CN101227193A - Method and device for encoding and decoding low density check code - Google Patents
Method and device for encoding and decoding low density check code Download PDFInfo
- Publication number
- CN101227193A CN101227193A CNA2008100575361A CN200810057536A CN101227193A CN 101227193 A CN101227193 A CN 101227193A CN A2008100575361 A CNA2008100575361 A CN A2008100575361A CN 200810057536 A CN200810057536 A CN 200810057536A CN 101227193 A CN101227193 A CN 101227193A
- Authority
- CN
- China
- Prior art keywords
- perforation
- code
- check
- sequence table
- repetitive sequence
- 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
Abstract
The invention discloses a device and a method for checking code, coding and decoding with low density, which is used to realize data transmission, wherein the device comprises an LDPC coding device, a boring/repeating pattern generating device which is used to generate boring or repeating pattern according to a boring/repeating sequence table, the sequence of quantitative nodal point of code word is marked by the boring/repeating sequence table after a boring or repeating LDPC coding device is coded, the device also comprises a boring/repeating device which is used to bore or repeat the code word after the LDPC coding device is coded by the boring or repeating pattern which is generated according to the boring/repeating pattern generating device, and a code word decoding device after changing code rate which is used to decode the code word which is changed code rate. The device can effectively increase coding and decoding property, and is applied to the LDPC code which is constructed by any methods which comprise regular codes and irregular codes, and can generate random code rate and code and decode.
Description
Technical field
Error correcting technique and encoding and decoding technique during Digital Transmission in the present invention relates to communicate by letter, particularly a kind of error correction encoding/decoding device of loe-density parity-check code and method.
Background technology
Low-density check (Low-Density Parity Check-LDPC) sign indicating number is a kind of code character with very strong error correction antijamming capability, performance is better than Turbo code, have bigger flexibility and the lower flat characteristic of mistake (Error Floors), describe simply, strict theory analysis is had verifiability, decoding complexity is lower than Turbo code, and can realize parallel work-flow completely, hardware complexity is low, thereby suitable hardware is realized, throughput is big, has the high-speed coding ability.Compare with Turbo code, the LDPC sign indicating number has all shown the performance that it is more superior good on apart from property, low complex degree and high concurrency interpretation method.Therefore, think that extensively the LDPC sign indicating number can be used as the error correcting code of next generation communication network.New European satellite broadcast system DVB-S2, the same cascaded code of forming by LDPC sign indicating number and BCH code that adopts.
The LDPC sign indicating number is a kind of linear block codes, defines by check matrix H, and H is a sparse matrix, promptly has only quantity element seldom to be " 1 " in the check matrix, and major part all is " 0 ".Gallager has provided the definition of regular LDPC sign indicating number the earliest, three conditions below the check matrix H of specifically regular LDPC sign indicating number satisfies:
(1), H's whenever shows j " 1 ";
(2), every row of H has k " 1 ", k 〉=3;
(3), compare with the line number of H matrix, k and j are very little with code length.
The coding of LDPC is exactly to construct check matrix.It is heavy that k is called row, and j is called column weight, all equates if the column weight of every row all equates and the row of every row is heavy, claims that then the LDPC sign indicating number is a regular code, otherwise, be non-regular code.
In communication system; because different business has different service quality (Quality of Service-QoS) requirement; and channel status has time-varying characteristics; so in transmission course, need the degree of protection of chnnel coding to be done real-time adjustment according to the difference of type of service in channel status and the information source information; to reach the reconstruction quality of expectation, so just need a kind of error correction coding of rate-compatible.So that encode speed self-adaption changes in the network bandwidth and error code.
Multi code Rate of Chinese character research for LDPC also has a lot, but multi-method concentrates in the change of check matrix greatly, and these class methods are all bigger to the change of coding side and decoding end, and overall complexity is higher.The mode that has the certain methods utilization that the code word of transmission is bored a hole realizes, such as document 1 (J.Ha, J.Kim, and S.W.McLaughlin, " Rate-compatible puncturing of low-densityparity-check codes ", IEEE Trans.Inf.Theory, vol.IT-50, no.11, pp.2824-2836, Nov.2004.), document 2 (J.Ha and S.W.McLaughlin, " Optimalpuncturing distributions for rate-compatible low-density parity-check codes ", in Proc.Int.Symp.Information Theory, Yokohama, Japan, Jun/Jul.2003, p.233.) and document 3 (" Optimal puncturing of irregular low-density parity-checkcodes ", in Proc.IEEE Int.Conf.Communications, Anchorage, AK, May 2003, pp.3110-3114.) at irregular LDPC codes a kind of method for punching that distributes based on degree has been proposed, yet, these methods are suitable for the code character of endless, previous research generally is any selection to the code character of moderate-length, adopts this selection at random can make that the performance loss of LDPC sign indicating number is bigger under certain code check, perhaps can cause some indivedual code words not recover.
Summary of the invention
The present invention seeks to improve the encoding and decoding performance, thereby a kind of coding and decoding device and method of loe-density parity-check code are provided.
