WO2013069887A1

WO2013069887A1 - Method for generating channel link quality-adaptive parity check matrix, and method and apparatus for encoding low-density parity check code using same

Info

Publication number: WO2013069887A1
Application number: PCT/KR2012/007732
Authority: WO
Inventors: 김상효; 장민
Original assignee: 성균관대학교 산학협력단
Priority date: 2011-11-09
Filing date: 2012-09-26
Publication date: 2013-05-16
Also published as: KR101221544B1

Abstract

The present invention relates to a design of a low-density parity-check (LDPC) code, which is used in various communication environments including data transmission and data compression, and relates to a method and an apparatus for designing an appropriate LDPC code in accordance with characteristics and requirements of an applied technology using the LDPC code and the quality of a channel link, and using same for error correction encoding/decoding. The method for designing the LDPC code, according to the present invention, is characterized by considering a connected-degree set, diversity of connectivity, and distribution of variable nodes that are connected to a check node on a Tanner graph, which visually displays a parity-check matrix of the LDPC code.

Description

Generation method of channel link quality adaptive parity check matrix, encoding method and encoding device of low density parity check code using the same

The present invention relates to a low-density parity-check (LDPC) code, and more particularly, to design a suitable LDPC code according to the characteristics of the technology to which the LDPC code is applied and the link state The present invention relates to a method of performing error correction encoding and an apparatus for performing the same.

In a communication system, the state of a link between a transmitter and a receiver is greatly degraded due to channel noise, fading and interference along multiple paths. Therefore, in order to implement a next generation mobile communication and multimedia transmission system requiring high data throughput and reliability, a technique for overcoming noise, fading, and interference must be accompanied. To this end, researches on error-correcting codes that efficiently restore distortion of information under the mathematical basis of information theory to improve communication reliability have been actively conducted.

Low-density parity-check (LDPC) codes have good performance close to Shannon channel capacity, which is a theoretical communication limit. Due to the excellent performance, various standard technologies such as IEEE 802.16m mobile WiMAX and IEEE 802.11 Wi-Fi adopt LDPC codes as error correction codes.

Each LDPC code is defined by a parity-check matrix, and the parity check matrix may be represented by a bipartite graph called a Tanner graph for visual and intuitive understanding. The bipartite graph includes a variable node set, a check node set, and a set of edges connecting elements of the variable node set and the check node set to each other. The variable node corresponds one-to-one with bits of a codeword that is a coded result, and the check node corresponds to a parity-check equation indicating an algebraic relationship between the bits of the codeword.

1 is an example of a parity check matrix of a binary LDPC code consisting of four rows and eight columns. Referring to FIG. 1, since the parity check matrix of the LDPC code has eight columns, a codeword having a length of eight is generated.

FIG. 2 is a Tanner graph corresponding to the parity check matrix of the LDPC code of FIG. 1.

The Tanner graph of FIG. 2 shows eight variable nodes v1 201, v2 202, v3 203, v4 204, v5 205, v6 206, v7 207, and v8 ( 208 and four test nodes c1 211, c2 212, c3 213, c4 214 and connecting lines connecting them. The connection line of the Tanner graph is connected according to the parity check matrix of FIG. 1. If the value of the element h _ij of the i th row and the j th column of the parity check matrix is not 0, the j th variable node and the i th check node are It is connected by connecting line.

The parity check matrix and Tanner graph of an LDPC code as shown in FIGS. 1 and 2 are characterized by basic elements including the length and code rate of a codeword. The length N of the codeword is the number of columns of the parity check matrix, and the number of variable nodes in a Tanner graph. The code rate R represents the ratio of the number of columns and the number of rows in the parity check matrix, and the ratio of the number N of variable nodes and the number M of check nodes in a Tanner graph. The code rate is calculated by the following equation.

[Revision 17.10.2012 under Rule 26]
Equation 1

In the Tanner graph, the degree of a variable node and an inspection node represents the number of connection lines connected to each node. The order of the variable node is the number of nonzero elements present in the corresponding column array in the parity check matrix, and the order of the check node is the number of nonzero elements present in the corresponding row array in the parity check matrix. For example, referring to FIG. 2, variable nodes v1 201, v2 202, v3 203, v4 204, v5 205, v6 206, v7 207 and v8 208. The orders of) are 2, 2, 2, 2, 3, 3, 3, and 4 in order, and the order of test nodes c1 211, c2 212, c3 213, and c4 214 is 6 in order. , 5, 5 and 5.

The node perspective degree distribution of each of the variable node and the check node of the LDPC code is an index indicating a probabilistic distribution of orders of nodes in the LDPC code. Where λ _i represents the ratio of the number of variable nodes of order i to the number N of all variable nodes, and ρ _j represents the ratio of the number of inspection nodes of degree j and the number M of all inspection nodes. For example, in the LDPC codes corresponding to FIGS. 1 and 2, since the total number of variable nodes is 8 and the number of variable

nodes having orders

2, 3, and 4 is 4, 3, and 1, λ ₂ = 4/8, λ ₃ = 3/8, λ ₄ = 1/8. In addition, since the total number of inspection nodes is 4, and the number of inspection nodes having orders 5 and 6 is 3 and 1, ρ ₅ = 3/4 and ρ ₆ = 1/4.

The order distribution of the LDPC code does not represent all the features of the LDPC code. That is, even if they have the same order distribution, different LDPC codes may be generated according to the actual connection configuration. However, LDPC codes with the same order distribution have the same asymptotic performance. The asymptotic performance of the LDPC code is the error correction capability of the LDPC code under the assumption that there is no cycle in the corresponding Tanner graph, that is, the length of the codeword is infinite. As described above, a process of calculating asymptotic performance using the order distribution of the LDPC code is called density evolution.

A general optimization process of LDPC codes is to find an order distribution with good asymptotic performance at a given code rate. In other words, the order distribution is obtained by changing the order distribution to obtain asymptotic performance through density evolution, and among these, select an order distribution having excellent asymptotic performance. Although this optimization process is performed under the assumption of asymptotic performance, i.e., the length of codeword is infinite, the performance tendency is similar in LDPC code with finite length of codeword. That is, an LDPC code of finite length with an order distribution of good asymptotic performance is generally superior to an LDPC code of equal length with an order distribution of poor asymptotic performance.

