CN115314156A - LDPC coding and decoding method and system based on self-coding network - Google Patents
LDPC coding and decoding method and system based on self-coding network Download PDFInfo
- Publication number
- CN115314156A CN115314156A CN202210837384.7A CN202210837384A CN115314156A CN 115314156 A CN115314156 A CN 115314156A CN 202210837384 A CN202210837384 A CN 202210837384A CN 115314156 A CN115314156 A CN 115314156A
- Authority
- CN
- China
- Prior art keywords
- check matrix
- sparse check
- reconstructed
- matrix
- original
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
- H04L1/0056—Systems characterized by the type of code used
- H04L1/0061—Error detection codes
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M13/00—Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
- H03M13/03—Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words
- H03M13/05—Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words using block codes, i.e. a predetermined number of check bits joined to a predetermined number of information bits
- H03M13/11—Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words using block codes, i.e. a predetermined number of check bits joined to a predetermined number of information bits using multiple parity bits
- H03M13/1102—Codes on graphs and decoding on graphs, e.g. low-density parity check [LDPC] codes
- H03M13/1105—Decoding
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M13/00—Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
- H03M13/03—Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words
- H03M13/05—Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words using block codes, i.e. a predetermined number of check bits joined to a predetermined number of information bits
- H03M13/11—Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words using block codes, i.e. a predetermined number of check bits joined to a predetermined number of information bits using multiple parity bits
- H03M13/1102—Codes on graphs and decoding on graphs, e.g. low-density parity check [LDPC] codes
- H03M13/1148—Structural properties of the code parity-check or generator matrix
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
- H04L69/04—Protocols for data compression, e.g. ROHC
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Probability & Statistics with Applications (AREA)
- Theoretical Computer Science (AREA)
- Computer Security & Cryptography (AREA)
- Mathematical Physics (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
The invention discloses an LDPC coding and decoding method and system based on a self-coding network, relating to the field of mobile communication. The method comprises the following steps: constructing an original sparse check matrix; compressing the original sparse check matrix by using a self-coding network, and reconstructing data of the compressed original sparse check matrix to obtain a reconstructed sparse check matrix; acquiring abnormal positions with differences according to a difference matrix of the acquired original sparse check matrix and the reconstructed sparse check matrix; calculating a correlation coefficient of the original sparse check matrix and the reconstructed sparse check matrix to determine a sparse check matrix; and compressing the determined sparse check matrix, and transmitting and decoding the compressed sparse check matrix and the signal to be transmitted. The invention combines the self-coding and the LDPC coding to code and decode the signal to be transmitted, thereby improving the efficiency of information transmission in the process of channel coding transmission.
Description
Technical Field
The application relates to the field of mobile communication, in particular to an LDPC coding and decoding method and system based on a self-coding network.
Background
Channel coding is a product of high-speed development of digital communication, and the main function is to perform error detection and correction processing on a transmitted signal, so that information loss in the signal transmission process is avoided. The occurrence of the error correcting code improves the transmission efficiency of signals, and error detection and correction of transmission signals are realized by adding a supervision code element on an information code element.
In the process of LDPC coding, as the code length is too long, the required check matrix is large, transmission occupies large channel capacity, and the longer the code length of the LDPC coding is, the better the performance is.
Aiming at the technical problem, the invention provides an LDPC coding and decoding method and system based on a self-coding network, which combines the self-coding network with LDPC coding.
Disclosure of Invention
Aiming at the technical problem, the invention provides an LDPC coding and decoding method and system based on a self-coding network.
In a first aspect, an embodiment of the present invention provides an LDPC coding and decoding method based on a self-coding network, including:
s1: constructing an original sparse check matrix in the LDPC coding process by using a finite geometry construction method;
s2: determining a reconstructed sparse check matrix, comprising:
s2.1: compressing the original sparse check matrix by using a self-coding network to obtain compressed original sparse check matrix data, and reconstructing the compressed original sparse check matrix data to obtain a reconstructed sparse check matrix;
s2.2: respectively extracting the column weight and the row weight of the original sparse check matrix and the reconstructed sparse check matrix, and judging whether the compression reconstruction process is qualified or not according to the column weight and the row weight of the original sparse check matrix and the reconstructed sparse check matrix;
s2.3: if the compression reconstruction process is unqualified, adjusting the self-coding network weight, and repeating S2.1-S2.2 by using the adjusted self-coding network weight until the compression reconstruction process is qualified;
s2.4: taking the reconstructed sparse check matrix obtained in the qualified compression reconstruction process as a reconstructed sparse check matrix;
s3: determining a final sparse check matrix, comprising:
s3.1: acquiring a difference matrix according to the original sparse check matrix and the reconstructed sparse check matrix, and acquiring an abnormal position of the difference between the original sparse check matrix and the reconstructed sparse check matrix according to the difference matrix;
s3.2: calculating correlation coefficients of the original sparse check matrix and the reconstructed sparse check matrix according to the neighborhood space of the abnormal positions of the original sparse check matrix and the reconstructed sparse check matrix, and judging whether the reconstructed sparse check matrix meets the requirements or not according to the correlation coefficients;
s3.3: if the reconstructed sparse check matrix does not meet the requirement, adjusting the weight of the self-coding network, and repeating S2.1-S3.3 by using the adjusted weight of the self-coding network until the reconstructed sparse check matrix meets the requirement;
s3.4: taking the reconstructed sparse check matrix which meets the requirements as a final sparse check matrix;
s4: and compressing the determined sparse check matrix by using a self-encoding network to obtain the LDPC code element of the signal to be transmitted, transmitting the compressed sparse check matrix and the LDPC code element of the signal to be transmitted, and decoding the LDPC code element of the signal to be transmitted according to the sparse check matrix.
