CN109495114B - LDPC decoder construction method based on Markov Monte Carlo method - Google Patents
LDPC decoder construction method based on Markov Monte Carlo method Download PDFInfo
- Publication number
- CN109495114B CN109495114B CN201811176885.5A CN201811176885A CN109495114B CN 109495114 B CN109495114 B CN 109495114B CN 201811176885 A CN201811176885 A CN 201811176885A CN 109495114 B CN109495114 B CN 109495114B
- Authority
- CN
- China
- Prior art keywords
- iteration
- binary information
- ldpc
- information vector
- vector
- 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.)
- Active
Links
Images
Classifications
-
- 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
- H03M13/1111—Soft-decision decoding, e.g. by means of message passing or belief propagation algorithms
- H03M13/1117—Soft-decision decoding, e.g. by means of message passing or belief propagation algorithms using approximations for check node processing, e.g. an outgoing message is depending on the signs and the minimum over the magnitudes of all incoming messages according to the min-sum rule
- H03M13/112—Soft-decision decoding, e.g. by means of message passing or belief propagation algorithms using approximations for check node processing, e.g. an outgoing message is depending on the signs and the minimum over the magnitudes of all incoming messages according to the min-sum rule with correction functions for the min-sum rule, e.g. using an offset or a scaling factor
-
- 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
Landscapes
- Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Error Detection And Correction (AREA)
Abstract
The invention discloses an LDPC decoder construction method based on a Markov Monte Carlo method, which comprises the following steps: s1: constructing an MCMC decoding algorithm of the LDPC through a generating matrix G, a receiving signal y and a preset maximum iteration number max; s2: constructing an LDPC MCMC-S decoding algorithm on the basis of the LDPC MCMC decoding algorithm; s3: the MCMC-L decoding algorithm of the LDPC is constructed by a generating matrix G, a receiving signal y and a preset expansion path number L. The invention solves the problem of poor performance when the code length is short in BP, and can be used for next generation mobile communication to meet the requirement of LDPC decoding.
Description
Technical Field
The invention relates to the technical field of signal processing, in particular to a construction method of an LDPC decoder based on a Markov Monte Carlo method.
Background
Low Density Parity Check Code (LDPC), proposed by Gallagar in 1962, is one of Linear Block Codes (LBC). With the proposal of the LDPC effective decoding algorithm, the LDPC is gradually popularized and is considered to have the characteristics of good performance and parallel friendliness close to Shannon limit. Because of these advantages, LDPC codes have been included in a variety of protocols since the 2G era (e.g., IEEE 802.11, IEEE 802.20). Nowadays, LDPC codes are adopted as coding schemes for long code blocks of data channels in 5G enhanced Mobile Broadband (eMBB) scenarios.
As an important component of LDPC codes, decoder modules are not a small number of. Decoding methods of LDPC code decoders can be roughly classified into two types: a decoding method based on hard decision and a decoding method based on soft decision. The first kind of decoding scheme has low computational complexity and is easy to implement, but has limited error correction capability. The second category of methods includes Belief Propagation (BP) algorithms that employ a posteriori probability information. It is computationally complex but is widely studied due to performance approaching the shannon limit. However, when the factor graph of the LDPC code contains small cycles, the BP decoding algorithm can only provide a sub-optimal solution, and particularly when the code length is short, the result is not ideal. Therefore, how to improve the performance of the BP algorithm in short codes while maintaining acceptable complexity becomes a key issue.
Disclosure of Invention
The invention aims to: the invention aims to provide an LDPC decoder construction method based on a Markov Monte Carlo method, which can improve the performance of a BP algorithm in short codes and maintain acceptable complexity.
The technical scheme is as follows: in order to achieve the purpose, the invention adopts the following technical scheme:
the LDPC decoder construction method based on the Markov Monte Carlo method comprises the following steps:
s1: constructing an MCMC decoding algorithm of the LDPC by generating a matrix G, a received signal y and a preset maximum iteration number max;
s2: constructing an LDPC MCMC-S decoding algorithm on the basis of the LDPC MCMC decoding algorithm;
s3: the MCMC-L decoding algorithm of the LDPC is constructed by a generating matrix G, a receiving signal y and a preset expansion path number L.
