CN106877980A - Hybrid Sparse Code Multiple Access Method - Google Patents

Abstract

The invention provides a hybrid sparse code multiple access method. The method includes: performing channel coding on the data of multiple user layers respectively to obtain multiple bit streams, and equally dividing the multiple bit streams into a first bit stream and a second bit stream; mapping the first bit stream to the first Constellation diagram, mapping the second bit stream to the second constellation diagram; respectively performing mapping processing on the constellation points of the first constellation diagram and the constellation points of the second constellation diagram through a mapping matrix to obtain the first codeword corresponding to the first bit stream , the second codeword corresponding to the second bit stream; the non-zero elements in the first codeword and the non-zero elements in the second codeword are arranged through an arrangement matrix to obtain the superposition mode of each user layer on the time-frequency resource ; Perform data transmission according to the superposition mode of each user layer on the time-frequency resource. The invention improves the quality of the initial information of the MPA receiver and the convergence reliability of the first detected signal in each iteration process of the MPA receiver.

Description

Hybrid Sparse Code Multiple Access Method

technical field

The invention relates to wireless communication technology, in particular to a hybrid sparse code multiple access method.

Background technique

Low latency, high reliability, low power consumption, and massive access are the main technical features of the future 5G wireless transmission network. In wireless communication systems, multiple access is a major technique. Generally, most communication systems adopt orthogonal multiple access technology. Orthogonal multiple access technology still cannot meet the requirements of 5G system. Non-orthogonal multiple access has become a research hotspot in recent years. Sparse Code Multiple Access (SCMA) is a non-orthogonal multiple access technology based on a multi-dimensional codebook, and its structure is similar to a low density signature (LDS). LDS is different from the traditional orthogonal spread spectrum sequence, it is a kind of sparse spread spectrum sequence. In the SCMA system, first the input data stream is mapped to a multi-dimensional constellation diagram. Then the mapping matrix maps the multi-dimensional constellation points to SCMA codewords. A codebook of SCMA consists of a certain number of codewords. The transmission data layer of SCMA has a corresponding codebook. In order to improve spectrum efficiency, two or more data layers will be superimposed on the same time-frequency resources. All overlay data layer constellations are the same size and length.

The SCMA receiver uses a joint decoding algorithm of Serial Interference Cancellation (SIC) and Message Passing Algorithms (MPA). The initial information of the MPA module comes from the output of the SIC module, but the initial information of the MPA module The quality of is susceptible to multipath and noise. Although SCMA increases the energy difference value of each data layer at each resource node, the energy difference between data layers may not be as expected. This will reduce the convergence reliability of the first detected data layer in each iteration of the MPA module. In summary, under the condition of multipath fading channel, the performance of SCMA receiver needs to be further improved.

Contents of the invention

The present invention provides a mixed sparse code multiple access method to solve the problem that the initial information of the MPA module in the SCMA receiver is easily affected by multipath and noise and the first detected data layer in each MPA module iteration process Unsatisfactory convergence reliability and other issues.

The present invention provides a hybrid sparse code multiple access method, including:

Perform channel coding on the data of multiple user layers respectively to obtain multiple bit streams, and equally divide the multiple bit streams into a first bit stream and a second bit stream;

mapping the first bit stream to a first constellation, and mapping the second bit stream to a second constellation, wherein the first constellation is an N1 _- dimensional constellation, and the second constellation is N ₂ -dimensional constellation diagram;

respectively performing mapping processing on the constellation points of the first constellation diagram and the constellation points of the second constellation diagram through a mapping matrix to obtain a first codeword corresponding to the first bit stream and a code word corresponding to the second bit stream A second codeword; wherein, the first constellation diagram corresponds to a first codebook, the second constellation diagram corresponds to a second codebook, and the first codeword is a codeword in the first codebook, so The second codeword is a codeword in the second codebook;

Arranging the non-zero elements in the first codeword and the non-zero elements in the second codeword through a permutation matrix to obtain the superposition mode of each user layer on the time-frequency resource; wherein, the permutation The matrix increases the dimensional Euclidean distance between the user layers superimposed on each resource node;

Data transmission is performed according to the superposition manner of each user layer on the time-frequency resource.