The method for judging whether the reconstructed sparse check matrix meets the requirements according to the correlation coefficient comprises the following steps:
the method for judging whether the reconstructed sparse check matrix is effective or not by setting the similarity threshold comprises the following steps:
wherein: rho is a correlation coefficient of the original sparse check matrix and the reconstructed sparse check matrix, and alpha is a similarity threshold.
The calculation process of the correlation coefficient of the original sparse check matrix and the reconstructed sparse check matrix is as follows:
acquiring abnormal positions of the original sparse check matrix and the reconstructed sparse check matrix, calculating the correlation of each abnormal position according to elements in a neighborhood space matrix of each abnormal position with the difference, and taking the sum of the correlations of all the abnormal positions with the difference in the original sparse check matrix and the reconstructed sparse check matrix as the correlation coefficient of the original sparse check matrix and the reconstructed sparse check matrix.
The method for calculating the correlation of each abnormal position according to the elements in the neighborhood space matrix of each abnormal position with difference comprises the following steps:
in the formula: rho k For the correlation of the k-th differenced abnormal position, F k Is a neighborhood space matrix, T, of the abnormal location in the kth differenced original sparse check matrix k A neighborhood space matrix, COV (F), for the abnormal location in the kth differenced reconstructed sparse check matrix k ,T k ) The covariance of the location of the anomaly is represented,
for the k-th differenced original sparse check matrixThe variance of the neighborhood spatial matrix of the anomaly location in the array,
and the variance of the neighborhood space matrix of the abnormal position in the k-th differential reconstruction sparse check matrix is obtained.
The method for obtaining the abnormal position of the difference between the original sparse check matrix and the reconstructed sparse check matrix according to the difference matrix comprises the following steps: and performing difference processing on the original sparse check matrix and the reconstructed sparse check matrix to obtain a difference matrix, wherein a non-zero position in the difference matrix is an abnormal position where the original sparse check matrix and the reconstructed sparse check matrix have difference.
The process of transmitting and decoding the compressed sparse check matrix and the information code element to be transmitted is as follows:
and compressing the final sparse check matrix by using a self-coding network to obtain compressed sparse check matrix data, encoding the signal to be transmitted by using LDPC coding to obtain a signal code element to be transmitted, uploading the compressed sparse check matrix and the signal code element to be transmitted to a transmission medium, and verifying and decoding the signal code element to be transmitted by using the compressed sparse check matrix data.
In a second aspect, an embodiment of the present invention provides an LDPC coding system based on a self-coding network, including:
a data processing unit: constructing an original sparse check matrix, and performing LDPC coding processing on a signal to be transmitted;
a data compression unit: compressing the original sparse check matrix in a self-coding mode, reconstructing the original sparse check matrix through compressed data to obtain a reconstructed sparse check matrix, calculating a correlation coefficient between the original sparse check matrix and the reconstructed sparse check matrix to determine a final sparse check matrix, and compressing the sparse check matrix;
transmission medium: LDPC coding is used for transmitting the compressed data of the obtained sparse check matrix and the signal to be transmitted;
receiving end: and decoding the LDPC codes of the signals to be transmitted by using the obtained sparse check matrix.
