CN106169935A - With reliability for the low density parity check code reliability propagation interpretation method guided - Google Patents
With reliability for the low density parity check code reliability propagation interpretation method guided Download PDFInfo
- Publication number
- CN106169935A CN106169935A CN201610551321.XA CN201610551321A CN106169935A CN 106169935 A CN106169935 A CN 106169935A CN 201610551321 A CN201610551321 A CN 201610551321A CN 106169935 A CN106169935 A CN 106169935A
- Authority
- CN
- China
- Prior art keywords
- variable point
- checkpoint
- reliability
- ldpc code
- code
- 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
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
- H03M13/1108—Hard decision decoding, e.g. bit flipping, modified or weighted bit flipping
Landscapes
- Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Error Detection And Correction (AREA)
Abstract
With reliability for the low density parity check code reliability propagation interpretation method guided, belong to communication data transfer and store with data.Comprise the following steps: carry out LDPC code decoded operation initial work, calculate channel initial information L of each variable pointch, and the value setting the Reliability Index of each variable point and checkpoint is equal to 1;The Reliability Index of each checkpoint in LDPC code is updated operation;The LLR value of LDPC code checkpoint is calculated;The Reliability Index of each variable point in LDPC code is updated operation;The LLR value of variable point in LDPC code is weighted;To variable point posteriority LLR value weighted calculation in LDPC code;According to the positive and negative situation of variable point posteriority LLR value, carry out the hard decision operation of code word bits successively, obtain the code word of a length of N;Decoding result is judged.
Description
Technical field
The invention belongs to communication data transfer and technical field of data storage, relate to a kind of with low density parity check code joint
Point reliability is that the reliability guided propagates (Reliability-Wise Belief Propagation is called for short RW-BP) decoding side
Method.
Background technology
Low-density checksum (Low Density Parity Check is referred to as LDPC) code is the earliest by R.Gallager
Proposed in 1962, but the present invention is not the most arousing attention.But at Mackay and Neal in 1996
(D.J.C.MacKay,R.M.Neal.Near Shannon limit performance of low-density parity-
Check codes.Electron.Lett., 1996,33 (8): 1645-1646.) rediscover LDPC code, they prove
Use reliability to propagate (Belief Propagation, referred to as BP) interpretation method in the case of code length is longer it is decoded
Time, this yard has the excellent properties of close shannon limit.The excellent error-correcting performance that LDPC code is had because of it since then, relatively low translates
The advantages such as code method complexity, have obtained research widely and have been applied in several scenes, such as Wimax standard, UWB
System, satellite communication system, DTV broadcasting system, disk storage system and solid hard disk storage system etc..
One LDPC code both can be represented by the check matrix of a M × N, it is also possible to is schemed by Tanner
(R.M.Tanner.A recursive approach to low complexity codes.IEEE
Trans.Inform.Theory, 1981,27 (5): 533-547) represent, wherein the annulus in Tanner figure represents variable point
With each row one_to_one corresponding of LDPC check matrix, and represent each row one_to_one corresponding of checkpoint and LDPC code with square.Now
The code length of this LDPC code is N, and code check is (N-M)/N.In applying due to reality, code length N of LDPC code is the most limited, so
Have led to the appearance necessarily having loop configuration in the Tanner figure corresponding with this code-phase.In iterative decoding process, likelihood is believed
Breath will be flowed between the individual nodes by these loops, the likelihood information that so each node sends after iteration several times
Between will be the most separate, during then this has just run counter to BP method, the likelihood information for different nodes should be separate
Requirement, its consequence be wrong information transmission with superpose, thus cause decode effect deterioration decoding error occurs then
Floor.
