apparatus when both the shortening and the puncturing are applied to the transmitting apparatus as in the embodiment of the present invention. The pattern determination or estimation may be performed first, or both the pattern determination or estimation of the shortening and the pattern determination or estimation of the perforation may occur.

The LDPC decoder 1240 performs decoding on the assumption that the probability that the punctured bits are 0 and the probability that the 1 is 1 are equal to 1/2, respectively. In addition, since the probability that the value of the shortened bit is 0 is 1 (that is, 100%), the shortened bits in the operation of the decoder may not participate in the operation of the decoder, or the decoding may be performed by using the probability 1 of the shortened bits. Decide if you want to participate.

When the LDPC decoder 1240 knows the length of the DVB-S2 LDPC code shortened by the shortened pattern determination or estimator 1220, the LDPC decoder 1240 restores data desired by the user from the received signal.

11, it can be seen that the shortening is performed before the input of the LDPC encoder 1160 and the puncturing is performed at the output stage of the LDPC encoder 1160. However, referring to the receiver of FIG. 12, the LDPC decoder 1240 needs to know information about puncturing and shortening simultaneously in the decoder so that decoding can be performed.

Claims (16)

In the channel decoding method in a system using a low density parity check code,
Demodulating a signal transmitted from a transmitter;
Determining position information of the punctured parity bits; And
Decoding data by using the location information of the punctured parity bits;
The determining of the position information of the punctured parity bit,
Determining the number of punctured parity bits;
Determining the number of punctured parity bit groups based on the number of punctured parity bits; And
And obtaining a sequence of a predetermined punctured parity bit group. The method of claim 1, wherein the order of the predetermined punctured parity bit group is 6, 4, 18, 9, 13, 8, 15, 20, 5, 17, 2, 24, 10 when the codeword length is 16200. , 22, 12, 3, 16, 23, 1, 14, 0, 21, 19, 7, 11 channel decoding method in a system using a low density parity check code, characterized in that. The method of claim 1, wherein the order of the predetermined punctured parity bit group is 27, 13, 29, 32, 5, 0, 11, 21, 33, 20, 25, 28, 18 when the codeword length is 16200. , 35, 8, 3, 9, 31, 22, 24, 7, 14, 17, 4, 2, 26, 16, 34, 19, 10, 12, 23, 1, 6, 30, 15 A channel decoding method in a system using a low density parity check code. The low density parity check code of claim 1, further comprising determining that the parity bit is punctured if the number of the punctured parity bits is not a multiple of the length of one parity bit group. Channel decoding method in your system. 2. The method of claim 1, further comprising receiving the remaining bits except for the punctured bits. The method of claim 1, wherein the number of parity bit groups is determined by the following equation,
&Lt; Equation &

A denotes the number of parity bit groups to be punctured, Np denotes the number of parity bits to be punctured, and M ₁ denotes the length of one parity bit group. Decryption method. The method of claim 6, wherein the parity bit group is determined by the following formula:
&Lt; Equation &

P _j represents the j th parity bit group, N ₁ is the length of the codeword, K ₁ is the length of the information word, M ₁ is the length of one parity bit group, q is q = (N ₁ / K ₁ ) / M ₁ is satisfied, K ₁ / M ₁ is an integer, and the equation satisfies 0 ≦ j <q in a channel decoding method in a system using a low density parity check code. The method of claim 7, wherein determining the position information of the punctured parity bit,
A parity bit group

Determining that all parity bits within are punctured; And

Of the parity bits in the parity bit group

And determining that two parity bits are additionally punctured. A channel decoding apparatus in a system using a low density parity check code,
A demodulator for demodulating a signal transmitted from a transmitter;
A puncturing pattern estimator for determining position information of punctured parity bits; And
A decoder which decodes data by using the position information of the punctured parity bits,
Position information of the perforated parity bit,
Determining the number of punctured parity bits, determining the number of punctured parity bit groups based on the number of punctured parity bits, and obtaining an order of a predetermined punctured parity bit group. Channel decoding apparatus in a system using a low density parity check code. The method of claim 9, wherein the order of the predetermined punctured parity bit group is 6, 4, 18, 9, 13, 8, 15, 20, 5, 17, 2, 24, 10 when the codeword length is 16200. , 22, 12, 3, 16, 23, 1, 14, 0, 21, 19, 7, 11 channel decoding apparatus in a system using a low density parity check code, characterized in that.
10. The method of claim 9, wherein the order of the predetermined punctured parity bit group is 27, 13, 29, 32, 5, 0, 11, 21, 33, 20, 25, 28, 18 when the codeword length is 16200. , 35, 8, 3, 9, 31, 22, 24, 7, 14, 17, 4, 2, 26, 16, 34, 19, 10, 12, 23, 1, 6, 30, 15 And a channel decoding apparatus in a system using a low density parity check code. 10. The method of claim 9, wherein if the number of punctured parity bits is not a multiple of the length of one parity bit group, the puncturing pattern estimator additionally determines that the parity bits are punctured. Channel decoding apparatus in the system used. 10. The apparatus of claim 9, further comprising a receiver for receiving the bits other than the punctured bits.
10. The method of claim 9, wherein the number of parity bit groups is determined by the following equation,
&Lt; Equation &

A denotes the number of parity bit groups to be punctured, Np denotes the number of parity bits to be punctured, and M ₁ denotes the length of one parity bit group. Decryption device.
The method of claim 14, wherein the parity bit group is determined by the following equation:
&Lt; Equation &

P _j represents the j th parity bit group, N ₁ is the length of the codeword, K ₁ is the length of the information word, M ₁ is the length of one parity bit group, q is q = (N ₁ / K ₁ ) / The apparatus for decoding a channel in a system using a low density parity check code, wherein M ₁ is satisfied, K ₁ / M ₁ is an integer, and the equation satisfies 0 ≦ j <q. The method of claim 15, wherein determining the position information of the punctured parity bit,
A parity bit group

Determine that all parity bits within are punctured,

Of the parity bits in the parity bit group

And a parity bit is additionally punctured. Channel decoding apparatus in a system using a low density parity check code, characterized in that it is determined.

Images