Compared with the prior art, the embodiment of the invention has the beneficial effects that:
the invention combines the self-coding network and the LDPC coding, can effectively reduce the problem of channel capacity waste caused by the fact that the code length of the LDPC coding is too long, the required check matrix is large, and the transmission occupies large channel capacity, and the low-density sparsity of the check matrix can lead the self-coding network to compress the check matrix, and the self-coding network is utilized to compress and transmit the low-density sparse check matrix, thereby realizing more efficient transmission of information in the channel coding transmission process.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
Fig. 1 is a block diagram of a system flow provided by an LDPC coding method based on a self-coding network in embodiment 1 of the present invention;
FIG. 2 is a flowchart of a method provided by an LDPC coding and decoding method based on a self-coding network according to embodiment 1 of the present invention;
FIG. 3 is a flowchart of a method provided by an LDPC coding and decoding method based on a self-coding network according to embodiment 2 of the present invention;
fig. 4 is a schematic diagram of a transmission process provided by an LDPC coding method based on a self-coding network in embodiment 2 of the present invention;
fig. 5 is a flowchart for obtaining a reconstructed sparse check matrix provided by an LDPC coding method based on a self-encoding network in embodiment 2 of the present invention;
fig. 6 is a flowchart for obtaining a final sparse check matrix provided by an LDPC coding method based on a self-encoding network in embodiment 2 of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In the following description, for purposes of explanation and not limitation, specific details are set forth, such as particular system structures, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.
The terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature; in the description of the present embodiment, "a plurality" means two or more unless otherwise specified.
Example 1
The embodiment of the invention provides an LDPC coding and decoding method based on a self-coding network, as shown in FIG. 1 and FIG. 2, comprising the following steps:
s101, constructing an original sparse check matrix
The LDPC coding rule is utilized to code the signal to be transmitted, because the core of the LDPC coding is to construct a low-density sparse check matrix, and the reduction of the coding complexity is realized through the sparsity of the sparse check matrix. Therefore, in this embodiment, an original sparse check matrix needs to be constructed before encoding a signal to be transmitted.
S102, obtaining a reconstructed sparse check matrix
The self-coding network is combined to compress and decompress the sparse check matrix to obtain a reconstructed sparse check matrix, so that the problems that the check matrix is too large and the length of a supervision code element is too large due to the overlong length of an LDPC coding code, and the information needing to be transmitted, namely the supervision code element, becomes redundant information in the final transmission process, and the waste of channel capacity is caused are solved.
S103, calculating correlation coefficients of the original sparse check matrix and the reconstructed sparse check matrix
Since the nature of the self-encoding network is not lossless compression reconstruction, errors may exist; and because the structure of the sparse check matrix is not directly related to the signal to be transmitted, and only the sparse check matrix accords with the low-density sparse characteristic, even if the reconstructed matrix has errors, the transmission effect of the final signal is not influenced as long as the errors are not large. Therefore, the difference between the original sparse check matrix and the reconstructed sparse check matrix is analyzed, and the correlation coefficient of the original sparse check matrix and the reconstructed sparse check matrix is calculated to determine the error between the reconstructed sparse check matrix and the original sparse check matrix.
S104, determining a sparse check matrix
And determining the similarity between the original sparse check matrix and the reconstructed sparse check matrix according to the obtained correlation coefficients of the original sparse check matrix and the reconstructed sparse check matrix, and judging the effectiveness of the reconstructed sparse check matrix in the final data compression according to the size of the similarity. If the reconstructed sparse check matrix is valid, it indicates that the original sparse check matrix is highly similar to the reconstructed sparse check matrix, and the reconstructed sparse check matrix can play a role in monitoring the code elements in subsequent operations.
S105, transmitting and decoding signals to be transmitted
And performing compression transmission and decoding processing on the signal to be transmitted according to the obtained sparse check matrix, and completing the process of performing compression transmission on the low-density sparse check matrix by using a self-coding network.
Example 2
The embodiment of the invention provides an LDPC coding and decoding method based on a self-coding network, as shown in FIG. 1 and FIG. 3, the specific contents include:
s201, constructing an original sparse check matrix
The LDPC coding rule is utilized to code the signal to be transmitted, because the core of the LDPC coding is to construct a low-density sparse check matrix, and the reduction of the coding complexity is realized through the sparsity of the sparse check matrix. Therefore, in this embodiment, an original sparse check matrix needs to be constructed before encoding a signal to be transmitted.
The method is designed based on a geometric theory, has low coding complexity and forms a linear relation with the code length, and can design codes with high code rate and large code length.
Therefore, a generating matrix can be determined by an encoding algorithm based on an approximate lower triangle according to the obtained original sparse check matrix, LDPC encoding is performed on the signal to be transmitted which needs to be compressed and transmitted in combination with the generating matrix, so as to obtain LDPC encoded data of the signal to be transmitted, and the transmission process is as shown in fig. 4.
Wherein, the relationship between the generated matrix and the original sparse check matrix satisfies the following formula:
GH T ＝0
in the formula: g is a generator matrix, H is an original sparse check matrix, H T Is the transpose of the original sparse check matrix.
S202, obtaining a reconstructed sparse check matrix
The self-coding network is combined to compress and decompress the sparse check matrix to obtain a reconstructed sparse check matrix, so that the problems that the check matrix is too large and the length of a supervision code element is too large due to the overlong length of an LDPC coding code, and the information needing to be transmitted, namely the supervision code element, becomes redundant information in the final transmission process, and the waste of channel capacity is caused are solved.