Further, the step S1 includes the steps of:
s1.1: initializing the binary information vector x through a hard decision result to obtain an initialized binary information vector x (0) ;
S1.2: setting the total times of the first iteration counter as max, and enabling an iteration variable j of the first iteration counter to be 1;
s1.3: setting the total times of the second iteration counter as K, and enabling an iteration variable i of the second iteration counter to be 1; k is the length of the binary information vector x;
s1.4: the calculation was performed according to formulas (1) to (3):
wherein x is i Representing the ith bit of a binary information vector x, x i=0 The ith bit representing the binary information vector x is 0, x i=1 The ith bit representing the binary information vector x is 1, sigma represents the standard deviation of white gaussian noise added by the signal in the process of space propagation, and L i Denotes x i Log-likelihood ratio of (2), P (x) i 0) represents x i Probability of 0, r is a random number uniformly distributed between 0 and 1;
s1.5: making i equal to i +1, and returning to step S1.4 until iterating to preset number K, and obtaining j-th iteration result x (j) Then entering the next operation;
s1.6: j is given as j +1, and the procedure returns to step S1.3 until the iteration reaches the preset number max, and the result x of the max-th iteration is obtained (max) Is an estimate of the transmitted information.
Further, the step S2 includes the steps of:
s2.1: initializing the binary information vector x through a hard decision result to obtain an initialized binary information vector x (0) ;
S2.2: setting the total times of the first iteration counter as K, and enabling an iteration variable i of the first iteration counter to be 1;
s2.3: the calculation is performed according to equation (4):
in the formula (4), the reaction mixture is,representing the result of inverting the ith bit of a binary information vector x, d i Is a measure of the flipped result, which is used to derive the bits in xUpdating the sequence;
s2.4: making i equal to i +1, returning to the step S2.3 until iteration reaches a preset number of times K, and then entering the next operation;
s2.5: for d obtained i (i-1, 2.. K) sorting in ascending order, storing the sorting result in vector s as the order of the subsequent update bits; the p-th element Sp in s is stored at d i Subscript of element at p-th order in ascending order;
s2.6: setting the total times of the second iteration counter as max, and enabling an iteration variable j of the second iteration counter to be 1;
s2.7: setting the total times of the third iteration counter as K, and enabling an iteration variable p of the third iteration counter to be 1;
s2.8: the calculation was performed according to the formulas (5) to (8):
where N denotes the length of the received signal y, σ * Represents the result of the standard deviation correction of the added noise in the channel,representing the modified log-likelihood ratio,s-th representing a binary information vector x p Each ratioIn particular to a method for preparing a high-purity sodium chloride solution,representProbability of 0, r is a random number uniformly distributed between 0 and 1;
s2.9: let p be p +1 and return to step S2.8 until the iteration reaches the preset number K, and obtain the jth iteration result x (j) Then, the next operation is carried out;
s2.10: j is equal to j +1 and returns to step S2.7 until the iteration reaches the preset number max, and the max-th iteration result x (max) Is an estimate of the transmitted information.
Further, the step S3 includes the steps of:
s3.1: initializing x with all-zero vectors 1 =x 2 =...=x L (ii) a Wherein x is u U is more than or equal to 1 and less than or equal to L which is a binary information vector on the u-th path, and L is the number of preset expansion paths;
s3.2: setting the total times of the first iteration counter as K, and enabling an iteration variable i of the first iteration counter to be 1;
s3.3: setting the total times of a second iteration counter as L, and enabling an iteration variable p of the second iteration counter to be 1;
s3.4: value is assigned according to equation (9):
in the formula (9), the reaction mixture is,represents the vector x p The result after the i-th bit flip, x p Representing a binary information vector, x, on the p-th path L+p Representing a binary information vector on the L + p path;
s3.5: setting the total times of the third iteration counter to be 2L, and enabling an iteration variable j of the third iteration counter to be 1;
s3.6: the calculation is performed according to equation (10):
d j '=||y-Gx j || 2 2 (10)
in the formula (10), d j ' is a measure of the binary information vector on each path to obtain the path that will eventually need to be reserved; x is the number of j Representing a binary information vector on the jth path;
s3.7: for d obtained j ' (i ═ 1, 2.., K) is sorted in ascending order, and the sorting result is stored in a vector s for obtaining L paths which need to be reserved finally;
s3.8: j is made to be j +1, the process returns to the step S3.6 until iteration is carried out for a preset time 2L, and then the next operation is carried out;
s3.9: the value is assigned according to equation (11):
in the formula (11), the reaction mixture is,denotes the S th p A binary information vector on a strip path;
s3.10: making p equal to p +1, returning to the step S3.4 until iteration reaches a preset number L, and then entering the next operation;
s3.11: i is equal to i +1, and the process returns to step S2.3 until the iteration reaches the preset number K, then the binary information vector x on the 1 st path 1 Is an estimate of the transmitted information.
