Channel Capacity Enhancement Method and Apparatus Based on Multi-Level LDPC Coding and Probabilistic Shaping

Abstract

This invention provides a channel capacity enhancement method and apparatus based on multi-level LDPC coding and probabilistic shaping, comprising: acquiring an input signal sequence, converting it from a serial-to-parallel input signal sequence to multiple parallel input signal sequences, constructing a parity check matrix corresponding to each level of LDPC coding, determining the coding at each level, and performing a base conversion on the coding at each level to obtain a first input sequence; acquiring a target probability distribution, output sequence length, and target amplitude set, determining available diversity, acquiring a chaotic sequence, dividing each element in the chaotic sequence into multiple value intervals according to its value range, and constructing a chaotic index mapping relationship between each value interval and the available diversity; dividing the first input sequence into multiple sub-sequences, determining the specific diversity corresponding to each sub-sequence, constructing a distribution-matched probability shaping mapping relationship corresponding to each available diversity, and determining the distribution-matched probability shaping output sequence corresponding to the input signal sequence. This method reduces code rate loss, increases channel capacity, and improves bit error rate performance.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a method and apparatus for improving channel capacity based on multi-level LDPC coding and probabilistic shaping. Background Technology

[0002] In recent years, the rapid development of multimedia technology has led to a continuous increase in network data traffic, and single-mode fiber can no longer meet the existing capacity requirements. Space division multiplexing (SDM) transmission systems have become a highly attractive solution due to the spatial diversity they offer. Therefore, how to further improve the capacity and reliability of communication systems based on SDM transmission systems has become a focus of attention for both academia and industry. According to Shannon's theorem, reliable transmission close to the channel capacity can be achieved through error correction coding, provided that the information transmission rate does not exceed the channel capacity. Currently, classic error correction coding schemes proposed by academia include Hamming codes, Golay codes, Reed-Solomon codes, and LDPC codes. Convolutional LDPC coding and traditional block coding (BC) rely on the parity-check matrix. Compared to traditional block coding, convolutional LDPC coding, due to the irregular nature of its parity-check matrix, exhibits performance close to the capacity limit under iterative belief propagation decoding.

[0003] While convolutional LDPC coding currently offers performance close to its capacity limit, it suffers from drawbacks such as significant rate loss, poor bit error rate performance, and limited improvement in channel capacity. Therefore, reducing rate loss, increasing channel capacity, and improving bit error rate performance are pressing technical challenges that need to be addressed. Summary of the Invention

[0004] In view of this, the present invention provides a channel capacity enhancement method and apparatus based on multi-level LDPC coding and probabilistic shaping to solve one or more problems existing in the prior art.

[0005] According to one aspect of the present invention, a channel capacity enhancement method based on multi-level LDPC coding and probabilistic shaping is disclosed, the method comprising:

[0006] The input signal sequence is obtained, and the input signal sequence is transformed into a multi-channel parallel input signal sequence. The parity check matrix corresponding to each level of LDPC encoding is constructed according to the degree distribution. The encoding at each level is determined based on the multi-channel parallel input signal sequence and the multiple parity check matrices. The encoding at each level is converted to a different base to obtain the first input sequence corresponding to the transmitting end.

[0007] Obtain the target probability distribution, output sequence length, and target amplitude set; determine the available subsets based on the target probability distribution, output sequence length, and target amplitude set; obtain the chaotic sequence; divide each element in the chaotic sequence into multiple value intervals according to the value range; and construct a chaotic index mapping relationship between each value interval and the available subsets.

[0008] The first input sequence is divided into multiple subsequences. Based on the chaotic index mapping relationship, the specific set corresponding to each subsequence is determined. The distribution matching probability shaping mapping relationship corresponding to each available set is constructed. Based on each subsequence, the specific set corresponding to each subsequence, and the distribution matching probability shaping mapping relationship, the distribution matching probability shaping output sequence corresponding to the input signal sequence is determined.

[0009] In some embodiments of the present invention, the method further includes:

[0010] The distributed matching probability shaped output sequence is inversely mapped based on the distributed matching probability shaped mapping relationship to obtain an inverse mapping sequence. The inverse mapping sequence is then subjected to multi-level LDPC decoding to obtain a multi-channel parallel output signal sequence. Finally, the multi-channel parallel output signal sequence is converted from parallel to serial to obtain an output signal sequence.

[0011] In some embodiments of the present invention, the method further includes:

[0012] Obtain a training sequence, and estimate the channel matrix based on the training sequence, the distribution matching probability shaped output sequence, and the output signal sequence;

[0013] The channel capacity is calculated based on the estimated channel matrix;

[0014] The code rate and degree distribution of each level of LDPC coding are updated according to the channel capacity and density evolution algorithm.

[0015] In some embodiments of the present invention, the parity-check matrix corresponding to each level of LDPC encoding is constructed according to the degree distribution, including:

[0016] Construct a prototype matrix based on the degree distribution;

[0017] The prototype matrix is ​​expanded to obtain the parity check matrix corresponding to the first-level LDPC encoding; wherein each element in the parity check matrix corresponding to the first-level LDPC encoding is a submatrix.

[0018] The check matrix corresponding to the second-level LDPC encoding is constructed based on the check matrix corresponding to the first-level LDPC encoding.

[0019] In some embodiments of the present invention, the expression of the first input sequence is:

[0020]

[0021] Where V represents the first input sequence, L represents the level of LDPC encoding, and v l This represents the level l encoding.

[0022] In some embodiments of the present invention, determining the available diversity based on the target probability distribution, the output sequence length, and the target amplitude set includes:

[0023] Determine the frequency set C of each amplitude in the target amplitude set. type ;

[0024] Traverse and search for matching Conditional diversity groups, where C represents θ complementary diversity;

[0025] Each set of subsets obtained through the traversal is taken as a usable subset.