The application scenario of the LDPC coding is that the longer the coding length is, the better the performance is, however, because of the characteristics of the LDPC coding, it is necessary to receive all symbols to perform the coding operation, and therefore, the longer the coding length means that the check matrix is larger, the length of the parity symbols is larger, and for information to be transmitted in the final transmission process, the parity symbols means redundant information, thereby causing a waste of channel capacity. The check matrix has the low-density sparse characteristic, the number of the zeros contained in the matrix is very large, the number of the zeros contained in the matrix is very small, the check matrix has regularity, the self-coding network can learn the inherent characteristics of a group of data, the data is compressed, efficient transmission of the data is achieved, data recovery is conducted at a terminal, and the characteristic of the low-density sparse matrix can be combined with the self-coding network to improve coding efficiency.
The self-coding network is utilized to compress an original sparse check matrix to obtain compressed data, the sparse check matrix has obvious and unique characteristics, most of elements in the matrix are 0, and fixed row weight and column weight are provided, namely, the number of the elements containing 1 in each row and each column is the same, the expression of the original sparse check matrix can be well learned by the self-coding network through the characteristics of the sparse check matrix, so that efficient decompression is performed, and the original sparse check matrix H is as follows:
most elements in the original sparse check matrix are 0, and the row weight of each row and each column is the same; meanwhile, the generation of the check matrix is not directly connected with the information source to be transmitted and only accords with the low-density characteristic, so that the final information transmission result cannot be influenced even if loss occurs in the decoding process.
S203, acquiring abnormal positions with difference in difference matrix
Since the nature of the self-encoding network is not lossless compression reconstruction, errors may exist; and because the construction of the sparse check matrix is not directly related to the signal to be transmitted, and only the sparse check matrix accords with the low-density sparse characteristic, even if the reconstructed matrix has errors, the transmission effect of the final signal is not influenced as long as the errors are not large. Therefore, a difference matrix of the original sparse check matrix and the reconstructed sparse check matrix needs to be obtained, abnormal positions with differences are extracted according to the difference matrix, and the abnormal positions are analyzed to judge the error between the original sparse check matrix and the reconstructed sparse check matrix.
Due to the characteristics of the self-coding network, information loss exists in the coding and decoding process, so that the check matrix obtained after the compressed information is reconstructed still has the sparse performance and can play the role of monitoring the code elements, and the validity of the compressed data is firstly checked.
H，H *
In the above formula, H represents the original matrix, H * Representing the restored matrix in m x n dimensions, the optimal result being that the restored matrix is the same as the original matrix
H＝H *
However, since the characteristics of the self-coding network are not lossless compression reconstruction, errors may exist, and because the construction of the sparse check matrix is not directly related to the signal to be transmitted, only the sparse check matrix conforms to the low-density sparse characteristics, even if the reconstructed sparse check matrix has errors, as long as the errors are not large, the transmission effect of the final signal is not affected. Therefore, the difference between the reconstructed sparse check matrix H and the original sparse check matrix H is determined in this step.
The foregoing indicates that the low-density sparse check matrix should have the same row weight and column weight, and therefore it is first checked whether the row weight and the column weight of the recovered and reconstructed sparse check matrix are consistent. And counting the number of 1 by traversing the rows and columns of the original sparse check matrix and the reconstructed sparse check matrix.
L＝{l 1 ，l 2 ，...，l m }
C＝{c 1 ，c 2 ，...，c n }
In the above formula, L represents the row-wise repeated set of the original sparse check matrix, and the matrix has m rows, so that m elements are shared, and L i Represents the row weight of the ith row; c represents a column-repeated set of an original sparse check matrix, wherein the matrix has n columns, so n elements are shared, and C i The column weight in the ith row is indicated. The row weight and the column weight of the sparse check matrix are consistent:
l 1 ＝l 2 ＝...＝l m
c 1 ＝c 2 ＝...＝c n
similarly, the reconstructed sparse check matrix after the compression reconstruction is H * Then the result obtained by traversal is:
same as above with L * *C * Respectively representing the row weight and column weight set of the reconstructed sparse check matrix, wherein the reconstructed check matrix firstly needs to meet the condition that the row weight and the column weight are the same and are the same as the original matrix;
the row weight and the column weight of the reconstructed sparse check matrix are verified in the above formula, if the above formula is not satisfied, it indicates that a large loss occurs in the compression process by using the self-coding network, the compression and reconstruction process is not qualified, and the weight of the self-coding network needs to be adjusted to perform the new compression and reconstruction so as to update the reconstructed sparse check matrix until the reconstructed sparse check matrix is obtained by satisfying the above formula, as shown in fig. 5.