For achieving the above object, according to an aspect of the present invention, provide a kind of code device of loe-density parity-check code, having comprised:
The LDPC encoder is used for the data of information source transmission are carried out low density check code;
Perforation/repetitions patterns generating apparatus is used for generating perforation or repetitions patterns according to perforation/repetitive sequence table, the order of the variable node of code word after its middle punch/repetitive sequence table has identified perforation or repeated the LDPC encoder encodes;
Perforation/duplicating device, it is used for the perforation that generates according to perforation/repetitions patterns generating apparatus or the repetitions patterns code word after to the LDPC encoder encodes and bores a hole or repeat.
In the above-mentioned code device, comprise perforation/repetitive sequence table calculation element, it is used for according to the influence degree calculation punch/repetitive sequence table of variable node to the code word recovery.
In the above-mentioned code device, the influence degree that variable node recovers code word is that the check-node number of degrees that are connected with this code word variable node interrelate.
In the above-mentioned code device, perforation/repetitive sequence table calculation element adopts offline mode calculation punch/repetitive sequence table.
In the above-mentioned code device, comprising:
Perforation/repetitions patterns generating apparatus is used for distributing and described perforation/repetitive sequence table generation perforation/repetitions patterns according to goodness;
Wherein, goodness distribution table has shown that the number of degrees of each code word optimum of variable node under different code checks distribute.
In the above-mentioned code device, comprise goodness distribution calculation element, be used for calculating goodness and distribute based on the density evolution algorithm.
In the above-mentioned code device, goodness distribution calculation element adopts offline mode to calculate described goodness to distribute.
According to a further aspect in the invention, provide a kind of loe-density parity-check code decoding device corresponding to above-mentioned code device, comprised the LDPC decoder, it is used for decoding according to the perforation pattern that perforation/repetitions patterns generating apparatus generates.
In the above-mentioned decoding device, the LDPC decoder is when reducing code check, and the LDPC decoder at first calculates posterior probability, and the perforation/repetitions patterns that generates according to perforation/repetitions patterns generating apparatus is decoded then.
In the above-mentioned decoding device, the LDPC decoder merges and calculates posterior probability by the code word that repeats is carried out code word according to power ratio.
According to another aspect of the invention, provide a kind of data transmission system, comprised above-mentioned code device and above-mentioned decoding device.
According to another aspect of the invention, provide a kind of coding method of loe-density parity-check code, comprised following steps:
At first, the data of information source transmission are carried out low density check code;
Then, generate perforation/repetitions patterns according to perforation/repetitive sequence table, its middle punch/repetitive sequence table has identified the order of the variable node of code word behind perforation or the repeated encoding;
According to perforation/repetitions patterns the back code word of encoding is bored a hole or repeated.
Above-mentioned coding method comprises according to variable node the influence degree of code word recovery is calculated described perforation/repetitive sequence table.
Above-mentioned coding method, the influence degree that variable node recovers code word is that the check-node number of degrees that are connected with this variable node interrelate.
Above-mentioned coding method adopts offline mode to calculate described perforation/repetitive sequence table.
Above-mentioned coding method is calculated described perforation/repetitive sequence table and is comprised the steps:
1) the described perforation of initialization/repetitive sequence table is empty set P;
2) select a variable node i at random from variable node set V, described variable node i is incorporated among the P, from check-node set C, remove the check-node r that links to each other with i
Ix, deletion and described check-node r from V
IxThe variable node that links to each other;
3) repeating step 2), be empty set up to V, then V=V P;
4) if V is an empty set, then obtain perforation/repetitive sequence table P, if not empty set, then turn back to step 2).
Above-mentioned coding method distributes and described perforation/repetitive sequence table generation perforation/repetitions patterns according to goodness;
Wherein, described goodness distribution table shows that the number of degrees of each code word optimum of variable node under different code checks distribute.
Above-mentioned coding method is calculated described goodness based on the density evolution algorithm and is distributed.
Above-mentioned coding method is adopted offline mode to calculate described goodness and is distributed.
Above-mentioned coding method generates perforation/repetitions patterns according to perforation/repetitive sequence table, specifically comprises the steps:
Code length according to the code check calculation expectation of expecting;
Calculate the difference p of the code length of female sign indicating number code length and described expectation;
P element before selecting from described perforation/repetitive sequence table is designated 1, and all the other elements are 0, obtain perforation/repetitions patterns.
According to another aspect of the invention, provide a kind of loe-density parity-check code coding/decoding method, decode according to perforation/repetitions patterns.
Above-mentioned coding/decoding method when reducing code check, at first calculates posterior probability, decodes according to perforation/repetitions patterns then.
Above-mentioned coding/decoding method merges and calculates posterior probability by the code word that repeats is carried out code word according to power ratio.
According to another aspect of the invention, provide a kind of data transmission method, it is characterized in that, may further comprise the steps:
Coding method is encoded according to the preferred embodiment of the invention;
Coding/decoding method is decoded according to the preferred embodiment of the invention.