In order to eliminate the loop configuration negative effect to interpretation method in LDPC code, Martin J.Wainwright teaches
Team proposes tree-shaped weighting BP interpretation method (Tree Reweighted BP, referred to as TRW-BP).The method is first repeatedly
Probability (the Edge Appearance that there is line on Tanner figure between every pair of variable point and checkpoint is calculated during Dai
Probabilities, referred to as EAPs), then the likelihood information that result of calculation is updated to node as weighted value was calculated
Cheng Zhong, carrys out the improvement of implementation method performance with this.The method can improve the performance of traditional BP interpretation method, but due to EAPs
Calculating sufficiently complex, limit the application of the method.In order to reduce the complexity of TRW-BP method, scholars propose one
Normalization weighting BP (Uniformly reweighted BP is called for short URW-BP) method.The method is by LDPC code, each becomes
The weighted value of amount point distributing uniform, thus avoid the calculating of EAPs.But regular LDPC code is only being decoded by the method
In time, show outstanding, and its performance is weaker than traditional BP interpretation method when decoding abnormal LDPC code, and this point limits it
Range of application.
Research shows, the reason that LDPC code decoding error floor occurs mainly has two: 1) coding codeword minimum range mistake
Little;2) LDPC check matrix exists a kind of special loop configuration trap collection (M.Ivkovic,
S.K.Chilappagari,and B.Vasic.Eliminating Trapping Sets in Low-Density Parity-
Check Codes by Using Tanner Graph Covers,IEEE Trans.Inform.Theory,2008,54(8):
3763-3768)。
Therefore to improve the performance of interpretation method, it is possible to the trap collection occurred during according to decoding, design interpretation method
Update rule, overcome the negative effect of trap set pair interpretation method with this, correct this trap collection the most further.Arizona is big
Learn Vasic and teach limited dictionary iterative decoding (Finite Alphabet Iterative Decoding, the abbreviation that team proposes
For FAID) a kind of embodiment of the most this thinking of method.Result shows, the method substantially shows in floor area and is better than BP decoding
The performance of method and be easy to hardware and realize.But it is the rule of 3 that the method is merely capable of on binary symmetric channel variable point degree
Then LDPC code decodes.
Understand from the angle of node, some error node occurred during LDPC code decoding can to connected other
Node sends error message, and owing to this node is difficult to be repaired, therefore cause the decline of interpretation method constringency performance.These are wrong
By mistake node definition is unreliable node and formulates corresponding more New Policy based on higher renewal priority, also is able to improvement undoubtedly
Method performance.Interpretation method based on this type of thought has propagates (Node-wise with node for the residue reliability guided
Residual Belief Propagation, referred to as NW-RBP) method.
The core of the method is the determination of unreliable node.First log-likelihood ratio (the Log to each node
Likelihood Rate, referred to as LLR) surplus of value calculates, shown in equation below:
R (m)=| | L (m)new-L(m)old| |,
Wherein m represents m-th node, L (m)oldIt is the LLR value before iteration, L (m)newIt it is the LLR value after iteration.When
During surplus maximum corresponding to one node, then it is assumed that this node is insecure, therefore is next to give this node more
High renewal priority.In order to unreliable node is determined more accurately, Zhongshan University professor Liu Xingcheng is to NW-RBP side
Method is further improved, it is proposed that RBP (Oscillating Variable nodes based based on concussion variable point
RBP, referred to as OV-RBP) method.
These RBP methods have obvious advantage in quickening LDPC code decoding convergence rate, and the method is by right simultaneously
The correction of part trap collection, it is also possible to be effectively improved decoding performance.But have a disadvantage in that the upper of method operation complexity
Rise, and do not determine these unreliable nodes from the angle of structure.
Summary of the invention
It is an object of the invention to provide the error-correcting performance that can improve BP method in high SNR region, can be reliable to data
Property the system that has higher requirements in be applied a kind of be that the low density parity check code reliability guided is propagated and translated with reliability
Code method (Reliability-Wise Belief Propagation method, referred to as RW-BP method).