Further, the transmission formula of the LDPC decoder is:
y=Gx+n (12)
in the formula (12), the vector n ∈ C N Each term obeys a mean value of 0 and a variance of σ 2 X is a K × 1-dimensional binary information vector, y is an N × 1-dimensional reception vector, N represents a code length, and K represents an information bit length.
Has the beneficial effects that:the invention discloses an LDPC decoder construction method based on a Markov Monte Carlo method, which has certain advantages on short codes compared with the BP decoding algorithm of the classical LDPC. The experimental result shows that compared with BP decoding, the optimized decoding method provided by the invention has the signal-to-noise ratio of about 10 -3 There is a performance gain of over 1 dB. The method solves the problem of poor performance when the code length is short in BP, and can be used for next generation mobile communication to meet the requirements of the next generation mobile communication on LDPC decoding.
Drawings
FIG. 1 is a detailed algorithm diagram of the LDPC-MCMC algorithm;
FIG. 2 is a detailed algorithm diagram of the LDPC-MCMC-S algorithm;
FIG. 3 is a detailed algorithm diagram of the LDPC-MCMC-L algorithm;
FIG. 4 is a graph comparing the error rate performance of the decoding results of the method of the present invention (MCMC, MCMC-S, MCMC-L) with the conventional BP decoding method when the received signal vector length and the information vector length are 20 and 10, respectively;
FIG. 5 is a graph comparing the error rate performance of the decoding results of the method of the present invention (MCMC, MCMC-S, MCMC-L) with the conventional BP decoding method when the received signal vector length and the information vector length are 32 and 16, respectively;
FIG. 6 is a hardware architecture diagram of the MCMC-L;
FIG. 7 is a detailed diagram of the computing modules in the hardware architecture diagram of the MCMC-L.
Detailed Description
The technical solution of the present invention will be further described with reference to the following detailed description and accompanying drawings.
The specific embodiment discloses an LDPC decoder construction method based on a Markov Monte Carlo method, which comprises the following steps:
s1: an LDPC MCMC decoding algorithm is constructed by a generating matrix G, a receiving signal y and a preset maximum iteration number max, and is called as an LDPC-MCMC algorithm, wherein the LDPC-MCMC algorithm is shown in figure 1;
s2: an LDPC MCMC-S decoding algorithm is constructed on the basis of the LDPC MCMC decoding algorithm, the LDPC MCMC-S decoding algorithm is called LDPC-MCMC-S algorithm, and the LDPC-MCMC-S algorithm is shown in figure 2;
s3: the LDPC MCMC-L decoding algorithm is constructed by the generation matrix G, the received signal y and the preset expansion path number L and is called LDPC-MCMC-L algorithm, and the LDPC-MCMC-L algorithm is shown in FIG. 3.
Step S1 includes the steps of:
s1.1: initializing the binary information vector x through a hard decision result to obtain an initialized binary information vector x (0) ;
S1.2: setting the total times of the first iteration counter as max, and enabling an iteration variable j of the first iteration counter to be 1;
s1.3: setting the total times of a second iteration counter as K, and enabling an iteration variable i of the second iteration counter to be 1; k is the length of the binary information vector x;
s1.4: the calculation was performed according to formulas (1) to (3):
wherein x is i Representing the ith bit of a binary information vector x, x i=0 The ith bit representing the binary information vector x is 0, x i=1 The ith bit representing the binary information vector x is 1, σ represents the standard deviation of white gaussian noise added by the signal during spatial propagation, and L i Represents x i Log-likelihood ratio of (2), P (x) i 0) represents x i Probability of 0, r is a random number uniformly distributed between 0 and 1;
s1.5: i is made to be i +1, and the process returns to step S1.4 until the iteration reaches the preset number K, and the jth iteration result x is obtained (j) Then, the next operation is carried out;
s1.6: j is given as j +1, and the procedure returns to step S1.3 until the iteration reaches the preset number max, and the result x of the max-th iteration is obtained (max) Is an estimate of the transmitted information.
Step S2 includes the steps of:
s2.1: initializing the binary information vector x through a hard decision result to obtain an initialized binary information vector x (0) ;
S2.2: setting the total times of the first iteration counter as K, and enabling an iteration variable i of the first iteration counter to be 1;
s2.3: the calculation is performed according to equation (4):
in the formula (4), the reaction mixture is,representing the result of inverting the ith bit of a binary information vector x, d i Is a measure of the result after the flip, used to get the update sequence of the bits in x;
s2.4: making i equal to i +1, returning to the step S2.3 until iteration reaches a preset number of times K, and then entering the next operation;
s2.5: for d obtained i (i-1, 2.. K) sorting in ascending order, storing the sorting result in vector s as the order of the subsequent update bits; p-th element S in S p Stored is at d i Subscripts of the elements at the p-th position in ascending order;
s2.6: setting the total times of the second iteration counter as max, and enabling an iteration variable j of the second iteration counter to be 1;
s2.7: setting the total times of a third iteration counter as K, and enabling an iteration variable p of the third iteration counter to be 1;
s2.8: the calculation was performed according to the formulas (5) to (8):
where N denotes the length of the received signal y, σ * Represents the result of the standard deviation correction of the added noise in the channel,representing the modified log-likelihood ratio,s-th representing a binary information vector x p One bit of the data is transmitted to the receiver,to representProbability of 0, r is a random number uniformly distributed between 0 and 1;
s2.9: p is equal to p +1, and the process returns to step S2.8 until the iteration reaches the preset number K, and the j-th iteration result x is obtained (j) Then, the next operation is carried out;
s2.10: j is given as j +1, and the process returns to step S2.7 until the iteration reaches the preset number max, and the result x of the max-th iteration is obtained (max) Is an estimate of the transmitted information.
Step S3 includes the following steps:
s3.1: initializing x with all-zero vectors 1 =x 2 =...=x L (ii) a Wherein x is u U is more than or equal to 1 and less than or equal to L which is a binary information vector on the u-th path, and L is the number of preset expansion paths;
s3.2: setting the total times of the first iteration counter as K, and enabling an iteration variable i of the first iteration counter to be 1;
s3.3: setting the total times of the second iteration counter to be L, and enabling an iteration variable p of the second iteration counter to be 1;
s3.4: value is assigned according to equation (9):
in the formula (9), the reaction mixture is,represents the vector x p The result after the i-th bit flip, x p Representing a binary information vector, x, on the p-th path L+p Representing a binary information vector on the L + p path;
s3.5: setting the total times of the third iteration counter to be 2L, and enabling an iteration variable j of the third iteration counter to be 1;
s3.6: the calculation is performed according to equation (10):
d j '=||y-Gx j || 2 2 (10)
in the formula (10), d j ' is a measure of the binary information vector on each path to obtain the path that needs to be reserved finally; x is a radical of a fluorine atom j Representing a binary information vector on the jth path;
s3.7: for d obtained j ' (i ═ 1, 2.., K) is sorted in ascending order, and the sorting result is stored in a vector s for obtaining L paths which need to be reserved finally;
s3.8: j is set as j +1, the process returns to the step S3.6 until iteration reaches the preset time 2L, and then the next operation is carried out;
s3.9: the value is assigned according to equation (11):
in the formula (11), the reaction mixture is,denotes the th S p A binary information vector on a strip path;
s3.10: making p equal to p +1, returning to the step S3.4 until iteration is carried out for a preset number of times L, and then entering the next operation;
s3.11: i is set to i +1, and the process returns to step S2.3 until the iteration reaches the preset number K, then the binary information vector x on the 1 st path 1 Is an estimate of the transmitted information.