The construction of the sparse check matrix has no direct relation with the signal itself, and only needs to satisfy the corresponding construction rule, and the final transmission result is not influenced as long as the characteristics of the low-density sparse check matrix are reserved even if some distortion exists after the constructed sparse check matrix is reconstructed before transmission. The row weight and the column weight of the reconstructed sparse check matrix are judged, and finally the fact that the row weight and the column weight of the reconstructed sparse check matrix are consistent and tangent to the original sparse check matrix is determined.
The optimal compression reconstruction result is the recurrence of the original sparse matrix, so that the difference between the two sparse check matrixes is made, and the larger the number of the matrixes which is not 0 is, the larger the difference between the reconstructed sparse check matrix and the original sparse check matrix is.
H 、 ＝H-H *
Obtaining a difference matrix H by making a difference between the original sparse check matrix and the reconstructed sparse check matrix 、 The more non-zero numbers in the difference matrix indicate the greater the difference between the reconstructed sparse check matrix and the original sparse check matrix.
Traverse difference matrix H 、 Adding the absolute values of the elements in the difference matrix to obtain the number of the nonzero elements as follows:
in the above formula h ` Represents the number of non-zero elements in the final difference matrix, h ` i,j The element representing the ith row and jth column in the matrix, the absolute value is taken because there are three values in the difference matrix: 0,1, -1, to avoid cancellation, a non-zero number of difference matrices is precisely counted.
1. If the following formula is satisfied:
h ` ＝0
the compressed data is shown to realize the reconstruction of the lossless check matrix, and the subsequent encoding operation can be directly carried out.
2. If in the above formula:
h ` >0
the distortion of the reconstructed check matrix is indicated, and the degree of distortion needs to be judged to determine whether the reconstructed matrix can be continuously used as the check matrix. The positions of the non-zero elements of the obtained difference matrix of the check matrix correspond to the positions of the difference between the original check matrix and the reconstructed check matrix.
In the above formula
h i,j And respectively representing the element positions in the reconstructed sparse check matrix and the original sparse check matrix corresponding to the nonzero position of the difference matrix.
S204, calculating correlation coefficients of the original sparse check matrix and the reconstructed sparse check matrix
And analyzing the difference between the original sparse check matrix and the reconstructed sparse check matrix, and calculating the correlation coefficient of the original sparse check matrix and the reconstructed sparse check matrix to determine the error between the reconstructed sparse check matrix and the original sparse check matrix.
The traditional method for judging the difference between two groups of data is generally judged by the correlation coefficient of the two groups of data, however, because of the particularity of LDPC coding and the longer code length, the formed sparse check matrix has very high dimensionality, and therefore, the method for directly calculating the correlation of the two matrixes is obviously not suitable. Moreover, since the distortion of the self-coding network is very small, it is not necessary to calculate the correlation of the two matrices as a whole, and the calculation waste is generated, so the embodiment quantifies the correlation of the two matrices by locating the correlation of the neighborhood space of the abnormal position in the matrix.
Respectively acquiring neighborhood space matrixes of abnormal positions in the original sparse check matrix and the reconstructed sparse check matrix:
wherein: f k For the k-th abnormal position in the original sparse check matrixOf a neighborhood space matrix, T k And reconstructing a neighborhood space matrix of the k-th abnormal position in the sparse check matrix.
In the above formula, the neighborhood space matrix is the pass h ` i,j And positioning the abnormal element points of the obtained original check matrix and the reconstructed matrix as centers, setting the step length in four directions of four neighborhoods as 2, performing sliding window to form a 5*5 matrix, and calculating the correlation between the original sparse check matrix and the reconstructed sparse check matrix by taking the matrix as a core.
The significance of selecting 5*5 matrix is that the distribution of outliers is uncertain, but since the check matrix is to conform to low density and sparsity, and the sparse check matrices before and after reconstruction should be as similar as possible, fewer outliers should appear in the same region, and the ideal result is that the outliers are uniformly distributed in the whole matrix. Normally, only one outlier exists in each region, so the correlation of the matrix is close; if a plurality of abnormal points appear in the same region, the correlation is greatly reduced, and the correlation between the neighborhood space matrix of the abnormal position in the original sparse check matrix and the neighborhood space matrix of the abnormal position in the reconstructed sparse check matrix is well calculated through the method.
In the formula: rho k For the correlation of the k-th differenced abnormal position, F k A neighborhood space matrix, T, for the abnormal location in the kth differenced original sparse check matrix k A neighborhood space matrix, COV (F), for the k-th differenced reconstructed sparse check matrix for the anomaly location k ,T k ) The covariance of the location of the anomaly is represented,
the variance of the neighborhood space matrix of the anomaly location in the kth differenced original sparse check matrix,
and the variance of the neighborhood space matrix of the abnormal position in the k-th differential reconstruction sparse check matrix is obtained.