For the ease of understanding following concrete formula, be first given will relate to english abbreviation, mathematical symbol and
The connotation of labelling:
1)vjRepresent jth variable point;
2)ciRepresent i-th checkpoint;
3)Sv(j i) represents variable node vjTo check-node ciThe node point reliability index of transmission;
4)Sc(i j) represents check-node ciTo variable node vjThe node point reliability index of transmission;
5)LchJ () represents and variable node vjLog-likelihood ratio (LLR) value of corresponding channel original state;
6)Lc(i j) represents check-node ciTo variable node vjThe LLR value of transmission;
7)Lv(j i) represents variable node vjTo check-node ciThe LLR value of transmission;
8)Represent and vjThe neighbours' checkpoint set being connected;
9)Represent and remove checkpoint ciWith variable point vjThe neighbours' checkpoint set being connected;
10)Represent and ciThe neighbours' variable point set being connected;
11)Represent and remove variable point vjWith variable point ciThe neighbours' variable point set being connected;
12)Sv(c)=k, k ∈ [1,2,3] represents the Reliability Index sending independent variable point (checkpoint), and k represents reliability
Magnitude;
13)Lv(c)Represent the LLR value sending independent variable point (checkpoint);
14)δk={ Sv(c)=k, Lv(c)Representing the external information of each node, k ∈ [1,2,3] represents taking of Reliability Index
Value, comprises δ the most in this method1, δ2And δ3Three class external informations;
15) sign () expression takes symbol manipulation;
16) tanh () represents tangent cosine function;
17)tanh-1() represents the inverse function of tangent cosine function;
18) min () represents that taking minima operates;
19) ρ (i, j) expression and Lc(i, j) corresponding weighter factor;
20) α represents stepped parameter.
The present invention comprises the following steps:
Step 1 carries out LDPC code decoded operation initial work, calculates channel initial information L of each variable pointch, and set
Fixed each variable point and Reliability Index (the i.e. S of checkpointvAnd Sc) value be equal to 1;
Step 2 is updated operation to the Reliability Index of each checkpoint in LDPC code;
The LLR value of LDPC code checkpoint is calculated by step 3, and concrete operation will be carried out according to equation below;
Step 4 is updated operation to the Reliability Index of each variable point in LDPC code;
The LLR value of variable point in LDPC code is weighted by step 5, and concrete operations will be carried out according to the following formula:
Wherein (s, value j) can have equation below to obtain to weighter factor ρ
ρ (s, j)=1.0+ [Sc(s,j)-3]×α
Wherein 0≤α < 0.5.
Step 6 is to variable point posteriority LLR value weighted calculation in LDPC code, and concrete operations are carried out according to the following formula:
Step 7, according to the positive and negative situation of variable point posteriority LLR value, carries out the hard decision operation of code word bits successively, obtains
Code word W=[the w of a length of N1,w2,...wj,...,wN], it may be assumed that
For jth (1≤j≤N), individual variable point, if sign is (Lv(j)) < 0, then wj=1;Otherwise, wj=0.
Decoding result is judged by step 8, if code word W and LDPC check matrix H meet W H=0, then should
Code word is correct, and output this code word of W also stops decoding;Otherwise, return step 2 and proceed decoded operation.
In step 2, the described concrete grammar that the Reliability Index of each checkpoint in LDPC code is updated operation
Can be: any one checkpoint c in definition LDPC codei, and the set that connected neighbours' variable point is formed
Definition node vjIt is belonging to setIn any one variable point, then this variable point sends to checkpoint ciReliability
Index is Sv(j,i).Definition node v simultaneouslysIt is belonging to setIn another aleatory variable point, the most corresponding thereto
Transmission to ciReliability Index be Sv(s,i).Therefore, from ciSend to vjCheckpoint Reliability Index can be by following public
Formula obtains:
In step 4, the described concrete grammar that the Reliability Index of each variable point in LDPC code is updated operation
Can be: assuming that LDPC code has aleatory variable point vj, and any neighbours checkpoint c being connected with this nodeiBroadly fall into setDefinition checkpoint c simultaneouslysIt is belonging to setAny neighbours' checkpoint, then it is sent to vjCheckpoint can
It is S by degree indexc(s, j), equidirectional on LLR value be Lc(s,j).With vjThe LLR value of corresponding channel original state is Lch
(j).From vjSend to ciReliability Index Sv(j, value i) can be calculated by below equation, it may be assumed that
For arbitrarilyIf Lc(s, symbol j) and LchJ the symbol of () is unequal, it may be assumed that
sign(Lc(s,j))≠sign(Lch(j))
Then Sv(j, value i) subtracts 1 until it is equal to 1;
Otherwise, Sv(j, value i) adds 1 until it is equal to 3.
Owing to RBP method has obvious advantage in quickening LDPC code decoding convergence rate, the method is by right simultaneously
The correction of part trap collection, it is also possible to be effectively improved decoding performance.But have a disadvantage in that the rising of method operation complexity,
And do not determine these unreliable nodes from the angle of structure.The present invention is just belonging to the variable point of trap collection in view of those
Being the variable point of those generation initial errors, they form loop configuration by coupled neighbours' checkpoint.Carry out
During decoding, these variable points ceaselessly send error message to neighbours' checkpoint, owing to the existence of loop configuration is then entered
One step exacerbates the transmission of mistake and adds up, thus hinders the realization of correct decoding.The most also can be by owning on trap collection
Node and the neighbor node being affected by are referred to as unreliable node.So update operation in conjunction with weighting, alleviate unreliable joint
Point weight in an iterative process, alleviates error propagation phenomenon with this, improves LDPC code at high s/n ratio (Signal to
Noise, referred to as SNR) performance in region.
Accompanying drawing explanation
Fig. 1 is the Reliability Index update status of T (3,2) trap collection and other coupled nodes;
Fig. 2 is that on awgn channel, code check is 0.5, when code length is the different interpretation method of regular LDPC code employing of 504 bits
BER curve figure;
Fig. 3 is that on awgn channel, code check is 0.5, and code length is that the abnormal LDPC code of 504 bits uses different interpretation method
Time BER curve figure;
Fig. 4 is that on disk storage channel, code check is 0.8, and code length is that the protograph LDPC code of 5632 bits uses difference decoding
BER curve figure during method;
Fig. 5 is that on disk storage channel, code check is 0.8, and code length is that the protograph LDPC code of 2816 bits uses difference decoding
BER curve figure during method.
Detailed description of the invention
Following example will the present invention is further illustrated in conjunction with accompanying drawing.
To unreliable node, Reliability Index and the definition of novel external information in the present invention given below.
First it is the definition to unreliable node:
1) for variable point, if external information of receiving is different from its initial prior information symbol during its decoding, then its
It it is unreliable variable point;
2) for all nodes (variable point and checkpoint), if it is connected with unreliable node, it is also unreliable
's.
In order to more clearly reflect the reliability of each node, this concept of Reliability Index is proposed.First, will
It is unreliable that the reliability of node is divided into, basic reliable and reliable three magnitudes.Corresponding with these 3 magnitudes is that reliability refers to
Number Sv(c)=k, k ∈ [1,2,3], represents the value of Reliability Index here with k, and it takes different integers between 1 to 3 respectively
With the reliability that corresponding above three is different.
The LLR value that the Reliability Index of different nodes sends with it is combined by RW-BP method be updated and
Transmission.Therefore, being different from traditional BP interpretation method, the external information of each node is represented by δ herek={ Sv(c)=k,
Lv(c)}.Because of Reliability Index round numbers between 1 to 3, add up to and have three kinds of value mode, therefore external information also can be divided three classes,
I.e. δ1, δ2And δ3.It should be noted that Reliability Index not only represents the reliability of node, it is also possible to represent that this node sends
The reliability of LLR value, the LLR value that unreliable node sends after all is it is also assumed that be insecure.So change through several times
In Dai Hou, Tanner figure, each node just can know the reliable of neighbor node and its LLR value by means of Reliability Index
Degree, thus tell low reliability node and LLR value, in order to reduce its negative effect to updating operation at next step.
In the present invention, the renewal of Reliability Index operates and is illustrated below:
The renewal operation of Reliability Index is divided into LDPC code checkpoint Reliability Index to update operation and variable point reliability
Index update two kinds of concrete operations of operation.
Checkpoint Reliability Index is updated and operates it essentially according to following method:
Any one checkpoint c in definition LDPC codei, and the set that connected neighbours' variable point is formedDefinition node vjIt is belonging to setIn any one variable point, then this variable point sends to checkpoint ciCan
It is S by degree indexv(j,i).Definition node v simultaneouslysIt is belonging to setIn another aleatory variable point, then with its phase
Corresponding transmission is to ciReliability Index be Sv(s,i).Therefore, from ciSend to vjCheckpoint Reliability Index can be by such as
Lower formula obtains:
Variable point reliability index update is operated it essentially according to following method:
Assuming that LDPC code has aleatory variable point vj, and any neighbours checkpoint c being connected with this nodeiBroadly fall into collection
CloseDefinition checkpoint c simultaneouslysIt is belonging to setAny neighbours' checkpoint, then it is sent to vjVerification
Point reliability index is Sc(s, j), equidirectional on LLR value be Lc(s,j).With vjThe initial LLR value of corresponding channel is Lch
(j).From vjSend to ciReliability Index Sv(j, value i) can be calculated by below equation, it may be assumed that
For arbitrarilyIf Lc(s, symbol j) and LchJ the symbol of () is unequal, it may be assumed that
sign(Lc(s,j))≠sign(Lch(j)) (3)
Then Sv(j, value i) subtracts 1 until it is equal to 1;Otherwise, Sv(j, value i) adds 1 until it is equal to 3.
Fig. 1 is shown that when carrying out the transmission of all-zero code word, T (3, a 2) trap collection and coupled its
The update status of his node external information when third time iteration.Wherein annulus represents the most correct variable point, and stain represents
The variable point of initial error, blank square represent meet verification checkpoint and black box with thumb down foot verification checkpoint.
As can be seen in the figure, a total of three classes of external information that RW-BP method sends at the 3rd iteration variations per hour point, i.e. δ1, δ2And δ3.And
Checkpoint then can judge neighbours' variable point and its LLR value sent according to the difference of the external information kind received
Reliability.Such as independent variable point v4Send to checkpoint c4LLR value because v4It is connected with T (3,2) trap collection, therefore from
c4Side looks up, and comes from v4LLR value be insecure;But for c1For, because v4It is the most correct variable point, and
It is able to receive that the correct information sent from other nodes, the most now v4This LLR value sent is reliable.Since
Have been able to distinguish node by Reliability Index and its LLR value sent is the most reliable, the most just can use for reference and add
Power BP method reduces unreliable node and its LLR value sent negative effect during information updating.
In the present invention, the formulation of LLR value calculating operation is as follows:
For variable point vjWith checkpoint ci, the formulation of the LLR value computational methods of checkpoint is as follows:
For variable point vjWith checkpoint ci, the LLR value weight computation method of variable point is as follows:
Variable point vjPosteriority LLR value weight computation method as follows:
The value of the weighter factor in above-mentioned formula (2) and (3) can be obtained by following equation:
ρ (i, j)=1.0+ [Sc(i,j)-3]×α (7)
Wherein the span of α is defined between 0 to 0.5, i.e. 0≤α < 0.5.
The present invention on awgn channel with disk storage channel on the performance comparison of log-domain BP (Log-BP) method such as
Under:
Fig. 2 Yu Fig. 3 illustrates the present invention and traditional Log-BP method bit error rate (Bit on awgn channel
Error Rate, referred to as BER) correlation curve figure.Fig. 4 Yu Fig. 5 illustrates the present invention to be believed in disk storage with Log-BP method
BER correlation curve figure on road.It is all wherein that the rule of 504 bits is entered with abnormal LDPC code respectively by code length on awgn channel
Row decoding performance test, being then respectively adopted code length on disk storage channel is 5632 bits and 2816 bits protographs
Abnormal LDPC code is as test pattern.The maximum times of LDPC code decoding iteration is all 20 times.Found by contrast, RW-BP
Multiple different types of LDPC code can be decoded in multiple systems by method, and can improve the error correction of LDPC code
Energy.
Specific embodiment given below, specifically comprises the following steps that
Step 1 carries out LDPC code decoded operation initial work, calculates channel initial information L of each variable pointch, and set
Fixed each variable point and Reliability Index (the i.e. S of checkpointvAnd Sc) value be equal to 1.
Step 2 is updated operation to the Reliability Index of each checkpoint in LDPC code.
Any one checkpoint c in definition LDPC codei, and the set that connected neighbours' variable point is formedDefinition node vjIt is belonging to setIn any one variable point, then this variable point sends to checkpoint ci's
Reliability Index is Sv(j,i).Definition node v simultaneouslysIt is belonging to setIn another aleatory variable point, then with it
Corresponding transmission is to ciReliability Index be Sv(s,i).Therefore, from ciSend to vjCheckpoint Reliability Index can be by
Equation below obtains:
The LLR value of LDPC code checkpoint is calculated by step 3, and concrete operation will be carried out according to equation below.
Step 4 is updated operation to the Reliability Index of each variable point in LDPC code.
Assuming that LDPC code has aleatory variable point vj, and any neighbours checkpoint c being connected with this nodeiBroadly fall into collection
CloseDefinition checkpoint c simultaneouslysIt is belonging to setAny neighbours' checkpoint, then it is sent to vjVerification
Point reliability index is Sc(s, j), equidirectional on LLR value be Lc(s,j).With vjThe LLR value of corresponding channel original state
For Lch(j).From vjSend to ciReliability Index Sv(j, value i) can be calculated by below equation, it may be assumed that
For arbitrarilyIf Lc(s, symbol j) and LchJ the symbol of () is unequal, it may be assumed that
sign(Lc(s,j))≠sign(Lch(j)) (11)
Then Sv(j, value i) subtracts 1 until it is equal to 1;
Otherwise, Sv(j, value i) adds 1 until it is equal to 3.
The LLR value of variable point in LDPC code is weighted by step 5, and concrete operations will be carried out according to the following formula.
Wherein (s, value j) can have equation below to obtain to weighter factor ρ
ρ (s, j)=1.0+ [Sc(s,j)-3]×α (13)
Wherein 0≤α < 0.5.
Step 6 is to variable point posteriority LLR value weighted calculation in LDPC code, and concrete operations are carried out according to the following formula.
Step 7, according to the positive and negative situation of variable point posteriority LLR value, carries out the hard decision operation of code word bits successively, obtains
Code word W=[the w of a length of N1,w2,...wj,...,wN], it may be assumed that
For jth (1≤j≤N), individual variable point, if sign is (Lv(j)) < 0, then wj=1;Otherwise, wj=0.
Decoding result is judged by step 8, if code word W and LDPC check matrix H meet W H=0, then should
Code word is correct, and output this code word of W also stops decoding;Otherwise, return step 2 and proceed decoded operation.
Claims (3)
1. with reliability for the low density parity check code reliability propagation interpretation method guided, it is characterised in that include following step
Rapid:
Step 1 carries out LDPC code decoded operation initial work, calculates channel initial information L of each variable pointch, and set every
Individual variable point and Reliability Index (the i.e. S of checkpointvAnd Sc) value be equal to 1;
Step 2 is updated operation to the Reliability Index of each checkpoint in LDPC code;
The LLR value of LDPC code checkpoint is calculated by step 3, and concrete operation will be carried out according to equation below;
Step 4 is updated operation to the Reliability Index of each variable point in LDPC code;
The LLR value of variable point in LDPC code is weighted by step 5, and concrete operations will be carried out according to the following formula:
Wherein weighter factor ρ (s, value j) can have equation below to obtain:
ρ (s, j)=1.0+ [Sc(s,j)-3]×α
Wherein 0≤α < 0.5;
Step 6 is to variable point posteriority LLR value weighted calculation in LDPC code, and concrete operations are carried out according to the following formula:
Step 7, according to the positive and negative situation of variable point posteriority LLR value, carries out the hard decision operation of code word bits successively, obtains length
Code word W=[w for N1,w2,...wj,...,wN], it may be assumed that
For jth (1≤j≤N), individual variable point, if sign is (Lv(j)) < 0, then wj=1;Otherwise, wj=0;
Decoding result is judged by step 8, if code word W and LDPC check matrix H meet W H=0, then this code word
Being correct, output this code word of W also stops decoding;Otherwise, return step 2 and proceed decoded operation.
