A coding and decoding method for fountain codes in a windowed double-layer unequal error protection system

Abstract

A coding and decoding method for a windowed double-layer unequal error protection system fountain code, the present invention relates to a coding and decoding method for a windowed double-layer unequal error protection system fountain code. The present invention is to solve the problems that the channel coding in the existing communication field adopts fountain code unequal error protection, which has the disadvantages of high complexity, non-system code, large overhead, etc. The process is: coding: 1. Divide the communication information source symbols into important information symbols and a sequence of secondary information symbols that are not set to zero; set the number of abnormal symbols E; 2. Get the check sequence of important information symbols; 3. Divide into RMIB1 and RMIB2 sequences; 4. Get the XLIB sequence; 5. Get the check sequence; 6. Restore the sequence of secondary information symbols; 7. Get the final check sequence; 8. Get the preliminary double-layer unequal error protection fountain code encoding; 9. Set the elements of the row corresponding to the abnormal symbol in the matrix in 5 to zero; and perform decoding. The present invention is used for channel coding technology in the communication field.

Description

Technical Field

[0001] The present invention relates to a channel coding technology in the communication field, and in particular to an encoding and decoding method of a fountain code of a windowed double-layer unequal error protection system. Background Art

[0002] In the ever-changing data transmission business, how to ensure efficient and reliable transmission of information is a hot and key research direction in modern communication. There are many kinds of communication services. For multimedia streaming services, information is divided into important information (MI) and secondary information (LI). Obviously, when receiving bit stream information, the recovery of important information directly affects the reliability of communication. For example, in the bit stream segmentation scheme of the latest video compression coding standard H.264, the importance level of I frame or IDR frame (immediate refresh frame), sequence parameter set (SPS) and picture parameter set (PPS) is very high. If these data cannot be recovered, it will have a significant impact on the decompression and reconstruction of the video. Therefore, it is hoped that these data can be restored first during communication. Traditional channel coding is equal error protection, that is, the protection level of all information codewords is the same, and priority protection of data cannot be achieved. Therefore, the academic community has designed an unequal error protection scheme to more efficiently recover important information under the same receiving overhead, that is, to reduce the bit error rate of important information. There are many application scenarios for digital fountain technology. It is mainly widely used in the following two scenarios with excellent performance-wireless multimedia broadcasting and deep space communication.

[0003] When transmitting data streams on traditional deletion channels, a file at the sender will usually be divided into multiple fragments and transmitted to the receiver. At the same time, multimedia streaming services are sensitive to data errors and have large data capacity. These characteristics make the general encoding and decoding methods very ineffective or too expensive for multimedia streaming services. Fountain codes have excellent characteristics such as anti-deletion, suitable for time-varying channels, low encoding and decoding complexity, and no need for feedback, which just meet the stringent requirements of multimedia streaming services. Due to its high efficiency in application, combined with the distinction of information importance levels in digital transmission, designing fountain codes with unequal error protection has become an important direction of fountain code research.

[0004] Raptor code is the most advanced fountain code technology today. R10 Raptor code is an international standard (RFC5053) with its simplicity and asymptotic linear complexity. Some communication standards such as 3GPP MBMS adopt it as the upper layer AL-FEC scheme, providing error correction performance for the system. The encoding and decoding of R10 Raptor code is the basis of the entire invention. It adopts three-layer encoding, with the outer code being the cascade of LDPC code and Half code, and the inner code being LT code encoding. The input symbol is passed through the outer code to obtain the intermediate symbol, and the intermediate symbol is LT encoded to obtain the output symbol Raptor code. The R10 Raptor code is specially designed to maintain its systematicity.

[0005] The encoding method of R10 Raptor code is as follows:

[0006] 1. Generate L intermediate symbols based on K input symbols, where the number of LDPC symbols is S and the number of Half symbols is H. So L = S + H + K, assign the values ​​of K symbols to the last K symbols of the L intermediate symbols, and set the first S + H symbols to zero.

[0007] 2. Calculate the generated matrix A, which is in the form of Figure 3 As shown, G_LDPC is the generator matrix of LDPC, G_Half is the generator matrix of Half code, G_LT is the generator matrix of LT code, I represents the unit matrix, and 0 represents the zero matrix;

[0008] 3. Through the matrix operation AC=D in step 2, find the inverse matrix of A, then C=A -1 D, thus obtaining L intermediate symbols;

[0009] 4. Perform LT encoding on the intermediate symbol according to the index to obtain the check symbol R, thereby obtaining the output symbol, which contains K input symbols and the following check symbol R.

[0010] The decoding method of R10 Raptor code is as follows:

[0011] 1. Calculate the total number of pre-coded symbols L = K + S + H according to K;

[0012] 2. Generate matrix A based on the precoding relationship. Calculate the inverse matrix A -1 ;

[0013] 3. According to C=A -1 D, get the intermediate symbol sequence;

[0014] 4. Perform LT encoding according to the intermediate symbol sequence to obtain K source symbols. Summary of the invention

[0015] The present invention proposes a coding and decoding method of a windowed double-layer unequal error protection system fountain code to solve the disadvantages of high complexity, non-systematic code, and large overhead in channel coding in the existing communication field.

[0016] The specific process of the encoding and decoding method of a windowed double-layer unequal error protection system fountain code is as follows:

[0017] The encoding method is carried out in the following steps:

[0018] Step 1: First, obtain the two-layer priority of the communication information source symbol, and divide the communication information source symbol into an important information symbol MIB sequence and a secondary information symbol LIB sequence that is not set to zero; set the number of abnormal symbols E, and evenly select E symbols from the secondary information symbol LIB sequence to set to zero; the specific process is:

[0019] Divide the K source symbols into K 1 MIB sequence of important information symbols and K 2 A sequence of non-zeroed minor information symbols LIB For E (E≤K 2 ) The secondary information symbol LIB sequence is set to zero to obtain a new LIB sequence

[0020] Step 2: Pre-encode the important information symbol MIB sequence in step 1 with R10 Raptor code to obtain an important intermediate symbol CMIB sequence;

[0021] Perform LT encoding of the R10 Raptor code on the CMIB sequence to obtain the check sequence of the important information symbol MIB; the specific process is:

[0022] The important information symbol MIB sequence obtained in step 1 is pre-coded with R10 Raptor code to obtain the important intermediate symbol CMIB sequence And the generator matrices are G_LDPC1 and G_Half1;

[0023] For important intermediate symbols CMIB sequence Perform LT encoding according to the index (1, 2, ..., N) to obtain N output symbol sequences and the generating matrices are G_LT_MIB, G_LT_RMIB1 and G_LT_RMIB2, where N represents the total number of generated coding symbols;

[0024] yes The check sequence of

[0025] Step 3: The N output symbol sequences obtained in step 2 It is divided into RMIB1 sequence and RMIB2 sequence, where the length of RMIB1 sequence is equal to that of LIB sequence;

[0026] Step 4: XOR the secondary information symbol LIB sequence in step 1 and the RMIB1 sequence in step 3 to obtain the XLIB sequence; the specific process is:

[0027] The secondary information symbol LIB sequence obtained in step 1 XOR operation is performed one by one with the RMIB1 sequence generated in step 3 to obtain the XLIB sequence

[0028] Step 5: Encode the XLIB sequence in step 4 by R10 Raptor code to obtain the XLIB sequence and the check sequence RXLIB of the XLIB sequence; the specific process is as follows:

[0029] For the XLIB sequence obtained in step 4 Perform R10Raptor code encoding and obtain an output length of NK 1 XLIB sequence and XLIB sequence check sequence RXLIB integration sequence and the generator matrices G_LDPC2, G_Half2, G_LT_XLIB, and G_LT_RXLIB;

[0030] Step 6: XOR the XLIB sequence obtained in step 4 and the RMIB1 sequence obtained in step 3 one by one to restore the secondary information symbol LIB sequence;

[0031] Step 7: XOR the check sequence RXLIB of the XLIB sequence obtained in step 5 and the RMIB2 sequence obtained in step 3 one by one to obtain the final check sequence R; the specific process is:

[0032] The check sequence RXLIB of the XLIB sequence obtained in step 5 is XOR operation is performed one by one with the RMIB2 sequence obtained in step 3 to obtain the final verification sequence

[0033] Step 8: Integrate the important information symbol MIB sequence of step 1, the secondary information symbol LIB sequence and the final check sequence R of step 7 to obtain a preliminary double-layer unequal error protection fountain code encoding; the specific process is:

[0034] Integrate the important information symbol MIB sequence in step 1, the secondary information symbol LIB sequence and the final check sequence R in step 7 to obtain the double-layer unequal error protection fountain code encoding

[0035] Step 9, setting the elements of the row corresponding to the abnormal symbol in the matrix G_LT_XLIB in step 5 to zero;

[0036] The decoding method is carried out in the following steps:

[0037] Step 1: Divide the received symbols into important information symbol MIB sequence received symbols, LIB received symbols and R check received symbols according to the index;

[0038] Step 2: Perform R10 Raptor code decoding on the received MIB sequence reception symbols and LIB reception symbols to obtain the correct important intermediate symbol CMIB sequence;

[0039] Step 3, determine whether there are unsolved intermediate symbols in the decoding in step 2, if not, go to step 4, otherwise go to step 8;

[0040] Step 4, performing LT encoding on the correct important intermediate symbol CMIB sequence solved in step 2 to obtain the important information symbol MIB sequence, RMIB1 sequence and RMIB2 sequence;

[0041] Step 5, performing an XOR operation on the RMIB1 sequence and the RMIB2 sequence obtained in step 4, the LIB reception symbol and the R check reception symbol received in step 1 according to the index of the symbol, to obtain the XLIB reception symbol sequence and the XLIB reception check sequence to be solved;

[0042] Step 6: Perform R10 Raptor decoding on the XLIB reception symbol sequence and XLIB reception check sequence to be decoded in step 5 to obtain a CXLIB symbol sequence;

[0043] Step 7, perform R10 Raptor LT encoding on the CXLIB symbol sequence obtained in step 6 to obtain an XLIB symbol sequence, perform XOR on the XLIB symbol sequence and the RMIB1 sequence obtained in step 4 one by one in index order, and obtain the LIB symbol, and the decoding is completed;

[0044] Step 8: If there are unsolved intermediate symbols in the decoding of step 2, the first level decoding is performed. If the CMIB is successfully solved, the process goes to step 3; otherwise, the process goes to step 9.

[0045] Step 9: Construct the second-level decoding matrix and solve the second-level decoding matrix. If it is completely decoded, go to step 10; if it is not completely decoded, set the undecoded symbols to zero to obtain the intermediate symbol CLIB sequence generated by the secondary symbols without abnormal symbols. and sequence CL t ;

[0046] Step 10: Perform R10 Raptor LT encoding on the intermediate symbol CLIB sequence generated by the secondary symbol without abnormal symbols obtained in step 9 to obtain a secondary LIB symbol sequence without abnormal symbols. t The MIB sequence is obtained by operation, and the MIB sequence of all important symbols is obtained, and the decoding is completed.

[0047] Beneficial effects of the present invention:

[0048] A windowed double-layer unequal error protection formalism fountain code is used to recover important symbols with low complexity.

[0049] 1. The present invention does not change the standard R10 Raptor code in the communication field, and the original degree distribution and coding structure are basically maintained unchanged, making full use of its good performance without destroying its systematicity; reducing complexity and overhead.

[0050] 2. The method of the present invention provides a simpler operation of secondary symbols in channel coding information in the communication field. It still uses the encoding and decoding principle of the R10 Raptor code, but adds some special matrix designs. By using clever matrix operations, the designed unequal error protection can also provide better recovery performance for secondary symbols in the communication information. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is a decoding flow chart of the present invention;

[0052] Figure 2 A diagram showing a generator matrix of the R10 Raptor code involved in the present invention;

[0053] Figure 3 A block matrix diagram of the coding generation matrix of the present invention;

[0054] Figure 4 This is a block matrix diagram of the decoding reception generation matrix of the present invention. DETAILED DESCRIPTION

[0055] Specific implementation method 1: The specific process of the encoding and decoding method of a windowed double-layer unequal error protection system fountain code in this implementation method is as follows:

[0056] The encoding method is carried out in the following steps:

[0057] Step 1: First, obtain the two-level priority of the communication information source symbol (binary symbol 0, 1), and divide the communication information source symbol into an important information symbol MIB sequence and a secondary information symbol LIB sequence that is not set to zero; set the number of abnormal symbols E, the larger the number of abnormal symbols, the higher the degree of protection of the important information symbol by the unequal error protection, and evenly select E symbols in the secondary information symbol LIB sequence to be set to zero; the specific process is:

[0058] Divide the K source symbols into K 1 MIB sequence of important information symbols and K 2 A sequence of non-zeroed minor information symbols LIB For E (E≤K 2 ) The secondary information symbol LIB sequence is set to zero to obtain a new LIB sequence

[0059] Step 2: Pre-encode the important information symbol MIB sequence in step 1 with R10 Raptor code to obtain an important intermediate symbol CMIB sequence;

[0060] Perform LT encoding of the R10 Raptor code on the CMIB sequence to obtain the check sequence of the important information symbol MIB; the specific process is:

[0061] The important information symbol MIB sequence obtained in step 1 is pre-coded with R10 Raptor code to obtain the important intermediate symbol CMIB sequence And the generator matrices are G_LDPC1 and G_Half1, with dimensions such as Figure 2 As shown;

[0062] For important intermediate symbols CMIB sequence Perform LT encoding according to the index (1, 2, ..., N) to obtain N output symbol sequences (Important Information Symbol MIB Sequence and the checksum sequence of the important information symbol MIB Combined sequence And the generated matrices are G_LT_MIB, G_LT_RMIB1 and G_LT_RMIB2, with dimensions like Figure 3 As shown, where L 1 and L 2 Respectively represent K 1 and K 2 The number of intermediate symbols calculated under R10 Raptor precoding, S 1 and S 2 Respectively represent the number of LDPC codes generated by MIB and LIB, H 1 and H 2 Respectively represent the number of Half codes generated by MIB and LIB, and N represents the total number of coded symbols generated;

[0063] yes The check sequence (each R is obtained by uniformly XORing K1 MIB sequences according to the degree distribution of R10Raptor code precoding);

[0064] Step 3: The N output symbol sequences obtained in step 2 It is divided into RMIB1 sequence and RMIB2 sequence, where the length of RMIB1 sequence is equal to that of LIB sequence;

[0065] Step 4: Submit the secondary information symbol LIB sequence (LIB sequence) in step 1 ) and the RMIB1 sequence in step 3 are XORed to obtain the XLIB sequence; the specific process is:

[0066] The secondary information symbol LIB sequence obtained in step 1 XOR operation with the RMIB1 sequence generated in step 3 one by one ( and the first XOR operation in the RMIB1 sequence, and the second XOR operation in the RMIB1 sequence), and obtain the XLIB sequence

[0067] Step 5: Encode the XLIB sequence in step 4 by R10 Raptor code to obtain the XLIB sequence and the check sequence RXLIB of the XLIB sequence; the specific process is as follows:

[0068] For the XLIB sequence obtained in step 4 Perform R10Raptor code encoding and obtain an output length of NK 1 XLIB sequence and XLIB sequence check sequence RXLIB integration sequence And the generator matrices G_LDPC2, G_Half2, G_LT_XLIB and G_LT_RXLIB have the dimensions shown in the figure;

[0069] Step 6: XOR the XLIB sequence obtained in step 4 with the RMIB1 sequence obtained in step 3 one by one (the first XOR operation of the XLIB sequence obtained in step 4 with the first XOR operation of the RMIB1 sequence, the second XOR operation of the XLIB sequence obtained in step 4 with the second XOR operation of the RMIB1 sequence), and restore the secondary information symbol LIB sequence (LIB sequence

[0070] Step 7: XOR the check sequence RXLIB of the XLIB sequence obtained in step 5 with the RMIB2 sequence obtained in step 3 one by one (the first XOR operation in RXLIB is performed with the first XOR operation in the RMIB2 sequence, and the second XOR operation in RXLIB is performed with the second XOR operation in the RMIB2 sequence) to obtain the final check sequence R; the specific process is as follows:

[0071] The check sequence RXLIB of the XLIB sequence obtained in step 5 is XOR operation is performed one by one with the RMIB2 sequence obtained in step 3 to obtain the final verification sequence

[0072] Step 8: Integrate the important information symbol MIB sequence and the secondary information symbol LIB sequence (LIB sequence) of step 1 And the final check sequence R in step 7, the preliminary double-layer unequal error protection fountain code is obtained; the specific process is:

[0073] Integrate the important information symbol MIB sequence of step 1 and the secondary information symbol LIB (LIB sequence ) sequence and the final check sequence R of step 7, and the double-layer unequal error protection fountain code is obtained.

[0074] Step 9: Set the elements of the row corresponding to the abnormal symbol in the matrix G_LT_XLIB in step 5 to zero. This step is for the convenience of decoding.

[0075] The decoding method is performed as follows:

[0076] Step 1: Divide the received symbols into important information symbol MIB sequence received symbols, LIB received symbols and R check received symbols according to the index;

[0077] Step 2: Perform R10 Raptor code decoding on the received MIB sequence reception symbols and LIB reception symbols to obtain the correct important intermediate symbol CMIB sequence;

[0078] Step 3, determine whether there are unsolved intermediate symbols in the decoding in step 2, if not, go to step 4, otherwise go to step 8;

[0079] Step 4, performing LT encoding on the correct important intermediate symbol CMIB sequence solved in step 2 to obtain the important information symbol MIB sequence, RMIB1 sequence and RMIB2 sequence;

[0080] Step 5, performing an XOR operation on the RMIB1 sequence and the RMIB2 sequence obtained in step 4, the LIB reception symbol and the R check reception symbol received in step 1 according to the index of the symbol, to obtain the XLIB reception symbol sequence and the XLIB reception check sequence to be solved;

[0081] Step 6: Perform R10 Raptor decoding on the XLIB reception symbol sequence and XLIB reception check sequence to be decoded in step 5 to obtain a CXLIB symbol sequence;

[0082] Step 7, perform R10 Raptor LT encoding on the CXLIB symbol sequence obtained in step 6 to obtain an XLIB symbol sequence, and XOR the XLIB symbol sequence and the RMIB1 sequence obtained in step 4 one by one in index order (the first XLIB symbol sequence and the first RMIB1 sequence, the second XLIB symbol sequence and the second RMIB1 sequence), and obtain the LIB symbol, and the decoding is completed;

[0083] Step 8: If there are unsolved intermediate symbols in the decoding of step 2, the first level decoding is performed. If the CMIB is successfully solved, the process goes to step 3; otherwise, the process goes to step 9.

[0084] Step 9: Construct the second-level decoding matrix and solve the second-level decoding matrix. If it is completely decoded, go to step 10; if it is not completely decoded, set the undecoded symbols to zero to obtain the intermediate symbol CLIB sequence generated by the secondary symbols without abnormal symbols. and sequence CL t ;

[0085] Step 10: Perform R10 Raptor LT encoding on the intermediate symbol CLIB sequence generated by the secondary symbol without abnormal symbols obtained in step 9 to obtain a secondary LIB symbol sequence without abnormal symbols. t The MIB sequence is obtained by operation, and the MIB sequence of all important symbols is obtained, and the decoding is completed.

[0086] Specific implementation method 2: This implementation method is different from specific implementation method 1 in that in step 1, the received symbols are divided into important information symbol MIB sequence received symbols, LIB received symbols and R check received symbols according to the index; the specific process is:

[0087] The receiving sequence is set to

[0088] Assume that the MIB receiving symbol is K r1 The LIB received symbol is K r2 The number of abnormal symbols received is Erd, and the total number of received symbols is M; the receiving matrix is ​​as follows Figure 3 As shown, the prefix Re represents the received matrix obtained after the corresponding row index processing and deletion;

[0089] Generate matrices Re_G_LT_MIB, Re_G_LT_XLIB, Re_G_LT_RMIB1, Re_G_LT_RMIB2, Re_G_LT_RXLIB;

[0090] Re_G_LT_MIB is the submatrix obtained by deleting the rows corresponding to the indices of the missing symbols in G_LT_MIB;

[0091] Re_G_LT_XLIB is the submatrix obtained by deleting the rows corresponding to the indices of the missing symbols in G_LT_XLIB;

[0092] Re_G_LT_RMIB1 is the submatrix obtained by deleting the rows corresponding to the indices of the missing symbols in G_LT_RMIB1;

[0093] Re_G_LT_RMIB2 is the submatrix obtained by deleting the rows corresponding to the indices of the missing symbols in G_LT_RMIB2;

[0094] Re_G_LT_RXLIB is the submatrix obtained by deleting the rows corresponding to the indices of the missing symbols in G_LT_RXLIB.

[0095] The other steps and parameters are the same as those in the first embodiment.

[0096] Specific implementation method three: This implementation method is different from specific implementation method one or two in that in step 2, the received MIB sequence reception symbols and LIB reception symbols are decoded by R10 Raptor code to obtain the correct important intermediate symbol CMIB sequence; the specific process is:

[0097] Based on the generation matrices Re_G_LT_MIB and Re_G_LT_RMIB1 in step 1, receive symbols of the MIB sequence in the received sequence in step 1 Decode with R10Raptor code decoder to get the correct important intermediate symbol CMIB sequence (middle symbol).

[0098] The other steps and parameters are the same as those in the first or second embodiment.

[0099] Specific implementation method 4: This implementation method is different from any one of specific implementation methods 1 to 3 in that in step 4, the correct important intermediate symbol CMIB sequence solved in step 2 is LT-encoded to obtain the important information symbol MIB sequence, RMIB1 sequence and RMIB2 sequence; the specific process is:

[0100] If there are no unsolved intermediate symbols, the complete CMIB sequence of important intermediate symbols is obtained

[0101] For the intermediate symbol sequence According to the index (1,2,...,L 1 ) performs LT encoding to obtain the important information symbol MIB sequence, RMIB1 sequence and RMIB2 sequence

[0102] The other steps and parameters are the same as those in Specific Embodiments 1 to 3.

[0103] Specific implementation mode 5: This implementation mode is different from any one of specific implementation modes 1 to 4 in that in step 5, the RMIB1 sequence and RMIB2 sequence obtained in step 4 are XOR-ed with the LIB reception symbol and R check reception symbol received in step 1 according to the index of the symbol to obtain the XLIB reception symbol sequence and XLIB reception check sequence to be solved; the specific process is:

[0104] The RMIB1 sequence and RMIB2 sequence obtained in step 4 are and the LIB received symbol and the R check received symbol received in step 1 Perform XOR operation according to the index of the symbol to obtain the XLIB received symbol sequence to be solved and XLIB receive check sequence

[0105] The other steps and parameters are the same as those in Specific Embodiments 1 to 4.

[0106] Specific implementation method 6: This implementation method is different from any one of specific implementation methods 1 to 5 in that in step 6, the XLIB reception symbol sequence and the XLIB reception check sequence to be decoded in step 5 are subjected to R10 Raptor decoding to obtain a CXLIB symbol sequence; the specific process is as follows:

[0107] For the XLIB received symbol sequence and XLIB received check sequence to be solved in step 4 Perform R10 Raptor decoding based on the generated matrices Re_G_LT_XLIB and Re_G_LT_RXLIB in step 1 to decode the correct XLIB intermediate symbol CXLIB symbol sequence If there are unresolved CXLIB symbols, force the unresolved CXLIB symbols to zero.

[0108] The other steps and parameters are the same as those in Specific Implementation Methods 1 to 5.

[0109] Specific implementation method 7: This implementation method is different from any one of specific implementation methods 1 to 6 in that in step 7, the CXLIB symbol sequence obtained in step 6 is subjected to R10 Raptor LT encoding to obtain an XLIB symbol sequence, and the XLIB symbol sequence and the RMIB1 sequence obtained in step 4 are XORed one by one in index order (the first XLIB symbol sequence and the first RMIB1 sequence, the second XLIB symbol sequence and the second RMIB1 sequence), and the LIB symbol is obtained, and the decoding is completed; the specific process is:

[0110] The CXLIB symbol sequence obtained in step 6 Perform R10Raptor LT encoding to obtain the XLIB symbol sequence The generated matrix is ​​Re_G_LT_XLIB;

[0111] XLIB symbol sequence and the RMIB1 sequence obtained in step 4 XOR one by one in index order to get the restored LIB symbols

[0112] The other steps and parameters are the same as those in Specific Embodiments 1 to 6.

[0113] Specific implementation eight: This implementation differs from any one of specific implementations one to seven in that, in step 8, when there are unsolved intermediate symbols in the decoding of step 2, the first level decoding is performed, and if the CMIB is successfully solved, the process proceeds to step 3, otherwise, the process proceeds to step 9; the specific process is as follows:

[0114] The first level decoding matrix is:

[0115]

[0116] Where G_LDPC1 is the R10 Raptor LDPC generator matrix of the MIB symbol, which constitutes the first S of the first-level decoding matrix. 1 Row and front K 1 Column (K 1 is to divide the K source symbols into K 1 K among the important information symbols 1 ), G_Half1 is the R10Raptor Half generation matrix of the MIB symbol, which constitutes the Sth layer of the first-level decoding matrix. 1 +1 to S 1 +H 1 Row (S 1 and H 1 According to K 1 Calculated, see background technology) and the previous K 1 +S 1 Column, I S1 For S 1 Re_G_LT_MIB is the received symbol of the MIB sequence in step 1. The matrix generated by decoding with the R10Raptor code decoder constitutes the Sth matrix of the first-level decoding matrix. 1 +H 1 +1 to S 1 +H 1 +K r1 Re_G_LT_RMIB1_Erd is a sub-matrix formed by extracting the rows where all LIB abnormal symbols are located in Re_G_LT_RMIB1, which constitutes the Sth sub-matrix of the first-level decoding matrix. 1 +H 1 +K r1 +1 to the last line; C-MIB indicates the important intermediate symbol CMIB sequence, 0 S1+L1 is a zero matrix, L 1 represents the number of intermediate symbols calculated under R10 Raptor precoding, S 1 Indicates the number of LDPC codes generated by MIB, and Re_D_MIB indicates the received MIB symbols Re_D_Erd ​​represents the received abnormal symbol. The position of Re_D_Erd ​​is the RMIB1 check symbol, and the corresponding check matrix is ​​Re_G_LT_RMIB1_Erd;

[0117] Solve the first-level decoding matrix. If the CMIB is successfully solved, go to step 3, otherwise go to step 9.

[0118] The other steps and parameters are the same as those in Specific Embodiments 1 to 7.

[0119] Specific implementation method 9: This implementation method is different from any one of specific implementation methods 1 to 8 in that in step 9, a second-level decoding matrix is ​​constructed and the second-level decoding matrix is ​​solved. If the decoding is complete, the process proceeds to step 10; if the decoding is not complete, the undecoded symbols are set to zero to obtain the intermediate symbol CLIB sequence generated by the secondary symbols without abnormal symbols. and sequence CL t ; The specific process is:

[0120] The important intermediate symbol CMIB sequence is represented by the matrix C-MIB, and the important intermediate symbol CMIB sequence is divided into blocks to obtain CL Lt and CL t , the block format is C-MIB = [CL Lt CL t ], Lt is the number of intermediate symbols for which the first-level decoding matrix is ​​successfully solved, and t is the number of intermediate symbols for which the first-level decoding matrix is ​​not successfully solved;

[0121] After removing the elements of Re_G_LT_RMIB1_Erd from Re_G_LT_RMIB1, merge Re_G_LT_RMIB2 in step 1 (the merge is to put the sequence after Re_G_LT_RMIB1 removes the elements of Re_G_LT_RMIB1_Erd on the top, and Re_G_LT_RMIB2 on the bottom) to form a new submatrix Re′_G_LT_RMIB. Let the number of rows of Re′_G_LT_RMIB be Kt. After dividing Re′_G_LT_RMIB into blocks, we get Gg Kt×Lt and Gg Kt×t , the block form is Re′_G_LT_RMIB=[Gg Kt×Lt Gt Kt×t ];

[0122] Step 8 failed to decode successfully, and a partial decoding matrix of the first-level decoding matrix was obtained, which is expressed as:

[0123]

[0124] In the formula, I Lt×Lt represents the identity matrix, F Lt×tIndicates the subarray that is not processed after the first level decoding fails, 0 t×Lt is a zero matrix, 0 t×t is a zero matrix, 0 t represents a vector of dimension t, D F Indicates the transformation of the received symbols according to the decoding plan during R10 Raptor decoding, and its transformation synchronizes the transformation of the decoding matrix; CF Lt Lt intermediate symbols are partially solved by the first-level decoding matrix;

[0125] The second level decoding matrix is:

[0126]

[0127] Where G_LDPC2 is the R10 Raptor LDPC generator matrix for LIB symbols, which constitutes the first S of the second-level decoding matrix. 2 Row and front K 2 Column (K 2 is to divide the K source symbols into K 2 K 2 ); G_Half2 is the R10 Raptor Half generation matrix of the LIB symbol, which constitutes the S of the second-level decoding matrix 2 +1 to H 2 Line and front S 2 +K 2 Column (S 2 and H 2 According to K 2 Calculated, see background technology); I S2 Indicates S 2 The first S of the second-level decoding matrix 2 Row and K 2 +1 to K 2 +S 2 Column; Re′_G_LT_RX+XLIB represents the new matrix formed by the Re_G_LT_XLIB in step 1 after removing the submatrix where the abnormal symbol is located, and the lower merging Re_G_LT_RXLIB (the lower merging is to put the submatrix where the abnormal symbol is located after Re_G_LT_XLIB is removed on the top, and Re_G_LT_RXLIB is placed on the bottom), which constitutes the H of the second-level decoding matrix 2 +S 2 +1 line to the last line and before S 2 +K 2 +H 2 Column; F Lt×t Indicates the subarray that was not processed after the first level decoding failed. Indicates exclusive OR; The Hth2 +S 2 +1 line to the last line and S 2 +K 2 +H 2 +1 column to the last column; C LIB Indicates the intermediate symbol CLIB sequence of LIB, 0 S2+L2 is a zero matrix, L 2 K 2 The number of intermediate symbols calculated under R10 Raptor precoding, S 2 Re_LIB represents the number of LDPC codes generated by LIB, Re_LIB represents the received LIB symbols and all check symbols after removing abnormal symbols; Kt is the number of rows of Re′_G_LT_RMIB, and Gg is obtained by dividing Re′_G_LT_RMIB into blocks. Kt×Lt and Gt Kt×t , the block form is Re′_G_LT_RMIB=[Gg Kt×Lt G Kt×t ], Re′_G_LT_RMIB represents the matrix formed by removing the elements of Re_G_LT_RMIB1_Erd from Re_G_LT_RMIB1 and merging it with Re_G_LT_RMIB2;

[0128] Solve the second-level decoding matrix. If it is not completely decoded, set the undecoded symbols to zero to obtain the intermediate symbol CLIB sequence generated by the secondary symbols without abnormal symbols. and sequence CL t ; If fully decoded C LIB and CL t Then go to step 10.

[0129] The other steps and parameters are the same as those in Specific Embodiments 1 to 8.

[0130] Specific implementation method 10: This implementation method is different from any one of specific implementation methods 1 to 9 in that in step 10, the intermediate symbol CLIB sequence generated by the secondary symbol without abnormal symbols obtained in step 9 is subjected to R10 Raptor LT encoding to obtain a secondary LIB symbol sequence without abnormal symbols, and the CL t The MIB sequence is obtained by operation, and the MIB sequence of all important symbols is obtained, and the decoding is completed; the specific process is:

[0131] The intermediate symbol CLIB sequence generated by the secondary symbol without abnormal symbols obtained in step 9 is encoded by R10Raptor LT. The LT encoding follows Degree distribution of the sequence, get the minor LIB symbol sequence without abnormal symbols About CL tPerform operations to obtain the MIB sequence, and solve the MIB sequence of all important symbols Decoding completed;

[0132] Among them, for CL t The MIB sequence obtained by operation is expressed as:

[0133]

[0134] The other steps and parameters are the same as those in Specific Embodiments 1 to 9.

[0135] The present invention may also have many other embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art may make various corresponding changes and modifications based on the present invention, but these corresponding changes and modifications should all fall within the scope of protection of the claims attached to the present invention.

Claims

1. A coding and decoding method for a windowed double-layer unequal error protection system fountain code, characterized in that: The specific process of the method is: The encoding method is carried out in the following steps: Step 1: First, obtain the two-layer priority of the communication information source symbol, and divide the communication information source symbol into an important information symbol MIB sequence and a secondary information symbol LIB sequence that is not set to zero; set the number of abnormal symbols E, and evenly select E symbols from the secondary information symbol LIB sequence to set to zero; the specific process is: Divide K source symbols into K1 important information symbol MIB sequences and K2 non-zero secondary information symbols LIB sequence Set the E (E≤K2) secondary information symbol LIB sequences to zero to obtain a new LIB sequence Step 2: Pre-encode the important information symbol MIB sequence in step 1 with R10 Raptor code to obtain an important intermediate symbol CMIB sequence; Perform LT encoding of the R10 Raptor code on the CMIB sequence to obtain the check sequence of the important information symbol MIB; the specific process is: The important information symbol MIB sequence obtained in step 1 is pre-coded with R10 Raptor code to obtain the important intermediate symbol CMIB sequence And the generator matrices are G_LDPC1 and G_Half1; For important intermediate symbols CMIB sequence Perform LT encoding according to the index (1, 2, ..., N) to obtain N output symbol sequences and the generating matrices are G_LT_MIB, G_LT_RMIB1 and G_LT_RMIB2, where N represents the total number of generated coding symbols; yes The check sequence of Step 3: The N output symbol sequences obtained in step 2 It is divided into RMIB1 sequence and RMIB2 sequence, where the length of RMIB1 sequence is equal to that of LIB sequence; Step 4: XOR the secondary information symbol LIB sequence in step 1 and the RMIB1 sequence in step 3 to obtain the XLIB sequence; the specific process is: The secondary information symbol LIB sequence obtained in step 1 XOR operation is performed one by one with the RMIB1 sequence generated in step 3 to obtain the XLIB sequence Step 5: Encode the XLIB sequence in step 4 by R10 Raptor code to obtain the XLIB sequence and the check sequence RXLIB of the XLIB sequence; the specific process is as follows: For the XLIB sequence obtained in step 4 Perform R10Raptor code encoding to obtain the integrated sequence of the XLIB sequence with an output length of N-K1 and the check sequence RXLIB of the XLIB sequence and the generator matrices G_LDPC2, G_Half2, G_LT_XLIB, and G_LT_RXLIB; Step 6: XOR the XLIB sequence obtained in step 4 and the RMIB1 sequence obtained in step 3 one by one to restore the secondary information symbol LIB sequence; Step 7: XOR the check sequence RXLIB of the XLIB sequence obtained in step 5 and the RMIB2 sequence obtained in step 3 one by one to obtain the final check sequence R; the specific process is: The check sequence of the XLIB sequence obtained in step 5 XOR operation is performed one by one with the RMIB2 sequence obtained in step 3 to obtain the final verification sequence Step 8: Integrate the important information symbol MIB sequence of step 1, the secondary information symbol LIB sequence and the final check sequence R of step 7 to obtain a preliminary double-layer unequal error protection fountain code encoding; the specific process is: Integrate the important information symbol MIB sequence in step 1, the secondary information symbol LIB sequence and the final check sequence R in step 7 to obtain the double-layer unequal error protection fountain code encoding Step 9, setting the elements of the row corresponding to the abnormal symbol in the matrix G_LT_XLIB in step 5 to zero; The decoding method is performed as follows: Step 1: Divide the received symbols into important information symbol MIB sequence received symbols, LIB received symbols and R check received symbols according to the index; Step 2: Perform R10 Raptor code decoding on the received MIB sequence reception symbols and LIB reception symbols to obtain the correct important intermediate symbol CMIB sequence; Step 3, determine whether there are unsolved intermediate symbols in the decoding in step 2, if not, go to step 4, otherwise go to step 8; Step 4, performing LT encoding on the correct important intermediate symbol CMIB sequence solved in step 2 to obtain the important information symbol MIB sequence, RMIB1 sequence and RMIB2 sequence; Step 5, performing an XOR operation on the RMIB1 sequence and the RMIB2 sequence obtained in step 4, the LIB reception symbol and the R check reception symbol received in step 1 according to the index of the symbol, to obtain the XLIB reception symbol sequence and the XLIB reception check sequence to be solved; Step 6: Perform R10 Raptor decoding on the XLIB reception symbol sequence and XLIB reception check sequence to be decoded in step 5 to obtain a CXLIB symbol sequence; Step 7, perform R10 Raptor LT encoding on the CXLIB symbol sequence obtained in step 6 to obtain an XLIB symbol sequence, perform XOR on the XLIB symbol sequence and the RMIB1 sequence obtained in step 4 one by one in index order, and obtain the LIB symbol, and the decoding is completed; Step 8: If there are unsolved intermediate symbols in the decoding of step 2, the first level decoding is performed. If the CMIB is successfully solved, the process goes to step 3; otherwise, the process goes to step 9. Step 9: Construct the second-level decoding matrix and solve the second-level decoding matrix. If it is completely decoded, go to step 10; if it is not completely decoded, set the undecoded symbols to zero to obtain the intermediate symbol CLIB sequence generated by the secondary symbols without abnormal symbols. and sequence CL t ; Step 10: Perform R10Raptor LT encoding on the intermediate symbol CLIB sequence generated by the secondary symbol without abnormal symbols obtained in step 9 to obtain a secondary LIB symbol sequence without abnormal symbols. t The MIB sequence is obtained by operation, and the MIB sequence of all important symbols is obtained, and the decoding is completed.

2. The encoding and decoding method of a windowed double-layer unequal error protection system fountain code according to claim 1 is characterized in that: In the step 1, the received symbols are divided into important information symbol MIB sequence received symbols, LIB received symbols and R check received symbols according to the index; The specific process is: The receiving sequence is set to Assume that the MIB receiving symbol is K r1 The LIB received symbol is K r2 The number of abnormal symbols received is Erd, and the total number of received symbols is M; Generate matrices Re_G_LT_MIB, Re_G_LT_XLIB, Re_G_LT_RMIB1, Re_G_LT_RMIB2, Re_G_LT_RXLIB; Re_G_LT_MIB is the submatrix obtained by deleting the rows corresponding to the indices of the missing symbols in G_LT_MIB; Re_G_LT_XLIB is the submatrix obtained by deleting the rows corresponding to the indices of the missing symbols in G_LT_XLIB; Re_G_LT_RMIB1 is the submatrix obtained by deleting the rows corresponding to the indices of the missing symbols in G_LT_RMIB1; Re_G_LT_RMIB2 is the submatrix obtained by deleting the rows corresponding to the indices of the missing symbols in G_LT_RMIB2; Re_G_LT_RXLIB is the submatrix obtained by deleting the rows corresponding to the indices of the missing symbols in G_LT_RXLIB.

3. The encoding and decoding method of a windowed double-layer unequal error protection system fountain code according to claim 2 is characterized in that: In step 2, the received MIB sequence reception symbols and LIB reception symbols are decoded by R10 Raptor code to obtain the correct important intermediate symbol CMIB sequence; the specific process is: Based on the generation matrices Re_G_LT_MIB and Re_G_LT_RMIB1 in step 1, receive symbols of the MIB sequence in the received sequence in step 1 Decode with R10Raptor code decoder to get the correct important intermediate symbol CMIB sequence 4. The encoding and decoding method of a windowed double-layer unequal error protection system fountain code according to claim 3 is characterized in that: In step 4, the correct important intermediate symbol CMIB sequence solved in step 2 is LT-encoded to obtain the important information symbol MIB sequence, RMIB1 sequence and RMIB2 sequence; the specific process is: If there are no unsolved intermediate symbols, the complete CMIB sequence of important intermediate symbols is obtained For the intermediate symbol sequence Perform LT encoding according to the index (1,2,...,L1) to obtain the important information symbol MIB sequence, RMIB1 sequence and RMIB2 sequence 5. The encoding and decoding method of a windowed double-layer unequal error protection system fountain code according to claim 4 is characterized in that: In step 5, the RMIB1 sequence and the RMIB2 sequence obtained in step 4 are XOR-ed with the LIB reception symbol and the R check reception symbol received in step 1 according to the index of the symbol to obtain the XLIB reception symbol sequence and the XLIB reception check sequence to be solved; the specific process is: The RMIB1 sequence and RMIB2 sequence obtained in step 4 are and the LIB received symbol and the R check received symbol received in step 1 Perform XOR operation according to the index of the symbol to obtain the XLIB received symbol sequence to be solved and XLIB receive check sequence 6. The encoding and decoding method of a windowed double-layer unequal error protection system fountain code according to claim 5, characterized in that: In step 6, the XLIB reception symbol sequence and the XLIB reception check sequence to be decoded in step 5 are subjected to R10Raptor decoding to obtain the CXLIB symbol sequence; the specific process is as follows: For the XLIB received symbol sequence and XLIB received check sequence to be solved in step 4 Perform R10 Raptor decoding based on the generated matrices Re_G_LT_XLIB and Re_G_LT_RXLIB in step 1 to obtain the correct XLIB intermediate symbol CXLIB symbol sequence If there are unresolved CXLIB symbols, set the unresolved CXLIB symbols to zero.

7. The encoding and decoding method of a windowed double-layer unequal error protection system fountain code according to claim 6 is characterized in that: In step 7, the CXLIB symbol sequence obtained in step 6 is encoded by R10 Raptor LT to obtain an XLIB symbol sequence, and the XLIB symbol sequence and the RMIB1 sequence obtained in step 4 are XORed one by one in index order to obtain the LIB symbol, and the decoding is completed; the specific process is: The CXLIB symbol sequence obtained in step 6 Perform R10Raptor LT encoding to obtain the XLIB symbol sequence The generated matrix is ​​Re_G_LT_XLIB; XLIB symbol sequence and the RMIB1 sequence obtained in step 4 XOR one by one in index order to get the restored LIB symbols 8. The encoding and decoding method of a windowed double-layer unequal error protection system fountain code according to claim 7, characterized in that: In step 8, when there are unsolved intermediate symbols in the decoding of step 2, the first level decoding is performed. If the CMIB is successfully solved, the process proceeds to step 3, otherwise, the process proceeds to step 9. The specific process is as follows: The first level decoding matrix is: Where G_LDPC1 is the R10 Raptor LDPC generator matrix of the MIB symbol, which constitutes the first S1 rows and the first K1 columns of the first-level decoding matrix. G_Half1 is the R10 Raptor Half generator matrix of the MIB symbol, which constitutes the S1+1 to S1+H1 rows and the first K1+S1 columns of the first-level decoding matrix. S1 is the S1-order unit matrix, Re_G_LT_MIB is the received symbol of the MIB sequence in step 1 The matrix generated by decoding with the R10Raptor code decoder constitutes the first-level decoding matrix from S1+H1+1 to S1+H1+K r1 Re_G_LT_RMIB1_Erd is a submatrix formed by extracting the rows where all LIB abnormal symbols are located in Re_G_LT_RMIB1, which constitutes the S1+H1+Kth submatrix of the first-level decoding matrix. r1 +1 to the last line; C-MIB indicates the important intermediate symbol CMIB sequence, 0 S1+L1 is a zero matrix, L1 represents the number of intermediate symbols calculated under R10 Raptor precoding, S1 represents the number of LDPC codes generated by MIB, and Re_D_MIB represents the received MIB symbols Re_D_Erd ​​represents the received abnormal symbol. The position of Re_D_Erd ​​is the RMIB1 check symbol, and the corresponding check matrix is ​​Re_G_LT_RMIB1_Erd; Solve the first-level decoding matrix. If the CMIB is successfully solved, go to step 3, otherwise go to step 9.

9. The encoding and decoding method of a windowed double-layer unequal error protection system fountain code according to claim 8, characterized in that: In step 9, the second level decoding matrix is ​​constructed and solved. If the decoding is complete, the process proceeds to step 10. If the decoding is not complete, the undecoded symbols are set to zero to obtain the intermediate symbol CLIB sequence generated by the secondary symbols without abnormal symbols. and sequence CL t ; The specific process is: The important intermediate symbol CMIB sequence is represented by the matrix C-MIB, and the important intermediate symbol CMIB sequence is divided into blocks to obtain CL Lt and CL t , the block format is C-MIB = [CL Lt CL t ], Lt is the number of intermediate symbols for which the first-level decoding matrix is ​​successfully solved, and t is the number of intermediate symbols for which the first-level decoding matrix is ​​not successfully solved; After removing the elements of Re_G_LT_RMIB1_Erd from Re_G_LT_RMIB1, merge Re_G_LT_RMIB2 in step 1 to form a new submatrix Re′_G_LT_RMIB. Let the number of rows of Re′_G_LT_RMIB be Kt. After dividing Re′_G_LT_RMIB into blocks, we get Gg Kt×Lt and Gg Kt×t , the block form is Re′_G_LT_RMIB=[Gg Kt×Lt Gt Kt×t ]; Step 8 failed to decode successfully, and a partial decoding matrix of the first-level decoding matrix was obtained, which is expressed as: In the formula, I Lt×Lt represents the identity matrix, F Lt×t Indicates the subarray that is not processed after the first level decoding fails, 0 t×Lt is a zero matrix, 0 t×t is a zero matrix, 0 t represents a vector of dimension t, D F Indicates that the received symbols are transformed according to the decoding plan during R10 Raptor decoding; CF Lt Lt intermediate symbols are partially solved by the first-level decoding matrix; The second level decoding matrix is: Where G_LDPC2 is the R10 Raptor LDPC generator matrix for LIB symbols, which constitutes the first S2 rows and the first K2 columns of the second-level decoding matrix; G_Half2 is the R10 Raptor Half generator matrix for LIB symbols, which constitutes the S2+1 to H2 rows and the first S2+K2 columns of the second-level decoding matrix; I S2 Re′_G_LT_RX+XLIB represents the S2-order unit matrix, which constitutes the first S2 rows and K2+1 to K2+S2 columns of the second-level decoding matrix; Re′_G_LT_RX+XLIB represents the new matrix formed by merging Re_G_LT_RXLIB with the submatrix after the row where the abnormal symbol is located in Re_G_LT_XLIB in step 1, which constitutes the H2+S2+1 row to the last row and the first S2+K2+H2 columns of the second-level decoding matrix; F Lt×t Indicates the subarray that was not processed after the first level decoding failed. Indicates exclusive OR; The H2+S2+1th row to the last row and the S2+K2+H2+1th column to the last column constitute the second-level decoding matrix; C LIB Indicates the intermediate symbol CLIB sequence of LIB, 0 S2+L2 is a zero matrix, L2 represents the number of intermediate symbols calculated by K2 under R10 Raptor precoding, S2 represents the number of LDPC codes generated by LIB, Re_LIB represents the received LIB symbols and all check symbols after the abnormal symbols are removed; Kt is the number of rows of Re′_G_LT_RMIB, and Gg is obtained by dividing Re′_G_LT_RMIB into blocks. Kt×Lt and Gt Kt×t , the block form is Re′_G_LT_RMIB=[Gg Kt×Lt G Kt×t ], Re′_G_LT_RMIB represents the matrix formed by removing the elements of Re_G_LT_RMIB1_Erd from Re_G_LT_RMIB1 and merging it with Re_G_LT_RMIB2; Solve the second-level decoding matrix. If it is not completely decoded, set the undecoded symbols to zero to obtain the intermediate symbol CLIB sequence generated by the secondary symbols without abnormal symbols. and sequence CL t ; If fully decoded C LIB and CL t Then go to step 10.

10. The encoding and decoding method of a windowed double-layer unequal error protection system fountain code according to claim 9, characterized in that: In step 10, the intermediate symbol CLIB sequence generated by the secondary symbol without abnormal symbols obtained in step 9 is encoded by R10 Raptor LT to obtain a secondary LIB symbol sequence without abnormal symbols. t The MIB sequence is obtained by operation, and the MIB sequence of all important symbols is obtained, and the decoding is completed; the specific process is: The intermediate symbol CLIB sequence generated by the secondary symbol without abnormal symbols obtained in step 9 is encoded by R10 RaptorLT. The LT encoding follows Degree distribution of the sequence, get the minor LIB symbol sequence without abnormal symbols About CL t Perform operations to obtain the MIB sequence, and solve the MIB sequence of all important symbols Decoding completed; Among them, for CL t The MIB sequence obtained by operation is expressed as:

Images