CN103944586A - Method for constructing code-rate compatibility QC-LDPC code - Google Patents

Abstract

The invention relate to the technical field of channel coding of a wireless communication system and a satellite communication system, and provides a method for constructing a code-rate compatibility QC-LDPC code on basis of matrix row deletion. The method includes the steps that firstly, on the basis of a GCDg8 algorithm, QC-LDPC codes with a large girth and a low code rate are constructed to serve as primary codes, then successive row deletion of an exponential matrix is carried out to obtain subcodes with a high code rate and an unchanged information bit length, and the short-length problem of code words with a high code rate is avoided; row deletion means that corresponding edges of Tanner graphs corresponding to the primary codes without the short-length problem are deleted, the short-length problem cannot occur, and the girth may increase; ultimately, the subcodes are processed through a hidden technique, the minimal distance of the code words is increased, and performance of the code words is improved. According to the method, the code-rate compatibility QC-LDPC code with structural design is adopted, and compared with an LDPC code constructed in a PEG algorithm, construction is easier to achieve, hardware achievement complexity is not high, and performance is more excellent.

Description

A kind of building method of code-rate-compatible QC-LDPC code

Technical field

The present invention relates to the channel coding technology field of wireless communication system and satellite communication system, particularly a kind of building method of code-rate-compatible QC-LDPC code.

Background technology

Be different from the channel coding technology of traditional wireless, satellite communication system, low-density checksum (Low-Density Parity-Check, LDPC) code is by a class linear block codes of sparse check matrix definition, adopt belief propagation (belief propagation, BP) algorithm decoding, not only there is the superperformance of approaching Shannon limit, and decoding complexity is lower, flexible structure, be the study hotspot of field of channel coding in recent years, be widely used at present the fields such as deep space communication, optical fiber communication, satellite digital video and audio broadcasting.

In recent years, researcher finds to introduce the coding techniques of code-rate-compatible LDPC codes under channel condition wireless, that satellite communication system becomes when residing, can make that the error correcting capability of chnnel coding is adaptive makes corresponding adjustment according to channel circumstance.Adopt the LDPC code of code-rate-compatible, can be under the prerequisite guaranteeing service quality, according to channel circumstance by adjusting real-time its error correcting capability of adjustment of code check, thereby the throughput of realizing channel maximizes, and has improved the efficiency of transmission of data.From the angle of implementation complexity, consider, code-rate-compatible LDPC codes is comprised of a nested structure, and it can adopt the work of single encoded device/decoder in certain range of code rates, thereby greatly reduces the complexity of system.In mixing automatic repeat request (HARQ), transmitting terminal can adopt code-rate-compatible LDPC codes to send the check digit increasing progressively, to guarantee the successfully decoding of receiving terminal decoder.In unequal error protection (UEP), it can carry out effective error protection according to the significance level of information.

At present, the conventional method of design code-rate-compatible LDPC codes mainly contains: (1) punching (Puncture), obtains to the LDPC code punching of low code check the subcode that code check is more and more higher.The people such as McLaughlin are at " Rate-compatible puncturing of low-density parity-check codes " [Information Theory, IEEE Transactions on, 2004,50 (11): 2824-2836] in article, adopt the method for Gaussian approximation for the optimum punching degree distribution expansion research of code-rate-compatible LDPC codes.Although this method is simple to operate, still there is very serious this drawback of its decoding performance decline of increase along with code check in the code word obtaining by punching.(2) method of expanding based on matrix ranks, to the LDPC code of high code check successively ranks expand and obtain the subcode that code check is more and more less.For subcode decoding performance in punching process, worsen this drawback, the people such as Jacobsen are at " Design of rate-compatible irregular LDPC codes based on edge growth and parity splitting " [Vehicular Technology Conference, 2007.VTC-2007Fall.2007IEEE66th.IEEE, 2007:1052-1056] adopt the degree distribution optimization method of external information transition diagram (EXIT) to launch research for the code-rate-compatible LDPC codes based on extended method design in article.Although this method can avoid subcode decoding performance in punching process to worsen this drawback, but, expanding course is equivalent to increase new check-node, variable node and limit corresponding to newly-increased nonzero element at its corresponding Tanner figure, therefore need to consider to avoid the nonzero element newly increasing to produce new becate in corresponding Tanner figure, thereby increased the complexity of expanding course degree moderate distribution optimization.Therefore, structure codeword performance code-rate-compatible LDPC codes good, that complexity is low becomes people in recent years and studies one of hot issue of code-rate-compatible LDPC codes application.