[0026] In some embodiments of the present invention, determining the specific subsets corresponding to each group of subsequences based on the chaotic index mapping relationship includes:

[0027] Calculate the chaotic index of each subsequence group;

[0028] The chaotic index mapping relationship is used to determine the subset corresponding to the chaotic index of each subsequence.

[0029] In some embodiments of the present invention, updating the code rate and degree distribution of each level of LDPC coding according to the channel capacity and density evolution algorithm includes:

[0030] Based on the estimated channel capacity and density evolution algorithm, the noise threshold of each coding level at the current code rate is calculated;

[0031] Determine whether the noise threshold of the calculated LDPC codes at each level is greater than a preset threshold;

[0032] When the values ​​exceed a preset threshold, the particle swarm optimization algorithm is used to adjust the code rate and degree distribution of each level of LDPC encoding.

[0033] According to another aspect of the present invention, a channel capacity enhancement system based on multi-level LDPC coding and probabilistic shaping is also disclosed. The system includes a processor and a memory, wherein computer instructions are stored in the memory, and the processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps of the method as described in any of the above embodiments.

[0034] According to another aspect of the present invention, a computer-readable storage medium is also disclosed having a computer program stored thereon that, when executed by a processor, implements the steps of the method as described in any of the above embodiments.

[0035] The channel capacity enhancement method and apparatus based on multi-level LDPC coding and probabilistic shaping disclosed in this invention first constructs a parity-check matrix corresponding to each level of LDPC coding according to the degree distribution. This involves introducing a multi-level LDPC coding scheme to encode the input signal sequence to obtain the first input sequence corresponding to the transmitter. Then, based on the constructed chaotic index mapping relationship, the specific diversity corresponding to each group of subsequences in the first input sequence is determined. Finally, based on the constructed distribution matching probability shaping mapping relationships, the distribution matching probability shaping output sequence corresponding to the input signal sequence is determined. This method introduces a multi-level convolutional LDPC coding scheme, reducing encoding / decoding latency and decoding complexity. By employing a multi-diversity distribution matching scheme with chaotic sequence indexing, it reduces code rate loss, increases channel capacity, and improves bit error rate performance.

[0036] Additional advantages, objects, and features of the invention will be set forth in part in the description which follows, and will also become apparent in part to those skilled in the art upon studying the text, or may be learned by practice of the invention. The objects and other advantages of the invention can be realized and obtained by means of the structures specifically pointed out in the written description, claims, and drawings.

[0037] Those skilled in the art will understand that the purposes and advantages that can be achieved by the present invention are not limited to the above specific descriptions, and the above and other purposes that can be achieved by the present invention will be more clearly understood based on the following detailed description. Attached Figure Description

[0038] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, are not intended to limit the scope of the invention. The components in the drawings are not drawn to scale but are merely illustrative of the principles of the invention. For ease of illustration and description of certain parts of the invention, corresponding portions in the drawings may be enlarged, i.e., may appear larger relative to other components in an exemplary device actually manufactured according to the invention. In the drawings:

[0039] Figure 1 This is a flowchart illustrating a channel capacity enhancement method based on multi-level LDPC coding and probabilistic shaping according to an embodiment of the present invention.

[0040] Figure 2 This is a flowchart illustrating another embodiment of the channel capacity enhancement method based on multi-level LDPC coding and probabilistic shaping according to the present invention.

[0041] Figure 3This is a schematic diagram of the multi-level LDPC encoding process according to an embodiment of the present invention.

[0042] Figure 4 This is a schematic diagram illustrating the process of constructing a multi-level LDPC-encoded parity-check matrix according to an embodiment of the present invention.

[0043] Figure 5 This is a schematic diagram of the process of multi-diversity distribution matching probability shaping mapping according to an embodiment of the present invention.

[0044] Figure 6 This is a schematic diagram of a chaotic index mapping relationship according to an embodiment of the present invention.

[0045] Figure 7 This is a schematic diagram of the process of multi-diversity distribution matching probability inverse mapping according to an embodiment of the present invention.

[0046] Figure 8 This is a schematic diagram of a multi-level decoding process for convolutional LDPC according to an embodiment of the present invention.

[0047] Figure 9 This is a schematic diagram of the code rate and degree distribution update process based on density evolution according to an embodiment of the present invention.

[0048] Figure 10 This is a schematic diagram of a communication system architecture based on multi-level LDPC coding and multi-diversity distribution matching according to an embodiment of the present invention.

[0049] Figure 11 This is a schematic diagram illustrating the simulation results of a communication system based on multi-level LDPC coding and multi-diversity distribution matching according to an embodiment of the present invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0051] It should be noted that, in order to avoid obscuring the invention with unnecessary details, only the structures and / or processing steps closely related to the solution according to the invention are shown in the accompanying drawings, while other details that are not closely related to the invention are omitted.

[0052] It should be emphasized that the term "including / comprises / has" as used herein refers to the presence of a feature, element, step, or component, but does not exclude the presence or addition of one or more other features, elements, steps, or components.

[0053] To reduce code rate loss, increase channel capacity, and improve bit error rate performance, this invention provides a channel capacity enhancement method and apparatus based on multi-level LDPC coding and probabilistic shaping. This method combines multi-level LDPC coding with probabilistic shaping. Probabilistic shaping reduces the average energy of transmitted symbols by remapping the probability distribution of transmitted constellation points, thereby increasing channel capacity at the same transmit power and reducing the system bit error rate. Distribution matching (DM), as a type of probabilistic shaping technique, can be integrated into coding and modulation systems. Multi-diversity distribution matching (MPDM) produces output sequences with equal lengths and target probability distributions. Its diverse mapping sets increase the capacity of the optional codebook, thus reducing code rate loss. The sequence length required for a specific code rate loss is significantly lower than that of traditional constant component distribution matching (CCDM).