Optionally, the mapping matrix is composed of non-zero rows and all-zero rows, the first codeword and the second codeword both contain non-zero elements and zero elements, and the non-zero elements of the first codeword Corresponding to the constellation points of the first constellation diagram, the non-zero elements of the second codeword correspond to the constellation points of the second constellation diagram.

Optionally, the time-frequency resource includes K resource nodes, where K=N ₁ +N ₂ ;

The number of non-zero elements of the user layer superimposed on each resource node is the same, and it satisfies

Wherein, d _f represents the number of non-zero elements of the user layer on each of the resource nodes;

The total number of user layers is J=2*d _f .

Optionally, the non-zero elements superimposed on each resource node correspond to the non-zero elements of the first codeword; or

The non-zero elements superimposed on each of the resource nodes correspond to the non-zero elements of the second codeword; or

A part of the non-zero elements superimposed on each resource node corresponds to the non-zero elements of the first codeword, and another part corresponds to the non-zero elements of the second codeword.

Optionally, each non-zero element of the first codeword corresponds to the first dimension on the first constellation diagram, and each non-zero element of the second codeword corresponds to the first dimension on the second constellation diagram. Two dimensions; the first dimension is any dimension on the first constellation diagram, and the second dimension is any dimension on the second constellation diagram.

Optionally, performing permutation processing on the non-zero elements in the first codeword and the non-zero elements in the second codeword through a permutation matrix, to obtain the superposition mode of each user layer on the time-frequency resource Previously, also included:

Acquiring all the allocation methods for allocating the user layer on each resource node;

For each allocation method, calculate the dimensional Euclidean distance sum between the user layers corresponding to each resource node;

Among the plurality of dimensional Euclidean distance sums, determine the minimum dimensional Euclidean distance and the corresponding target resource node;

Obtain the target dimension Euclidean distance sum corresponding to all allocation methods corresponding to the target resource node;

Among all target dimension Euclidean distance sums, select the largest dimension Euclidean distance sum;

The permutation matrix is determined according to the maximum dimensional Euclidean distance and a corresponding allocation manner.

Optionally, the calculating the dimensional Euclidean distance sum between user layers corresponding to each resource node includes:

Obtain the Euclidean distance between the dimensions of any two user layers on each of the resource nodes in the target constellation diagram, where the target constellation diagram is the first constellation corresponding to each of the user layers in each of the resource nodes under this allocation method diagram or second constellation diagram;

Sum the multiple obtained Euclidean distances to obtain the sum of Euclidean distances in the dimension.

Optionally, before determining the permutation matrix according to the maximum dimensional Euclidean distance and the corresponding allocation method, it also includes:

Determine that the maximum dimensional Euclidean distance sum is greater than 1, and obtain the maximum dimensional Euclidean distance and the corresponding target allocation method;

Calculate the variance of the sum of the corresponding dimensional Euclidean distances on all resource points under each target allocation method;

Determine the maximum dimensional Euclidean distance sum corresponding to the maximum variance;

The said permutation matrix is determined according to the largest dimensional Euclidean distance and the corresponding distribution method, including:

The permutation matrix is determined according to the maximum dimensional Euclidean distance corresponding to the maximum variance and the corresponding target allocation manner.

The hybrid sparse code multiple access method of the present invention obtains the equally divided first bit stream and second bit stream by performing channel coding on multiple user layers, and maps the first bit stream and the second bit stream to two different In the first constellation diagram and the second constellation diagram of the dimension, the constellation points of the first constellation diagram and the constellation points of the second constellation diagram are respectively mapped through the mapping matrix, and the first codeword and the second codeword corresponding to the first bit stream are obtained. The second codeword corresponding to the two-bit stream helps to increase the Euclidean distance between the transmission signal dimensions of each user superimposed on each resource point. Next, the non-zero elements in the first codeword and the non-zero elements in the second codeword are permuted through the permutation matrix to obtain the superimposition mode of each user layer on each resource node of the time-frequency resource. The present invention further increases the sum of the Euclidean distances between the transmission signal dimensions of the user layer of each resource node, and improves the difference between the Euclidean distance sums of the transmission signal dimensions of the user layer of all resource nodes, which further increases the The distance between the transmission signal dimensions of each user at the point improves the quality of the initial information of the MPA receiver, and also improves the convergence reliability of the first detected user transmission signal in each iteration of the MPA receiver.