Compared with prior art, the present invention can reach following technique effect:
The present invention can improve the encoding and decoding performance effectively.When reducing code word, some code words of boring a hole make less to the performance impact of code word, and when increasing code word, increase some code words and make the performance of code word be greatly improved.Not changing check matrix, increasing under the situation of less encoding and decoding parts, can obtain the coding and decoding of the arbitrary bit rate of any LDPC code word, comprise regular code and non-regular code.
Description of drawings
Below, describe embodiments of the invention in conjunction with the accompanying drawings in detail, wherein:
Fig. 1 is based on the system construction drawing of the multi code Rate of Chinese character generation method that goodness distributes.
Fig. 2 is an arbitrary bit rate generation system structure chart of the present invention.
Fig. 3 is that goodness distribution table of the present invention generates schematic diagram.
Fig. 4 is that perforation of the present invention (repetition) sequence table generates schematic diagram.
Fig. 5 is a regular code coded system schematic diagram of the present invention.
Fig. 6 is a non-regular code coded system schematic diagram of the present invention.
Embodiment
The code check r that supposes a LDPC is
K is the length of information bit, and n is a code length, and then m=n-k is the length of check digit.K information bit i=[i
1, i
2... i
k] expression, through obtaining C=[c behind the basic LDPC sign indicating number coding
1, c
2... c
n] codeword vector, i.e. C=iG, G is a generator matrix, at this moment, can according to the state of channel and transmit data type adjust the redundancy of coding, to reach certain performance requirement.If channel status is relatively poor or the important increase degree of protection that needs of the data of transmission, then can increase redundancy, reduce code check; If channel status is better or the data of transmission are inferior important, then can reduce redundancy, increase code check.Under same channel condition, can design a kind of irregular LDPC sign indicating number with the density evolution algorithm.
Under the situation that the hypothesis of Infinite Cyclic is set up, the density evolution algorithm can pass through all possible density function of graph nodes tracking message, usually, under the condition that the number of degrees distribute and signal to noise ratio limits, carries out optimization to the channel thresholding or to code check.The channel thresholding is the value of channel parameter, and is infinite if number of iterations trends towards, and the probability that then is higher than this value trends towards 0.Optimization will reduce thresholding exactly as far as possible or improve code check as far as possible.
The optimization of LDPC sign indicating number goes for various types of channels, and verified be effective solution in following channel:
Partial response channels (having the Gaussian Profile characteristic);
Frequency selective fading (OFDM) (having the Gaussian Profile characteristic);
Many access channels and associating Multiuser Detection;
Associating AWGN and similar rayleigh fading channel (having the Gaussian Profile characteristic);
The high efficient band width modulation;
Minimum Shift Keying Modulation;
MIMO in the OFDM multi-carrier modulation imports more, multiple-output channel.
Utilize the density evolution algorithm, can can realize the progressive threshold value of reliable transmission in the hope of LDPC sign indicating number with degree of certainty distribution function, conversely, also can utilize the density evolution algorithm to optimize the degree distribution design of LDPC sign indicating number, thus the degree distribution function that acquisition has bigger progressive threshold value.
Based on above theory, document 1,2 and 3 proposes a kind of multi code Rate of Chinese character generation system that distributes based on goodness, and wherein goodness distribution table has shown the number of degrees of the optimum of each variable node of code word.Fig. 1 is the transmission system structure chart of this method, and this structure mainly is suitable for irregular LDPC codes.The data of an information source 100 of this system's transmitting terminal through a LDPC encoder 102 after, by the encoder bit rate that obtains expecting after punching machine 104 perforation, again by behind the modulator 106, be transmitted into receiving terminal through channel 108, receiving terminal comprises demodulator 110, the LDPC decoder 112 and the stay of two nights 114.Code check calculation element 120 receives the information source information of sending from information source; the protection demand that shows these information source data needs; code check calculation element 120 also receives the channel condition information that feeds back from receiving terminal simultaneously, and this channel condition information is used for reflecting the characteristic and the quality information of channel 108.Code check calculation element 120 comprehensive information source information and channel condition informations, calculation code speed is to satisfy the transmission demand of data.The generation of the multi code Rate of Chinese character of LDPC coding specifically is to be distributed by the goodness that goodness distribution calculation element 118 obtains this code check under, the pattern generating apparatus 122 of boring a hole then generates the perforation pattern each number of degrees under at random, and the foundation pattern of boring a hole is bored a hole and realized.In the decoding end decoding also is to realize according to the perforation pattern.In this prior art, adopt the perforation pattern of selecting at random to obtain can make that the performance loss of LDPC sign indicating number is bigger under certain code check, perhaps can cause some indivedual code words not recover.
The influence degree that the present invention recovers code word according to variable node is selected the perforation order.According to the density evolution algorithm, be d for the variable node number of degrees
v, the check-node number of degrees are d
cRegular code, variable node message average m
uUpdate rule as shown in Equation 1:
Wherein: m
U0It is the initial value of variable node message average;
m
u (l-1)It is the variable node message average that the l-1 time iteration obtains;
m
u (l)It is the variable node message average that the l time iteration obtains;
Because φ (x) is a decreasing function, and 0≤φ (x)≤1, as can be seen from the above equation,
So d
cBig more, then
More little, then
Big more, because φ (x) is continuous monotone decreasing, m as a result
u (l)More little, so when selecting the perforation order, should preferentially select the variable node that links to each other with the big check-node of the number of degrees.Though be that example draws this conclusion above with the Gaussian channel, this conclusion is applicable to other channel model too.