Summary of the invention

At present, there is the very serious drawback of its decoding performance decline of increase along with code check in the code-rate-compatible LDPC codes word by punching art designs, and the code-rate-compatible LDPC codes based on the design of matrix expanding method proposing for this drawback, great majority due to needs carry out a large amount of Optimization Works with the nonzero element of avoiding newly increasing in corresponding Tanner figure, produce new becate with and non-structured design and cannot guarantee the encoder complexity that it is low.

For above deficiency of the prior art, the object of the present invention is to provide a kind of code-rate-compatible QC-LDPC code that produces multi code Rate of Chinese character, superior performance; Technical scheme of the present invention is as follows: a kind of building method of code-rate-compatible QC-LDPC code, and it comprises the following steps:

101, initialization sequence S is that { columns L, obtains sequence S={a according to GCDg8 algorithm for 0,1}, the line number J of input exponential matrix E ₀, a ₁..., a _j-1, then according to calculating formula P>=(a _j-1-a ₀) (L-1)+1, the dimension that draws cyclic permutation matrices is P * P, element a _ij circulates to the right and moves a corresponding to every row of P * P unit matrix _ij position,, i=0,1 ..., J-1; J=0,1 ..., L-1, described exponential matrix E is

$E(a_0,a_1,...,a_{J-1})=\left(\begin{array}{cccc}{a}_0\&CenterDot;0& a_0\&CenterDot;1& ...& a_0\&CenterDot;(L-1)\\ a_1\&CenterDot;0& a_1\&CenterDot;1& ...& a_1\&CenterDot;(L-1)\\ & & & .\\ ...& ...& ...& .\\ & & & .\\ a_{J-1}\&CenterDot;0& a_{J-1}\&CenterDot;1& ...& a_{J-1}\&CenterDot;(L-1)\end{array}\right);$ J and L are two integers, J>=3, L>=3, and 0≤a ₀<a ₁< ... <a _j-1, a ₀..., a _j-1for integer;

102, according to the sequence S={a obtaining in step 101 ₀, a ₁..., a _j-1and the dimension P * P of cyclic permutation matrices, according to the structure principle of 101 Exponential matrixes, obtain its check matrix, and enclosing of this check matrix is longly more than or equal to 8;

103, exponential matrix last column and last row of the female code of the quasi-circulating low-density parity check QC-LDPC code of delete step 102 gained LDPC0 obtain subcode LDPC1, last column and last row of deleting the exponential matrix of subcode LDPC1 obtain subcode LDPC2 again, delete by that analogy the high code check subcode that obtains a plurality of code checks, through the code check of the high rate codewords that obtains after ranks delet method successively, be

n=1,2 ..., represent the row/column number that the exponential matrix of female code is deleted altogether;

104, to the exponential matrix E of female code LDPC0 of gained in step 103 and subcode LDPC1, LDPC2, adopt the hidden matrix M of same dimension to process respectively, hidden matrix M (J, L, J ')=(m _i,j) _{0≤i<J, 0≤j<L}be the binary matrix of a J * L, hidden operation be defined as follows: if m _i,j=1, m _i,je _i,j=E _i,j, otherwise m _i,j=0 o'clock, m _i,je _i,jthe full null matrix of corresponding P * P, the hidden matrix of binary of employing and exponential matrix same dimension is processed the exponential matrix of the QC-LDPC code of phase code rate, be in corresponding exponential matrix, to adopt the null matrix of equal dimension to replace with 0 corresponding position units in hidden matrix, remain unchanged with 1 corresponding position units in hidden matrix, obtain the QC-LDPC code of code-rate-compatible.

Further, when setting J=6, during L=12, female code LDPC0, the subcode LDPC1 in step 103, the information bit length of subcode LDPC2 are 2514, and code check is respectively 1/2,6/11,3/5, and the sequence obtaining by GCDg8 algorithm search is (a ₀, a ₁..., a ₅)={ 0,1,12,13,35,38}, P=419.