Description of drawings

Fig. 1 is the flow chart one of hybrid sparse code multiple access method of the present invention;

Fig. 2 is the factor graph of hybrid sparse code multiple access method of the present invention;

3 is a schematic diagram of the transmission process of the hybrid sparse code multiple access method of the present invention;

4 is a schematic diagram of two codebooks of the hybrid sparse code multiple access method of the present invention;

Fig. 5 is the second flow chart of the mixed sparse code multiple access method of the present invention;

FIG. 6 is the third flowchart of the hybrid sparse code multiple access method of the present invention.

detailed description

Fig. 1 is the flow chart one of hybrid sparse code multiple access method of the present invention, as shown in Fig. 1, the method of this embodiment comprises:

Step 101: Perform channel coding on data of multiple user layers respectively to obtain multiple bit streams, and equally divide the multiple bit streams into a first bit stream and a second bit stream.

Specifically, multiple bit streams obtained after performing channel coding on multiple user layers in the MSCMA in this embodiment are equally divided into two parts, the first bit stream and the second bit stream.

Step 102. Map the first bit stream to a first constellation, and map the second bit stream to a second constellation, wherein the first constellation is an N ₁ -dimensional constellation, and the second constellation is an N ₂ -dimensional constellation .

Specifically, the first bit stream is mapped to the constellation points of the N ₁ -dimensional constellation diagram, and the second bit stream is mapped to the constellation points of the N ₂ -dimensional constellation diagram.

Step 103: Perform mapping processing on the constellation points of the first constellation diagram and the constellation points of the second constellation diagram respectively through the mapping matrix to obtain the first codeword corresponding to the first bit stream and the second codeword corresponding to the second bit stream; Wherein, the first constellation diagram corresponds to the first codebook, the second constellation diagram corresponds to the second codebook, the first codeword is a codeword in the first codebook, and the second codeword is a codeword in the second codebook.

Specifically, the constellation points of the first constellation diagram and the constellation points of the second constellation diagram are respectively mapped through a mapping matrix. The first constellation diagram and the second constellation diagram correspond to two different codebooks, namely the first codebook and the second codebook. The constellation points of the first constellation diagram correspond to the first codeword of the first codebook, and the constellation points of the second constellation diagram correspond to the second codeword of the second codebook. According to step 102, the first bit stream is mapped to constellation points of the first constellation diagram, and the second bit stream is mapped to constellation points of the second constellation diagram. It can be seen that the first bit stream corresponds to the first codeword, and the second bit stream corresponds to the second codeword.

Further, in the MSCMA of this embodiment, the mapping matrix is a sparse matrix. Optionally, the mapping matrix consists of non-zero rows and all-zero rows. Therefore, both the first codeword and the second codeword contain non-zero elements and zero elements. Wherein, the non-zero elements of the first codeword correspond to the constellation points of the first constellation diagram, and the non-zero elements of the second codeword correspond to the constellation points of the second constellation diagram.

Optionally, the mapping matrix of MSCMA needs to meet the following characteristics:

1) Time-frequency resources include K resource nodes, K=N ₁ +N ₂ ;

2) The number of non-zero elements of the user layer superimposed on each resource node is the same, and satisfies

where d _f represents the number of non-zero elements of the user layer on each resource node;

3) The total number of user layers is J=2*d _f .

Specifically, the number of occupied time-frequency resource nodes is determined through the dimensions of the first constellation diagram and the second constellation diagram. And the number of rows and columns in the mapping matrix is determined by the number of resource nodes and the dimensions of the two constellation diagrams, the first codeword corresponding to the first bit stream of the user layer and the second bit stream of the user layer After corresponding to the second codeword, the mapping process of the non-zero elements of the first codeword corresponding to the constellation points of the first constellation diagram and the non-zero elements of the second codeword corresponding to the constellation points of the second constellation diagram can be realized through the mapping matrix.

In a specific embodiment, the mapping process of steps 101-103 can be described by an MSCMA that contains J user layers, and the specific content is as follows:

First, after user layer j undergoes channel coding, the log ₂ (M _j ) bits of user layer j are mapped to M _j point N _j dimensional constellation C _j ;

Secondly, the binary mapping matrix V _j of the user layer j maps the constellation points of the N _j -dimensional constellation C _j into a K-dimensional codeword. And all the K-dimensional codewords of the user layer j form the codebook of the user layer j.

Among them, the binary mapping matrix V _j contains KN _j all zero rows, and after removing the all zero rows of V _j , V _j can be expressed as an identity matrix That is to say, the K-dimensional codeword of user layer j contains N _j non-zero elements and KN _j zero elements.

Finally, there are two different types of codebooks in MSCMA, which are respectively mapped by M ₁ -point N ₁ -dimensional constellation C ₁ and M ₂ -point N ₂ -dimensional constellation C ₂ to generate N _j ∈ {N ₁ ,N ₂ },M _j ∈{M ₁ ,M ₂ }, C _j ∈{C ₁ ,C ₂ },

Fig. 2 is a factor diagram of the hybrid sparse code multiple access method of the present invention, wherein the circles in Fig. 2 represent user layers, and the squares represent resource nodes. For example, for an MSCMA system with 6 transmission user layers and 5 resource nodes, as shown in Figure 2, when N ₁ =2, N ₂ =3, df=3, the factor graph matrix F corresponding to MSCMA represents as follows:

The transmission structure of the hybrid sparse code multiple access method of the present invention can be represented by a factor graph matrix.

Fig. 3 is a schematic diagram of the transmission process of the hybrid sparse code multiple access method of the present invention, and Fig. 4 is a schematic diagram of two codebooks of the hybrid sparse code multiple access method of the present invention, as shown in Fig. 3 and Fig. 4, in one containing The MSCMA transmission structure 6000 of six user layers includes codebooks 6100, 6200, 6300, 6400, and 6600. Codebooks 6100, 6200, 6300, 6400, 6500, and 6600 correspond to layer 1, layer 2, layer 3, layer 4, layer 5, and layer 6, respectively. As shown in Figure 4, MSCMA contains two different types of codebooks, wherein codebooks 6100, 6200, and 6300 belong to one type of codebook, and codebooks 6400, 6500, and 6600 belong to another type of codebook. Among them, the codebook 6100 includes codewords 6101, 6102, 6103, ..., 6164, the binary bit stream '000000' is mapped to codeword 6101, and the binary bit stream '000001' is mapped to codewords 6102, ..., binary bit stream' 111111' is mapped to codeword 6164. Codebook 6200 includes codewords 6201, 6202, 6203, ..., 6264, binary bit stream '000000' is mapped to codeword 6201, binary bit stream '000001' is mapped to codeword 6202, ..., binary bit stream '111111' Mapped to codeword 6264. Codebook 6300 contains codewords 6301, 6302, 6303, ..., 6364, binary bitstream '000000' is mapped to codeword 6301, binary bitstream '000001' is mapped to codeword 6302, ..., binary bitstream '111111' Mapped to codeword 6364. Codebook 6400 contains codewords 6401, 6402, 6403, ..., 6416, binary bit stream '0000' is mapped to codeword 6401, binary bit stream '0001' is mapped to codeword 6402, ..., binary bit stream '1111' Mapped to codeword 6416. Codebook 6500 contains codewords 6501, 6502, 6503, ..., 6516, binary bitstream '0000' is mapped to codeword 6501, binary bitstream '0001' is mapped to codeword 6502, ..., binary bitstream '1111' Mapped to codeword 6516. Codebook 6600 includes codewords 6601, 6602, 6603, ..., 6616, binary bit stream '0000' is mapped to codeword 6601, binary bit stream '0001' is mapped to codeword 6602, ..., binary bit stream '1111' Mapped to codeword 6616. Assuming that the transmission bit stream corresponding to layer 1, layer 2, and layer 3 is '000000', and the transmission bit stream corresponding to layer 4, layer5, and layer 6 is '0000', then the codewords 6101, 6201, 6301, 6401, 6501, and 6601 are superimposed The transmitted structure 6000 is formed on the same time-frequency resource.

In this way, multiple user layers obtain the first bit stream and the second bit stream through channel coding, the first bit stream is mapped to the first constellation diagram and the second bit stream is mapped to the second constellation diagram, and then the second bit stream is mapped to the second constellation diagram through the mapping matrix The constellation points of the first constellation diagram and the constellation points of the second constellation diagram are mapped to realize the first codeword in the first codebook corresponding to the first bit stream and the first codeword in the second codebook corresponding to the second bit stream. Two code words.