Then the calculation formula of the correlation coefficient of the original sparse check matrix and the reconstructed sparse check matrix is as follows:
in the formula: rho is the correlation coefficient of the original sparse check matrix and the reconstructed sparse check matrix, rho k The k-th abnormal position correlation is the k-th abnormal position correlation, k is the serial number of the abnormal position, and h is the number of the abnormal positions.
S205, determining a sparse check matrix
And determining the similarity between the original sparse check matrix and the reconstructed sparse check matrix according to the obtained correlation coefficient of the original sparse check matrix and the reconstructed sparse check matrix, and judging the effectiveness of the reconstructed sparse check matrix in finally compressing data according to the size of the similarity. If the reconstructed sparse check matrix is valid, it indicates that the original sparse check matrix is highly similar to the reconstructed sparse check matrix, and the reconstructed sparse check matrix can play a role in monitoring the code elements in subsequent operations.
Judging whether the original sparse check matrix is similar to the reconstructed sparse check matrix according to the correlation coefficient of the original sparse check matrix and the reconstructed sparse check matrix:
the method for judging whether the reconstructed sparse check matrix is effective or not by setting the similarity threshold comprises the following steps:
wherein: rho is a correlation coefficient of the original sparse check matrix and the reconstructed sparse check matrix, and alpha is a similarity threshold.
Determining a sparse check matrix according to the similarity of the original sparse check matrix and the reconstructed sparse check matrix:
if the original sparse check matrix is highly similar to the reconstructed sparse check matrix, taking the reconstructed sparse check matrix as a sparse check matrix;
if the original sparse check matrix is not similar to the reconstructed sparse check matrix, the weight of the self-coding network is adjusted, iteration is performed in the process of reconstructing the sparse check matrix again until the reconstructed sparse check matrix is highly similar to the original sparse check matrix, and the iteration is stopped, and the reconstructed sparse check matrix of the last iteration is used as the sparse check matrix, as shown in fig. 6.
S206, transmitting and decoding signals to be transmitted
And performing compression transmission and decoding processing on the signal to be transmitted according to the obtained sparse check matrix, and completing the process of performing compression transmission on the low-density sparse check matrix by using a self-coding network.
And transmitting the compressed sparse check matrix and the signal to be transmitted into a channel, transmitting the signal to a receiving end by using a transmission medium, and decoding the information code element to be transmitted by using the compressed sparse check matrix at the receiving end.
Based on the same inventive concept as the method, the embodiment further provides an LDPC coding system based on a self-encoding network, and the LDPC coding system based on the self-encoding network in the embodiment includes a data processing unit, a data compression unit, a transmission medium, and a receiving end, where the data processing unit, the data compression unit, the transmission medium, and the receiving end are configured to implement that an original sparse check matrix is constructed by using a finite geometry method as described in the embodiment of the LDPC coding method based on the self-encoding network, and the original sparse check matrix is converted into a generator matrix; compressing the original sparse check matrix by using a self-coding network, and reconstructing data of the compressed original sparse check matrix to obtain a reconstructed sparse check matrix; acquiring abnormal positions with differences according to a difference matrix of the acquired original sparse check matrix and the reconstructed sparse check matrix; calculating correlation coefficients of the original sparse check matrix and the reconstructed sparse check matrix to determine a sparse check matrix; and compressing the determined sparse check matrix, and transmitting and decoding the compressed sparse check matrix and the signal to be transmitted.
