An Adaptive Non-Uniform Quantization Decoding Method

Abstract

The present invention discloses an adaptive non-uniform quantization decoding method, belonging to the field of electronic communication technology, which includes: obtaining the variable node initialization message and the received message, performing the quantization decoding of the first iteration on the received message by using the variable node initialization message, and obtaining the first iteration decision codeword matrix composed of the first iteration decision codewords; performing discrimination by using the above matrix, if the preset iteration end requirement is satisfied, outputting the first iteration decision codeword, if not, calculating the adaptive quantization parameter according to the first iteration decision codeword and performing non-uniform quantization on the message transmitted by the variable node by using it, then increasing the iteration number by one and performing the quantization decoding of the next iteration on the previous iteration decision codeword by using the message transmitted by the variable node to the check node in the previous iteration until the preset iteration end requirement is satisfied, outputting the decision codeword, and completing the decoding; improving the problem that the traditional method has a low adaptability to the rapid change of parameters in the decoding.

Description

Technical Field

[0001] The present invention relates to an adaptive non-uniform quantization decoding method, belonging to the technical field of electronic communication. Background Art

[0002] At present, the decoding methods of Low Density Parity Check (LDPC) codes mainly include Belief Propagation (BP) decoding algorithm, Log-Likelihood Ratio Belief Propagation (LLR-BP) algorithm based on log-likelihood ratio, and Minimum-sum (MS) decoding algorithm.

[0003] The BP algorithm needs to use multiplication operations to judge the confidence of messages. The LLR-BP algorithm introduces hyperbolic tangent function and inverse hyperbolic tangent function to reduce the multiplication operations in the BP algorithm. However, the inverse hyperbolic tangent function is difficult to implement on hardware. The minimum-sum algorithm uses minimum value operations to replace the hyperbolic tangent function and inverse hyperbolic tangent function operations in the LLR-BP algorithm, sacrificing some decoding performance, but greatly reducing the algorithm complexity and being conducive to hardware implementation. It is a commonly used method for hardware implementation of LDPC codes at present.

[0004] In hardware implementation, messages also need to be quantized. The general idea of quantization is to convert floating-point operations during decoding into fixed-point operations, which can reduce the storage cost of hardware, but will generate quantization noise and reduce decoding performance.

[0005] Conventional quantization methods include uniform quantization and non-uniform quantization; uniform quantization uses a fixed quantization interval, which is simple to implement, but is not suitable for sequences with uneven distributions and has large quantization noise; the quantization interval of non-uniform quantization will increase continuously, effectively improving the problem of quantization noise, and the quantization range is much larger than that of uniform quantization under the same number of bits.

[0006] However, some studies have pointed out that during the iterative decoding process, variable messages will grow exponentially. Traditional quantization methods have fixed quantization ranges and precisions and cannot well adapt to LDPC decoding algorithms. Summary of the Invention

[0007] The purpose of the present invention is to provide an adaptive non-uniform quantization decoding method to solve the problem that the quantization method in the prior art cannot well adapt to the rapid change of parameters in the decoding algorithm.

[0008] To achieve the above object, the present invention is implemented by the following technical solutions:

[0009] The present invention provides an adaptive non-uniform quantization decoding method, including:

[0010] Obtain the variable node initialization message and the received message, and perform the first iteration of quantization decoding on the received message by using the variable node initialization message to obtain a first iteration decision codeword matrix composed of the first iteration decision codewords;

[0011] Use the first iteration decision codeword matrix for discrimination. If the preset iteration end requirement is met, output the first iteration decision codeword to complete the decoding. If the preset iteration end requirement is not met, calculate the adaptive quantization parameter according to the first iteration decision codeword, perform non-uniform quantization on the message transmitted by the variable node by using the adaptive quantization parameter, then increment the iteration count by one, and perform the quantization decoding of the next iteration on the previous iteration decision codeword by using the message transmitted from the variable node to the check node in the previous iteration until the decision codeword matrix meets the preset iteration end requirement, output the decision codeword, and complete the decoding.

[0012] Further, the received message is obtained by the following method:

[0013] Transmit the transmitted codeword through an AWGN channel under BPSK modulation to obtain the received message, and the received message is expressed as:

[0014]

[0015] where x i is the transmitted codeword and N is Gaussian white noise.