Step 104, arrange the non-zero elements in the first codeword and the non-zero elements in the second codeword through permutation matrix to obtain the superposition mode of each user layer on the time-frequency resource; wherein, the permutation matrix makes superimposition on each The dimensional Euclidean distance between user layers on resource nodes increases.

Specifically, there are many ways to assign each user layer to each resource node, and there are also many ways to correspond to the dimensions of the constellation points on the constellation diagram for each user layer on a resource node. On the one hand, it can maximize the dimensional Euclidean distance sum between the user layers superimposed on each resource node, and on the other hand, it can also make the dimensional Euclidean distance sum between the user layers on each resource node. maximum.

Step 105: Perform data transmission according to the superimposition manner of each user layer on the time-frequency resource.

Specifically, the data in each user layer is transmitted according to the finally determined superposition manner. The traditional SCMA codebook design mainly utilizes the SIC characteristic to separate the superimposed users' transmission signals. Traditional SCMA adopts SIC-MPA receiver. The initial information of MPA comes from the soft information output by SIC, but the soft information output by SIC is easily affected by multipath fading and noise. In addition, the arrangement criterion of the traditional SCMA codebook improves the energy difference between the dimensions of the user's transmission signal superimposed on each resource node, but the energy difference between the transmission signals of each user may not reach the expected value. This will reduce the convergence reliability of the transmission signal of the first detected user in each iteration of the MPA receiver.

For a single MPA receiver, the initial information of the MPA receiver is related to the distance between the transmission signal dimensions of each user layer superimposed on each resource point. In the MSCMA of this embodiment, by using two types of codebooks corresponding to constellation diagrams of different dimensions, it is helpful to increase the Euclidean distance between the transmission signal dimensions of each user layer superimposed on each resource point. In addition, the permutation matrix transmission method not only increases the dimensional Euclidean distance sum of the user layer transmission signals on each resource node, but also further improves the difference between the dimensional Euclidean distance sums of the user layer transmission signals of all resource nodes , thereby further increasing the Euclidean distance between the transmission signal dimensions of each user layer superimposed on each resource point. This not only improves the quality of the initial information of the MPA receiver, but also improves the convergence reliability of the first detected user transmission signal in each iteration of the MPA receiver. Finally, the bit error rate (BitError Rate, BER) performance of MSCMA will be better than the BER performance of SCMA.

In the mixed sparse code multiple access method of this embodiment, the first bit stream and the second bit stream after equalization are obtained by performing channel coding on multiple user layers, and the first bit stream and the second bit stream are mapped to two In the first constellation diagram and the second constellation diagram of different dimensions, the constellation points of the first constellation diagram and the constellation points of the second constellation diagram are respectively mapped through the mapping matrix, and the first codeword and the corresponding first bit stream are obtained. The second codeword corresponding to the second bit stream helps to increase the Euclidean distance between the transmission signal dimensions of each user superimposed on each resource point. Next, the non-zero elements in the first codeword and the non-zero elements in the second codeword are permuted through the permutation matrix to obtain the superimposition mode of each user layer on each resource node of the time-frequency resource. In this embodiment, by increasing the sum of the Euclidean distances between the transmission signal dimensions of the user layer of each resource node, and improving the difference between the sums of the Euclidean distances of the transmission signal dimensions of the user layer of all resource nodes, it further increases the The Euclidean distance between the transmission signal dimensions of each user on the resource point, thereby improving the quality of the initial information of the MPA receiver, and also improving the convergence of the first detected user transmission signal in each iteration of the MPA receiver. reliability.

How to obtain the permutation matrix described in the above-mentioned embodiment will be described below with reference to FIG. 5 . FIG. 5 is the second flow chart of the hybrid sparse code multiple access method of the present invention. As shown in Figure 5, the method includes:

Step 201. Obtain all allocation modes for allocating user layers on each resource node.

Specifically, in the MSCMA of this embodiment, the non-zero elements superimposed on each resource point may correspond to the non-zero elements of the first codeword, and may also correspond to the non-zero elements of the second codeword. Optionally, the non-zero elements superimposed on each resource node correspond to the non-zero elements of the first codeword; or the non-zero elements superimposed on each resource node correspond to the non-zero elements of the second codeword; or superimposed on each resource node A part of the non-zero elements on the node corresponds to the non-zero elements of the first codeword, and another part corresponds to the non-zero elements of the second codeword. In this way, there are many ways to allocate each user layer superimposed on each resource node.

Step 202. For each allocation method, calculate the dimensional Euclidean distance sum between user layers corresponding to each resource node.

Specifically, since there are many ways to allocate the user layers on each resource node, after obtaining all the allocation ways, for each allocation mode, it is necessary to calculate the sum of the dimensional Euclidean distances between the user layers corresponding to each resource node.

Optionally, obtain the Euclidean distance between the dimensions of any two user layers on each resource node in the target constellation diagram, where the target constellation diagram is the first constellation diagram or the second constellation diagram corresponding to each user layer in each resource node under this allocation method. Constellation diagram; sum the obtained multi-dimensional Euclidean distances to obtain the sum of dimensional Euclidean distances.

Specifically, randomly select an allocation method, and calculate the Euclidean distance between any pair of user layers mapped to the constellation point dimensions on the corresponding target constellation diagram for all user layers superimposed on a resource node, and then calculate the distance between all the dimensions The sum of the Euclidean distances of the dimensions is obtained by adding the Euclidean distances of . For the rest of the allocation methods, use the same calculation process to calculate the dimensional Euclidean distance between all user layers superimposed on the remaining resource nodes, and sum them to obtain the dimensional Euclidean distance sum. In this way, the dimensional Euclidean distance sum of all user layers superimposed on each resource node under all allocation methods is obtained.

Step 203 , among the multidimensional Euclidean distance sums, determine the minimum dimensional Euclidean distance and the corresponding target resource node.

Specifically, a distribution method is randomly selected, and the minimum dimensional Euclidean distance sum is selected by comparing the dimensional Euclidean distance sum corresponding to each resource node, and then the corresponding target resource node is selected.

Step 204, acquiring the Euclidean distance sum of the target dimension corresponding to all allocation modes corresponding to the target resource node.

Step 205, among all target dimension Euclidean distance sums, select the largest dimension Euclidean distance sum.

Specifically, according to the above-mentioned process of step 203, the target dimension Euclidean distance sum corresponding to all allocation modes corresponding to the target resource node can be obtained, and the largest dimension Euclidean distance sum is selected among them.

Step 206: Determine the permutation matrix according to the largest dimensional Euclidean distance and the corresponding distribution method.

Specifically, the maximum dimensional Euclidean distance and the corresponding allocation method are used as the way to finally arrange and superimpose the user layers of each resource node, which not only determines which user layer's non-zero elements are superimposed on which resource node, but also determines each user layer. Corresponding to the dimensions of the constellation points on the corresponding constellation diagram, the permutation matrix is then determined.

Further, there may be many kinds of selected maximum dimensional Euclidean distances and corresponding allocation methods. Optionally, each non-zero element of the first codeword corresponds to the first dimension on the first constellation diagram, and the second codeword Each non-zero element of corresponds to the second dimension on the second constellation diagram; the first dimension is any dimension on the first constellation diagram, and the second dimension is any dimension on the second constellation diagram.

Specifically, if the dimension of the first constellation diagram is 2, the non-zero elements of the first codeword may correspond to the first dimension of the constellation points of the first constellation diagram, or may correspond to the second dimension of the constellation points of the first constellation diagram , if the dimension of the second constellation diagram is 3, the non-zero elements of the second codeword may correspond to the first dimension of the constellation points of the second constellation diagram, or may correspond to the second dimension of the constellation points of the second constellation diagram, or It may correspond to the third dimension of the constellation points of the second constellation diagram. In this way, the non-zero elements of the first codeword can correspond to different dimensions of the constellation points of the first constellation diagram, and the non-zero elements of the second codeword can correspond to the different dimensions of the constellation points of the second constellation diagram, so that the first codeword There are many ways to allocate the constellation points corresponding to the first constellation diagram, and there are also many ways for the second codeword to correspond to the constellation points in the second constellation diagram.

For the embodiment shown in FIG. 5 , when there are multiple distribution methods, one distribution method may be selected from the multiple distribution methods. A possible implementation manner of selecting an allocation manner among multiple allocation manners is given below with reference to FIG. 6 .

Fig. 6 is a flowchart three of the mixed sparse code multiple access method of the present invention, as shown in Fig. 6, the method includes:

Step 301. Determine that the maximum dimensional Euclidean distance sum is greater than 1, and obtain the maximum dimensional Euclidean distance and the corresponding target allocation method.

Step 302. Calculate the variance of the sum of the corresponding dimensional Euclidean distances on all resource points under each target allocation mode.

Step 303. Determine the maximum dimensional Euclidean distance sum corresponding to the maximum variance.

Step 304: Determine the permutation matrix according to the largest dimensional Euclidean distance and the corresponding distribution method, including: determining the permutation matrix according to the largest dimensional Euclidean distance corresponding to the maximum variance and the corresponding target distribution method.

Specifically, the number of target allocation methods is judged by the number of maximum dimensional Euclidean distance sums. In all target allocation methods, the maximum variance is selected by calculating the variance of the corresponding dimensional Euclidean distance sum of all resource nodes. Since the maximum variance corresponds to the maximum dimensional Euclidean distance sum, the maximum dimensional Euclidean distance and the corresponding target distribution method, the corresponding target distribution method can be determined according to the maximum variance, and finally the permutation matrix is determined.

In a specific embodiment, in order to better describe the permutation matrix, assuming that the operation on the multi-dimensional constellation of user layer j is only permutation matrix π _j , then the codeword of user layer j can be expressed as x _j =q _j = V _j π _j z.

Among them, z(z ¹ ,z ² ,...,z ^N ), N=max(N ₁ ,N ₂ ) represents a constellation point of multidimensional constellation C ₁ or C ₂ , z ⁿ ∈ { ⁿ C ₁ , ⁿ C ₂ }, represents the nth dimension of the multidimensional constellation C ₁ , Represents the nth dimension of the multidimensional constellation _C2 .

Under the AWGN channel condition, the received signal can be expressed as Among them, p(z)=(p ₁ (z),p ₂ (z),...,p _K (z)) ^T , p _k (z)=d _k1 z+d _k2 z ² +... +d _kN z ^N is the interference polynomial of resource node k, J represents the number of superimposed users, and K represents the number of resource nodes occupied by users. Because the number of non-zero elements of the user layer superimposed on each resource node is the same, so Suppose N ₁ =2, N ₂ =3, d _f =4, the interference polynomial of resource node 1 is p ₁ (z)=2z+2z ² (that is, non-zero elements of four users are superimposed on resource node 1, where The non-zero elements of two users come from the 1st dimension of the multidimensional constellation, and the rest come from the 2nd dimension of the multidimensional constellation). That is, p(z) and the set of permutation matrices It is a one-to-one relationship. For a single MPA receiver, the initial information of the MPA receiver is mainly related to the dimensional Euclidean distance between the transmission user layers, and the criterion of the arrangement matrix is to increase the dimensional Euclidean distance between the transmission user layers as much as possible. The column descriptions are as follows:

First, assuming that d _f =m, then defining the dimensional Euclidean distance sum n(p _k (z)) between codewords superimposed on resource node k can be expressed as:

in, is the real part of the nth dimension of the user layer j superimposed on the resource node k, and is the imaginary part of the nth dimension of the user layer j superimposed on the resource node k.

Secondly, use formula (4) to select a certain number of permutation matrix sets ∏ ^* ={∏ ^1* ,∏ ^2* ,...}, and their corresponding n(p(z))={n(p ₁ (z )),...,n(p _K (z))} the smallest n(p _k (z) is the largest.

Finally, use the formula (5) to select the final permutation matrix ∏ ^l* , and its corresponding n(p(z))={n(p ₁ (z)),...,n(p _K (z)) } has the largest variance.

Those of ordinary skill in the art can understand that all or part of the steps for implementing the above method embodiments can be completed by program instructions and related hardware. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above-mentioned method embodiments; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other various media that can store program codes.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than limiting them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: It is still possible to modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the technical solutions of the various embodiments of the present invention. scope.