In the embodiment of the LDPC coding and decoding method based on the self-coding network, an original sparse check matrix is constructed by using a finite geometry method, and the original sparse check matrix is converted into a generating matrix; compressing the original sparse check matrix by using a self-coding network, and reconstructing data of the compressed original sparse check matrix to obtain a reconstructed sparse check matrix; acquiring abnormal positions with differences according to a difference matrix of the acquired original sparse check matrix and the reconstructed sparse check matrix; calculating correlation coefficients of the original sparse check matrix and the reconstructed sparse check matrix to determine a sparse check matrix; the method for compressing the determined sparse check matrix and transmitting and decoding the compressed sparse check matrix and the signal to be transmitted is explained, and the details are not repeated here.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (7)
1. The LDPC coding and decoding method based on the self-coding network is characterized by comprising the following steps:
s1: constructing an original sparse check matrix in the LDPC coding process by using a finite geometry construction method;
s2: determining a reconstructed sparse check matrix, comprising:
s2.1: compressing the original sparse check matrix by using a self-coding network to obtain compressed original sparse check matrix data, and reconstructing the compressed original sparse check matrix data to obtain a reconstructed sparse check matrix;
s2.2: respectively extracting the column weight and the row weight of the original sparse check matrix and the reconstructed sparse check matrix, and judging whether the compression reconstruction process is qualified or not according to the column weight and the row weight of the original sparse check matrix and the reconstructed sparse check matrix;
s2.3: if the compression reconstruction process is unqualified, adjusting the self-coding network weight, and repeating the steps S2.1-S2.2 by using the adjusted self-coding network weight until the compression reconstruction process is qualified;
s2.4: taking the reconstructed sparse check matrix obtained in the qualified compression reconstruction process as a reconstructed sparse check matrix;
s3: determining a final sparse check matrix, comprising:
s3.1: acquiring a difference matrix according to the original sparse check matrix and the reconstructed sparse check matrix, and acquiring an abnormal position of the difference between the original sparse check matrix and the reconstructed sparse check matrix according to the difference matrix;
s3.2: calculating correlation coefficients of the original sparse check matrix and the reconstruction sparse check matrix according to neighborhood spaces of abnormal positions of the original sparse check matrix and the reconstruction sparse check matrix, and judging whether the reconstruction sparse check matrix meets requirements or not according to the correlation coefficients;
s3.3: if the reconstructed sparse check matrix does not meet the requirement, adjusting the weight of the self-coding network, and repeating S2.1-S3.3 by using the adjusted weight of the self-coding network until the reconstructed sparse check matrix meets the requirement;
s3.4: taking the reconstructed sparse check matrix which meets the requirements as a final sparse check matrix;
s4: and compressing the determined sparse check matrix by using a self-encoding network to obtain the LDPC code element of the signal to be transmitted, transmitting the compressed sparse check matrix and the LDPC code element of the signal to be transmitted, and verifying and decoding the LDPC code element of the signal to be transmitted according to the sparse check matrix.
2. The LDPC coding and decoding method based on self-coding network according to claim 1, wherein the method for determining whether the reconstructed sparse check matrix meets the requirement according to the correlation coefficient comprises:
the method for judging whether the reconstructed sparse check matrix is effective or not by setting the similarity threshold comprises the following steps:
wherein: rho is a correlation coefficient of the original sparse check matrix and the reconstructed sparse check matrix, and alpha is a similarity threshold.
3. The LDPC coding and decoding method based on self-coding network according to claim 1, wherein the calculation process of the correlation coefficient of the original sparse check matrix and the reconstructed sparse check matrix is as follows:
acquiring abnormal positions of the original sparse check matrix and the reconstructed sparse check matrix, calculating the correlation of each abnormal position according to elements in a neighborhood space matrix of each abnormal position with the difference, and taking the sum of the correlations of all the abnormal positions with the difference in the original sparse check matrix and the reconstructed sparse check matrix as the correlation coefficient of the original sparse check matrix and the reconstructed sparse check matrix.
4. The LDPC coding method based on self-coding network according to claim 3, wherein the method of calculating the correlation of each abnormal position according to the elements in the neighborhood space matrix of each abnormal position having a difference is as follows:
in the formula: rho k For the correlation of the k-th differenced abnormal position, F k Is a neighborhood space matrix, T, of the abnormal location in the kth differenced original sparse check matrix k A neighborhood space matrix, COV (F), for the abnormal location in the kth differenced reconstructed sparse check matrix k ,T k ) The covariance of the location of the anomaly is represented,
the variance of the neighborhood space matrix of the anomaly location in the kth differenced original sparse check matrix,
and the variance of the neighborhood space matrix of the abnormal position in the k-th differential reconstruction sparse check matrix is obtained.
5. The LDPC coding and decoding method based on self-coding network according to claim 1, wherein the method for obtaining the abnormal position where the original sparse check matrix and the reconstructed sparse check matrix have difference according to the difference matrix comprises: and performing difference processing on the original sparse check matrix and the reconstructed sparse check matrix to obtain a difference matrix, wherein a non-zero position in the difference matrix is an abnormal position where the original sparse check matrix and the reconstructed sparse check matrix have difference.
6. The LDPC coding and decoding method based on self-encoding network as claimed in claim 1, wherein the process of transmitting and decoding the compressed sparse check matrix and the information symbol to be transmitted is as follows:
and compressing the final sparse check matrix by using a self-coding network to obtain compressed sparse check matrix data, encoding the signal to be transmitted by using LDPC coding to obtain a signal code element to be transmitted, uploading the compressed sparse check matrix and the signal code element to be transmitted to a transmission medium, and decoding the signal code element to be transmitted by using the compressed sparse check matrix data.
7. The LDPC coding and decoding system based on the self-coding network comprises: data processing unit, data compression unit, transmission medium and receiving terminal, its characterized in that:
a data processing unit: constructing an original sparse check matrix, and performing LDPC coding processing on a signal to be transmitted;
a data compression unit: compressing the original sparse check matrix in a self-coding mode, reconstructing the original sparse check matrix through compressed data to obtain a reconstructed sparse check matrix, calculating a correlation coefficient between the original sparse check matrix and the reconstructed sparse check matrix to determine a final sparse check matrix, and compressing the sparse check matrix;
transmission medium: the LDPC code element is used for transmitting the compressed data of the obtained sparse check matrix and the signal to be transmitted;
receiving end: and decoding the LDPC coded code element of the signal to be transmitted by using the obtained sparse check matrix data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210837384.7A CN115314156B (en) | 2022-07-15 | 2022-07-15 | LDPC coding and decoding method and system based on self-coding network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210837384.7A CN115314156B (en) | 2022-07-15 | 2022-07-15 | LDPC coding and decoding method and system based on self-coding network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115314156A true CN115314156A (en) | 2022-11-08 |
| CN115314156B CN115314156B (en) | 2023-04-25 |
Family
ID=83856166
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210837384.7A Active CN115314156B (en) | 2022-07-15 | 2022-07-15 | LDPC coding and decoding method and system based on self-coding network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115314156B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009060288A (en) * | 2007-08-30 | 2009-03-19 | Nippon Telegr & Teleph Corp <Ntt> | Information transmitting/receiving method, coder, and decoder |
| CN109919864A (en) * | 2019-02-20 | 2019-06-21 | 重庆邮电大学 | A kind of compression of images cognitive method based on sparse denoising autoencoder network |
| CN112702070A (en) * | 2020-12-29 | 2021-04-23 | 厦门大学 | Decoding optimization method of distributed joint source-channel coding system |
-
2022
- 2022-07-15 CN CN202210837384.7A patent/CN115314156B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009060288A (en) * | 2007-08-30 | 2009-03-19 | Nippon Telegr & Teleph Corp <Ntt> | Information transmitting/receiving method, coder, and decoder |
| CN109919864A (en) * | 2019-02-20 | 2019-06-21 | 重庆邮电大学 | A kind of compression of images cognitive method based on sparse denoising autoencoder network |
| CN112702070A (en) * | 2020-12-29 | 2021-04-23 | 厦门大学 | Decoding optimization method of distributed joint source-channel coding system |
Non-Patent Citations (3)
| Title |
|---|
| JIANG XIAOBO， LI HONGYUAN: "High efficiency and low power multi-rate LDPC decoder design for CMMB" * |
| THIERRY DUMAS，ALINE ROUMY，CHRISTINE GUILLEMOT: "Shallow sparse autoencoders versus sparse coding algorithms for image compression" * |
| 杜凤强;叶润;闫斌;: "一种基于PEG的构造矩阵在压缩感知中的应用" * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115314156B (en) | 2023-04-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107832837B (en) | Convolutional neural network compression method and decompression method based on compressed sensing principle | |
| Farsad et al. | Deep learning for joint source-channel coding of text | |
| CN106888026B (en) | Segmented polarization code coding and decoding method and system based on LSC-CRC (least significant likelihood-Cyclic redundancy check) decoding | |
| US12010352B2 (en) | Coding method and related device | |
| US8219890B2 (en) | Denoising and error correction for finite input, general output channel | |
| CN109075805B (en) | Apparatus and method for implementing polarization code | |
| CN110999093A (en) | Hybrid architecture for check node processing for Extended Minimum Sum (EMS) decoding of non-binary LDPC codes | |
| CN110545162B (en) | Multivariate LDPC decoding method and device based on code element reliability dominance degree node subset partition criterion | |
| Wang et al. | Efficient compression of encrypted binary images using the Markov random field | |
| CN103026636B (en) | Orthogonal multiple description coded | |
| Pham et al. | Sublinear compressive sensing reconstruction via belief propagation decoding | |
| Yeh et al. | On the optimality of code options for a universal noiseless coder | |
| Levina et al. | Algorithms of constructing linear and robust codes based on wavelet decomposition and its application | |
| Gabay et al. | Real BCH codes as joint source channel codes for satellite images coding | |
| Al-Hashemi et al. | A new lossless image compression technique based on Bose, Chandhuri and Hocquengham (BCH) codes | |
| Gabay et al. | Spectral Interpolation Coder for impulse noise cancellation over a binary symetric channel | |
| CN115314156B (en) | LDPC coding and decoding method and system based on self-coding network | |
| Berezkin et al. | Data compression methods based on Neural Networks | |
| Yeh et al. | On the optimality of a universal noiseless coder | |
| Gabay et al. | Joint source-channel coding using structured oversampled filters banks applied to image transmission | |
| CN115037415A (en) | Error correction coding method, device and terminal based on CRC | |
| CN114024551A (en) | Data lossless compression method, system, electronic device and medium | |
| Shahnas | Performance analysis of regular and irregular ldpc codes on spiht coded image data | |
| CN106572093A (en) | Wireless sensor array data compression method and wireless sensor array data compression system | |
| CN108599775B (en) | Construction method of hybrid check LDPC code |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |