A decoding verification based LDPC code encoding identification method

Abstract

The present application belongs to the technical field of channel coding and decoding, and particularly relates to a LDPC code encoding identification method based on decoding verification. The method is based on the properties of LDPC codes, and in the verification of identification, the product of the codeword and the transposed check matrix, i.e. the property that the determinant (syndrome) is 0, is used to verify the decoding result of the LDPC, and the accurate identification result is output. Further, in order to increase fault tolerance, the threshold of the syndrome is set, so that correct encoding identification results can still be obtained under a small amount of uncorrectable errors. The method can also dynamically set the threshold of the syndrome according to different Signal Noise Ratios (SNRs).

Description

Technical Field

[0001] This invention belongs to the field of channel coding and decoding technology, specifically relating to an LDPC code encoding and recognition method based on decoding verification. Background Technology

[0002] Channel coding, as one of the core technologies of communication systems, plays a crucial role in the reliability and stability of signals during transmission. By adding redundancy to the signal source, channel coding can effectively combat noise, interference, and fading in the channel, thereby achieving a balance between transmission efficiency and reliability. The goal of blind identification technology in channel coding is to automatically identify the channel coding scheme and its technical parameters used by the transmitter based solely on the received signal or bitstream, even without knowing the coding parameters (such as coding type, code length, code rate, etc.). This not only supports Adaptive Modulation and Coding (AMC) technology but also improves the performance of wireless communication systems. Therefore, it is widely used in non-cooperative communication, adaptive communication systems, and intelligent jamming countermeasures.

[0003] Low-density parity-check (LDPC) codes are characterized by: using a sparse parity-check matrix to reduce decoding complexity; and employing a Tanner graph-based belief propagation algorithm for efficient error correction through message passing between nodes. Their advantages include: performance approaching the Shannon limit for long code lengths; faster bit error rate reduction in high signal-to-noise ratio regions, making them suitable for high-reliability scenarios; and high parallelism and throughput in decoding, making them suitable for hardware acceleration. LDPC codes are widely used due to their strong error correction capabilities and efficient encoding / decoding performance. Therefore, encoding and recognizing LDPC codes can provide crucial information for further signal processing. Summary of the Invention

[0004] To address the issues of misjudgment and omission in existing LDPC code recognition technologies, this invention performs a judgment based on the decoded result using the determinant (syndrome) of the LDPC code. The closed-set condition represents the use of known parity-check matrices for multiple LDPC codes. After decoding a series of LDPC codes with the same code length but different code rates or encoding types, the parity-check matrices are used for verification, ultimately achieving the recognition objective. To increase error tolerance, a threshold is set for the syndrome, ensuring correct code recognition results even with a small number of uncorrectable errors.

[0005] The technical solution adopted in this invention is as follows:

[0006] A method for identifying LDPC codes based on decoding verification is proposed. The method sets the closed-set condition as the known parity-check matrices of multiple LDPC codes. Under this condition, the method includes the following steps:

[0007] S1. Using soft-decision data from multiple LDPC codes with the same code length but different code rates or encoding types, perform LLR BP decoding to obtain the decoding result, specifically:

[0008] 1) Calculate the number of bits transmitted by the sending end. or The initial posterior probability is set as the initial posterior probability of the bits sent by the sender, based on the initial message passed from the variable node to the verification node.

[0009] 2) Calculate the extrinsic information passed from the verification node to the variable node;

[0010] 3) Calculate the external information passed from the variable node to the verification node;

[0011] 4) Hard decision is made based on the posterior probabilities of the variable nodes after iteration to obtain the decoding result;

[0012] S2. Combine the decoding result with the transpose of the LDPC code parity-check matrix. Multiplying them together gives the determinant;

[0013] S3. Determine if the determinant is a vector of all zeros. If so, identify the code length and code rate parameters of LDPC based on the number of rows and columns of the parity check matrix and output them; otherwise, proceed to S4.

[0014] S4. Determine whether the determinant is within the set threshold. If so, identify the LDPC code length and code rate parameters based on the number of rows and columns of the parity check matrix and output them. Otherwise, determine that the identification has failed and the data is not an LDPC codeword under the parity check matrix.