Based on prior art and above-mentioned theory analysis, the present invention proposes a kind of coding and decoding device of loe-density parity-check code, as shown in Figure 2.In the native system, the generation of the coding of arbitrary bit rate specifically is to be realized by perforation/repetitive sequence table 224, perforation/repetitions patterns generating apparatus 222 and perforation/duplicating device 204, and the decoding of rate-compatible code word is to be realized according to perforation/repetitions patterns by decoder 212.The code check that obtains according to code check calculation element 220 and perforation/repetitive sequence table 224 generate perforation or the repetitions patterns under this code check, its middle punch/repetitive sequence table obtains according to the number of degrees of the check-node that links to each other with variable node, and has represented the order of perforation or repetition variable node.For non-regular code, each node number of degrees difference.After calculate obtaining suitable code check by code check, can be according to perforation or the repetitions patterns under a predefined degree and this code check of perforation/repetitive sequence table 224 generation, preferably, increase goodness distribution table 218, distributing according to the goodness under this code check generates perforation or repetitions patterns again, as shown in phantom in FIG..Above-mentioned repeating is the situation that is applied to reduce code check.
Fig. 3 is that goodness distribution table of the present invention calculates schematic diagram.Owing to calculate complexity that goodness distributes than higher, need the long time of cost, and to a given LDPC sign indicating number, goodness under its each code check distributes and fixes, can calculate in advance, so in the codeword transmission process, carry out online calculation mode based on the density evolution algorithm except adopting, the present invention preferably adopts the goodness of 301 pairs of given LDPC sign indicating numbers of goodness distribution calculation element to distribute to carry out calculated off-line in advance, can in transmission course, directly inquire about obtaining like this, reduce time-delay.
Fig. 4 is that perforation of the present invention/repetitive sequence table calculates schematic diagram.Calculate same reason with goodness distribution table, the present invention is to the mode of a given LDPC sign indicating number except employing online calculation punch/repetitive sequence table in the codeword transmission process, preferably, perforation/repetitive sequence table shifts to an earlier date calculated off-line by perforation/repetitive sequence table calculation element 403, when transmission, directly inquire about, reduce time-delay.
Fig. 5 is a regular code coding side structure chart of the present invention.The transfer of data of an information source 500 is behind LDPC encoder 502, and code check calculation element 520 calculates the code rate that protect according to the type and the channel status of data source, and the data of 502 pairs of information sources 500 of encoder are carried out the LDPC coding.If increase the situation of code check, the then code check and the perforation sequence table 524 of calculating according to code check calculation element 520, perforation/repetition generating apparatus 522 generates corresponding perforation pattern, then the code word behind the coding is bored a hole and transmitted according to perforation pattern 522, if code word is designated 1 just do not pass in the perforation pattern, be designated 0 just be transferred to channel.If reduce the situation of code check, then according to the code check and the repetitive sequence table 524 of 520 calculating of code check calculation element, perforation/repetitions patterns generating apparatus 522 generates corresponding repetitions patterns.If in the repetitions patterns code word be designated 1 just repeat transmission, be designated 0 just do not repeat transmission.
Fig. 6 is the non-regular code coding side of a present invention structure chart.The transfer of data of an information source 600 is behind LDPC encoder 602, and code check calculation element 620 calculates the code rate that protect according to the type and the channel status of data source, and 602 pairs of data from information source 600 of encoder are carried out the LDPC coding.The goodness of searching under this code check according to the code check of code check calculation element 620 distributes 618 then.If increase the situation of code check, according to the perforation sequence table 624 under each number of degrees, the perforation pattern that perforation/repetitions patterns generating apparatus 622 generates under this code check, at last the code word behind the coding is bored a hole and transmitted according to the perforation pattern, if code word is designated 1 just do not pass in the perforation pattern, be designated 0 just be transferred to channel.If reduce the situation of code check, according to the repetitive sequence table 624 under each number of degrees, the repetitions patterns 622 that perforation/repetitions patterns generating apparatus 622 generates under this code check.If in the repetitions patterns code word be designated 1 just repeat transmission, be designated 0 just do not repeat transmission.
The performing step of component function involved in the present invention is described in detail in detail respectively below.
The code check calculation element is according to the channel status that feeds back, and for example the type of service that will transmit of interchannel noise parameter σ and information source calculates and satisfies the degree of protection that desired reconstruction quality condition lower channel should provide at least, just chnnel coding code check r.Generally speaking, the data based professional difference of information source has different transmission requirements, and the transmission requirement of information source is just transmitted the lowest bit error rate requirement that must guarantee, is designated as P
SeThe error rate P of chnnel coding
eBe the function of channel signal to noise ratio and code check, i.e. P
e=f (r, σ), think the requirement that guarantees information source, should make P
e≤ P
SeAnd chnnel coding is under different code checks and signal to noise ratio condition, and the error rate that can reach is also different.For the error rate P of LDPC sign indicating number under different code checks and different signal to noise ratios
e, the complexity of calculate is bigger, can be at first obtains each P under different code checks and the different signal to noise ratios by emulation
e, store in the look-up table, during transmission, can be in order to reach real-time by tabling look-up according to P
eObtain corresponding code check r.