Further, described code-rate-compatible QC-LDPC code is comprised of a nested codeword structure, and it can work under single encoder/decoder.

Further, in radio communication and satellite communication system under time varying channel environment, information to be sent is after information source coding, adaptively according to channel circumstance, adopt code-rate-compatible QC-LDPC encoder to carry out chnnel coding to transmit in channel, when channel circumstance is good, adopt high rate codewords to encode, otherwise when channel circumstance is undesirable, adopt low rate codewords to encode.

Advantage of the present invention and beneficial effect:

The building method of the code-rate-compatible QC-LDPC code of deleting based on exponential matrix ranks that the present invention proposes has been considered time variation wireless, satellite communication environment lower channel, propose the building method with a kind of code-rate-compatible QC-LDPC code, its error correcting capability can adaptively be made corresponding adjustment according to channel circumstance.Adopt this code-rate-compatible QC-LDPC code can be under the prerequisite guaranteeing service quality, according to channel circumstance by adjusting real-time its error correcting capability of adjustment of code check, thereby improve the throughput of channel.This based on exponential matrix ranks deletion thought structure code-rate-compatible QC-LDPC code, can avoid subcode to occur becate, because ranks are deleted is to delete corresponding limit in the Tanner figure corresponding without female code of becate, not only can not produce becate, but also can increase, enclose length.Therefore the building method of, deleting thought based on ranks can be good at overcoming the shortcoming that occurs becate in the degradation of subcode decoding performance after punching and expanding course.

Accompanying drawing explanation

Fig. 1 is the structure flow chart of code-rate-compatible QC-LDPC code in the present invention;

Fig. 2 is the exponential matrix structure chart of code-rate-compatible QC-LDPC code in the present invention;

Fig. 3 is the exponential matrix structure chart of the code-rate-compatible QC-LDPC code based on GCDg8 algorithm in the present invention;

Fig. 4 is wireless, the satellite communication system illustraton of model that adopts code-rate-compatible QC-LDPC code in the present invention;

Fig. 5 is the BER performance simulation figure of three kinds of code-rate-compatible QC-LDPC codes in the present invention;

Fig. 6 is that the BER Performance Ratio of three kinds of code-rate-compatible QC-LDPC codes and PEG algorithm construction LDPC code in the present invention is compared with analogous diagram.

Embodiment

The invention will be further elaborated below in conjunction with accompanying drawing, to provide the embodiment of an indefiniteness.

Shown in Fig. 1-Fig. 3, structure for the QC-LDPC code of code-rate-compatible, the present invention chooses is that QC-LDPC code based on GCDg8 algorithm construction is as being female code, then to the exponential matrix of its J * L successively ranks delete the constant subcode of information bit length obtain high code check, finally to it, adopt hidden matrix to process.

Be illustrated in figure 1 the structure flow chart of code-rate-compatible QC-LDPC code in the present invention.Comprise the following steps: first based on one of GCDg8 algorithm construction, enclose greatly long low code check QC-LDPC code as female code, then to the exponential matrix of the female code of above-mentioned QC-LDPC successively ranks delete the constant high code check subcode of obtaining information bit length, finally adopt concealing technology respectively the code word of each code check to be processed, obtain the code-rate-compatible LDPC codes that a kind of codeword performance is good, complexity is low.

Specifically comprise the following steps:

1. based on GCDg8 algorithm construction, enclose greatly the female code of long low code check QC-LDPC: this step is chosen be QC-LDPC code based on GCDg8 algorithm construction as female code, also can choose other method and construct the female code of LDPC that meets above-mentioned requirements.Based on being constructed as follows of GCDg8 algorithm QC-LDPC code:

Each element in exponential matrix E can be expressed as: P _r,c=f (r, c).According to people such as Wang X at " Construction of girth-eight QC-LDPC codes from greatest common divisor " [Communications Letters, IEEE, 2013,17 (2): 369-372] the GCDg8 algorithm providing, P _r,c=g (r) h (c), wherein g (r _i)=a _i(i=0,1 ..., J-1), h (c _j)=j (j=0,1 ..., L-1).So, the exponential matrix correspondence of a J * L one without the piece cyclic check matrix of zero location:

$E(a_0,a_1,...,a_{J-1})=\left(\begin{array}{cccc}{a}_0\&CenterDot;0& a_0\&CenterDot;1& ...& a_0\&CenterDot;(L-1)\\ a_1\&CenterDot;0& a_1\&CenterDot;1& ...& a_1\&CenterDot;(L-1)\\ & & & .\\ ...& ...& ...& .\\ & & & .\\ a_{J-1}\&CenterDot;0& a_{J-1}\&CenterDot;1& ...& a_{J-1}\&CenterDot;(L-1)\end{array}\right)---\left(1\right)$

Wherein J and L(J>=3, L>=3) be two integers, and 0≤a ₀<a ₁< ... <a _j-1(a ₀..., a _j-1for integer).When exponential matrix is converted into check matrix, element a _ij (i=0,1 ..., J-1; J=0,1 ..., L-1) corresponding to every row of P * P unit matrix, circulate to the right and move a _ij(mod P) position.

In above-mentioned exponential matrix, we need search sequence a _idimension with definite cyclic permutation matrices.Structure encloses long 8 the QC-LDPC code that is at least, and first, we carry out search sequence a according to GCDg8 algorithm _i.The step of GCDg8 algorithm is as follows:

Input: the heavy L of column weight J and row.

Output: sequence S={a ₀, a ₁..., a _j-1.

Initialization: S set is initialized as { 0,1}.Initialization j=0.

Step 1: if j< (J-2) goes to step 2.Otherwise finish, search completes.

Step 2: initialization: value+1 of last element of Y=S.

Step 3: in current S set traversal choose two number S (i) and S (j), and S (j) > S (i), if for any two S (i) and S (j), if all met:

(Y-S (i))/gcd (Y-S (i), S (j)-S (i)) >=L(gcd (a, b) represents a, the greatest common factor of b) (2) arrange check=1 so; Otherwise, check=0 is set.

Step 4: if check=0, Y=Y+1, and be back to step 3; If check=1, S=S ∪ Y, j=j+1, and be back to step 1.

Then we are again according to P>=(a _j-1-a ₀) (L-1)+1(3), the dimension of determining cyclic permutation matrices is P * P.

Finally, the check matrix H of the QC-LDPC code based on GCDg8 algorithm can be passed through formula (1), (2), (3) according to unique the determining of the structure principle of exponential matrix.

2. successively ranks deletion of the exponential matrix based on the female code of QC-LDPC: J * L exponential matrix of GCDg8 algorithm or other algorithm constructions, from last column and row, is carried out successively ranks deletion and obtains high code check subcode.This method of successively deleting is that exponential matrix is deleted to a line one row each time, and what the information that guaranteed was length is constant, and the exponential matrix structure of its code-rate-compatible QC-LDPC code as shown in Figure 2.Owing to carrying out successively ranks on the basis of enclosing greatly long LDPC code and delete constructing, be in Tanner figure, to delete variable node and check-node and corresponding limit with it, so not only can avoid occurring becate at expanding course, and the limit of reducing can contribute to increase and enclose length, thereby protected the performance of card code word.

3. adopt concealing technology respectively the code word of each code check to be processed: to the exponential matrix of each code check subcode, to adopt the hidden matrix of same dimension to process respectively.Concealing technology is simply used for adopting some full null matrix to replace the cyclic permutation matrices of relevant position conventionally.We establish M (J, L, J ')=(m _i,j) _{0≤i<J, 0≤j<L}be the binary matrix of a J * L, we carry out the operation of concealing technology to exponential matrix E at definition as follows:

$M\&CircleTimes;E\stackrel{\mathrm{\&Delta;}}{=}(m_{i,j}\&CenterDot;E_{i,j}{)}_{0\&le;i<J,0\&le;j<L}---\left(4\right)$

If m _i,j=1, m _i,je _i,j=E _i,j, otherwise m _i,j=0 o'clock, m _i,je _i,jthe full null matrix of corresponding P * P.

If J * L matrix that the check matrix of QC-LDPC code is comprised of cyclic permutation matrices and null matrix, this class code word that column weight is identical so has larger minimum range.Hidden matrix (Masking Matrix) and exponential matrix are enclosed greatly to long QC-LDPC in conjunction with designing, by by " 0 " of diverse location in hidden matrix and " 1 " element respectively with exponential matrix in the element of opposite position multiply each other obtain new exponential matrix acquired can better yard.

Below by embodiment, further illustrate by reference to the accompanying drawings the present invention, but the scope not limiting the present invention in any way.

Below elaborate the building method that utilizes the present invention to state, the QC-LDPC code of 3 kinds of code-rate-compatibles of structure, its information bit length is all k=2514, code check is respectively 1/2,6/11,3/5.

1) first, by GCDg8 algorithm search, obtain (a ₀, a ₁..., a ₅)={ 0,1,12,13,35,38} determines P=419 according to (3) simultaneously, according to a J=6 of (1) structure, and L=12, the exponential matrix of R=1/2 is female code.Thus, obtain the female code of (5028,2514) QC-LDPC that 1 column weight is 3.

2) then, through deleting last column one row for the 1st time, can obtain the 1st (4609 of code check R=6/11,2514) QC-LDPC subcode, then delete through the 2nd time the 2nd (4190,2514) QC-LDPC subcode that last column one row have obtained code check R=3/5.Successively ranks delete procedure is as shown in Figure 3 for it.

3) last, adopt the hidden matrix of corresponding dimension to process the exponential matrix of each code check.Adopting last minute and provide M (6,12,3), M (5,11,3), M (4,10,3) is 1/2,6/11,3/5 to process to code check respectively.For the exponential matrix of J * L, set corresponding hidden matrix M (J, L, J ') and equal J * L binary matrix of (being similar to) J ' for column weight.The hidden matrix of the QC-LDPC code of code fetch rate R=1/2 is as follows:

$M_{1/2}\left(\mathrm{6,12,3}\right)=\left(\begin{array}{cccccccccccc}0& 1& 1& 1& 0& 0& 0& 1& 1& 1& 0& 0\\ 0& 0& 1& 1& 1& 0& 0& 0& 1& 1& 1& 0\\ 0& 0& 0& 1& 1& 1& 0& 0& 0& 1& 1& 1\\ 1& 0& 0& 0& 1& 1& 1& 0& 0& 0& 1& 1\\ 1& 1& 0& 0& 0& 1& 1& 1& 0& 0& 0& 1\\ 1& 1& 1& 0& 0& 0& 1& 1& 1& 0& 0& 0\end{array}\right)---\left(5\right)$

The hidden matrix of the QC-LDPC code of code fetch rate R=6/11 is as follows:

$M_{6/11}(5,11,3)=\left(\begin{array}{ccccccccccc}0& 1& 1& 1& 0& 0& 1& 1& 1& 0& 0\\ 0& 0& 1& 1& 1& 0& 0& 1& 1& 1& 0\\ 1& 0& 0& 1& 1& 1& 0& 0& 1& 1& 1\\ 1& 1& 0& 0& 1& 1& 1& 0& 0& 1& 1\\ 1& 1& 1& 0& 0& 1& 1& 1& 0& 0& 1\end{array}\right)---\left(6\right)$

The hidden matrix of the QC-LDPC code of code fetch rate R=3/5 is as follows:

$M_{3/5}(4,10,3)=\left(\begin{array}{cccccccccc}0& 1& 1& 1& 0& 1& 1& 1& 1& 1\\ 1& 0& 1& 1& 1& 0& 1& 1& 1& 1\\ 1& 1& 0& 1& 1& 1& 0& 1& 1& 1\\ 1& 1& 1& 0& 1& 1& 1& 0& 1& 1\end{array}\right)---\left(7\right)$

Matrix after above-mentioned a series of hidden matrix disposals is a series of check matrix H of enclosing greatly the code-rate-compatible QC-LDPC code of length, larger minimum range that this part proposes.

Coding for above-mentioned code-rate-compatible QC-LDPC code, employing utilizes Gaussian elimination method to the parity check matrix H of above-mentioned structure, produce a lower triangular matrix, then further elementary transformation obtains the right unit matrix form: H=[P|I] (the unit matrix that I is (N-K) * (N-K), P is the matrix of (N-K) * K), by G=[I|P ^t] obtain generator matrix G, thus by C=MG, directly encode and obtain coding codeword (M is that length is the information vector of K, and N is the code length after encoding).

In radio communication and this class of satellite communication, have in the system of time varying channel, information to be sent is after information source coding, adaptively according to channel circumstance, adopt single code-rate-compatible QC-LDPC encoder to carry out chnnel coding to transmit in channel, thereby greatly improved the throughput of channel, at the single code-rate-compatible QC-LDPC decoder of the same employing of receiving terminal, coding&decoding adopts a set of encoder/decoder simultaneously, has greatly reduced the complexity of system.Adopt wireless, the satellite communication system model of this code-rate-compatible QC-LDPC code as shown in Figure 4.

Next, provide in the embodiment of the present invention based on matrix ranks delet method, construct 3 kinds of code-rate-compatible QC-LDPC code performance simulation results.Emulation is carried out under awgn channel, adopts the mode of BPSK to modulate, and adopts BP algorithm to carry out decoding, greatest iteration time is set and equals 100.

Fig. 5 is the code word BER performance curve simulation result of the lower 3 kinds of code-rate-compatibles of additive white Gaussian noise channel.As can be seen from the figure in embodiments of the invention, the performance curve of 3 kinds of code-rate-compatible QC-LDPC codes from right to left, along with code check reduces, become better and better by codeword performance.

For the advantage of the code-rate-compatible QC-LDPC code constructing method of deleting based on matrix ranks of the present invention under wireless, satellite communication channel is described, the present invention has chosen the code-rate-compatible code word based on GCDg8 algorithm construction that LDPC code based on PEG algorithm construction and the present invention propose and has contrasted.

Fig. 6 is under additive white Gaussian noise channel, in the embodiment of the present invention 3 kinds of code-rate-compatible QC-LDPC codes with adopt PEG algorithm construction etc. code length, etc. code check, etc. code weight, etc. enclose long LDPC code BER Performance Ratio compared with simulation result.The performance of 3 kinds of code-rate-compatible QC-LDPC codes in the embodiment of the present invention slightly promotes with the code length such as corresponding, code check, a code performance heavy, that enclose the LDPC of long PEG algorithm construction, illustrated that the present invention proposes based on enclosing greatly long LDPC code, adopt successively the performance that the code-rate-compatible QC-LDPC code of ranks delet method structure can highly effective each rate codewords of assurance.The building method of deleting based on exponential matrix ranks can construct with lower implementation complexity the suitable code-rate-compatible LDPC codes of random code performance of being constructed with PEG algorithm.

Exponential matrix successively ranks deletion refers to that the exponential matrix of female code is successively deleted to last column one is listed as (deleting a line one row or identical ranks number) to obtain high code check code at every turn.If adopting the dimension of the female code exponential matrix enclose greatly long construction algorithm structure is J * L, through the code check of the high rate codewords that obtains after ranks delet method successively, be:

(n=1,2 ..., represent the row/column number that the exponential matrix of female code is deleted altogether) and (8)

The method of this structure code-rate-compatible QC-LDPC code has avoided high rate codewords to occur becate, this is to be to delete corresponding limit in Tanner figure corresponding to the high rate codewords without becate because the ranks of matrix are deleted, not only can not produce becate, but also may increase, enclose length.Therefore the building method of, deleting thought based on ranks can be good at overcoming the shortcoming that occurs becate in the degradation of subcode decoding performance after punching and expanding course.

The code-rate-compatible QC-LDPC code constructing method of deleting based on matrix ranks of above the present invention being stated is described in detail and illustrates.Above-mentioned concrete implementation can be used for helping to understand core concept of the present invention.The present invention maintains close ties with wireless, the feature becoming during satellite communication system channel, consider that high rate codewords decoding performance that the punching technology of structure code-rate-compatible LDPC codes produces easily worsens and expand because needs carry out a large amount of Optimization Works and non-structured design and cannot guarantee these some shortcomings of its low encoder complexity based on matrix ranks, drawback, the present invention proposes a kind of code-rate-compatible QC-LDPC code constructing method of deleting based on matrix ranks, with punching, expanding the implementation of these traditional code-rate-compatibles compares, this method can effectively overcome these drawbacks that easily occur becate after punching in the degradation of high rate codewords decoding performance and expanding course.Compare with the same code check of PEG algorithm construction, the code word of code length, this method is used structurized building method to adjust the block matrix of the check matrix of code-rate-compatible QC-LDPC code, thereby has guaranteed the lower complexity of construction process.

These embodiment are interpreted as only for the present invention is described, is not used in and limits the scope of the invention above.After having read the content of record of the present invention, technical staff can make various changes or modifications the present invention, and these equivalences change and modification falls into the inventive method claim limited range equally.

Claims (4)

1. a building method for code-rate-compatible QC-LDPC code, is characterized in that comprising the following steps:

101, initialization sequence S is that { columns L, obtains sequence S={a according to GCDg8 algorithm for 0,1}, the line number J of input exponential matrix E ₀, a ₁..., a _j-1, then according to calculating formula P>=(a _j-1-a ₀) (L-1)+1, the dimension that draws cyclic permutation matrices is P * P, element a _ij circulates to the right and moves a corresponding to every row of P * P unit matrix _ij position,, i=0,1 ..., J-1; J=0,1 ..., L-1, described exponential matrix E is

$E(a_0,a_1,...,a_{J-1})=\left(\begin{array}{cccc}{a}_0\&CenterDot;0& a_0\&CenterDot;1& ...& a_0\&CenterDot;(L-1)\\ a_1\&CenterDot;0& a_1\&CenterDot;1& ...& a_1\&CenterDot;(L-1)\\ & & & .\\ ...& ...& ...& .\\ & & & .\\ a_{J-1}\&CenterDot;0& a_{J-1}\&CenterDot;1& ...& a_{J-1}\&CenterDot;(L-1)\end{array}\right);$ J and L are two integers, J>=3, L>=3, and 0≤a ₀<a ₁< ... <a _j-1, a ₀..., a _j-1for integer;

102, according to the sequence S={a obtaining in step 101 ₀, a ₁..., a _j-1and the dimension P * P of cyclic permutation matrices, according to the structure principle of 101 Exponential matrixes, obtain its check matrix, and enclosing of this check matrix is longly more than or equal to 8;

103, exponential matrix last column and last row of the female code of the quasi-circulating low-density parity check QC-LDPC code of delete step 102 gained LDPC0 obtain subcode LDPC1, last column and last row of deleting the exponential matrix of subcode LDPC1 obtain subcode LDPC2 again, delete by that analogy the high code check subcode that obtains a plurality of code checks, through the code check of the high rate codewords that obtains after ranks delet method successively, be

n=1,2 ..., represent the row/column number that the exponential matrix of female code is deleted altogether;

104, to the exponential matrix E of female code LDPC0 of gained in step 103 and subcode LDPC1, LDPC2, adopt the hidden matrix M of same dimension to process respectively, hidden matrix M (J, L, J ')=(m _i,j) _{0≤i<J, 0≤j<L}be the binary matrix of a J * L, hidden operation be defined as follows: if m _i,j=1, m _i,je _i,j=E _i,j, otherwise m _i,j=0 o'clock, m _i,je _i,jthe full null matrix of corresponding P * P, the hidden matrix of binary of employing and exponential matrix same dimension is processed the exponential matrix of the QC-LDPC code of phase code rate, be in corresponding exponential matrix, to adopt the null matrix of equal dimension to replace with 0 corresponding position units in hidden matrix, remain unchanged with 1 corresponding position units in hidden matrix, obtain the QC-LDPC code of code-rate-compatible.

2. the building method of code-rate-compatible QC-LDPC code according to claim 1, it is characterized in that: when setting J=6, during L=12, female code LDPC0, the subcode LDPC1 in step 103, the information bit length of subcode LDPC2 are 2514, code check is respectively 1/2,6/11,3/5, and the sequence obtaining by GCDg8 algorithm search is (a ₀, a ₁..., a ₅)={ 0,1,12,13,35,38}, P=419.

3. the building method of code-rate-compatible QC-LDPC code according to claim 1, is characterized in that: described code-rate-compatible QC-LDPC code is comprised of a nested codeword structure, and it can work under single encoder/decoder.

4. according to building method and the coding and decoding mode of the code-rate-compatible QC-LDPC code one of claim 1～3 Suo Shu, it is characterized in that: in radio communication and satellite communication system under time varying channel environment, information to be sent is after information source coding, adaptively according to channel circumstance, adopt code-rate-compatible QC-LDPC encoder to carry out chnnel coding to transmit in channel, when channel circumstance is good, adopt high rate codewords to encode, otherwise when channel circumstance is undesirable, adopt low rate codewords to encode.