[0015] Furthermore, the sending end is set to send bits as or The initial posterior probability is or The initial message that the variable node sends to the verification node is: and , Represents variable nodes Passed to the verification node The initial message;

[0016] The verification node is at the The external information passed to the variable node in the next iteration is and The calculation formula is:

[0017] ,

[0018] in, Indicates except the first Outside of the variable node, with the first variable node The set of other variable nodes connected to each verification node;

[0019] The variable node is at the 1st The external information passed to the verification node in the next iteration is and The calculation formula is:

[0020] ,

[0021] in, Indicates except the first In addition to the first verification node, with the first The set of other verification nodes connected to each variable node; LLR decision formula ;

[0022] The posterior probability of the variable node after iteration is:

[0023] ,

[0024] If the posterior probability Then the bits will be sent. The judgment is Conversely, the judgment is .

[0025] The parameters in the text are uniformly defined as follows:

[0026] For the first One variable node;

[0027] For the first One verification node;

[0028] For the first Sub-iteration verification node Passed to variable node External information, ;

[0029] For the first Next iteration variable node Passed to the verification node External information, ;

[0030] In order to be with the first The set of all check nodes connected to each variable node. ;

[0031] In order to be with the first The set of all variable nodes connected to each verification node. ;

[0032] To exclude the first In addition to the first verification node, with the first The set of other check nodes connected to each variable node. ;

[0033] To exclude the first Outside of the variable node, with the first variable node The set of other variable nodes connected to each verification node. ;

[0034] For the receiving end to receive Then, the corresponding endcode word sent. The posterior probability, ;

[0035] For the first The second iteration Posterior probability information of each variable node .

[0036] This method, which re-encodes the decoded LDPC code and compares it with the input hard-decision data, suffers from inherent errors due to the input bitstream being a noisy bitstream. Because LDPC codes have strong error correction capabilities, even successful decoding and error correction may result in the code being judged as "incorrect" due to differences in the comparison, thus participating in the threshold determination and leading to missed judgments. Setting a broad threshold increases the risk of misjudgments. The proposed solution can set a single threshold through simulation, or a dynamic threshold based on the demodulated Signal Noise Ratio (SNR). When the SNR is high, a lower threshold is set to achieve more accurate threshold recognition results.

[0037] The beneficial effects of this invention are as follows: Based on the properties of LDPC codes, in the verification of recognition, the product of the codeword and the transpose of the parity check matrix, i.e., the property that the determinant (syndrome) is 0, is used to verify the decoding result of LDPC and output an accurate recognition result. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of the LDPC code encoding and recognition method based on decoding verification in an embodiment of the present invention;

[0039] Figure 2This is a schematic diagram of the LDPC LLR BP decoding process in an embodiment of the present invention;

[0040] Figure 3 The figures show the simulation results of the performance of two LDPC code specifications selected in the embodiments of the present invention.

[0041] Figure 4 In this embodiment of the invention, after using one type of LDPC encoding, two types of matrices are used for decoding, and the SNR corresponds to the syndrome error-rate result diagram. Detailed Implementation

[0042] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings and examples.

[0043] Example:

[0044] like Figure 1 As shown, this example provides an LDPC code encoding recognition method based on decoding verification, including the following steps S1 to S4:

[0045] S1. Select two LDPC codes with the same code length. and The specific parameters for the two LDPC methods are shown in Table 1, and the performance simulation curves of the matrix are shown in Table 1. Figure 3 .

[0046] Table 1 Simulation parameters for two LDPC methods

[0047]

[0048] S2, Use The matrix will contain the information source Encoded into LDPC codewords Soft bit information is obtained by adding noise at a range of 5.3~5.9dB. ( The length of the codeword. The source length is the length of the encoded information bits. (Length of the check bits after encoding).

[0049] Note: Selection The addition of noise at 5.3~5.9dB after encoding is because, at this SNR, the LDPC contains both correctable and uncorrectable codewords. Furthermore, simulation results show that this SNR also... Within the scope of code correction.

[0050] S3. Soft bit information using two different LDPC code specifications. and Perform decoding using LLR Belief Propagation (BP) and output the decoding result. and The decoding flowchart is as follows: Figure 2 As shown;

[0051] In this embodiment, step S3 specifically includes the following sub-steps:

[0052] S31. Calculate the bits sent by the sender. or The initial posterior probability is set as the initial posterior probability of the bits sent by the sender, based on the initial message passed from the variable node to the verification node: and ;

[0053] S32. Calculate the external information passed from the check node to the variable node: and ;

[0054] S33. Calculate the external information passed from the variable node to the verification node: and ;

[0055] S34. Decoding Decision: The posterior probability of the variable nodes after iteration. and Hard decision is made based on the probabilities of 0 and 1 to obtain the decoded output. .

[0056] S4. Decode the result and respectively with the transpose of the parity-check matrix and Multiply to get syndrome and :

[0057]

[0058] S5. Find the syndrome vector that is all zeros. Then, based on the number of rows and columns in the parity-check matrix, identify the LDPC code length, code rate, and other parameters. The decoded output is as follows:

[0059]

[0060]

[0061] Otherwise, the syndrome error rate is calculated and the threshold determination process begins.

[0062]

[0063] The syndrome error rate is calculated as follows:

[0064]

[0065] The simulated SNR-syndrome error-rate curve is as follows: Figure 4 .

[0066] S6. If the syndrome error-rate is not an all-zero vector, then determine whether the syndrome error-rate is within the set threshold. If so, the LDPC code length, code rate, and other parameters are identified based on the number of rows and columns in the parity check matrix; otherwise, the identification fails, and the data is not an LDPC codeword under this parity check matrix.

[0067] This completes the entire process of implementing the present invention.

[0068] In the embodiments of the present invention, two LDPC specifications were simulated, and the specific parameters are shown in Table 1. The simulation results are as follows: Figure 3 After processing S1 to S5, the syndrome error-rate data under different SNRs are shown in Table 2, and the simulation results are as follows. Figure 4 .

[0069] Table 2. Syndrome error-rate results under different SNRs using different LDPC decoding methods.

[0070]

[0071] Under the threshold determination, a confidence level needs to be added to the determination result to distinguish it from accurate recognition. This ensures that a recognition result can still be output even with a small number of uncorrectable errors.

[0072] The threshold can be dynamically set based on the demodulated SNR. Simulation results show that under high signal-to-noise ratio conditions, setting a lower threshold can significantly reduce the risk of misjudgment.

Claims

1. A method for encoding and recognizing LDPC codes based on decoding verification, wherein the closed-set condition is defined as the known parity-check matrices of multiple LDPC codes, and under the closed-set condition, the method is characterized in that... The method includes the following steps: S1. Using soft-decision data from multiple LDPC codes with the same code length but different code rates or encoding types, perform LLR BP decoding to obtain the decoding result, specifically: 1) Calculate the number of bits transmitted by the sending end. or The initial posterior probability is set as the initial posterior probability of the bits sent by the sender, based on the initial message passed from the variable node to the verification node. 2) Calculate the extrinsic information passed from the verification node to the variable node; 3) Calculate the external information passed from the variable node to the verification node; 4) Hard decision is made based on the posterior probabilities of the variable nodes after iteration to obtain the decoding result; S2. Combine the decoding result with the transpose of the LDPC code parity-check matrix. Multiplying them together gives the determinant; S3. Determine if the determinant is a vector of all zeros. If so, identify the code length and code rate parameters of LDPC based on the number of rows and columns of the parity check matrix and output them; otherwise, proceed to S4. S4. Determine whether the determinant is within the set threshold. If so, identify the LDPC code length and code rate parameters based on the number of rows and columns of the parity check matrix and output them. Otherwise, determine that the identification has failed and the data is not an LDPC codeword under the parity check matrix.

2. The LDPC code encoding and recognition method based on decoding verification according to claim 1, characterized in that, In S1, the transmitting end is set to transmit bits as follows: or The initial posterior probability is or The initial message that the variable node sends to the verification node is: and , Represents variable nodes Passed to the verification node The initial message; The verification node is at the The external information passed to the variable node in the next iteration is and The calculation formula is: ， in, Indicates except the first Outside of the variable node, with the first variable node The set of other variable nodes connected to each verification node; The variable node is at the 1st The external information passed to the verification node in the next iteration is and The calculation formula is: ， in, Indicates except the first In addition to the first verification node, with the first The set of other verification nodes connected to each variable node; LLR decision formula ; The posterior probability of the variable node after iteration is: ， If the posterior probability Then the bits will be sent. The judgment is Conversely, the judgment is .

Images