The asymptotic performance of LDPC codes with the same order distribution is all the same. However, the actual performance of LDPC codes having the same order distribution may vary. That is, in the case of the LDPC code having a finite length, even if they have the same order distribution, the performance may vary depending on the actual connection configuration. Therefore, it is important to design an LDPC code having excellent performance that is suitable for the characteristics and requirements of a system which intends to use the LDPC code even among LDPC codes having the same order distribution in the actual LDPC code design.

In order to determine the optimized LDPC code, various LDPC codes have been proposed. However, in the conventional LDPC code encoding / decoding technique, there is a problem in that the optimal LDPC code changes as the channel quality changes. In other words, the quality of the channel link is relatively poor, and the LDPC code that shows excellent performance in the falling water region in which the slope of the error rate curve increases rapidly as the quality of the channel link improves is a range in which the quality of the channel link is relatively good. In the error-floor area, where the slope of the error rate curve is drastically lowered compared to the fall-off area, a poor quality fall / error-floor conflict is observed. However, the factors causing the fall / error-floor conflict between LDPC codes having the same order distribution have not been identified.

Therefore, the method of encoding / decoding using the LDPC code which has the excellent error correction capability regardless of the quality of the channel link by identifying the factors causing the fall / error-floor conflict and selecting the LDPC code adaptively to the quality of the channel. There is a need for research.

A first object of the present invention for overcoming the above disadvantages is to provide a method of designing a parity check matrix of an LDPC code having optimal performance according to the characteristics of a communication system in which an LDPC code is to be used and the quality of a channel link. .

It is also a second object of the present invention to provide an LDPC code encoding method having an optimal error correction capability by selecting a parity check matrix adaptively to the quality of a channel link on which an encoded codeword is transmitted.

It is also a third object of the present invention to provide an LDPC code encoding apparatus having an optimal error correction capability by selecting a parity check matrix adaptively to the quality of a channel link on which an encoded codeword is transmitted.

A method of generating a parity check matrix of a Low Density Parity Check (LDPC) code according to an embodiment of the present invention for achieving the first object of the present invention is that of the communication system using the LDPC code A condition setting step of determining at least one of a connected-degree set condition and a connectivity condition of the inspection nodes based on the channel link quality and based on at least one of the determined connection order set condition and the connectivity condition Generating a parity check matrix. Here, the connection order set condition may be set so that the number of different connection order sets of the check nodes is greater than a preset connection order set number criteria when the channel link quality is worse than a preset quality criterion. If the channel link quality is superior to a preset quality criterion, the number of different connection order sets of the check nodes may be set to be smaller than a preset connection order set number criterion. The connectivity condition may be set so that the distribution value of the connectivity of the check nodes is greater than a preset connectivity dispersion value criterion when the channel link quality is worse than a preset quality criterion. When the quality is superior to a preset quality criterion, the dispersion value of the connectivity of the check nodes may be set to be smaller than the preset connectivity dispersion value criterion. The channel link quality may be determined based on a bit energy to noise ratio (Eb / No). The quality criterion of the LDPC code is that the frame error rate curve of the LDPC code belongs to a waterfall region when the channel link quality is poorer than the quality criterion, and the channel link quality is superior to the quality criterion. The frame error rate curve may be characterized as belonging to an error-floor region.

According to an embodiment of the present invention for encoding a low density parity check (LDPC) code according to an embodiment of the present invention, a method of receiving information data and a set of connection orders of inspection nodes ( generating a parity check matrix of a predetermined number of LDPC codes having different connected-degree set conditions, or connectivity conditions, selecting one parity check matrix from the generated predetermined number of parity check matrices; And encoding the information data based on the selected parity check matrix to generate a codeword. The selecting of the parity check matrix may include measuring channel link quality of a communication system in which the LDPC code is used and having an error correction capability of the predetermined number of parity check matrices based on the channel link quality. And selecting an excellent parity check matrix. The measuring of the channel link quality may include estimating the channel link quality based on a stochastic characteristic of the channel link quality, which may include a signal-to-noise ratio (SNR). Can be. The selecting of the parity check matrix may include selecting a number of sets of different connection order sets among the predetermined number of parity check matrices when the channel link quality is worse than a preset quality criterion. A parity check matrix may be selected that is larger than a set number criterion, and when the channel link quality is superior to a preset quality criterion, the number of different connection order sets of the check nodes among the generated predetermined number of parity check matrices is preset. It may be characterized by selecting a parity check matrix less than the connection order set number criterion. The selecting of the parity check matrix may include: when the channel link quality is worse than a preset quality criterion, a variance value of the connectivity set by the variance value of the connectivity of the check nodes among the generated predetermined number of parity check matrices. A parity check matrix larger than a reference may be selected, and if the channel link quality is superior to a preset quality criterion, the variance value of the connectivity of the check nodes among the generated predetermined number of parity check matrices is preset. The parity check matrix smaller than the reference may be selected. The quality criterion is that the frame error rate curve of the LDPC code belongs to a waterfall region when the channel link quality is poorer than the quality criterion, and the frame error rate of the LDPC code when the channel link quality is superior to the quality criterion. The curve may be characterized as belonging to an error-floor region.

A low density parity check (LDPC) encoding apparatus according to an embodiment of the present invention for achieving the third object of the present invention includes an input unit for receiving information data and a set of connection order of check nodes. a parity check matrix generator for generating a parity check matrix of a predetermined number of LDPC codes having at least one of a connected-degree set condition and a connectivity condition, and a parity of one of the generated number of parity check matrices; And a control unit for selecting a parity check matrix, and an encoding unit for generating a codeword by encoding the information data based on the selected parity check matrix. The encoding apparatus may further include a measuring unit measuring a channel link quality of a communication system in which the LDPC code is used, wherein the controller is configured to correct an error in a predetermined number of parity check matrices based on the channel link quality. The best parity check matrix may be selected. The measurement unit may estimate the channel link quality based on a stochastic characteristic of the channel link quality, which may include a signal-to-noise ratio (SNR). The controller may be further configured to determine that the number of different connection order sets of the check nodes is greater than the predetermined number of connection order sets based on the predetermined number of parity check matrices when the channel link quality is worse than a preset quality criterion. A check matrix can be selected, and if the channel link quality is superior to a predetermined quality criterion, the number of different connection order sets of the check nodes among the generated predetermined number of parity check matrices is greater than the preset connection order set number criteria. You can choose a smaller parity check matrix. The control unit may further include a parity check matrix having a variance value of the connectivity of the check nodes greater than a preset variance criterion of the predetermined number of parity check matrices when the channel link quality is worse than a preset quality criterion. When the channel link quality is better than the predetermined quality criterion, the parity check matrix of the generated number of parity check matrix is smaller than the dispersion value of the connectivity of the preset connection node of the predetermined number of parity check matrix Can be selected. The quality criterion is that the frame error rate curve of the LDPC code belongs to a waterfall region when the channel link quality is poorer than the quality criterion, and the frame error rate of the LDPC code when the channel link quality is superior to the quality criterion. The curve may be characterized as belonging to an error-floor region.

According to the method of generating a parity check matrix of an LDPC code according to an embodiment of the present invention, and a method and apparatus for encoding an LDPC code using the same, a connected order set means a set of orders of a variable node connected to a check node. A parity check matrix can be generated in accordance with the characteristics of a communication system using an LDPC code in consideration of the diversity and distribution of connectivity, meaning -degree set) and order sum. Further, in performing error correction encoding in a communication and compression system using an LDPC code, a parity check matrix may be selected according to the quality of a channel link through which information encoded with an LDPC code is transmitted.

Therefore, the fall / error-floor conflict can be adjusted in the error correction coding using the LDPC code. In other words, using a single LDPC code solves the problem of having relatively poor error correction ability as the channel link quality changes, and uses error correction coding that always has an optimal error correction capability regardless of the channel link quality. Can be done.

1 is an exemplary diagram of a parity check matrix of an LDPC code having a length of a codeword of 8 and a code rate of 1/2.

FIG. 2 is an exemplary diagram of a Tanner graph corresponding to the example of the parity check matrix of FIG. 1.

3 is an illustration of a circular ring in a Tanner graph of an LDPC code.

4 is an exemplary diagram of a Tanner graph of an LDPC code under design and an extended tree rooted at an arbitrary variable node corresponding to the Tanner graph.

5 is a frame error rate curve graph of an LDPC code having different downpour / error-floor conflict performance.

6 is an exemplary diagram of a set of connection orders of test nodes in a Tanner graph.

7 is an exemplary diagram of connectivity of test nodes in a Tanner graph.

8 is a connectivity distribution diagram of a first LDPC code having 129 different sets of connection orders.

9 is a connectivity distribution diagram of a second LDPC code having 2021 sets of different link orders.

10 is a connectivity distribution diagram of a third LDPC code having 2446 different order sets.

11 is a frame error rate curve graph of the first LDPC code, the second LDPC code, and the third LDPC code of FIGS. 8, 9, and 10.

12 is a frame error rate curve graph of a channel quality adaptive LDPC encoding / decoding method according to an embodiment of the present invention.

13 is a flowchart of a method of generating a parity check matrix of an LDPC code according to an embodiment of the present invention.

14 is a detailed flowchart of the condition determination step of FIG. 13.

15 is a flowchart of a method of encoding an LDPC code according to an embodiment of the present invention.

16 is a block diagram of an apparatus for encoding an LDPC code, according to an embodiment of the present invention.

As the present invention allows for various changes and numerous embodiments, particular embodiments will be illustrated in the drawings and described in detail in the written description.

However, this is not intended to limit the present invention to specific embodiments, it should be understood to include all modifications, equivalents, and substitutes included in the spirit and scope of the present invention.

Terms such as first and second may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as the second component, and similarly, the second component may also be referred to as the first component. The term and / or includes a combination of a plurality of related items or any item of a plurality of related items.

When a component is said to be "connected" or "connected" to another component, it may be directly connected to or connected to that other component, but it may be understood that another component may be present in the middle. Should be. On the other hand, when a component is said to be "directly connected" or "directly connected" to another component, it should be understood that there is no other component in between.

The terminology used herein is for the purpose of describing particular example embodiments only and is not intended to be limiting of the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this application, the terms "comprise" or "have" are intended to indicate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, and one or more other features. It is to be understood that the present disclosure does not exclude the possibility of the presence or the addition of numbers, steps, operations, components, components, or combinations thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art. Terms such as those defined in the commonly used dictionaries should be construed as having meanings consistent with the meanings in the context of the related art and shall not be construed in ideal or excessively formal meanings unless expressly defined in this application. Do not.

Hereinafter, with reference to the accompanying drawings, it will be described in detail a preferred embodiment of the present invention. In the following description of the present invention, the same reference numerals are used for the same elements in the drawings and redundant descriptions of the same elements will be omitted.

The design of the LDPC code is based on design factors. The length N and code rate R of the codeword are determined according to the characteristics of the communication system in which the LDPC code is to be used, and a degree distribution having excellent asymptotic performance at the determined code rate is selected. Based on these design elements, an important point in designing a practically usable LDPC code, that is, a finite length LDPC code, is to generate a short cycle of minimum length. A circular ring is a set of connecting lines that start from one node and then through other nodes to its own node.

FIG. 2 is an exemplary diagram of a Tanner graph corresponding to the example of the parity check matrix of FIG. 1, and FIG. 3 is an exemplary diagram of a circular ring in the Tanner graph of FIG. 2.

Referring to FIG. 3, a connection line connected from the variable node v1 301 to the variable node v1 301 through the test node c1 311, the variable node v5 305, and the test node c2 312 is again connected to a circular loop ( 320). The length of the purified ring is the number of connecting lines constituting the circulation ring, and the length of the circulation ring 320 of FIG. 3 is four. The more loops in the Tanner graph of the LDPC code, and the shorter the length of the loops, the greater the probability of occurrence of a phenomenon called error-floor. The LDPC code has an error correction capability that is exponentially improved as the quality of the channel link is improved. The error-floor phenomenon is a phenomenon in which the error correction capability is not significantly improved compared to the degree of improvement in the quality of the channel link. . Stopping sets and trapping sets, which are extended concepts of circular loops, also affect the error-floor behavior of LDPC codes.

In designing LDPC codes with finite codeword lengths, stop sets and trap sets, including circular loops, are necessarily generated. Therefore, it is important to reduce the size of the stop set and trap set including the circular loop in the designed LDPC code in designing the LDPC code based on the given design elements such as code length, code rate and order distribution.

4 is a diagram illustrating an example of progressive-edge-growth (PEG), which is one of methods for increasing the size of a circular ring in designing an LDPC code.

Referring to FIG. 4, the order of a variable node is determined according to a degree distribution given as a design element, and then a connection line corresponding to a predetermined order for each variable node is sequentially connected to the inspection node. Referring to the state of the Tanner graph 401 of FIG. 4, the variable nodes v1 411, v2 412, v3 413, and v4 414 connect connection lines of a predetermined order to the test node. It is a state. The variable node v5 415 connects one connection line to the inspection node c1 421, and it is time to determine which inspection node to connect the second connection line to. The goal of the design is to maximize the length of the loop that is created by placing new connections. Accordingly, the variable node v5 415 constructs a spreading tree for looking at the distance of each variable node and the check node from itself. Referring to the extension tree 402 of FIG. 4, the test node farthest from the variable node v5 415 is c3 423. Therefore, if the new connection line is connected between the variable nodes v5 415 and c3 423, the size of the circular ring formed by the new connection line is maximized. This process is inevitably performed in the design process of the LDPC code.

5 is a frame error rate curve graph of an LDPC code having different downpour / error-floor conflict performance. That is, a diagram showing a frame error rate (FER) curve graph indicating the performance of two LDPC codes having the same order distribution but different connection configurations.

Referring to the graph of FIG. 5, the horizontal axis 501 is a bit energy-to-noise ratio indicating the quality of the channel link. The larger the value, the better the quality of the channel link, and the smaller the value, the worse the quality of the channel link. In addition, the vertical axis 502 represents the average probability of restoration failure when the LDPC code is used in the quality of each channel link. The smaller the value, the better the performance of the LDPC code. The higher the quality of the channel link, the lower the average probability of restoration failure when the LDPC code is used. The worse the quality of the channel link, the higher the average probability of restoration failure when the LDPC code is used.

Referring to the graph of FIG. 5, the performance graph of the LDPC code is largely divided into a waterfall region 510 and an error-floor region 520. The downpour area 510 is an area where the slope of the error rate curve increases rapidly as the quality of the channel link improves in a range where the quality of the channel link is relatively poor. The error-floor area 520 is an area where the slope of the error rate curve is drastically lower than that of the downfall area in a range where the quality of the channel link is relatively good.

Referring to the graph of FIG. 5, even though different LDPC codes have the same length of the same codeword and the same order distribution, different LDPC codes have different error performances according to the concatenated configuration of the Tanner graph. In particular, when the performance in the downpour area 510 is excellent, the performance is poor in the error-floor area 520, and when the performance in the error-floor area 520 is excellent, the performance in the downpour area 510 is poor. A tradeoff between waterfall and error-floor is observed. That is, when the graph 531 representing the LDPC code 1 and the graph 532 representing the LDPC code 2 are observed, the falling water region 510 shows better performance when the LDPC code 2 is used, and the error-floor area 520 is used. In the case of using LDPC code 1, it shows better performance.

Drop / error-floor conflicts between finite length LDPC codes with different order distributions can be controlled by varying the ratio of order-2 variable nodes. However, the factors causing the fall / error-floor conflict between LDPC codes having the same order distribution have not yet been identified.

If it is possible to control the fall / error-floor conflict in the finite length LDPC code with the same order distribution, the LDPC code can be designed according to the characteristics of the communication system using the LDPC code and the quality of the channel link. In other words, the order distribution of LDPC codes is optimized through asymptotic performance analysis using density evolution under a given code rate of LDPC codes, and then suitable drop / error-according to the characteristics and requirements of the communication system and the quality of the channel link. Designing LDPC codes with floor conflict features can improve the overall performance of the system.

For example, if the quality of the channel link between the transmitter and the receiver is generally low, the average reliability of the link can be improved by using an LDPC code having better performance in the downfall region 510. On the other hand, if the channel link quality between the transmitter and the receiver is generally good, the average reliability of the link can be improved by using an LDPC code having excellent performance in the error-floor region 520.

Accordingly, the method of designing an LDPC code according to an embodiment of the present invention provides a method of designing an LDPC code when a length, a code rate, and an order distribution of a codeword are given. In addition, the LDPC code design method according to an embodiment of the present invention provides a design method of the LDPC code having a suitable performance according to the characteristics of the communication system even if the same codeword length, code rate, and order distribution.

Prior to describing an embodiment of the present invention, a connected-degree set and connectivity will be described below as a concept used by the present invention.

6 is an exemplary diagram of a set of connection orders of test nodes in a Tanner graph. That is, a diagram showing an example of a connected-degree set in an extension tree drawn with one test node as a root in an arbitrary Tanner graph.

Referring to FIG. 6, in the Tanner graph of an LDPC code, a connection order set of one test node is a set of orders of variable nodes connected to the test node and one connection line. For example, in FIG. 6, the number of variable nodes connected to the test node c1 601 as one connection line is v1 611, v2 612, v3 613, and v4 614. The order of this variable node is v1 611 is 2, v2 612 is 3, v3 613 is 3 and v4 614 is 4, respectively. Accordingly, the connection order set of the check node c1 601 is {2, 3, 3, 4}.

7 is an exemplary diagram of connectivity of test nodes in a Tanner graph. That is, an example of connectivity is shown in an extension tree drawn with one test node as a root in a Tanner graph as shown in FIG. 6.

Referring to FIG. 7, the connectivity of a test node in a Tanner graph of an LDPC code is a sum of subtracting 1 from orders of variable nodes connected to the test node and one connection line. For example, in FIG. 7, four variable nodes are connected to the test node c1 701 with one connection line, that is, v1 711, v2 712, v3 713, and v4 714. The order of this variable node is v1 711 is 2, v2 712 is 3, v3 713 is 3 and v4 714 is 4, respectively. Therefore, the connectivity of the check node c1 701 is 1 + 2 + 2 + 3 = 8 plus the values of (2-1), (3-1), (3-1) and (4-1). Also, connectivity is the number of test nodes 703 located at depth 2 when drawing an extension tree rooted at test node c1 701 under the premise that there are no circular rings of length 4.

In the LDPC code design method according to an embodiment of the present invention, in designing an LDPC code when a length, a code rate, and an order distribution of a codeword are given, the connection order set and connectivity of a check node, which have not been previously considered, are considered. . LDPC code according to an embodiment of the present invention, even if the same codeword length, the same code rate and the same or similar order distribution according to the fall / error-floor conflict according to the diversity of the connection order set of the test node and the variance of the connectivity This has an adjusted performance.

If the form of the connection order set of each test node constituting the Tanner graph of the LDPC code is uniform or the variance of the connectivity is small, the performance in the error-floor region is excellent, but the performance in the falling water region is poor. For example, i) the connection order set of check nodes constituting the Tanner graph of the LDPC code is {2, 3, 3, 4} and the connectivity is all equal to 12, or ii) the connection order of some check nodes. The set is {2, 3, 3, 4}, the connectivity is 12, and the connection order set of the test nodes other than some of the check nodes is {2, 2, 3, 4}, and the connectivity is 11 per check node. The connection order set and connectivity are somewhat different, but generally similar, they have good performance in the error-floor domain.

On the other hand, if the form of the connection order set of each check node constituting the Tanner graph of the LDPC code is varied or the dispersion of connectivity is large, the performance in the error-floor region is poor, but the performance is excellent in the downpour region. For example, among the check nodes constituting the Tanner graph of the LDPC code, a check order set of some check nodes is {2, 2, 2, 2}, and has a connectivity of 8, and other check nodes other than the check nodes. If the connection order set of these is {4, 4, 4, 4}, and the connectivity is 16, if the connection order set is varied for each test node constituting the Tanner graph and the distribution of connectivity is large, it has excellent performance in the falling water region. .

FIG. 8 is a connectivity distribution diagram of a first LDPC code having 129 different connection order sets, FIG. 9 is a connectivity distribution diagram of a second LDPC code having 2021 different connection order sets, and FIG. 10 is a 2446 different connection order set. A connectivity distribution diagram of individual third LDPC codes.

A method of designing an LDPC code according to an embodiment of the present invention will be described with reference to the first to third LDPC codes shown in FIGS. 8 to 10. The first to third LDPC codes have the same length of the codeword, the same code rate, the same variable node order distribution, and the same check node order distribution. On the other hand, the connection order set and the connectivity distribution of the check nodes of the first to third LDPC codes are different from each other. Table 1 below shows the number of different connection order sets appearing in the first to third LDPC codes.

Table 1

8 to 10, the connectivity is widely spread in the order of the first LDPC code, the second LDPC code, and the third LDPC code.

11 is a diagram illustrating a frame error rate curve of each of the first to third LDPC codes.

Referring to FIG. 11, the third LDPC code has the best performance in the falling water region, and the second LDPC code shows the best performance, and the first LDPC code shows the poor performance. That is, the LDPC code having various connection order sets of check nodes, that is, a large number of different connection order sets, has a better performance in a falling water region. On the other hand, in the error-floor region, the first LDPC code has the best performance, the second LDPC code has the best performance, and the third LDPC code has the worst performance. That is, the LDPC code having a relatively uniform connection order set of the check node, that is, a small number of different connection order sets, has better performance in the error-floor region. As described above, in designing the LDPC code, a fall / error-floor conflict is adjusted in consideration of the connection order set and the connectivity of the test node.

12 is a frame error rate curve graph of a channel quality adaptive LDPC encoding / decoding method according to an embodiment of the present invention. That is, it is a graph illustrating a frame error rate curve of a communication device that selects an LDPC code having an optimal fall / error-floor conflict performance based on a stochastic characteristic on channel link quality.

The encoder of the LDPC code according to an embodiment of the present invention selects and encodes an LDPC code having excellent fall / error-floor conflict performance in the channel link quality in a situation where the quality of the channel link is known. In many communication systems, it is difficult for encoders and decoders to accurately grasp information about instantaneous channel links, but the signal-to-noise ratio (SNR), etc., makes it easier for the stochastic nature of channel link quality. I can figure it out. Therefore, the LDPC coder according to the present invention has the same or similar codeword length, code rate, and order distribution, and has a different number of connected order set distributions and connectivity distributions, thereby showing different drop / error-floor conflict performance. An LDPC code to be used for encoding is selected within the LDPC code group based on a stochastic characteristic on the quality of channel link using the group.

As shown in FIG. 12, an encoder of a communication device using an LDPC code according to an embodiment of the present invention has three LDPC code groups having different falling / error-floor conflict performances. The LDPC code group may include first to third LDPC codes of FIGS. 8 to 10. When an encoder of a communication system using an LDPC code knows the stochastic characteristics of channel link quality, selecting and encoding an LDPC code having an optimal fall / error-floor conflict performance, error correction such as a frame error rate curve 1201 Have the ability. As a result, overall performance improvement can be expected in data transmission throughput and final error rate of the communication system. That is, by adaptively selecting the LDPC code without being affected by the change in channel link quality, it always has an optimal error correction capability.

패리티 검사 행렬 생성 방법Parity Check Matrix Generation Method

13 is a flowchart of a parity check matrix generation method of an LDPC code according to an embodiment of the present invention, and FIG. 14 is a detailed flowchart of a condition determination step of FIG. 13.

As shown in FIG. 13, a method of generating a parity check matrix of a low density parity check (LDPC) code according to an embodiment of the present invention is based on a channel link quality of a communication system using the LDPC code. A condition setting step (S1310) of determining at least one of a connected-degree set condition and a connectivity condition of the check nodes based on the parity based on at least one of the determined connection order set condition and the connectivity condition A check matrix may be generated (S1320).

That is, using the characteristics of the communication system in which the LDPC code is to be used and the average channel link quality of the communication system, connection of check nodes to generate a parity check matrix of the LDPC code having the best error correction capability in the communication system. Use order aggregation conditions, or connectivity conditions.