2. as claimed in claim 1 with reliability for the low density parity check code reliability propagation interpretation method guided, its feature
It is in step 2, described the Reliability Index of each checkpoint in LDPC code is updated operation method particularly includes: be fixed
Any one checkpoint c in justice LDPC codei, and the set that connected neighbours' variable point is formedDefinition node
vjIt is belonging to setIn any one variable point, then this variable point sends to checkpoint ciReliability Index be Sv
(j,i);Definition node v simultaneouslysIt is belonging to setIn another aleatory variable point, transmission the most corresponding thereto
To ciReliability Index be Sv(s, i), from ciSend to vjCheckpoint Reliability Index obtained by equation below:
3. as claimed in claim 1 with reliability for the low density parity check code reliability propagation interpretation method guided, its feature
It is in step 4, described the Reliability Index of each variable point in LDPC code is updated operation method particularly includes: be false
Determine that LDPC code has aleatory variable point vj, and any neighbours checkpoint c being connected with this nodeiBroadly fall into setWith
Shi Dingyi checkpoint csIt is belonging to setAny neighbours' checkpoint, then it is sent to vjCheckpoint Reliability Index
For Sc(s, j), equidirectional on LLR value be Lc(s, j), with vjThe LLR value of corresponding channel original state is LchJ (), from vj
Send to ciReliability Index Sv(j, value i) is calculated by below equation, it may be assumed that
For arbitrarilyIf Lc(s, symbol j) and LchJ the symbol of () is unequal, it may be assumed that
sign(Lc(s,j))≠sign(Lch(j))
Then Sv(j, value i) subtracts 1 until it is equal to 1;
Otherwise, Sv(j, value i) adds 1 until it is equal to 3.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610551321.XA CN106169935B (en) | 2016-07-14 | 2016-07-14 | It take reliability as the low density parity check code reliability propagation interpretation method of guiding |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610551321.XA CN106169935B (en) | 2016-07-14 | 2016-07-14 | It take reliability as the low density parity check code reliability propagation interpretation method of guiding |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106169935A true CN106169935A (en) | 2016-11-30 |
CN106169935B CN106169935B (en) | 2019-07-09 |
Family
ID=58066310
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610551321.XA Expired - Fee Related CN106169935B (en) | 2016-07-14 | 2016-07-14 | It take reliability as the low density parity check code reliability propagation interpretation method of guiding |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106169935B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110690906A (en) * | 2019-09-29 | 2020-01-14 | 中国科学院微电子研究所 | Dynamic self-correction minimum sum decoding method and decoder based on same |
CN110830050A (en) * | 2019-11-27 | 2020-02-21 | 武汉虹信通信技术有限责任公司 | LDPC decoding method, system, electronic device and storage medium |
CN112702070A (en) * | 2020-12-29 | 2021-04-23 | 厦门大学 | Decoding optimization method of distributed joint source-channel coding system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101345601A (en) * | 2007-07-13 | 2009-01-14 | 华为技术有限公司 | Interpretation method and decoder |
US20140281823A1 (en) * | 2013-03-15 | 2014-09-18 | Pmc-Sierra Us, Inc. | System and method with reference voltage partitioning for low density parity check decoding |
US20150026541A1 (en) * | 2013-07-22 | 2015-01-22 | Nec Laboratories America, Inc. | Iterative Decoding for Cascaded LDPC and TCM Coding |
CN104467874A (en) * | 2014-12-24 | 2015-03-25 | 中山大学 | LDPC code dynamic scheduling decoding method based on vibration variable nodes |
-
2016
- 2016-07-14 CN CN201610551321.XA patent/CN106169935B/en not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101345601A (en) * | 2007-07-13 | 2009-01-14 | 华为技术有限公司 | Interpretation method and decoder |