[0054] In the following description, embodiments of the invention will be illustrated with reference to the accompanying drawings. In the drawings, the same reference numerals represent the same or similar parts, or the same or similar steps.

[0055] Figure 1 This is a flowchart illustrating a channel capacity enhancement method based on multi-level LDPC coding and probabilistic shaping according to an embodiment of the present invention. (Refer to...) Figure 1 The channel capacity enhancement method includes at least steps S10 to S30.

[0056] Step S10: Obtain the input signal sequence, convert the input signal sequence from serial to parallel to a multi-channel parallel input signal sequence, construct the parity check matrix corresponding to each level of LDPC encoding according to the degree distribution, determine the encoding at each level based on the multi-channel parallel input signal sequence and the multiple parity check matrices, and perform base conversion on the encoding at each level to obtain the first input sequence corresponding to the transmitting end.

[0057] In this step, the input signal sequence is the original binary information sequence input from the input terminal. That is, this step transforms the original binary information sequence from serial to parallel into multiple signals. Then, the multiple signals are encoded level by level by the parity check matrix corresponding to the multi-level LDPC encoding and output. Each level of LDPC encoding has a parity check matrix. Therefore, the number of LDPC encoding levels is the same as the number of parallel input signal sequences, and the number of parity check matrices is also the same as the number of parallel input signal sequences.

[0058] Figure 3 This is a schematic diagram of the multi-level LDPC encoding process according to an embodiment of the present invention. (Refer to...) Figure 3First, the code rate, code length, and number of encoding levels are input, and then a prototype matrix is ​​constructed. Based on the prototype matrix, level 0 parity-check matrices, level 1 parity-check matrices, level 2 parity-check matrices, and so on, are constructed sequentially. In this embodiment, the input bit sequence represents one original binary information sequence, and the number of LDPC encoding levels in this embodiment is 3. The input bit sequence is divided into three parallel input signal sequences, such as information sequence 0, information sequence 1, and information sequence 2. It is understood that setting the number of LDPC encoding levels to 3 in this embodiment is only an example; in other embodiments, the number of LDPC encoding levels may be set to two or more.

[0059] Furthermore, constructing the check matrix corresponding to each level of LDPC encoding based on the degree distribution includes: constructing a prototype matrix based on the degree distribution; expanding the prototype matrix to obtain the check matrix corresponding to the first level LDPC encoding; wherein each element in the check matrix corresponding to the first level LDPC encoding is a submatrix; and constructing the check matrix corresponding to the second level LDPC encoding based on the check matrix corresponding to the first level LDPC encoding.

[0060] Figure 4 This is a schematic diagram illustrating the process of constructing a multi-level LDPC-encoded parity-check matrix according to an embodiment of the present invention, as shown below. Figure 4 As shown, firstly, based on the degree distribution (the degree distribution represents the prototype matrix B)... SC Neutron matrices B0, B1, ..., B u Construct the prototype matrix B using the number of non-zero elements in each row or column. SC The method of constructing the prototype matrix is ​​not limited; each submatrix in the prototype matrix is ​​a b×c matrix. Further, a level 0 parity check matrix is ​​constructed based on the prototype matrix. Specifically, during the expansion construction, the submatrixes B0, B1, ..., B... of the prototype matrix are... u The zero elements in the matrix are replaced with an M×M matrix of all zeros, while the submatrices B0, B1, ..., B... are... u The non-zero elements in the matrix are randomly replaced with a permutation matrix of size M×M; the permutation matrix is ​​a matrix in which only one element in each row or column is 1, for example... The prototype matrix, after the above substitutions, yields the level 0 parity check matrix. Each submatrix in the matrix is ​​derived from the prototype matrix B. SC The submatrix at the same position in the matrix is ​​expanded to obtain the result, and The submatrices in the equation have a cyclic period of T and satisfy the following condition: When t max =W, Let be a matrix of size (W+μ)bM×WcM, and The level 0 coding code rate is (cb) / c; where b and c are the number of rows and columns of the submatrix in the prototype matrix, respectively, M is the number of rows or columns of the all-zero matrix or permutation matrix, and μ is... The number of elements in each column.

[0061] After constructing the level 0 parity check matrix based on the prototype matrix, the level j (j≥1) parity check matrix is ​​further constructed based on the level 0 parity check matrix. Submatrix in a matrix The submatrix at the corresponding position in the (j-1)th level parity check matrix The first m j Okay, m j The value of is less than the number of rows in the submatrix of the (j-1)th level parity check matrix, and The matrix code rate is cm j / c; where i takes values ​​from 0, 1…μ. For example, suppose... submatrix in When m1 = 2, we can obtain Submatrices in the verification matrix

[0062] After constructing the parity-check matrices for each level of LDPC encoding, the parallel signal sequences are further encoded, with information sequence 0 undergoing level 0 LDPC encoding, signal sequence 1 undergoing level 1 LDPC encoding, and signal sequence 2 undergoing level 2 LDPC encoding.

[0063] Specifically, the t-th legal codeword of level l, v l (t)∈{0,1} n satisfy Where n represents the length of a single codeword, a l (t) and s l (t) are all adjoint expressions, and the adjoint expressions are... Accompanying Where T represents the cycle period, and μ is The number of elements in each column. Indicates v l The transpose of (t).

[0064] For example, taking LDPC coding level L=3 as an example for detailed explanation, firstly, the input bit sequence is transformed into a parallel binary sequence, converting the one-way binary sequence transmitted bit by bit in sequence into three parallel binary sequences transmitted simultaneously, resulting in information sequence 0, information sequence 1, and information sequence 2, each with a length of one-third of the original. At level 0, information sequence 0 is divided into a subsequence u of length (cb)M.0 (0),u 0 (1),u 0 (2), …, u 0 (t),…u 0 (t max (), b and c are the number of rows and columns of the submatrix in the prototype matrix, respectively, and M is the number of rows or columns of the all-zero matrix or permutation matrix. Because a systematic code form is used, the encoded codeword is composed of an information subsequence u of length (cb)M from the input. l (t) and redundant information w l (t) is formed by connecting them in series, i.e., v l (t)=[u l (t) w l (t)], l∈[0,L], based on the above formula, we can calculate w. l (t) can then be used to obtain the encoding of level l.

[0065] Furthermore, It can be represented as In this expression, l = 0, therefore s can be obtained. 0 (t) = 0, that is Substitute v 0 (0) = [u 0 (0)w 0 (0)] can be solved to find w 0 (0), thus obtaining v 0 (0). Similarly, v can be obtained based on the following formula. 0 (1) v 0 (2), ..., v 0 (t max ):

[0066]

[0067]

[0068] ...

[0069]

[0070] ...

[0071] At this point, the total output of the l=0 level encoding is: v 0 =[v 0 (0),v 0 (1),v 0 (2),…,v 0 (t),…,v 0 (t max )).

[0072] At level l=1, the accompanying expression s 1=[s 1 (0),s 1 (1),s 1 (2),...,s 1 (t),...,s 1 (t max )],at this time Indicates v 0 Similarly, the transpose of (t) can be obtained as follows:

[0073]

[0074] Therefore, we can further solve for v. 1 =[v 1 (0),v 1 (1),v 1 (2),…,v 1 (t),…v 1 (t max It is understandable that at level l=2, based on the adjoint s 2 v can also be obtained 2 .

[0075] After obtaining the LDPC codes at each level using the above method, the first input sequence can be further obtained, and the expression of the first input sequence is:

[0076]

[0077] Where V represents the first input sequence, L represents the total number of LDPC encoding levels, and v l This represents the level l encoding.

[0078] Step S20: Obtain the target probability distribution, output sequence length, and target amplitude set; determine the available subsets based on the target probability distribution, output sequence length, and target amplitude set; obtain the chaotic sequence; divide each element in the chaotic sequence into multiple value intervals according to the value range; and construct a chaotic index mapping relationship between each value interval and the available subsets.

[0079] In this step, the available subsets are further determined, and the chaotic sequence is obtained. A chaotic index mapping relationship is established between the chaotic sequence and the available subsets, so that the first input sequence obtained in step S10 can be distributed and mapped through the chaotic sequence index, thereby realizing probability shaping.

[0080] Figure 5 This is a flowchart illustrating the multi-diversity distribution matching probability shaping mapping according to an embodiment of the present invention. (Refer to...) Figure 5 In this step, the target probability distribution P is first determined. AThe output sequence length n and the target amplitude set A are determined based on the target probability distribution, the target amplitude set A, and the number of diversity N. pairs Determine available subsets

[0081] Furthermore, determining the available diversity based on the target probability distribution, the output sequence length, and the target amplitude set includes: determining the frequency set C of each amplitude occurrence in the target amplitude set. type ; Traverse and search for matching Conditional diversity groups, where C represents θ The complementary sets; each set obtained by traversal is used as a usable set.

[0082] For example, the output sequence length n = 10, and the target amplitude distribution P A = [0.4, 0.3, 0.2, 0.1], target amplitude set A = [1, 2, 3, 4], P A =[0.4,0.3,0.2,0.1] means that in a sequence of length 10, the frequency set of the four amplitudes in amplitude set A = [1,2,3,4] is C. type = [4,3,2,1], meaning amplitude a1=1 appears 4 times, amplitude a2=2 appears 3 times, amplitude a3=3 appears 2 times, and amplitude a4=4 appears 1 time. To achieve the target probability distribution P in the output signal... A There are two types of frequency diversity that can be used, the first being C. type The second condition is satisfied. θ takes values ​​of 1, 2, ..., N pairs For the set C θ Complementary diversity The combination of the two satisfies P A For example, C θ =[5,3,3,1] and satisfy That is, C θ The value a1 = 1 appears 5 times. In the first set, a1 = 1 appears 3 times, and in the two subsets, a1 = 1 appears an average of 4 times. The same applies to a2, a3, and a4. The final traversal satisfies... The number of diversity logs N pairs =49, because Finally, 97 valid subsets can be identified, and the number of permutations of available amplitudes, N, can be determined. perms = 164214 types, and the numbers of all available subsets are denoted as follows:

[0083] Step S30: Divide the first input sequence into multiple subsequences, determine the specific set corresponding to each subsequence based on the chaotic index mapping relationship, construct the distribution matching probability shaping mapping relationship corresponding to each available set, and determine the distribution matching probability shaping output sequence corresponding to the input signal sequence based on each subsequence, the specific set corresponding to each subsequence, and the distribution matching probability shaping mapping relationship.

[0084] In this step, the distribution matching probability integer mapping relationship corresponding to each available subset is constructed, and the first input sequence is further divided into multiple subsequences, so as to find the distribution matching probability integer output corresponding to each subsequence from the corresponding distribution matching probability integer mapping relationship.

[0085] Specifically, determining the specific subsets corresponding to each group of subsequences based on the chaotic index mapping relationship includes: calculating the chaotic index of each group of subsequences; and determining the subsets corresponding to the chaotic index of each group of subsequences through the chaotic index mapping relationship.

[0086] In one embodiment, before mapping, the first input sequence is first grouped into groups with a k-bit interval, where k satisfies the condition L represents the coding level, N perms This represents the number of permutations of available amplitudes. When L = 3, then the first input sequence... t max Let Z represent the total length of the first input sequence, and let Z represent the set of integers, that is, each bit in the output sequence has 8 possibilities (integers in the range of 0 to 7 can be selected). Therefore, in the specific distribution matching probability integer mapping, the mapping is performed in groups of 5 bits, and the subsequences after grouping are denoted as g = [g(0), g(1), g(2), ...], where g(i) = [v(k·i), v(k·i+1), ... v(k·i+4)], i represents the i-th subsequence, k represents the number of bits in the subsequence, and the number of bits in the subsequence can also be understood as the number of intervals used when grouping the first input sequence, and v represents the first input sequence.

[0087] Furthermore, the chaotic index corresponding to each group of subsequences is calculated based on the following formula:

[0088] Among them, x n Let represent the nth term in the chaotic sequence, with a value range of (0,1), where n represents the ordinal number of the term; u is a control parameter of the expression, and u can be selected within the range (0,4]. For example, after 999 iterations, x can be... 1000 x 1001 , respectively, serve as the chaotic indices of g(0) and g(1), that is, the subsequence corresponding to g(i) is x. 1000+i index.

[0089] Figure 6 This is a schematic diagram of a chaotic index mapping relationship according to an embodiment of the present invention, as shown below. Figure 6 As shown, when the total number of episodes is 97, then x will be... n The value range is divided into 97 equal parts, each indexed into a different subset to achieve subset selection. For example, we select u = 3.7, with an initial value x0 = 0.3. x can be obtained through iterative chaotic sequences. 1000 ≈0.0401, that is Therefore, based on this chaotic index mapping relationship, g(0) corresponds to... Diversification. At this point, further based on the target amplitude set A = [1,2,3,4] and... Build The distribution matching probability integer mapping relationship corresponding to the diversity is obtained by arranging and combining the elements in the amplitude set, and the number of identical elements in the output sequence satisfies the distribution probability corresponding to "using diversity". The distribution matching probability integer mapping relationship corresponding to the subset is shown in the table below:

[0090]

[0091] In some embodiments of the present invention, the channel capacity enhancement method based on multi-level LDPC coding and probabilistic shaping further includes the following steps: performing inverse mapping on the distributed matching probability shaping output sequence based on the distributed matching probability shaping mapping relationship to obtain an inverse mapping sequence, performing multi-level LDPC decoding on the inverse mapping sequence to obtain a multi-path parallel output signal sequence, and performing parallel-to-serial transformation on the multi-path parallel output signal sequence to obtain an output signal sequence.

[0092] Figure 7 This is a flowchart illustrating the multi-diversity distribution matching probability inverse shaping mapping according to an embodiment of the present invention, as shown below. Figure 7 As shown, when performing multi-diversity distribution matching probability shaping inverse mapping, the same distribution matching probability shaping mapping relationship table as the sending end is first established at the receiving end, and the distribution matching inverse mapping is performed on the received distribution matching probability shaping output sequence at the output end, and the inverse mapping sequence is output.

[0093] After obtaining the inverse mapping sequence at the output, further multi-level LDPC decoding is performed on the inverse mapping sequence based on convolutional LDPC multi-level decoding units. For example, refer to... Figure 8 This decoding method corresponds to the convolutional LDPC multi-level coding method. Specifically, it first calculates the log-likelihood ratio (LLR) based on the inverse mapping sequence v' received at the receiver: Where r(j) represents the j-th signal received by the receiver; then, the LLR information of level 0 is sent in batches to the level 0 sliding window decoder, with a sliding window length of WMc symbols; confidence propagation iterative decoding is performed inside the sliding window decoder until the maximum number of iterations is reached, outputting the first Mc symbols, while moving the sliding window to receive the last Mc symbols. Before level 1 decoding, the information output from level 0 decoding and the syndrome corresponding to level 1 encoding are used to send the log-likelihood ratio information and syndrome together to the level 1 sliding window decoder; the same applies to level 2 decoding. Finally, the decoded information output from levels 0, 1, and 2 is converted from parallel to serial as the system output. Where c is the number of rows and columns of the submatrix in the prototype matrix, M is the number of rows or columns of the all-zero matrix or permutation matrix, and W = t max .

[0094] Figure 2 This is a flowchart illustrating another embodiment of the channel capacity enhancement method based on multi-level LDPC coding and probabilistic shaping according to the present invention, as shown below. Figure 2 As shown, this method first performs three-level LDPC encoding on the raw data input to obtain the first input sequence. Then, the first input sequence is subjected to multi-diversity distribution matching probability shaping mapping to obtain the distribution matching probability shaped output sequence. The distribution matching probability shaped output sequence is then modulated using 16QAM to obtain the modulated output S. d Modulated output S d Further training sequences are added and spatially multiplexed. The receiver first performs spatial multiplexing and demultiplexing, then 16QAM demodulation. The demodulated information is then subjected to multi-diversity distribution matching probability integer inverse mapping to obtain a 2-mapping sequence. The inverse mapping sequence is finally used for multi-level LDPC decoding to obtain the decoded received data.

[0095] In one embodiment of the present invention, the channel capacity enhancement method based on multi-level LDPC coding and probabilistic shaping further includes the following steps: obtaining a training sequence; estimating a channel matrix based on the training sequence, a distribution-matched probabilistic shaping output sequence, and an output signal sequence; calculating the channel capacity based on the estimated channel matrix; and updating the code rate and degree distribution of each level of LDPC coding according to the channel capacity and density evolution algorithm.

[0096] In this step, the channel state is estimated based on the training sequence, the channel capacity is calculated, and the result is fed back to the transmitter to adjust the transmitter's coding rate. Specifically, when estimating the channel state, let the transmitter's training sequence be denoted as S. p The information sequence is denoted as S. d The total transmission sequence is [S] p S d ], where S dThe modulated output is obtained by modulating the probability-matched distribution-shaped output sequence with 16QAM; the sequences received by the receiver corresponding to the training sequence and the information sequence are Y, respectively. p and Y d The total received sequence is [Y p Y d ]; where the training sequence is known at both the transmitting and receiving ends, let the channel matrix within the coherence time be G, and the Gaussian random variable be N, then the following relationship is satisfied: [Y p Y d ] = G[S p S d ]+N. For example, the channel matrix can be estimated based on the EM algorithm (Expectation-Maximization algorithm), which iterates multiple times based on expectation calculation to achieve channel matrix estimation. First, for Random initialization and expected calculation:

[0097]

[0098]

[0099] Where S represents the set of all possible constellation points, and s[n] and y[n] represent the transmitted signal S, respectively. d and received signal Y d The nth column in the array. Furthermore, based on the expected value... Update: S p H Indicates the training sequence S sent from the transmitting end. p The conjugate transpose of , where Y represents the total received sequence and H represents the conjugate transpose. This represents the estimated value of the channel matrix G in stage l.

[0100] Furthermore, the channel capacity C is calculated based on the estimated channel matrix. Where λ i G·G H The non-zero eigenvalues ​​of , where r represents the rank of the channel matrix. Here, E0 represents the average signal energy transmitted per channel mode per symbol period, L represents the mode average propagation loss, and N0 represents the noise power spectral density, G·G H This represents the product of the channel matrix and its conjugate transpose.

[0101] In one embodiment, updating the code rate and degree distribution of each level of LDPC coding according to the channel capacity and density evolution algorithm includes: calculating the noise threshold of each level of coding at the current code rate based on the estimated channel capacity and density evolution algorithm; determining whether the calculated noise threshold of each level of LDPC coding is greater than a preset threshold; and if it is greater than the preset threshold, adjusting the code rate and degree distribution of each level of LDPC coding using a particle swarm optimization algorithm.

[0102] In the above embodiments, the channel capacity enhancement method based on multi-level LDPC coding and probabilistic shaping firstly uses multi-level LDPC coding to transmit information between different coding layers, thereby improving the bit error rate performance at the receiver; by using chaotic sequences to provide an index for diversity selection in multi-diversity distribution matching, the code rate loss is further reduced; by adding training sequences to estimate the channel capacity, and by using a density evolution algorithm to optimize the code rate of each encoder level; this method has low complexity, adopts a modular design, and can be integrated into existing communication systems.

[0103] Figure 9 This is a schematic diagram of a code rate and degree distribution update process based on density evolution according to an embodiment of the present invention, as shown below. Figure 9 As shown, when updating the code rate and degree distribution, the noise threshold of each level of coding at the current code rate is calculated based on the estimated channel capacity and density evolution algorithm. When the noise threshold difference is greater than the specified threshold, the particle swarm optimization algorithm is used to adjust the code rate and degree distribution of each level of coding until the noise threshold difference of each level is less than the noise threshold tolerance.

[0104] For example, let the degree distribution pair be (d v ,d c If ), then for a regular LDPC code, the code rate is . There is a corresponding relationship between the degree distribution and the code rate of regular LDPC codes, and the channel capacity determines the upper limit of the sum of the code rates of each level of coding. By employing the density evolution algorithm, the decoding noise threshold of each level of LDPC coding under a given degree distribution can be calculated. By employing the particle swarm optimization algorithm, the degree distribution of each level of LDPC coding can be optimized. According to the multi-level coding rate design criteria, each level of coding should have an approximate noise threshold. The specific algorithm flow is as follows:

[0105] Step (1), Initialization: Determine the noise threshold tolerance σ th Spatial multiplexing channel capacity, maximum number of iterations, and optimal particle number N p And particle motion parameters. Particle motion parameters include the particle's initial velocity e. ij Inertia factor w. From the above analysis, when the coding level is L=3, the number of variables is 6 (three levels in total, each level corresponding to a degree distribution pair). Therefore, N is randomly generated.p A 6-dimensional particle vector X i =(x i1 ,x i2 ,x i3 ,x i4 ,x i5 ,x i6 ), 0≤i≤N p Where (x) i1 ,x i2 ) is the degree distribution pair of the level 0 encoder, (x i3 ,x i4 ) is the degree distribution pair of the Level 1 encoder, (x i5 ,x i6 ) represents the degree distribution pairs of the second-level encoder. The optimal position vector for each particle is initialized to P. i =X i According to the density evolution algorithm, the noise thresholds for the three levels can be calculated as σ. i1 ,σ i2 ,σ i3 The difference between the noise thresholds is calculated as σ. id =max{σ i1 ,σ i2 ,σ i3}-min{σ i1 ,σ i2 ,σ i3}; Calculate the noise threshold difference for all particles, and find the position vector of the particle with the smallest threshold difference among all particles, denoted as P. g .

[0106] Step (2), change the particle position: x ij =x' ij +e ij .

[0107] Where x ij Let x' represent the coordinates of the i-th particle in the j-th dimension after the update. ij Let e ​​represent the coordinates of the i-th particle in the j-th dimension before the update. ij This represents the velocity of the i-th particle in the j-th dimension.

[0108] Calculate the total bitrate, i.e. If R tot If the value is greater than the channel capacity, proceed to step (4) to recalculate e. ij .

[0109] Step (3), calculate the particle threshold difference: if the difference is less than the currently recorded minimum difference for that particle, then update the optimal position vector P. i Otherwise P iRemain unchanged. If the minimum threshold difference among all particles is less than the minimum difference among all particles before the motion, then update P. g Otherwise P g constant.

[0110] Step (4), update particle velocity: e ij =w·e′ ij +f1(p ij -x ij )+f2(p gj -x ij Acceleration factors f1 and f2 are uniformly distributed random numbers in the interval [0,1], w is the inertia factor, and p ij For P i The coordinates of the j-th dimension, p gj For P g The coordinates of the j-th dimension, x ij Let e ​​represent the coordinates of the i-th particle in the j-th dimension. i ' j Let e ​​be the velocity of the i-th particle in the j-th dimension before the update. ij Let be the updated velocity of the i-th particle in the j-th dimension.