Claims (8)

1. A hybrid sparse code multiple access method, comprising:

respectively carrying out channel coding on data of a plurality of user layers to obtain a plurality of bit streams, and equally dividing the bit streams into a first bit stream and a second bit stream;

mapping the first bit stream to a first constellation diagram and mapping the second bit stream to a second constellation diagram, wherein the first constellation diagram is N₁Dimension constellation diagram, the second constellation diagram is N₂Dimension constellation diagram;

mapping the constellation points of the first constellation diagram and the constellation points of the second constellation diagram through a mapping matrix to obtain a first code word corresponding to the first bit stream and a second code word corresponding to the second bit stream; the first constellation map corresponds to a first codebook, the second constellation map corresponds to a second codebook, the first codeword is a codeword in the first codebook, and the second codeword is a codeword in the second codebook;

arranging non-zero elements in the first code word and non-zero elements in the second code word through an arrangement matrix to obtain a superposition mode of each user layer on time-frequency resources; the permutation matrix enables the dimension Euclidean distance between user layers superposed on the resource nodes to be increased;

and carrying out data transmission according to the superposition mode of each user layer on the time frequency resource.

2. The method of claim 1, wherein the mapping matrix is composed of non-zero rows and all-zero rows, wherein the first codeword and the second codeword each include non-zero elements and zero elements, wherein the non-zero elements of the first codeword correspond to constellation points of the first constellation, and wherein the non-zero elements of the second codeword correspond to constellation points of the second constellation.

3. The method of claim 2, wherein the time-frequency resource comprises K resource nodes, and wherein K is N₁+N₂；

The number of the non-zero elements of the user layer superposed on each resource node is the same and meets the requirement

Wherein d is_fA number of non-zero elements representing a user layer on each of the resource nodes;

the total number of the user layers is J-2 x d_f。

4. The method of claim 3, wherein the non-zero elements superimposed on each of the resource nodes correspond to non-zero elements of a first codeword; or

The non-zero elements superposed on each resource node correspond to the non-zero elements of the second code word; or

A portion of the non-zero elements superimposed on each of the resource nodes corresponds to non-zero elements of the first codeword, and another portion corresponds to non-zero elements of the second codeword.

5. The method of claim 4, wherein each non-zero element of the first codeword corresponds to a first dimension on the first constellation and each non-zero element of the second codeword corresponds to a second dimension on the second constellation; the first dimension is any dimension on the first constellation diagram, and the second dimension is any dimension on the second constellation diagram.

6. The method according to claim 5, wherein before the arranging the non-zero elements of the first codeword and the non-zero elements of the second codeword by the arrangement matrix to obtain the superposition mode of each of the user layers on the time-frequency resource, the method further comprises:

acquiring all allocation modes of the user layer allocated on each resource node;

calculating the sum of the dimension Euclidean distances between user layers corresponding to the resource nodes aiming at each distribution mode;

determining a minimum dimension Euclidean distance and a corresponding target resource node in the dimension Euclidean distance sums;

acquiring target dimension Euclidean distance sums corresponding to all distribution modes corresponding to the target resource node;

selecting a maximum dimension Euclidean distance sum from all the target dimension Euclidean distance sums;

and determining the arrangement matrix according to the maximum dimension Euclidean distance and the corresponding distribution mode.

7. The method of claim 6, wherein calculating the sum of Euclidean distances in dimension between user layers corresponding to the resource nodes comprises:

obtaining Euclidean distance of dimensionalities of any two user layers on each resource node in a target constellation diagram, wherein the target constellation diagram is a first constellation diagram or a second constellation diagram corresponding to each user layer in each resource node under the distribution formula;

and summing the obtained Euclidean distances to obtain the dimensionality Euclidean distance sum.

8. The method according to claim 6, wherein before determining the permutation matrix according to the maximum dimension euclidean distance and the corresponding allocation manner, the method further comprises:

determining that the number of the maximum dimension Euclidean distance sums is larger than 1, and obtaining the maximum dimension Euclidean distance and a corresponding target distribution mode;

calculating the variance of the Euclidean distance sum of corresponding dimensions on all the resource points under each target distribution mode;

determining the maximum dimension Euclidean distance sum corresponding to the maximum variance;

the determining the permutation matrix according to the maximum dimension Euclidean distance and the corresponding distribution mode comprises:

and determining the arrangement matrix according to the maximum dimension Euclidean distance corresponding to the maximum variance and the corresponding target distribution mode.