US20140281823A1 (en) * | 2013-03-15 | 2014-09-18 | Pmc-Sierra Us, Inc. | System and method with reference voltage partitioning for low density parity check decoding |
US20150026541A1 (en) * | 2013-07-22 | 2015-01-22 | Nec Laboratories America, Inc. | Iterative Decoding for Cascaded LDPC and TCM Coding |
CN104467874A (en) * | 2014-12-24 | 2015-03-25 | 中山大学 | LDPC code dynamic scheduling decoding method based on vibration variable nodes |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110690906A (en) * | 2019-09-29 | 2020-01-14 | 中国科学院微电子研究所 | Dynamic self-correction minimum sum decoding method and decoder based on same |
CN110690906B (en) * | 2019-09-29 | 2023-06-02 | 中国科学院微电子研究所 | Dynamic self-correction minimum sum decoding method and decoder based on same |
CN110830050A (en) * | 2019-11-27 | 2020-02-21 | 武汉虹信通信技术有限责任公司 | LDPC decoding method, system, electronic device and storage medium |
CN110830050B (en) * | 2019-11-27 | 2023-09-29 | 武汉虹旭信息技术有限责任公司 | LDPC decoding method, system, electronic equipment and storage medium |
CN112702070A (en) * | 2020-12-29 | 2021-04-23 | 厦门大学 | Decoding optimization method of distributed joint source-channel coding system |
Also Published As
Publication number | Publication date |
---|---|
CN106169935B (en) | 2019-07-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Angarita et al. | Reduced-complexity min-sum algorithm for decoding LDPC codes with low error-floor | |
CN110572163B (en) | Method and apparatus for encoding and decoding LDPC code | |
US8370711B2 (en) | Interruption criteria for block decoding | |
Lian et al. | Learned belief-propagation decoding with simple scaling and SNR adaptation | |
US8291279B2 (en) | Memory-efficient LDPC decoder and method | |
US8504895B2 (en) | Using damping factors to overcome LDPC trapping sets | |
Xu et al. | XJ-BP: Express journey belief propagation decoding for polar codes | |
US20090319860A1 (en) | Overcoming ldpc trapping sets by decoder reset | |
US8099645B2 (en) | LDPC codes and stochastic decoding for optical transmission | |
US20200091933A1 (en) | Iterative decoding with early termination criterion that permits errors in redundancy part | |
Zhao et al. | Spatially coupled codes via partial and recursive superposition for industrial IoT with high trustworthiness | |
CN106169935B (en) | It take reliability as the low density parity check code reliability propagation interpretation method of guiding | |
Grinchenko et al. | Improving performance of multithreshold decoder over binary erasure channel | |
Cheng et al. | Encoder and list decoder of Reed-Solomon kernel based polar codes | |
CN104184480B (en) | A kind of LDPC for reducing complexity improves interpretation method | |
CN112491422A (en) | Gauss optimization-based bit flipping serial elimination list decoding method and system | |
Vellambi et al. | Results on the improved decoding algorithm for low-density parity-check codes over the binary erasure channel | |
Ghaffari et al. | Probabilistic gradient descent bit-flipping decoders for flash memory channels | |
Dupraz et al. | Evaluation of the robustness of LDPC encoders to hardware noise | |
Song et al. | A novel low-complexity joint coding and decoding algorithm for NB-LDPC codes | |
Kelley | Codes over graphs | |
Lechner et al. | Optimization of binary LDPC codes for the q-ary symmetric channel with moderate q | |
Das et al. | Multiplicity assignments for algebraic soft-decoding of Reed-Solomon codes using the method of types | |
Healy | Short-length low-density parity-check codes: construction and decoding algorithms | |
CN112470406B (en) | Ordering apparatus and method for basic check node processing for message passing decoding of non-binary codes |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190709 Termination date: 20200714 |