Calculation procedure for regular code perforation/repetitive sequence table is as follows:
The LDPC of a rule of setting (n, j, k) sign indicating number, n is a code length, and j is the variable node number of degrees, and k is the check-node number of degrees.
Step 1: initialization, set set: V={1,2 ... n} represents the set of alternative variable node sequence number, C={1,2 ... m} represents the set of check-node sequence number, set P represents the set of the variable node sequence number selected, and just perforation/repetitive sequence table is initialized as empty set.
Step 2: select a variable node sequence number i at random from V, sequence number i is incorporated among the set P, promptly { i}, the set of the check-node sequence number that links to each other with variable node i is R to P=P ∪
i, from set C, remove set R
iIn element, promptly C=C R
i, simultaneously, corresponding to R
iIn each element r
Ix, with check-node r
IxThe set of the variable node that links to each other is S
Ix, with S
IxRemove the V from set, promptly V=V S
Ix
Step 3: repeating step 2 is an empty set up to V, then with V={1, and 2 ..., n} P.
Step 4: if V is an empty set, then obtain gathering P, finish; If not empty set, then turn back to step 2.
When regular code increased code check, the P that obtains was the perforation sequence table; When reducing code check, the P that obtains is the repetitive sequence table.
Increase the code check situation for regular code, the method that perforation/repetitions patterns generating apparatus generates the perforation pattern is as follows: the code check of the female sign indicating number of LDPC is
According to the top step sequence table P that obtains boring a hole
After, if the code check that reaches is r
p, then this moment, code length was
By calculating
The quantity that obtains boring a hole, wherein p is for needing the number of perforation, and p element before selecting from set P is designated 1, and these elements will not transmit, and remaining component identification is 0 and transmits.Promptly is the perforation pattern in this formation by 1 and 0 Serial No. of forming.
Reduce the code check situation for regular code, the method that perforation/repetitions patterns generating apparatus generates repetitions patterns is as follows: the code check of the female sign indicating number of LDPC is
After obtaining repetitive sequence table P according to top step, if the code check that reaches is r
p, then this moment, code length was
By calculating
Obtain the quantity of repetition, the code word number of p wherein for need repeating, p element before selecting from set P is designated 1 and follow after basic code word and repeating to transmit, and remaining component identification is 0.Promptly is repetitions patterns in this formation by 1 and 0 Serial No. of forming.
Non-regular code changes code check and calculates except relating to code check mentioned above, also relates to the perforation different with regular code/repetitive sequence table and calculates.Step for non-regular code calculation punch/repetitive sequence table is as follows:
Step 1: initialization: set set: V={1,2 ..., n} represents alternative variable node sequence number, C={1, and 2 ..., m} represents the set of check-node sequence number, set P1, and P2 ... Pd
vBe illustrated in the set of the variable node sequence number of having selected under each number of degrees, also promptly bore a hole or the repetitive sequence table, all be initialized as empty set.
Step 2: select a variable node v at random, if the number of degrees of node v are i, then v is incorporated among the set Pi, promptly { v}, the set of the check-node sequence number that links to each other with variable node v is R to Pi=Pi ∪
v, from set C, remove set R
vIn element, promptly C=C R
v, simultaneously, corresponding to R
vIn each element r
Vx, with check-node r
VxThe set of the variable node that links to each other is S
Vx, with S
VxRemove the V from set, promptly V=V S
Vx
Step 3: repeating step 2 is an empty set up to V, then with V={1, and 2 ..., n} (P1 ∪ P2 ∪ ... ∪ Pd
v).
Step 4: if V is an empty set, then obtain perforation or repetitive sequence table P1 under all number of degrees, P2 ... Pd
v, finish; If not empty set, then turn back to step 2.
For the situation of non-regular code increase code check, P1, P2 ... Pd
vBe the perforation sequence table, reduce the situation of code check for non-regular code, P1, P2 ... Pd
vIt is the repetitive sequence table.
For the situation of non-regular code increase code check, promptly the code check of the female sign indicating number of LDPC is
Required code check is r
p, r wherein
p＞r.When code check is r
pThe time, distribute according to goodness, calculate the variable node number of the required perforation of each number of degrees, p
i=V
i* δ
i, p
iFor the number of degrees are the number of the required perforation of variable node of i, V
iFor the number of degrees is total number of the variable node of i, δ
iFor the number of degrees is the ratio of the variable node perforation of i.According to the number of needs perforation under each number of degrees, from perforation sequence table P1, P2 ... Pd
vIn select the bit of required perforation separately, then perforation/repetitions patterns generating apparatus has generated the perforation pattern under this code check.
The situation that increases code check with non-regular code is similar, reduces the situation of code check for non-regular code, and promptly the code check of the female sign indicating number of LDPC is
Required code check is r
r, r wherein
r＜r.When code check is r
rThe time, distribute according to goodness and to calculate the variable node number of required repetition under each number of degrees, p
i=V
i* γ
i, p
iFor the number of degrees are the number of the required repetition of variable node of i, V
iFor the number of degrees is total number of the variable node of i, γ
iFor the number of degrees is the ratio of the variable node repetition of i.Number p according to required repetition under each number of degrees
i, from repetitive sequence table P1, P2 ... Pd
vIn select the bit of required repetition separately, then perforation/repetitions patterns generating apparatus has generated the repetitions patterns under this code check.
Situation for reducing code check has increased redundancy, has some code words to repeat transmission.For the codeword decoding that reduces behind the code check, the LDPC decoder at first carries out yard hand to the code word that repeats according to power ratio and merges.Be described to be transmitted as example in Gaussian channel (AWGN) below, but the present invention is not limited to AWGN, other channel is analogized equally.The i bit word need transmit twice, and interchannel noise is σ when transmitting for the first time
1 i, interchannel noise is σ when transmitting for the second time
2 i, then the posterior probability of transmission is for the first time
Then the posterior probability of transmission is for the second time
Then last total posterior probability is
Adopt belief propagation algorithm (BP) decoding then.
Should be noted that and understand, under the situation that does not break away from the desired the spirit and scope of the present invention of accompanying Claim, can make various modifications and improvement the present invention of foregoing detailed description.Therefore, the scope of claimed technical scheme is not subjected to the restriction of given any specific exemplary teachings.
Claims (24)
1. the code device of a loe-density parity-check code is characterized in that, described device comprises:
The LDPC encoder is used for the data of information source transmission are carried out low density check code;
Perforation/repetitions patterns generating apparatus is used for generating perforation or repetitions patterns according to perforation/repetitive sequence table, the order of the variable node of code word after wherein said perforation/repetitive sequence table has identified perforation or repeated described LDPC encoder encodes;
Perforation/duplicating device, it is used for the perforation that generates according to perforation/repetitions patterns generating apparatus or the repetitions patterns code word after to described LDPC encoder encodes and bores a hole or repeat.
2. code device according to claim 1 is characterized in that, comprises perforation/repetitive sequence table calculation element, and it is used for calculating described perforation/repetitive sequence table according to the influence degree that variable node recovers code word.
3. code device according to claim 2 is characterized in that, the influence degree that described variable node recovers code word is that the check-node number of degrees that are connected with described variable node interrelate.
4. code device according to claim 2 is characterized in that, described perforation/repetitive sequence table calculation element adopts offline mode to calculate described perforation/repetitive sequence table.
5. according to each described code device in the claim 1 to 4, it is characterized in that, comprising:
Perforation/repetitions patterns generating apparatus is used for distributing and described perforation/repetitive sequence table generation perforation/repetitions patterns according to goodness;
Wherein, described goodness distribution table has shown that the number of degrees of each code word optimum of variable node under different code checks distribute.
6. code device according to claim 5 is characterized in that, comprises goodness distribution calculation element, is used for calculating described goodness based on the density evolution algorithm and distributes.
7. code device according to claim 6 is characterized in that, described goodness distribution calculation element adopts offline mode to calculate described goodness to distribute.
8. the loe-density parity-check code decoding device corresponding to the described code device of claim 1 is characterized in that described decoding device comprises the LDPC decoder, and it is used for decoding according to the perforation pattern that described perforation/repetitions patterns generating apparatus generates.
9. decoding device according to claim 8 is characterized in that, described LDPC decoder is when reducing code check, and the LDPC decoder at first calculates posterior probability, and the perforation/repetitions patterns that generates according to described perforation/repetitions patterns generating apparatus is decoded then.
10. decoding device according to claim 9 is characterized in that, described LDPC decoder merges and calculates described posterior probability by the code word that repeats is carried out code word according to power ratio.
11. a data transmission system is characterized in that, comprises code device according to claim 1 and decoding device according to claim 8.
12. the coding method of a loe-density parity-check code comprises following steps:
At first, the data of information source transmission are carried out low density check code;
Then, generate perforation/repetitions patterns according to perforation/repetitive sequence table, wherein said perforation/repetitive sequence table has identified the order of the variable node of code word behind perforation or the repeated encoding;
Described coding back code word is bored a hole or repeat according to described perforation/repetitions patterns.
13. coding method according to claim 12 is characterized in that, comprises according to variable node the influence degree of code word recovery is calculated described perforation/repetitive sequence table.
14. coding method according to claim 13 is characterized in that, the influence degree that described variable node recovers code word is that the check-node number of degrees that are connected with described variable node interrelate.
15. coding method according to claim 13 is characterized in that, adopts offline mode to calculate described perforation/repetitive sequence table.
16. coding method according to claim 15 is characterized in that, calculates described perforation/repetitive sequence table and comprises the steps:
1) the described perforation of initialization/repetitive sequence table is empty set P;
2) select a variable node i at random from variable node set V, described variable node i is incorporated among the P, from check-node set C, remove the check-node r that links to each other with I
Ix, deletion and described check-node r from V
IxThe variable node that links to each other;
3) repeating step 2), be empty set up to V, then V=V P;
4) if V is an empty set, then obtain perforation/repetitive sequence table P, if not empty set, then turn back to step 2).
17. according to each described coding method in the claim 12 to 16, it is characterized in that, distribute and described perforation/repetitive sequence table generation perforation/repetitions patterns according to goodness;
Wherein, described goodness distribution table shows that the number of degrees of each code word optimum of variable node under different code checks distribute.
18. coding method according to claim 17 is characterized in that, calculates described goodness based on the density evolution algorithm and distributes.
19. coding method according to claim 18 is characterized in that, adopts offline mode to calculate described goodness and distributes.
20. coding method according to claim 12 is characterized in that, and is described according to perforation/repetitive sequence table generation perforation/repetitions patterns, specifically comprises the steps:
Code length according to the code check calculation expectation of expecting;
Calculate the difference p of the code length of female sign indicating number code length and described expectation;
P element before selecting from described perforation/repetitive sequence table is designated 1, and all the other elements are 0, obtain perforation/repetitions patterns.
21. the loe-density parity-check code coding/decoding method corresponding to the described coding method of claim 12 is characterized in that, decodes according to described perforation/repetitions patterns.
22. coding/decoding method according to claim 21 is characterized in that, when reducing code check, at first calculates posterior probability, decodes according to described perforation/repetitions patterns then.
23. coding/decoding method according to claim 22 is characterized in that, merges and calculates described posterior probability by the code word that repeats is carried out code word according to power ratio.
24. a data transmission method is characterized in that, may further comprise the steps:
Coding method according to claim 12 is encoded;
Coding/decoding method according to claim 21 is decoded.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008100575361A CN101227193B (en) | 2008-02-02 | 2008-02-02 | Method and device for encoding and decoding low density check code |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008100575361A CN101227193B (en) | 2008-02-02 | 2008-02-02 | Method and device for encoding and decoding low density check code |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101227193A true CN101227193A (en) | 2008-07-23 |
| CN101227193B CN101227193B (en) | 2010-06-02 |
Family
ID=39858987
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2008100575361A Active CN101227193B (en) | 2008-02-02 | 2008-02-02 | Method and device for encoding and decoding low density check code |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101227193B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101753195A (en) * | 2008-11-28 | 2010-06-23 | Tcl集团股份有限公司 | Method of rate matching in TD-SCDMA (Time Division-Synchronous Code Division Multiple Access) |
| CN102118231A (en) * | 2011-04-13 | 2011-07-06 | 厦门大学 | Code rate self-adaptive encoding method for multi-edge type low-density parity check code |
| WO2011088791A1 (en) * | 2010-01-25 | 2011-07-28 | Huawei Technologies Co., Ltd. | System and method for digital communications with unbalanced codebooks |
| CN104981978A (en) * | 2013-02-13 | 2015-10-14 | 高通股份有限公司 | LDPC design using quasi-cyclic constructions and puncturing for high rate, high parallelism, and low error floor |
| CN105356890A (en) * | 2015-11-30 | 2016-02-24 | 华侨大学 | LDPC (Low Density Parity Check) code perforating algorithm based on importance sampling technology |
| CN106656409A (en) * | 2015-10-30 | 2017-05-10 | 华为技术有限公司 | Method and device for perforation and unperforation of check code |
| CN108599891A (en) * | 2017-03-17 | 2018-09-28 | 华为技术有限公司 | Coding method, code device and communication device |
| WO2019134676A1 (en) * | 2018-01-05 | 2019-07-11 | 中兴通讯股份有限公司 | Passive optical network coding processing method and device |
| US10355711B2 (en) | 2014-12-30 | 2019-07-16 | Huawei Technologies Co., Ltd. | Data processing method and system based on quasi-cyclic LDPC |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100364238C (en) * | 2005-01-21 | 2008-01-23 | 清华大学 | Coding and decoding scheme for Turbo code and multi-dimensional modulating cascade system |
| CN100486150C (en) * | 2005-01-23 | 2009-05-06 | 中兴通讯股份有限公司 | Non-regular low intensity parity code based coder and its creation method |
-
2008
- 2008-02-02 CN CN2008100575361A patent/CN101227193B/en active Active
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101753195B (en) * | 2008-11-28 | 2013-12-04 | Tcl集团股份有限公司 | Method of rate matching in TD-SCDMA (Time Division-Synchronous Code Division Multiple Access) |
| CN101753195A (en) * | 2008-11-28 | 2010-06-23 | Tcl集团股份有限公司 | Method of rate matching in TD-SCDMA (Time Division-Synchronous Code Division Multiple Access) |
| WO2011088791A1 (en) * | 2010-01-25 | 2011-07-28 | Huawei Technologies Co., Ltd. | System and method for digital communications with unbalanced codebooks |
| CN102511141A (en) * | 2010-01-25 | 2012-06-20 | 华为技术有限公司 | System and method for using nonequilibrium codebook for digital communication |
| CN102511141B (en) * | 2010-01-25 | 2014-11-05 | 华为技术有限公司 | System and method for using nonequilibrium codebook for digital communication |
| CN102118231A (en) * | 2011-04-13 | 2011-07-06 | 厦门大学 | Code rate self-adaptive encoding method for multi-edge type low-density parity check code |
| CN102118231B (en) * | 2011-04-13 | 2013-04-17 | 厦门大学 | Code rate self-adaptive encoding method for multi-edge type low-density parity check code |
| CN104981978A (en) * | 2013-02-13 | 2015-10-14 | 高通股份有限公司 | LDPC design using quasi-cyclic constructions and puncturing for high rate, high parallelism, and low error floor |
| CN104981978B (en) * | 2013-02-13 | 2017-12-08 | 高通股份有限公司 | Using quasi- looping construct and perforate to realize that the LDPC of high-speed, high concurrency and low error floor is designed |
| US10355711B2 (en) | 2014-12-30 | 2019-07-16 | Huawei Technologies Co., Ltd. | Data processing method and system based on quasi-cyclic LDPC |
| CN106656409A (en) * | 2015-10-30 | 2017-05-10 | 华为技术有限公司 | Method and device for perforation and unperforation of check code |
| CN106656409B (en) * | 2015-10-30 | 2020-02-14 | 华为技术有限公司 | Check code puncturing and de-puncturing method and device |
| CN105356890A (en) * | 2015-11-30 | 2016-02-24 | 华侨大学 | LDPC (Low Density Parity Check) code perforating algorithm based on importance sampling technology |
| CN105356890B (en) * | 2015-11-30 | 2019-03-12 | 华侨大学 | A kind of LDPC code drilling method based on importance sampling technology |
| CN108599891A (en) * | 2017-03-17 | 2018-09-28 | 华为技术有限公司 | Coding method, code device and communication device |
| CN108599891B (en) * | 2017-03-17 | 2020-05-08 | 华为技术有限公司 | Encoding method, encoding device, and communication device |
| US10979076B2 (en) | 2017-03-17 | 2021-04-13 | Huawei Technologies Co., Ltd. | Encoding method, encoding apparatus, and communications apparatus |
| WO2019134676A1 (en) * | 2018-01-05 | 2019-07-11 | 中兴通讯股份有限公司 | Passive optical network coding processing method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101227193B (en) | 2010-06-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101227193B (en) | Method and device for encoding and decoding low density check code | |
| CN106100794B (en) | Coding cooperation method based on punched polarization code | |
| US7729437B2 (en) | Method and apparatus for space-time coding using lifting low density parity check codes in a wireless communication system | |
| CN101453297B (en) | Encoding method and apparatus for low density generation matrix code, and decoding method and apparatus | |
| US8369448B2 (en) | Bit mapping scheme for an LDPC coded 32APSK system | |
| CN101611552B (en) | Method of matching codeword size and transmitter thereof in mobile communications system | |
| US7966548B2 (en) | Method and system for encoding data using rate-compatible irregular LDPC codes based on edge growth and parity splitting | |
| CN106100795B (en) | Polar code coding cooperation method based on Plotkin construction and information bit re-dormancy | |
| CN101459430B (en) | Encoding method and apparatus for low density generation matrix code | |
| CN102244922B (en) | Power control method applicable to Raptor Codes under additive white Gaussian noise channel | |
| CN101615913A (en) | The quick convergence decoding algorithm of LDPC sign indicating number | |
| CN106992844A (en) | The degree distribution optimization method of digital fountain code based on M QAM | |
| CN105162552A (en) | Ka frequency range deep space communication method and system of q-LDPC-LT cascade fountain code | |
| CN1933336B (en) | Construction of LDPC (low density parity check) codes using GRS (generalized reed-solomon) code | |
| CN103338046A (en) | Encoding and decoding method of LDPC-RS (Low Density Parity Check-Reed-Solomon) two-dimensional product code with compatible code rate | |
| CN111245571B (en) | Receiver design method combined with coding modulation technology under noise aggregation system | |
| CN106209305A (en) | A kind of fountain codes interpretation method under access channel | |
| KR101503656B1 (en) | Apparatus and method for channel encoding and decoding in communication system using low-density parity-check codes | |
| CN110601699B (en) | Method for realizing code rate dynamic variable multi-element LDPC code | |
| Li et al. | Generalized low-density parity-check coding scheme with partial-band jamming | |
| Liu et al. | Adaptive polar coding with high order modulation for block fading channels | |
| KR101145673B1 (en) | The method of encoding using lt code with specific structure and the method of encoding using raptor code based thereof | |
| CN101150377A (en) | Bit mapping scheme of 32APSK system for low-density checksum coding | |
| Wang et al. | Blind Recognition of 5G LDPC Codes Over a Candidate Set | |
| Liang et al. | Novel joint encoding/decoding algorithms of fountain codes for underwater acoustic communication |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |