Generalized low-density parity check codes (GLDPC)

Patent metadata

Field	Value
Publication number	US-11075651-B2
Application number	US-202016741199-A
Country	US
Kind code	B2
Filing date	Jan 13, 2020
Priority date	Jul 13, 2017
Publication date	Jul 27, 2021
Grant date	Jul 27, 2021

Abstract

Provided is a system and method for determining a generalized Low-Density Parity-Check (LDPC) code for forward error correction channel coding that has a repeat-accumulate code structure.

First claim

What is claimed is:

1. A system for determining a generalized Low-Density Parity-Check (LDPC) code for forward error correction channel coding, the system including a processor being configured to:
determine 2k parity check matrix columns of size k, wherein k is an integer;
label components of a first parity check matrix with n columns and k rows, wherein n is an integer, and wherein the first parity check matrix includes the 2k parity check matrix columns of size k; and
select Cordaro-Wagner component code check matrices, wherein each of the selected Cordaro-Wagner component code check matrices has two rows which replace one row of the first parity check matrix to derive a second parity check matrix defining the generalized LDPC code;
wherein the determination of the 2k parity check matrix columns of size k and the selection of the Cordaro-Wagner component code check matrices are constrained to 2k parity check matrix columns of size k and Cordaro-Wagner component code check matrices which allow that rows and columns of a parity part consisting of 2k columns of the second parity check matrix which correspond to the 2k parity check matrix columns of size k, can be brought in an order in which the ordered rows and columns form a parity part which has a repeat-accumulate code structure.

2. The system of claim 1 , the processor being configured to:
split/duplicate each entry of the 2k parity check matrix columns of size k into a vector of size two, wherein each vector of size two having a non-zero weight requires a corresponding non-zero entry, to determine the 2k columns of the second parity check matrix which correspond to the 2k parity check matrix columns of size k, and can be brought in an order in which the ordered rows and columns form a parity part which has a repeat-accumulate code structure.

3. The system of claim 2 , the processor being configured to:
iteratively label components of n−k unlabeled columns of the first parity check matrix based on a performance measure.

4. The system of claim 1 , the processor being configured to:
compare multiple alternatives for labelling different components of the n−k columns with non-zero entries; and
select one alternative achieving a highest performance score.

5. The system of claim 1 , wherein a column of the Cordaro-Wagner component code check matrix has a zero weight if a corresponding component of the row of the first parity check matrix is zero.

6. The system of claim 1 , wherein each of the 2k parity check matrix columns of size k has weight one or two.

7. The system of claim 6 , wherein k−1 parity check matrix columns of size k of the 2k parity check matrix columns of size k have a weight of one and the remaining parity check matrix columns of size k of the 2k parity check matrix columns of size k have weight two.

8. The system of claim 7 , wherein the 2k parity check matrix columns of size k are linearly independent.

9. The system of claim 1 , wherein the selection of the Cordaro-Wagner component code check matrices includes replacing each non-zero entry in a row of the first parity check matrix with a non-zero column of a Cordaro-Wagner component code check matrix, wherein:
a row having exactly three non-zero entries in components which correspond to the 2k parity check matrix columns, is replaced with a Cordaro-Wagner component code check matrix having columns which correspond to the 2k parity check matrix columns, wherein said columns are linearly independent; and
a row having exactly four non-zero entries in components which correspond to the 2k parity check matrix columns is replaced with a Cordaro-Wagner component code check matrix having columns which correspond to the 2k parity check matrix columns, wherein three of said columns are linearly independent.

10. A method of determining a generalized Low-Density Parity-Check (LDPC) code for forward error correction channel coding, the method comprising:
determining 2k parity check matrix columns of size k, wherein k is an integer;
labeling components of a first parity check matrix with n columns and k rows, wherein n is an integer, and wherein the first parity check matrix includes the 2k parity check matrix columns of size k; and
selecting Cordaro-Wagner component code check matrices, wherein each of the selected Cordaro-Wagner component code check matrices has two rows which replace a row of the first parity check matrix to derive a second parity check matrix defining the generalized LDPC code;
wherein the determining of the 2k parity check matrix columns of size k and the selecting of the Cordaro-Wagner component code check matrices are constrained to 2k parity check matrix columns of size k and Cordaro-Wagner component code check matrices which allow that rows and columns of a parity part consisting of 2k columns of the second parity check matrix which correspond to the 2k parity check matrix columns of size k can be brought in an order in which the ordered rows and columns form a parity part which has a repeat-accumulate code structure.

11. The method of claim 10 , the method comprising:
splitting/duplicating each entry of the 2k parity check matrix columns of size k into a vector of size two, wherein each vector of size two having a non-zero weight requires a corresponding non-zero entry, to determine the 2k columns of the second parity check matrix which correspond to the 2k parity check matrix columns of size k, and can be brought in an order in which the ordered rows and columns form a parity part which has a repeat-accumulate code structure.

12. The method of claim 11 , comprising:
iteratively labeling components of n−k unlabeled columns of the first parity check matrix based on a performance measure.

13. The method of claim 10 , comprising:
comparing multiple alternatives for labelling different components of the n−k columns with non-zero entries; and
selecting one alternative achieving a highest performance score.

14. The method of claim 10 , wherein a column of the Cordaro-Wagner component code check matrix has zero weight if at least one of the following is satisfied:
a) a corresponding component of the row of the first parity check matrix is zero;
b) each of the 2k parity check matrix columns of size k has weight one or two; and
c) the 2k parity check matrix columns of size k are linearly independent.

15. The method of claim 10 , wherein selecting the Cordaro-Wagner component code check matrices includes replacing each non-zero entry in a row of the first parity check matrix with a non-zero column of a Cordaro-Wagner component code check matrix, wherein:
a row having exactly three non-zero entries in components which correspond to the 2k parity check matrix columns, is replaced with a Cordaro-Wagner component code check matrix having columns which correspond to the 2k parity check matrix columns, wherein said columns are linearly independent; and
a row having exactly four non-zero entries in components which correspond to the 2k parity check matrix columns is replaced with a Cordaro-Wagner component code check matrix having columns which correspond to the 2k parity check matrix columns, wherein three of said columns are linearly independent.

Assignees

• Huawei Tech Co Ltd

Inventors

• Usatyuk Vasily Stanislavovich
• Polianskii Nikita Andreevich
• Vorobyev Ilya Viktorovich
• Gaev Vladimir Anatolyevich
• Svistunov German Viktorovich
• Kamenev Mikhail Sergeevich
• Kameneva Yulia Borisovna

Classifications

• H03M13/1174 Primary

  Parity-check or generator matrices built from sub-matrices representing known block codes such as, e.g. Hamming codes, e.g. generalized LDPC codes · CPC title

• H03M13/036 Primary

  Heuristic code construction methods, i.e. code construction or code search based on using trial-and-error · CPC title

• H03M13/616

  Matrix operations, especially for generator matrices or check matrices, e.g. column or row permutations · CPC title

• H03M13/1182

  wherein the structure of the parity-check matrix is obtained by reordering of a random parity-check matrix · CPC title