As shown in FIG. 14, the condition setting step (S1310) includes determining a communication system to which the LDPC code is applied (S1410), and the average quality of the channel link of the determined communication system is higher than a predetermined quality criterion. Determining whether it is excellent (S1420), if the quality of the channel link is superior to a predetermined quality standard, as a result of the determination, minimizing the number of different connection order sets by uniformizing the connection order sets (S1430), Minimizing the variance of connectivity when the quality of the channel link is better than the preset quality criterion (S1440); when the quality of the channel link is worse than the predetermined quality criterion, varying the connection order set to diversify the different connection order Maximizing the number of sets (S1450); and maximizing the connectivity variance value if the quality of the channel links is worse than the preset quality criteria. It may include (S1450).

In more detail, in the determining of the communication system to which the LDPC code is applied (S1410), the communication system is determined and the average channel link quality of the determined communication system is determined. For example, it is determined whether the quality of the channel link between the transmitter and the receiver included in the communication system is generally low, or whether the quality of the channel link between the transmitter and the receiver is generally good.

The average quality of the determined channel link is compared with a predetermined quality criterion, and a connection order set or condition of connectivity is determined based on the comparison result. As described above with reference to FIG. 5, the error rate curve of the LDPC code is divided into the downfall area 510 and the error-floor area 520 by the quality of the channel link, and the dispersion of the connectivity of the test nodes is large or the connection order set is high. These various cases tend to have good error correction capability in the downfall region 510, and tend to have good error correction capability in the error-floor region 520 when the distribution of connectivity of check nodes is small or the set of connection orders is generally similar. There is this.

Referring to FIG. 5 in more detail, the channel link quality may be determined based on a bit energy-to-noise ratio (Eb / No), and is represented on the horizontal axis 501 of the graph of FIG. 5. In addition, when the channel link quality is worse than the quality criterion 540, the preset quality criterion 540 includes a frame error rate curve of the LDPC code belonging to a waterfall region 510, and the quality criterion 540. When the channel link quality is higher than), the frame error rate curve of the LDPC code may belong to an error-floor region 520.

The quality criterion 540 is compared with the quality of the channel link to determine whether the quality of the channel link is superior to the quality criterion 540 (S1420). The connection order set condition is that if the channel link quality is worse than a preset quality criterion 540, the number of connection order sets of the test nodes is set to be larger than a preset connection order set criterion. If the channel link quality is superior to a preset quality criterion 540, the number of different connection order sets of the check nodes is less than the preset number of different connection order sets. It is possible to set the connection order set to be uniform (S1430). The preset different sets of connection order sets may be used to find and use different sets of connection order sets having similar error correction capability in the falling water region or the error-floor region through simulation.

In addition, the connectivity condition may be set such that when the channel link quality is worse than a preset quality criterion 540, the variance value of the connectivity of the test nodes is greater than a preset connectivity variance criterion (S1460). If the link quality is superior to the preset quality criterion 540, the distribution value of the connectivity of the check nodes may be set to be smaller than the preset connectivity dispersion value criterion (S1440). The predetermined connectivity variance criterion may be used to find and use different connection order sets having similar error correction capability in the falling water region or the error-floor region through simulation.

Accordingly, in consideration of the characteristics of the communication system using the LDPC code and the tendency of the quality of the channel link, it is possible to generate a parity check matrix of the LDPC code having an optimal error correction capability.

LDPC 부호의 부호화 방법LDPC code encoding method

In certain communication systems, the quality of the channel link can vary depending on the situation, so that using the parity check matrix of the first LDPC code has excellent error correction capability in the error-floor region when the quality of the channel link is good. In case of poor link quality, it has poor error correction capability in the falling down region, and using the parity check matrix of the second LDPC code has poor error correction capability in the error-floor region when the quality of the channel link is excellent. A situation with excellent error correction capability may occur in the downpour area when the channel link quality is poor. Accordingly, there is a need for an LDPC code encoding method that selects a parity check matrix of an LDPC code adaptively to the quality of a channel link and always has an optimal error correction capability.

As shown in FIG. 15, a method of encoding a low density parity check (LDPC) code according to an embodiment of the present invention includes inputting information data (S1501) and connecting order sets of check nodes. generating a parity check matrix of a predetermined number of LDPC codes different from at least one of a degree set condition and a connectivity condition (S1502), and selecting one parity check matrix from the generated number of parity check matrices And a step S1505 of generating a codeword by encoding the information data based on the selected parity check matrix. The step of selecting the parity check matrix (S1510) may include measuring channel link quality of a communication system in which the LDPC code is used (S1503), and generating the predetermined number of parity check matrices based on the channel link quality. The method may include selecting one parity check matrix having the best error correction capability (S1504).

Referring to FIGS. 11 and 15 again, a method of encoding a low density parity check code according to an embodiment of the present invention will be described in more detail.

Information data including information to be transmitted through a channel link of the communication system is input (S1501), and a predetermined number of parity check matrices are generated to encode the information data (S1502). As described above, the parity check matrix of the LDPC code may generate a predetermined number of parity check matrices having different connection order aggregation conditions or connectivity conditions for adaptively selecting the quality of the channel link.

For example, a parity check matrix of three LDPC codes is generated, and as described through Table 1, the parity check matrix of the first LDPC code has the number of 129 different connection order sets, and the parity of the second LDPC code. The parity check matrix may have the number of 2021 different connection order sets, and the parity check matrix of the third LDPC code may have the number of 2445 different connection order sets. As shown in FIGS. 8 to 10, the connectivity variance of the parity check matrices of the LDPC code is the smallest when the parity check matrix of the first LDPC code is used, and the parity check matrix of the second LDPC code and the third LDPC code. The parity check matrix grows in order.

As a selection criterion of the parity check matrix, the channel link quality of the communication system in which the LDPC code is used is measured (S1503). In a typical communication system, it is difficult for an encoder to grasp information about instantaneous channel links, but the stochastic characteristics of channel link quality such as signal-to-noise (SNR) can be easily grasped. Accordingly, measuring the channel link quality (S1503) may include estimating the channel link quality based on a stochastic characteristic of the channel link quality, which may include a signal-to-noise ratio (SNR). It can be characterized.

Based on the measured or estimated channel link quality, one of the generated parity check matrices having the best error correction capability is selected (S1504). As shown in FIG. 11, a frame error rate curve can be obtained for encoding using a parity matrix of each LDPC code through simulation. Accordingly, the parity check matrix having the best error correction capability may be selected with reference to the frame error rate curve based on the measured or estimated channel link quality. For example, referring to FIG. 11, when the quality of the measured or estimated channel link is 1.0, the smallest frame error rate is shown when the parity check matrix of the second LDPC code is used (1120). The parity check matrix can be selected.

Here, the selection criterion of the parity check matrix may use any criterion including not only the quality of the channel link described above but also the characteristics, state and type of the channel link.

In some cases, a quality criterion may be set in advance, and one of the generated predetermined number of parity check matrices may be selected based on a result of comparing the quality criterion with the quality of the measured or estimated channel link.

That is, when the measured or estimated channel link quality is inferior to a predetermined quality criterion, the number of different connection order sets of the check nodes among the generated predetermined number of parity check matrices is larger than the preset connection order set number criterion. A parity check matrix is selected, and if the measured or estimated channel link quality is superior to a preset quality criterion, the number of sets of different connection order sets of the check nodes among the generated predetermined number of parity check matrices is preset A parity check matrix may be selected that is smaller than the set number criterion.

Further, when the measured or estimated channel link quality is inferior to a preset quality criterion, the parity check of the generated variance of the connectivity of the test nodes is greater than the predetermined variance criterion of the connectivity among the predetermined number of parity check matrices. When a matrix is selected and the measured or estimated channel link quality is superior to a preset quality criterion, the variance value of the connectivity of the check nodes among the generated predetermined number of parity check matrices is greater than the preset variance criterion of the connectivity. You can select a small parity check matrix.

The quality criterion of the LDPC code is that the frame error rate curve of the LDPC code belongs to a waterfall region when the channel link quality is poorer than the quality criterion, and the channel link quality is superior to the quality criterion. The frame error rate curve may be characterized as belonging to an error-floor region.

For example, referring to FIG. 11, when the quality of the channel link is 1.05, the quality criterion 1140 may be set, and the number of different connection order sets may be set to 1000. Therefore, when the quality of the channel link exceeds 1.05, the parity check matrix of the first LDPC code having the number of 129 different connection order sets less than 1000, which is based on the preset number of connection order sets, is selected, and the quality of the channel link is selected. If less than 1.05, it is possible to select a parity check matrix of the third LDPC code having the number of 2446 different connection order sets greater than 1000, which is a preset number of connection order sets.

In consideration of the parity check matrix of the second LDPC code, more specific criteria may be additionally set. A plurality of quality standards may be set like the first quality standard and the second quality standard, and the interval may be set using the plurality of quality criteria to select a parity check matrix corresponding to each quality interval.

A codeword may be generated by encoding the information data based on the selected parity check matrix (S1505). The codeword generation method may use a conventional LDPC code encoding method.

According to the LDPC code encoding method described above, as shown in FIG. 12, a frame error rate curve 1201 having an optimal error correction capability can be obtained. That is, by selecting the parity check matrix adaptively to the quality of the channel link, it is possible to solve the fall / error-floor conflict and obtain the encoding result of the LDPC code having the optimal error correction capability.

LDPC 부호의 부호화 장치LDPC code encoding device

As shown in FIG. 16, the LDPC code encoding apparatus according to an embodiment of the present invention includes an input unit 1610 for inputting information data, a connected-degree set condition of test nodes, or connectivity. Parity check matrix generator 1620 for generating a parity check matrix of a predetermined number of LDPC codes having different conditions; a measurer 1630 for measuring channel link quality of a communication system in which the LDPC code is used; A controller 1640 for selecting one parity check matrix having the best error correction capability among the generated predetermined number of parity check matrices based on quality, and encoding the information data based on the selected parity check matrix. The encoder 1650 to generate and the communicator 1660 to transmit the generated codeword to the decoding apparatus through a channel link.

The measurement unit 1630 may estimate the channel link quality based on a stochastic characteristic of the channel link quality, which may include a signal-to-noise ratio (SNR).

If the channel link quality is worse than a preset quality criterion, the controller 1640 may have a larger number of sets of different connection order sets of the check nodes among the generated predetermined number of parity check matrices. If a parity check matrix is selected and the channel link quality is superior to a preset quality criterion, the number of different connection order sets of the check nodes among the generated predetermined number of parity check matrices is greater than the preset connection order set number criterion. You can choose a smaller parity check matrix.

In addition, when the channel link quality is poorer than a preset quality criterion, the controller 1640 may have a variance value of the connectivity of the check nodes greater than a predetermined variance criterion of the connectivity among the predetermined number of parity check matrices. A parity check matrix is selected, and if the channel link quality is superior to a preset quality criterion, the variance value of the connectivity of the check nodes among the generated predetermined number of parity check matrices is smaller than the preset variance criterion of connectivity. You can select a check matrix.

The quality criterion is that the frame error rate curve of the LDPC code belongs to a waterfall region when the channel link quality is poorer than the quality criterion, and the frame error rate of the LDPC code when the channel link quality is superior to the quality criterion. The curve may be characterized as belonging to an error-floor region.

A detailed operation method of the encoding apparatus of the LDPC code according to the embodiment of the present invention is the same as the description of the encoding method of the LDPC code according to the embodiment of the present invention described above.

(Explanation of the sign)

510: downfall area 520: error floor area

531 error rate curve of the first LDPC code 532 error rate curve of the second LDPC code

1600: LDPC code encoding device 1610: input unit

1620: parity check matrix generation unit 1630: measurement unit

1640: controller 1650: encoder

1660: communication

Claims

In the method for generating a parity check matrix of the Low Density Parity Check (LDPC) code,

A condition setting step of determining at least one of a connected-degree set condition and a connectivity condition of test nodes based on channel link quality of a communication system in which the LDPC code is used; And

And generating a parity check matrix based on at least one of the determined link order aggregation condition and the connectivity condition.
The method of claim 1, wherein the connection order aggregation condition is

If the channel link quality is worse than a predetermined quality criterion, the channel link quality adaptive parity check matrix is set such that the number of different connection order sets of the check nodes is greater than a preset connection order set number criterion. How to produce.
The method of claim 1, wherein the connection order aggregation condition is

If the channel link quality is higher than a predetermined quality criterion, the channel link quality adaptive parity check matrix is set such that the number of different connection order sets of the check nodes is smaller than a preset connection order set number criterion. How to produce.
The method of claim 1, wherein the connectivity condition is

And if the channel link quality is worse than a preset quality criterion, setting a distribution value of the connectivity of the check nodes to be greater than a preset connectivity dispersion value criterion.
The method of claim 1, wherein the connectivity condition is

And if the channel link quality is better than a preset quality criterion, setting the variance of the connectivity of the check nodes to be smaller than a preset connectivity variance criterion.
The method of claim 1, wherein the channel link quality is

A method for generating a channel link quality adaptive parity check matrix, characterized in that it is determined based on a bit energy to noise ratio (Eb / No).
The method of claim 2, wherein the quality criterion is

If the channel link quality is poorer than the quality criterion, the frame error rate curve of the LDPC code belongs to a waterfall region, and if the channel link quality is better than the quality criterion, the frame error rate curve of the LDPC code is error-. A channel link quality adaptive parity check matrix generation method characterized by belonging to an error-floor region.
In the encoding method of the low density parity check (LDPC, Low Density Parity Check) code,

Receiving information data;

Generating a parity check matrix of a predetermined number of LDPC codes in which at least one of a connected-degree set condition and a connectivity condition of the check nodes are different from each other;

Selecting one parity check matrix from the generated number of parity check matrices; And

And encoding the information data based on the selected parity check matrix to generate a code word.
9. The method of claim 8, wherein selecting the parity check matrix

Measuring channel link quality of a communication system in which the LDPC code is used; And

Selecting one of the generated parity check matrices having the best error correction capability based on the channel link quality, and encoding the channel link quality adaptive low density parity check code. Way.
10. The method of claim 9, wherein measuring channel link quality comprises:

Coding of channel link quality adaptive low density parity check code, wherein the channel link quality is estimated based on a stochastic characteristic of channel link quality, which may include a signal-to-noise ratio (SNR). Way.
10. The method of claim 9, wherein selecting the parity check matrix

When the channel link quality is worse than a preset quality criterion, selecting a parity check matrix having a larger number of different connection order sets of the check nodes than a predetermined number of connection order sets based on the generated number of parity check matrices. And a channel link quality adaptive low density parity check code encoding method.
10. The method of claim 9, wherein selecting the parity check matrix

When the channel link quality is superior to a preset quality criterion, selecting a parity check matrix of the generated predetermined number of parity check matrices, wherein the number of different connection order sets of the check nodes is smaller than a preset connection order set number criterion; And a channel link quality adaptive low density parity check code encoding method.
10. The method of claim 9, wherein selecting the parity check matrix

When the channel link quality is poorer than a preset quality criterion, selecting a parity check matrix having a variance value of the connectivity of the check nodes greater than a predetermined variance criterion of the connectivity among the generated predetermined number of parity check matrices. A method of encoding a channel link quality adaptive low density parity check code.
10. The method of claim 9, wherein selecting the parity check matrix

When the channel link quality is superior to a preset quality criterion, a parity check matrix is selected from among the generated predetermined number of parity check matrices, the variance value of the connectivity of the test nodes being smaller than the preset variance criterion of the connectivity. A method of encoding a channel link quality adaptive low density parity check code.
The method of claim 11, wherein the quality criterion is

If the channel link quality is poorer than the quality criterion, the frame error rate curve of the LDPC code belongs to a waterfall region, and if the channel link quality is better than the quality criterion, the frame error rate curve of the LDPC code is error-. A method of encoding a channel link quality adaptive low density parity check code, characterized by belonging to an error-floor region.
In the encoding device of the Low Density Parity Check (LDPC) code,

An input unit for receiving information data;

A parity check matrix generator for generating a parity check matrix of a predetermined number of LDPC codes having at least one of a connected-degree set condition and a connectivity condition of the check nodes;

A control unit for selecting one parity check matrix from the generated number of parity check matrices; And

And an encoder for generating a codeword by encoding the information data based on the selected parity check matrix.
The method of claim 16,

Further comprising a measuring unit for measuring the channel link quality of the communication system in which the LDPC code is used,

The controller selects one parity check matrix having the best error correction capability among the generated predetermined number of parity check matrices based on the channel link quality, wherein the control unit encodes the channel link quality adaptive low density parity check code. .
The method of claim 17, wherein the measuring unit

Coding of channel link quality adaptive low density parity check code, wherein the channel link quality is estimated based on a stochastic characteristic of channel link quality, which may include a signal-to-noise ratio (SNR). Device.
The method of claim 17, wherein the control unit

When the channel link quality is worse than a preset quality criterion, selecting a parity check matrix having a larger number of different connection order sets of the check nodes than a predetermined number of connection order sets based on the generated number of parity check matrices. And a channel link quality adaptive low density parity check code encoding apparatus.
The method of claim 17, wherein the control unit

When the channel link quality is superior to a preset quality criterion, selecting a parity check matrix of the generated predetermined number of parity check matrices, wherein the number of different connection order sets of the check nodes is smaller than a preset connection order set number criterion; And a channel link quality adaptive low density parity check code encoding apparatus.
The method of claim 17, wherein the control unit

When the channel link quality is poorer than a preset quality criterion, selecting a parity check matrix having a variance value of the connectivity of the check nodes greater than a predetermined variance criterion of the connectivity among the generated predetermined number of parity check matrices. An apparatus for encoding a channel link quality adaptive low density parity check code.
The method of claim 17, wherein the control unit

When the channel link quality is superior to a preset quality criterion, a parity check matrix is selected from among the generated predetermined number of parity check matrices, the variance value of the connectivity of the test nodes being smaller than the preset variance criterion of the connectivity. An apparatus for encoding a channel link quality adaptive low density parity check code.
23. The method of claim 19, wherein the quality criterion is

If the channel link quality is poorer than the quality criterion, the frame error rate curve of the LDPC code belongs to a waterfall region, and if the channel link quality is better than the quality criterion, the frame error rate curve of the LDPC code is error-. An apparatus for encoding a channel link quality adaptive low density parity check code, which belongs to an error-floor region.