[0111] Step (5), Termination condition: If the maximum number of iterations is reached or σ exists. id <σ th Then the output is P. g Otherwise, jump to step (2) and repeat (2) to step (5).

[0112] Step (6), P g The degree distribution pairs corresponding to the three encoder levels are split and fed back to the transmitter encoder respectively.

[0113] In the channel capacity method based on multi-level LDPC coding and probabilistic shaping disclosed in this invention, at the transmitting end, the original binary information sequence is first converted into a multi-ary information sequence through convolutional multi-level LDPC coding, the number of ary levels being determined by the number of coding levels used. The output information sequence is then probabilistically shaped and remapped through a multi-diversity distribution matching unit, the diversity used being indexed by the values ​​at corresponding positions in the chaotic sequence. The signal groups generated by multi-diversity distribution matching are then modulated with 16QAM and transmitted through a standard single-mode fiber optic channel. At the receiving end, the received signal is first demodulated with 16QAM and then sent to a multi-diversity distribution matching probabilistic shaping demapping unit. The chaotic sequence used for indexing diversity uses the same initial values ​​and parameters as at the transmitting end to select the same diversity, and is then sent to a convolutional LDPC multi-level decoder for decoding. Simultaneously, a training sequence is used to estimate the channel capacity, and the encoder code rate is optimized based on the channel capacity.

[0114] Accordingly, the present invention also discloses a channel capacity enhancement system based on multi-level LDPC coding and probabilistic shaping. The system includes a processor and a memory. The memory stores computer instructions. The processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps of the method as described in any of the above embodiments.

[0115] Specifically, the system includes a convolutional multi-level LDPC coding unit, a multi-diversity distribution matching probability shaping mapping unit, a multi-diversity distribution matching probability shaping demapping unit, a convolutional LDPC multi-level decoding unit, and a channel capacity calculation and code rate update unit.

[0116] Figure 10 This is a schematic diagram of a communication system architecture based on multi-level LDPC coding and multi-diversity distribution matching according to an embodiment of the present invention, as shown below. Figure 10 As shown, in this system, the signal after convolutional LDPC encoding and distribution matching at the transmitting end is modulated by 16QAM to form a single signal. The output signal undergoes serial-to-parallel conversion, is fed into an arbitrary waveform generator for digital-to-analog conversion, and then linearly amplified by an electrical amplifier. The output electrical signal is then fed into a Mach-Zehnder modulator, and a variable optical attenuator adjusts the transmitting end optical power to a suitable range. The multiple signals are coupled and sent into a few-mode fiber channel, amplified by a bidirectional amplifier, and then arrive at the receiving end. The receiving end signal is first demultiplexed, then the receiving end optical power is adjusted by a variable optical attenuator. Next, a photodiode is used to detect the optical signal, and a 50G Sa / s mixed-signal oscilloscope is used to complete the analog-to-digital conversion. Then, offline digital signal processing is used to recover the original information, including multi-diversity distribution matching demapping and convolutional LDPC multi-level decoding. The code rate update module calculates the channel capacity and updates the code rate according to the training sequence of each channel, and feeds it back to the transmitting end encoder through reverse transmission.

[0117] Taking a coding level of L=3 as an example, the simulation results show the bit error rate performance of a communication system based on convolutional multi-level LDPC coding and multi-diversity distribution matching as follows: Figure 11 As shown, since each decoding stage can utilize the a posteriori information provided by the previous stage and the a priori information provided by the channel information, the bit error rate decreases progressively with the increase of the number of stages. In addition, the use of a multi-diversity distributed matcher reduces the average power of the transmitted signal, further improving system reliability and channel capacity without compromising the system code rate.

[0118] In addition, the present invention also discloses a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method as described in any of the above embodiments.

[0119] As can be seen from the above embodiments, the channel capacity enhancement method and apparatus based on multi-level LDPC coding and probabilistic shaping disclosed in this application utilizes the synod of the previous level of convolutional LDPC coding in each level. At the receiving end, the decoding module uses prior channel information and posterior information from the previous level for decoding, enabling information transfer between different levels. By employing chaotic sequence indexing of the target diversity in the multi-diversity distribution matching module, rapid diversity selection and probabilistic shaping mapping are achieved. Specifically, a convolutional LDPC parity-check matrix is ​​first constructed level by level based on the degree distribution of the parity-check matrix. Then, coding is performed from low to high levels based on the parity-check matrix, improving the correlation between levels, and the code rate gradually increases with the number of levels. At the receiving end, decoding is performed first on the lower levels, and then, based on the posterior information provided by the lower levels, decoding is performed level by level from low to high. Ultimately, the bit error rate decreases level by level from low to high, while the code rate increases level by level, achieving an overall performance improvement. After encoding, multi-diversity distribution matching probability shaping mapping is used, and chaotic sequences are used to index the used diversity, reducing the code rate loss caused by adding index tags in traditional schemes, facilitating parallel processing, and improving diversity lookup speed. By combining encoding and distribution matching, the average power of the transmitted signal is reduced, bit error rate performance is improved, and the channel capacity of the communication system is optimized. Those skilled in the art will understand that the various exemplary components, systems, and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software, or a combination of both. Whether implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention. When implemented in hardware, it can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this invention are programs or code segments used to perform the desired tasks. Programs or code segments can be stored in a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, and so on. Code segments can be downloaded via computer networks such as the Internet and intranets.

[0120] It should also be noted that the exemplary embodiments mentioned in this invention describe methods or systems based on a series of steps or apparatus. However, this invention is not limited to the order of the steps described above; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0121] In the present invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with or replace features of other embodiments.

[0122] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, various modifications and variations can be made to the embodiments of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A channel capacity enhancement method based on multi-level LDPC coding and probabilistic shaping, characterized in that, The method includes: The input signal sequence is obtained, and the input signal sequence is transformed into a multi-channel parallel input signal sequence. The parity check matrix corresponding to each level of LDPC encoding is constructed according to the degree distribution. The encoding at each level is determined based on the multi-channel parallel input signal sequence and the multiple parity check matrices. The encoding at each level is converted to a different base to obtain the first input sequence corresponding to the transmitting end. Obtain the target probability distribution, output sequence length, and target amplitude set; determine the available subsets based on the target probability distribution, output sequence length, and target amplitude set; obtain the chaotic sequence; divide each element in the chaotic sequence into multiple value intervals according to the value range; and construct a chaotic index mapping relationship between each value interval and the available subsets. The first input sequence is divided into multiple subsequences. Based on the chaotic index mapping relationship, the specific set corresponding to each subsequence is determined. The distribution matching probability shaping mapping relationship corresponding to each available set is constructed. Based on each subsequence, the specific set corresponding to each subsequence, and the distribution matching probability shaping mapping relationship, the distribution matching probability shaping output sequence corresponding to the input signal sequence is determined. Determining the available subsets based on the target probability distribution, output sequence length, and target amplitude set includes: determining the frequency set C of each amplitude occurrence in the target amplitude set. type ; Traverse and search for matching Conditional diversity groups, where C represents θ Complementary sets; each set obtained through traversal is taken as a usable set; Determining the specific subsets corresponding to each group of subsequences based on the chaotic index mapping relationship includes: calculating the chaotic index of each group of subsequences; and determining the subsets corresponding to the chaotic index of each group of subsequences through the chaotic index mapping relationship. First, construct the prototype matrix B based on the degree distribution. SC Furthermore, based on the prototype matrix, a level 0 parity check matrix is ​​constructed. Specifically, during the expansion construction, the submatrices B0, B1, ..., B in the prototype matrix are... u The zero elements in the matrix are replaced with an M×M matrix of all zeros, while the submatrices B0, B1, ..., B... are... u The non-zero elements in the matrix are randomly replaced with a permutation matrix of size M×M; the permutation matrix is ​​a matrix in which only one element in each row or column is 1; the prototype matrix is ​​then subjected to the above replacements to obtain the level 0 parity matrix. Each submatrix in the matrix is ​​derived from the prototype matrix B. SC The submatrix at the same position in the matrix is ​​expanded to obtain the result, and The submatrices in the equation have a cyclic period of T and satisfy the following condition: When t max =W, Let be a matrix of size (W+μ)bM×WcM, and The level 0 coding code rate is (cb) / c; where b and c are the number of rows and columns of the submatrix in the prototype matrix, respectively, M is the number of rows or columns of the all-zero matrix or permutation matrix, and μ is... The number of elements in each column; After constructing the level 0 parity check matrix based on the prototype matrix, the level j parity check matrix is ​​further constructed based on the level 0 parity check matrix. Submatrix in a matrix The submatrix at the corresponding position in the (j-1)th level parity check matrix The first m j Okay, m j The value of is less than the number of rows in the submatrix of the (j-1)th level parity check matrix, and The matrix code rate is cm j / c; where i takes values ​​from 0, 1…μ; After constructing the parity-check matrices for each level of LDPC encoding, the parallel signal sequences are further encoded. Before mapping, the first input sequence is first grouped into k-bit intervals, where k satisfies the condition. L represents the coding level, N perms Indicates the number of permutations of available amplitudes; The chaotic index corresponding to each group of subsequences is calculated based on the following formula: Among them, x n The nth term in the chaotic sequence is represented by the value (0,1), where n represents the ordinal number of the term; u is the control parameter of the expression, and u is selected in the range (0,4].

2. The channel capacity enhancement method based on multi-level LDPC coding and probabilistic shaping according to claim 1, characterized in that, The method further includes: The distributed matching probability shaped output sequence is inversely mapped based on the distributed matching probability shaped mapping relationship to obtain an inverse mapping sequence. The inverse mapping sequence is then subjected to multi-level LDPC decoding to obtain a multi-channel parallel output signal sequence. Finally, the multi-channel parallel output signal sequence is converted from parallel to serial to obtain an output signal sequence.

3. The channel capacity enhancement method based on multi-level LDPC coding and probabilistic shaping according to claim 2, characterized in that, The method further includes: Obtain a training sequence, and estimate the channel matrix based on the training sequence, the distribution matching probability shaped output sequence, and the output signal sequence; The channel capacity is calculated based on the estimated channel matrix; The code rate and degree distribution of each level of LDPC coding are updated according to the channel capacity and density evolution algorithm.

4. The channel capacity enhancement method based on multi-level LDPC coding and probabilistic shaping according to claim 1, characterized in that, The expression for the first input sequence is: Where V represents the first input sequence, L represents the level of LDPC encoding, and v l This represents the level l encoding.

5. The channel capacity enhancement method based on multi-level LDPC coding and probabilistic shaping according to claim 3, characterized in that, The code rate and degree distribution of each level of LDPC coding are updated according to the channel capacity and density evolution algorithm, including: Based on the estimated channel capacity and density evolution algorithm, the noise threshold of each coding level at the current code rate is calculated; Determine whether the noise threshold of the calculated LDPC codes at each level is greater than a preset threshold; When the values ​​exceed a preset threshold, the particle swarm optimization algorithm is used to adjust the code rate and degree distribution of each level of LDPC encoding.

6. A channel capacity enhancement system based on multi-level LDPC coding and probabilistic shaping, the system comprising a processor and a memory, characterized in that, The memory stores computer instructions, and the processor executes the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps of the method as described in any one of claims 1 to 5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method as described in any one of claims 1 to 5.

Images