The transmission formula of the LDPC decoder is:
y=Gx+n (12)
in the formula (12), the vector n ∈ C N Each term obeys a mean value of 0 and a variance of σ 2 X is a K × 1-dimensional binary information vector, y is an N × 1-dimensional reception vector, N represents a code length, and K represents an information bit length.
The results of comparing the method of this embodiment with the classical BP decoding algorithm are shown in fig. 4 and 5.
As can be seen from fig. 4 and 5, when the code length is short: under the (20,10) and (32,16) scenarios, the performance is obviously improved compared with the traditional BP algorithm. Compared with BP decoding, the optimized decoding method provided by the invention has the signal-to-noise ratio of about 10 -3 There is a performance gain of over 1 dB. The method solves the problem of poor performance when the code length is short in BP, and can be used for next generation mobile communication to meet the requirements of the next generation mobile communication on LDPC decoding.
In terms of computational complexity, the algorithms of the invention all comprise operations of multiplication, addition, comparison and sequencing, so that the computational complexity is improved compared with the traditional BP algorithm. However, since these methods are all designed for short codes, the required complexity is still relatively low, and the performance improvement obtained by using the present solution is combined, the cost of the calculation amount can be accepted.
Finally, the present invention provides a hardware architecture diagram of an example MCMC-L of the present invention, FIG. 6, and a detailed diagram of a specific computing module, FIG. 7, for reference.
FIG. 6 shows an MCMC-L hardware architecture diagram, where initialization is performed first, followed by the calculation of d j ' and d j+L ' (j-1, 2, …, L) and then sorted and re-iterated until a maximum number of iterations max is reached.
FIG. 7 is a pair d j ' and d j+L ' concrete presentation of calculation module. Where #and #, including '2' are addition and binary addition, respectively. Suppose G is stored in column vector form, and G i The ith column vector of the G matrix is represented. D can be implemented with less complexity using the algorithm of FIG. 7 j ' and d j+L ' calculation.
Claims (2)
1. An LDPC decoder construction method based on a Markov Monte Carlo method is characterized in that: the method comprises the following steps:
s1: constructing an MCMC decoding algorithm of the LDPC by generating a matrix G, a received signal y and a preset maximum iteration number max; the MCMC decoding algorithm for constructing the LDPC specifically comprises the following steps:
s1.1: initializing the binary information vector x through a hard decision result to obtain the initialized binary information vector x (0) ;
S1.2: setting the total times of the first iteration counter as max, and enabling an iteration variable j of the first iteration counter to be 1;
s1.3: setting the total times of the second iteration counter as K, and enabling an iteration variable i of the second iteration counter to be 1; k is the length of the binary information vector x;
s1.4: the calculation was performed according to formulas (1) to (3):
wherein x is i Representing the ith bit of a binary information vector x, x i=0 The ith bit representing the binary information vector x is 0, x i=1 The ith bit representing the binary information vector x is 1, sigma represents the standard deviation of white gaussian noise added by the signal in the process of space propagation, and L i Denotes x i Log likelihood ratio of (c), P (x) i 0) represents x i Probability of 0, r is a random number uniformly distributed between 0 and 1;
s1.5: i is made to be i +1, and the process returns to step S1.4 until the iteration reaches the preset number K, and the jth iteration result x is obtained (j) Then entering the next operation;
s1.6: j is given as j +1, and the procedure returns to step S1.3 until the iteration reaches the preset number max, and the result x of the max-th iteration is obtained (max) An estimate of the transmitted information;
s2: constructing an LDPC MCMC-S decoding algorithm on the basis of the LDPC MCMC decoding algorithm; the MCMC-S decoding algorithm for constructing the LDPC specifically comprises the following steps:
s2.1: initializing the binary information vector x through a hard decision result to obtain the initialized binary information vector x (0) ;
S2.2: setting the total times of the first iteration counter as K, and enabling an iteration variable i of the first iteration counter to be 1;
s2.3: the calculation is performed according to equation (4):
in the formula (4), the reaction mixture is,representing the junction after inverting the ith bit of a binary information vector xFruit, d i Is a measure of the result after the flip, used to get the update sequence of the bits in x;
s2.4: making i equal to i +1, returning to the step S2.3 until iteration reaches a preset number of times K, and then entering the next operation;
s2.5: for d obtained i (i 1, 2.. K) sorting in ascending order, storing the sorting result in a vector s as the order of the subsequent update bits; p-th element s of s p Stored is at d i Subscript of element at p-th order in ascending order;
s2.6: setting the total times of the second iteration counter as max, and enabling an iteration variable j of the second iteration counter to be 1;
s2.7: setting the total times of the third iteration counter as K, and enabling an iteration variable p of the third iteration counter to be 1;
s2.8: the calculation was performed according to formulas (5) to (8):
where N denotes the length of the received signal y, σ * Represents the result of the standard deviation correction of the added noise in the channel,representing the modified log-likelihood ratio,s-th representing a binary information vector x p One bit of the data is transmitted to the receiver,to representProbability of 0, r is a random number uniformly distributed between 0 and 1;
s2.9: let p be p +1 and return to step S2.8 until the iteration reaches the preset number K, and obtain the jth iteration result x (j) Then, the next operation is carried out;
s2.10: j is equal to j +1 and returns to step S2.7 until the iteration reaches the preset number max, and the max-th iteration result x (max) An estimate of the transmitted information;
s3: constructing an MCMC-L decoding algorithm of the LDPC by a generating matrix G, a receiving signal y and a preset expansion path number L; the MCMC-L decoding algorithm for constructing the LDPC specifically comprises the following steps:
s3.1: initializing x with all-zero vectors 1 =x 2 =...=x L (ii) a Wherein x is u The binary information vector on the u-th path is represented by u being more than or equal to 1 and less than or equal to L, and L is the number of preset expansion paths;
s3.2: setting the total times of the first iteration counter as K, and enabling an iteration variable i of the first iteration counter to be 1;
s3.3: setting the total times of a second iteration counter as L, and enabling an iteration variable p of the second iteration counter to be 1;
s3.4: the value is assigned according to equation (9):
in the formula (9), the reaction mixture is,represents the vector x p The ith bit-flipped result of (1), x p Representing a binary information vector, x, on the p-th path L+p Representing a binary information vector on the L + p path;
s3.5: setting the total times of the third iteration counter to be 2L, and enabling an iteration variable j of the third iteration counter to be 1;
s3.6: the calculation is performed according to equation (10):
d j '=||y-Gx j || 2 2 (10)
in the formula (10), d j ' is a measure of the binary information vector on each path to obtain the path that needs to be reserved finally; x is the number of j Representing a binary information vector on the jth path;
s3.7: for d obtained j ' (i ═ 1, 2.., K) is sorted in ascending order, and the sorting result is stored in a vector s for obtaining L paths which need to be reserved finally;
s3.8: j is set as j +1, the process returns to the step S3.6 until iteration reaches the preset time 2L, and then the next operation is carried out;
s3.9: the value is assigned according to equation (11):
in the formula (11), the reaction mixture is,denotes the s th p A binary information vector on a path;
s3.10: making p equal to p +1, returning to the step S3.4 until iteration reaches a preset number L, and then entering the next operation;
s3.11: i is equal to i +1, and the process returns to step S2.3 until the iteration reaches the preset number K, then the binary information vector x on the 1 st path 1 Is an estimate of the transmitted information.
2. The method of constructing an LDPC decoder according to claim 1 based on a markov monte carlo method, wherein: the transmission formula of the LDPC decoder is as follows:
y=Gx+n (12)
in the formula (12), the vector n ∈ C N Each term obeys a mean value of 0 and a variance of σ 2 X is a K × 1-dimensional binary information vector, y is an N × 1-dimensional reception vector, N represents a code length, and K represents an information bit length.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811176885.5A CN109495114B (en) | 2018-10-10 | 2018-10-10 | LDPC decoder construction method based on Markov Monte Carlo method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811176885.5A CN109495114B (en) | 2018-10-10 | 2018-10-10 | LDPC decoder construction method based on Markov Monte Carlo method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109495114A CN109495114A (en) | 2019-03-19 |
CN109495114B true CN109495114B (en) | 2022-07-26 |
Family
ID=65690253
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811176885.5A Active CN109495114B (en) | 2018-10-10 | 2018-10-10 | LDPC decoder construction method based on Markov Monte Carlo method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109495114B (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101322328A (en) * | 2005-11-18 | 2008-12-10 | 高通股份有限公司 | Reduced complexity detection and decoding for a receiver in a communication system |
CN103746731A (en) * | 2014-01-21 | 2014-04-23 | 电子科技大学 | Probability calculation-based multiple input multiple output detector and detection method |
WO2017100689A1 (en) * | 2015-12-10 | 2017-06-15 | University Of Utah Research Foundation | Soft-information moderation for mimo detectors |
-
2018
- 2018-10-10 CN CN201811176885.5A patent/CN109495114B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101322328A (en) * | 2005-11-18 | 2008-12-10 | 高通股份有限公司 | Reduced complexity detection and decoding for a receiver in a communication system |
CN103746731A (en) * | 2014-01-21 | 2014-04-23 | 电子科技大学 | Probability calculation-based multiple input multiple output detector and detection method |
WO2017100689A1 (en) * | 2015-12-10 | 2017-06-15 | University Of Utah Research Foundation | Soft-information moderation for mimo detectors |
Non-Patent Citations (2)
Title |
---|
Approaching MIMO capacity using bitwise Markov Chain Monte Carlo detection;Rong-rong Chen 等;《IEEE Transactions on Communications》;20100208;第58卷(第2期);全文 * |
Markov Chain Monte Carlo Based Detection for Two-Dimensional Intersymbol Interference Channels;Mahdi Shaghaghi 等;《IEEE Transactions on Magnetics》;20101203;第47卷(第2期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN109495114A (en) | 2019-03-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8326213B2 (en) | Encoding low density parity check (LDPC) codes through an LDPC decoder | |
US7539920B2 (en) | LDPC decoding apparatus and method with low computational complexity algorithm | |
CN108847848B (en) | BP decoding algorithm of polarization code based on information post-processing | |
US8099645B2 (en) | LDPC codes and stochastic decoding for optical transmission | |
CN108964669B (en) | LDPC code quadratic programming decoding method based on degree decomposition and alternative multiplier method | |
CN105763203B (en) | Multi-element LDPC code decoding method based on hard reliability information | |
CN107204780B (en) | Merging BP decoding algorithm and device of polar-LDPC (Low Density parity check) concatenated code | |
US20070094568A1 (en) | Method for updating check node in low density parity check decoder | |
US10892783B2 (en) | Apparatus and method for decoding polar codes | |
JP2012151839A (en) | Method for performing soft decision decoding of euclidean space reed-muller code | |
CN114244375A (en) | LDPC normalized minimum sum decoding method and device based on neural network | |
CN102811065B (en) | Mini-sum decoding correcting method based on linear minimum mean error estimation | |
CN110166171A (en) | Multielement LDPC code compensates high-performance decoding scheme based on the segmented of EMS | |
US20070083802A1 (en) | Broadcast message passing decoding of low density parity check codes | |
CN106656209B (en) | Cascade code method for correcting synchronous errors by adopting iterative decoding | |
CN110995279A (en) | Polarization code combined SCF spherical list overturning decoding method | |
CN106998240A (en) | A kind of interpretation method and decoder | |
CN109495114B (en) | LDPC decoder construction method based on Markov Monte Carlo method | |
CN101106437A (en) | A decoding method for limited geometrical low density checksum code | |
CN112165338A (en) | Estimation method for interleaving relation of convolutional code random interleaving sequence | |
CN115276668A (en) | LDPC code hybrid decoding method based on CRC | |
CN114448570B (en) | Deep learning decoding method of distributed joint information source channel coding system | |
CN113014271B (en) | Polarization code BP decoding method for reducing turnover set | |
CN114374397A (en) | Method for three-dimensional Turbo product code decoding structure and optimizing iteration weight factor | |
CN110730006B (en) | LDPC code error correction method and error correction module for MCU |
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 |