[0016] Further, the variable node initialization message is obtained by the following method:

[0017] Initialize the variable node by using the received message to obtain the variable node initialization message, specifically as follows:

[0018]

[0019] where M (l) (v ij ) represents the message transmitted from the variable node v i to the check node c j in the l-th iteration, l is 1, L(y i ) is the maximum likelihood ratio of the received message, yi is the received message, and σ is the variance.

[0020] Further, the quantization decoding includes:

[0021] Update the check node by using the message transmitted from the variable node to the check node to obtain the message transmitted from the check node to the variable node, specifically as follows:

[0022]

[0023] Among them, E (l) (c ji ) is the check node c in the l-th iteration j transmitting the message to the variable node v i . V(j) / i represents all variable nodes adjacent to the check node c i except the variable node v j . M (l) (v i'j ) represents the message transmitted by the variable node v in the l-th iteration i’ to the check node c j . sign is the sign function, and Min is the minimum value function.

[0024] Furthermore, the quantization decoding includes:

[0025] Updating the variable node by using the initial message of the variable node and the message transmitted from the check node to the variable node to obtain the message transmitted by the variable node to the check node, specifically as follows:

[0026]

[0027] Among them, M (l) (v ij ) represents the message transmitted by the variable node v in the l-th iteration i to the check node c j . N(i) / j is the set of all check nodes adjacent to the variable node v j except the check node c i . E l (c j’i ) is the message transmitted by the check node c in the l-th iteration j’ to the variable node v i . L(y i ) is the maximum likelihood ratio of the received message, and yi is the received message.

[0028] Furthermore, the quantization decoding includes calculating the posterior probability of the variable node:

[0029]

[0030] Among them, L l (v i ) is the posterior probability of the variable node v i , is the set of all check nodes adjacent to the variable node v i . E (l) (c ji ) is the message transmitted by the check node c in the l-th iteration j to the variable node vi Regarding the message, L(y i ) is the maximum likelihood ratio of the received message, and y i is the received message.

[0031] Furthermore, the quantization decoding includes performing a hard decision using the posterior probability. If the posterior probability is greater than 0, the decision codeword is 0; otherwise, the decision codeword is 1. The decision codewords form a decision codeword matrix.

[0032] Furthermore, the preset requirement for ending the iteration is as follows:

[0033] The number of iterations reaches the preset maximum number of iterations or H·z T = 0, where H is the LDPC code parity-check matrix and z T is the transpose of the decision codeword matrix.

[0034] Furthermore, the adaptive quantization parameter is obtained by the following method:

[0035]

[0036] where w is the set of error codewords in the decision codewords, n is the total number of codewords to be decoded, and η 1 to η k-1 are all constants, 0 < η 1 < η 2 < … < η k-1 < 1, and k is a preset integer.

[0037] Furthermore, non-uniform quantization of the messages passed by variable nodes using the adaptive quantization parameter includes:

[0038]

[0039] where x is the message passed by the variable node to the check node, d is the quantization interval, s is the adaptive quantization parameter, n is an integer taking values from 1 to N - 1, N = 2 q-1 - 1, and q is the number of bits of the non-uniform quantizer.

[0040] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0041] An adaptive non-uniform quantization decoding method provided by the present invention improves the problem that traditional uniform quantization and non-uniform quantization cannot well adapt to the too-fast change of parameters in the iterative decoding process by means of quantization decoding iteration and introducing an adaptive quantization parameter in the iterative process, and enlarges the quantization step size as the number of iterations increases, thereby improving the decoding performance; the messages transmitted by variable nodes are non-uniformly quantized by using the adaptive quantization parameter, greatly improving the utilization rate of quantization levels in the quantizer. Compared with traditional uniform and non-uniform quantization methods, the increased complexity is low and the performance is good. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is one of the flowcharts of an adaptive non-uniform quantization decoding method provided by an embodiment of the present invention;

[0043] Figure 2 is the second flowchart of an adaptive non-uniform quantization decoding method provided by an embodiment of the present invention;

[0044] Figure 3 is one of the comparison diagrams of decoding results of the adaptive non-uniform quantization decoding method, traditional floating-point BP decoding, traditional min-sum decoding algorithm, traditional min-sum decoding algorithm with uniform quantization, and traditional min-sum decoding algorithm with non-uniform quantization according to the present invention under a QC-LDPC code with a code length of 576 provided by an embodiment of the present invention;

[0045] Figure 4 is the second comparison diagram of decoding results of the adaptive non-uniform quantization decoding method, traditional floating-point BP decoding, traditional min-sum decoding algorithm, traditional min-sum decoding algorithm with uniform quantization, and traditional min-sum decoding algorithm with non-uniform quantization according to the present invention under a QC-LDPC code with a code length of 576 provided by an embodiment of the present invention;

[0046] Figure 5 is one of the comparison diagrams of decoding results of the adaptive non-uniform quantization decoding method, traditional floating-point BP decoding, traditional min-sum decoding algorithm, traditional min-sum decoding algorithm with uniform quantization, and traditional min-sum decoding algorithm with non-uniform quantization according to the present invention under a QC-LDPC code with a code length of 960 provided by an embodiment of the present invention;

[0047] Figure 6 is the second comparison diagram of decoding results of the adaptive non-uniform quantization decoding method, traditional floating-point BP decoding, traditional min-sum decoding algorithm, traditional min-sum decoding algorithm with uniform quantization, and traditional min-sum decoding algorithm with non-uniform quantization according to the present invention under a QC-LDPC code with a code length of 960 provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0048] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention and cannot be used to limit the protection scope of the present invention.

[0049] Embodiment 1

[0050] As Figure 1 shown, an adaptive non-uniform quantization decoding method provided by an embodiment of the present invention includes:

[0051] S1. Obtain the variable node initialization message and the received message, and perform quantization decoding of the first iteration on the received message by using the variable node initialization message to obtain a first iteration decision codeword matrix composed of the first iteration decision codewords.

[0052] An adaptive non-uniform quantization decoding method provided by an embodiment of the present invention is implemented based on LDPC codes.

[0053] Using an AWGN (Additive White Gaussian Noise) channel under BPSK (Binary Phase Shift Keying) modulation, the transmitted codeword x i After passing through the AWGN channel, it can be expressed as:

[0054]

[0055] where y i is the received message, x i is the transmitted codeword, and N is Gaussian white noise.

[0056] The range of the received message y i can be represented by the variance:

[0057] (-1 - 3σ, 1 + 3σ)

[0058] where σ is the variance, and σ is n 0 is the power spectral density of Gaussian white noise.

[0059] Since the value range is fixed, uniform quantization is performed on the received message y i .

[0060] Initialize the variable node to obtain the variable node initialization message, specifically as follows:

[0061]

[0062] where M (l) (v ij ) represents the variable node v at the l-th iteration iThe message passed to the check node c j , where l = 1, and L(y i ) is the maximum likelihood ratio of the received message, yi is the received message, and σ is the variance.

[0063] Set the maximum number of iterations l max to 50.

[0064] Update the check node. Use the message passed from the variable node to the check node to update the check node, and obtain the message passed from the check node to the variable node, as follows:

[0065]

[0066] where l = 1, and E (l) (c ji ) is the message passed from the check node c j to the variable node v i in the l-th iteration. V(j) / i represents the set of all variable nodes adjacent to the check node c i except the variable node v j . M (l) (v i'j ) represents the message passed from the variable node v i’ to the check node c j in the l-th iteration, which represents the initialization message of the variable node when l = 1. sign is the sign function and Min is the minimum function.

[0067] Update the variable node. Use the initialization message of the variable node and the message passed from the check node to the variable node to update the variable node and obtain the message passed from the variable node to the check node, as follows:

[0068]

[0069] where l = 1, and M (l) (v ij ) represents the message passed from the variable node v i to the check node c j in the l-th iteration. N(i) / j is the set of all check nodes adjacent to the variable node v j except the check node c i . E l (c j’i ) is the message passed from the check node c j’ to the variable node v i in the l-th iteration. L(y i ) is the maximum likelihood ratio of the received message and also the initialization message of the variable node, and y i is the received message.

[0070] Calculate the posterior probability of the variable node, and the calculation formula is:

[0071]

[0072] where l = 1, L l (v i ) is the posterior probability of variable node v i , and is the set of all check nodes adjacent to variable node v i , E (l) (c ji ) is the message passed from check node c j to variable node v i at the l-th iteration, L(y i ) is the maximum likelihood ratio of the received message and also the initial message of the variable node, and y i is the received message.

[0073] Perform hard decision using the posterior probability. If the posterior probability is greater than 0, the decoded codeword is considered to be 0, otherwise the decoded codeword is considered to be 1. The decoded codewords form the decoded codeword matrix.

[0074] The generated decoded codeword is z i (l) :

[0075]

[0076] where L l (v i ) is the posterior probability of variable node v i .

[0077] S2. Use the decoded codeword matrix of the first iteration for discrimination. If the preset iteration end requirement is met, output the decoded codeword of the first iteration to complete decoding. If the preset iteration end requirement is not met, calculate the adaptive quantization parameter based on the decoded codeword of the first iteration, perform non-uniform quantization on the message passed by the variable node using the adaptive quantization parameter, then increment the iteration count and perform quantization decoding for the next iteration of the decoded codeword of the previous iteration using the message passed from the variable node to the check node in the previous iteration until the decoded codeword meets the preset iteration end requirement after a certain iteration, output the decoded codeword, and complete decoding.

[0078] Use the decoded codeword matrix of the first iteration for discrimination. If the preset iteration end requirement is met, that is, if l = l max or Hz T = 0, output the decoded codeword of the first iteration to complete decoding, where H is the parity-check matrix of the LDPC code and z T is the transpose of the decoded codeword matrix.

[0079] If the above preset iteration end requirement is not met, calculate the adaptive quantization parameter according to the first iteration decision codeword:

[0080]

[0081] Where w is the set of error codewords in the decision codeword, n is the total number of codewords to be decoded, η 1 to η k-1 are all constants, 0 < η 1 < η 2 < … < η k-1 < 1, and k is a preset integer.

[0082] Then perform non-uniform quantization on the message passed by the variable node using the adaptive quantization parameter, including:

[0083]

[0084] Where x is the message passed by the variable node to the check node, d is the quantization interval, s is the adaptive quantization parameter, n is an integer with values from 1 to N - 1, N = 2 q-1 - 1, and q is the number of bits of the non-uniform quantizer.

[0085] After non-uniform quantization, increment the number of iterations by one, and perform quantization decoding of the previous iteration decision codeword for the next iteration using the message passed by the variable node to the check node in the previous iteration, until the decision codeword meets the preset iteration end requirement after a certain iteration, output the decision codeword, and complete the decoding.

[0086] Embodiment 2

[0087] As Figure 2 shown, an adaptive non-uniform quantization decoding method provided by an embodiment of the present invention includes:

[0088] S1. Channel initialization, perform uniform quantization on the initial sequence: Use an AWGN (Additive White Gaussian Noise) channel under BPSK (Binary Phase Shift Keying) modulation, and the transmitted codeword x i After passing through the AWGN channel, it can be expressed as:

[0089]

[0090] Where y i is the received message, x i is the transmitted codeword, and N is Gaussian white noise.

[0091] The received message y iThe range can be represented by variance:

[0092] (-1 - 3σ, 1 + 3σ)

[0093] where σ is the variance, and σ is n 0 the power spectral density of Gaussian white noise.

[0094] Since the value range is fixed, the received message y i is uniformly quantized, and the received message is the initial sequence.

[0095] S2. Variable node initialization: The message passed from the variable node v i at the l-th iteration to the check node c j is initialized as:

[0096]

[0097] where M (l) (v ij ) represents the message passed from the variable node v i at the l-th iteration to the check node c j , L(y i ) is the maximum likelihood ratio of the received message, yi is the received message, and σ is the variance.

[0098] Set the maximum number of iterations l max to 50.

[0099] S3. Check node update: The update formula for the message passed from the check node c j at the l-th iteration to the variable node v i is:

[0100]

[0101] where E (l) (c ji ) is the message passed from the check node c j at the l-th iteration to the variable node v i , V(j) / i represents the set of all variable nodes adjacent to the check node c i except the variable node v j , M (l) (v i'j ) represents the message passed from the variable node v i’ at the l-th iteration to the check node c j , which is updated during each iteration (in step S4), sign is the sign function, and Min is the minimum function.

[0102] S4. Variable node update: The variable node v at the l-th iterationi The update formula for the message passed to the check node c j is as follows:

[0103]

[0104] where M (l) (v ij ) represents the message passed from the variable node v i to the check node c j in the l-th iteration, N(i) / j is the set of all check nodes adjacent to the variable node v j except the check node c i , E l (c j’i ) is the message passed from the check node c j’ to the variable node v i in the l-th iteration, L(y i ) is the maximum likelihood ratio of the received message and also the initial message of the variable node, and yi is the received message.

[0105] S5. Calculate the posterior probability of the variable node. The calculation formula is:

[0106]

[0107] where L l (v i ) is the posterior probability of the variable node v i , is the set of all check nodes adjacent to the variable node v i , E (l) (c ji ) is the message passed from the check node c j to the variable node v i in the l-th iteration, L(y i ) is the maximum likelihood ratio of the received message and also the initial message of the variable node, and y i is the received message.

[0108] S6. Decoding decision: Perform a hard decision using the posterior probability. If the posterior probability is greater than 0, the decoded codeword is considered to be 0; otherwise, the decoded codeword is considered to be 1. The decoded codewords form a decoded codeword matrix.

[0109] The generated decoded codeword is z i (l) :

[0110]

[0111] where L l (v i ) is the posterior probability of the variable node v i .

[0112] The specific process of decoding decision is as follows:

[0113] First, it is judged whether the decoding is successful. If H·z T = 0, then the decoding is successful, and the current transmission iteration ends, and the codeword z i (l) is output; where H is the parity-check matrix of the LDPC code, and z T is the transpose of the decision codeword matrix.

[0114] If the decoding is not successful, then the determination of the number of iterations is carried out. If the maximum number of iterations is reached, then the current transmission iteration also ends, and the codeword z i (l) is output; otherwise, continue to step S7.

[0115] S7. Adaptive quantization parameter update: Calculate the adaptive quantization parameter according to the decision codeword of the first iteration. The specific calculation formula is as follows:

[0116]

[0117] where w is the set of error codewords in the decision codeword, n is the total number of codewords to be decoded, and η 1 to η k-1 are all constants, 0 < η 1 < η 2 <... < η k-1 < 1, and k is a preset integer.

[0118] S8. Variable node adaptive quantization: Use the adaptive quantization parameter to perform non-uniform quantization on the message transmitted by the variable node. For a q-bit non-uniform quantizer, the quantization formula is:

[0119]

[0120] where x is the message (i.e., the input signal) transmitted by the variable node to the check node, d is the quantization interval, s is the adaptive quantization parameter, n is an integer taking values from 1 to N - 1, N = 2 q-1 - 1, and q is the number of bits of the non-uniform quantizer.

[0121] S9. Increase the number of iterations by one, return to step S3, and perform a new round of iteration until the iteration ends.

[0122] From Figure 3 and Figure 5From the test results, it can be seen that when decoding the QC-LDPC code under the 802.16e standard with a code length of 576, the method of the present invention starts to have better performance (characterized by BER, i.e., bit error probability) than the traditional non-uniform quantization min-sum decoding algorithm from 2 dB, and saves 2 bits of storage space. When the signal-to-noise ratio is 3 dB, the performance of the method of the present invention starts to be better than the traditional floating-point BP decoding. In terms of the average number of iterations, the method of the present invention is always less than that of the traditional uniform quantization min-sum decoding algorithm and the traditional non-uniform quantization min-sum decoding algorithm.

[0123] From Figure 4 and Figure 6 From the test results, it can be seen that when decoding the QC-LDPC code under the 802.16e standard with a code length of 960, the method of the present invention starts to have better performance than the traditional non-uniform quantization min-sum decoding algorithm from 1.5 dB, and saves 2 bits of storage space. When the signal-to-noise ratio is 2.5 dB, the decoding performance of the method of the present invention is better than that of the traditional floating-point BP decoding with only about one more iteration.

[0124] Figures 3 to 6 In the figure, float BP is the traditional floating-point BP decoding, float MSA is the traditional min-sum decoding algorithm, Uniform quantization is the traditional uniform quantization min-sum decoding algorithm, Nouniform quantization is the traditional non-uniform quantization min-sum decoding algorithm, and Adaptive quantization represents the method of the present invention.

[0125] Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0126] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be realized by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, so that the instructions executed by the processors of the computer or other programmable data processing devices generate for realizing the processes Figure 1One process or multiple processes and / or boxes Figure 1 Apparatus for the functions specified in one box or multiple boxes.

[0127] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction device that implements the functions in the process Figure 1 One process or multiple processes and / or boxes Figure 1 The functions specified in one box or multiple boxes.

[0128] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operational steps are performed on the computer or other programmable device to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in the process Figure 1 One process or multiple processes and / or boxes Figure 1 The functions specified in one box or multiple boxes.

[0129] The above is only the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the technical principle of the present invention, several improvements and modifications can be made, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. An adaptive non-uniform quantization decoding method, characterized in that, it includes: Obtain the variable node initialization message and the received message, and use the variable node initialization message to perform the first iterative quantization decoding on the received message to obtain a first iterative decision codeword matrix composed of the first iterative decision codewords; Use the first iterative decision codeword matrix for discrimination. If the preset iteration end requirement is met, output the first iterative decision codeword to complete the decoding. If the preset iteration end requirement is not met, calculate the adaptive quantization parameter according to the first iterative decision codeword, use the adaptive quantization parameter to perform non-uniform quantization on the message transmitted by the variable node, then increment the iteration count by one, and use the message transmitted by the variable node to the check node in the previous iteration to perform the next iterative quantization decoding on the previous iterative decision codeword until the decision codeword matrix meets the preset iteration end requirement, output the decision codeword, and complete the decoding.

2. The adaptive non-uniform quantization decoding method according to claim 1, characterized in that, the received message is obtained by the following method: Transmit the transmitted codeword through an AWGN channel under BPSK modulation to obtain the received message, and the received message is expressed as: where x i is the transmitted codeword and N is the additive white Gaussian noise.

3. The adaptive non-uniform quantization decoding method according to claim 1, characterized in that, the variable node initialization message is obtained by the following method: Use the received message to initialize the variable node to obtain the variable node initialization message, specifically as follows: Among them, M (l) (v ij ) represents the message passed from the variable node vi to the check node cj in the l-th iteration, where l ranges from 1 to L(y i ) is the maximum likelihood ratio of the received message, yi is the received message, and σ is the variance.

4. The adaptive non-uniform quantization decoding method according to claim 3, characterized in that, the quantization decoding includes: Use the message transmitted by the variable node to the check node to update the check node to obtain the message transmitted by the check node to the variable node, specifically as follows: Among them, E (l) (c ji ) is the message passed from the check node cj to the variable node vi in the l-th iteration. V(j) / i represents all variable nodes adjacent to the check node c i except for the variable node v j . M (l) (v i'j ) represents the message passed from the variable node v i’ to the check node c j in the l-th iteration. sign is the sign function, and Min is the minimum function.

5. The adaptive non-uniform quantization decoding method according to claim 3, characterized in that, the quantization decoding includes: Use the variable node initialization message and the message transmitted by the check node to the variable node to update the variable node to obtain the message transmitted by the variable node to the check node, specifically as follows: Among them, M (l) (v ij ) represents the variable node v at the l-th iteration i transmitting the message to the check node c j . N(i) / j is the set of all check nodes adjacent to the variable node v j except for the check node c i . E l (c j’i ) is the message transmitted from the check node cj’ to the variable node vi at the l-th iteration. L(y i ) is the maximum likelihood ratio of the received message, and y i is the received message.

6. The adaptive non-uniform quantization decoding method according to claim 4, characterized in that, the quantization decoding includes calculating the posterior probability of the variable node: where L l (v i ) is the posterior probability of variable node v i , and is the set of all check nodes adjacent to variable node v i . E (l) (c ji ) is the message passed from check node cj to variable node vi in the l-th iteration. L(y i ) is the maximum likelihood ratio of the received message, and y i is the received message.

7. The adaptive non-uniform quantization decoding method according to claim 6, characterized in that, the quantization decoding includes performing a hard decision using the posterior probability. If the posterior probability is greater than 0, the decision codeword is 0, otherwise the decision codeword is 1, and the decision codewords form a decision codeword matrix.

8. The adaptive non-uniform quantization decoding method according to claim 7, characterized in that, the preset iteration end requirement is: The number of iterations reaches the preset maximum number of iterations or H·z T = 0, where H is the LDPC code parity-check matrix and z T is the transpose of the decision codeword matrix.

9. The adaptive non-uniform quantization decoding method according to claim 1, characterized in that, the adaptive quantization parameter is obtained by the following method: Where \(w\) is the set of error codewords in the decoded codeword, \(n\) is the total number of codewords to be decoded, \(\eta\) 1 to \(\eta\) k-1 are all constants, \(0 \lt \eta\) 1 \lt \eta\) 2 \lt \cdots \lt \eta\) k-1 \lt 1\), and \(k\) is a preset integer.

10. The adaptive non-uniform quantization decoding method according to claim 9, characterized in that, Using the adaptive quantization parameter to perform non-uniform quantization on the message transmitted by the variable node includes: Among them, x is the message passed from the variable node to the check node, d is the quantization interval, s is the adaptive quantization parameter, n is an integer taking values from 1 to N - 1, and N = 2 q-1 - 1, and q is the number of bits of the non-uniform quantizer.

Images