A method, medium and program product for blind identification of non-systematic convolutional code parameters

Abstract

The present invention relates to the technical field of blind channel coding identification, and discloses a method, medium, and program product for blind identification of non-systematic convolutional code parameters. The method comprises: constructing a first linear equation group and determining a first mapping vector, performing segmentation processing on the first mapping vector to obtain a first segmentation vector; then constructing a second linear equation group and obtaining a second mapping vector, performing segmentation processing on the second mapping vector to obtain a second segmentation vector, and finally generating a generator matrix and outputting a final identification result. The method can realize blind identification of non-systematic convolutional code parameters when any parameters are unknown. In non-cooperative communication, information can be decoded according to the coding identification parameters, thereby restoring the original information. In addition, segmentation processing is used to decompose the problem of transforming and solving a high-dimensional system of equations into the problem of solving two systems of equations with lower dimensions, which is easy to implement in engineering.

Description

Technical Field

[0001] The present invention relates to the technical field of blind recognition of channel coding, and is concerned with a method, medium and program product for blind recognition of parameters of a non-systematic convolutional code. Background Art

[0002] In digital communication systems, various interference and noise conditions exist in the channel, leading to information errors during transmission. Channel coding is divided into two categories: block codes and convolutional codes, based on the relationship between the parity elements and the information groups. Convolutional codes offer excellent performance due to their full utilization of the correlation between code groups. Furthermore, their code length and information bits are relatively small. Therefore, given the same code rate and equipment complexity, convolutional codes outperform block codes and are increasingly widely used. Compared to systematic convolutional codes, non-systematic convolutional codes output all check bits along with the information bits, resulting in stronger error correction capabilities. Therefore, non-systematic convolutional codes play an important role in modern satellite communications, deep space communications, mobile communications, and other fields, and blind identification of non-systematic convolutional codes is of great significance. Therefore, correctly identifying the coding parameters of non-systematic convolutional codes, particularly blindly identifying these parameters, and then recovering the original information from the intercepted signal is crucial.

[0003] However, there is still room for improvement in the blind recognition method of non-systematic convolutional codes in the prior art, which makes it difficult to implement in engineering. Summary of the Invention

[0004] The purpose of the present invention is to overcome the deficiencies of the prior art and provide a method, medium and program product for blind identification of non-systematic convolutional code parameters. The method can quickly identify parameters when no parameters are known and is easier to implement in engineering.

[0005] The object of the present invention is achieved through the following technical solutions:

[0006] In the first aspect, the present application discloses a method for blind identification of non-systematic convolutional code parameters, including: obtaining code stream data, initial code length, and initial constraint length of the non-systematic convolutional code to be identified to construct a first linear equation group and determine a first mapping vector; performing segmentation processing on the first mapping vector to obtain a first segmentation vector, constructing a first transformation matrix based on the first segmentation vector, and then performing an orthogonal transformation on the first transformation matrix to obtain a first spectrum distribution, and determining a first parameter according to the peak value of the first spectrum distribution; wherein the first parameter includes a correct code length, a correct constraint length, and a correct information bit; constructing a second linear equation group based on the determined first parameter and obtaining a second mapping vector, performing segmentation processing on the second mapping vector to obtain a second segmentation vector, constructing a second transformation matrix based on the second segmentation vector, performing an orthogonal transformation to obtain a second spectrum distribution, and determining the correct starting position according to the peak value of the second spectrum distribution; after obtaining a generating matrix based on the first parameter and the correct starting position, the final identification result is output.

[0007] Furthermore, it also includes: after obtaining the first spectrum distribution, detecting the peak of the non-zero point in the spectrum distribution as the valid peak, so as to determine the first parameter according to the position of the peak; otherwise, incrementally adjusting the initial code length and re-executing: constructing the first linear equation group and determining the first mapping vector and subsequent steps.

[0008] Further, the method specifically includes the following steps: S1, obtaining the code stream data, initial code length, and initial constraint length of the non-systematic convolutional code to be identified; S2, constructing the first linear equation group based on the code stream data, initial code length, and initial constraint length, and mapping the row coefficients of the first linear equation group to decimal data to obtain the first mapping vector V ; S3, the first mapping vector is segmented to obtain a first segment vector Ψ , based on the first segment vector Constructing a Hadamard matrix; S4, constructing the first segment vector of the Hadamard matrix ΨPerform a Walsh-Hadamard transform to obtain the first spectrum distribution of the Walsh spectrum; S5, detect the peak of the non-zero point in the first spectrum distribution, including: searching for the peak of the first spectrum distribution as a valid peak outside the zero point position, so as to determine the first parameter according to the peak position of the first spectrum distribution, and then execute S6; otherwise, increase the initial code length by 1, and then re-execute S2 and subsequent steps; S6, construct the second linear equation group based on the determined first parameter and obtain the second mapping vector, perform segmentation processing on the second mapping vector to obtain a second segmented vector of the second mapping vector, construct a Hadamard matrix based on the second segmented vector of the second mapping vector, and then perform a Walsh-Hadamard transform on the second segmented vector of the second mapping vector to obtain the correct starting position and the correct check matrix; S7, in combination with the first parameter and the correct starting position, solve the generating matrix based on the coding structure of the non-systematic convolutional code of the corresponding code rate, and output the parameter result of the non-systematic convolutional code.

[0009] Furthermore, solving the generator matrix based on the coding structure of the non-systematic convolutional code of the corresponding code rate includes: calculating the coding rate of the non-systematic convolutional code according to the length of the correct information bit and the code length, and then selecting different generator matrix identification algorithms according to the coding rate.

[0010] Furthermore, the coding rate includes: 1 / 2 code rate, 3 / 4 code rate, and 7 / 8 code rate.

[0011] Furthermore, the step S1, obtaining the code stream data, initial code length, and initial constraint length of the non-systematic convolutional code to be identified, includes: performing the binary code stream data of the non-systematic convolutional code according to 2 n Segmentation is performed, where n is an integer between 1 and 12, the code length range is 2 to 8, and traversal is performed within the code length range to form the first segment vector .

[0012] Further, the S3, segmenting the first mapping vector to obtain a first segment vector , based on the first segment vector Construct a Hadamard matrix, where: according to the first segment vector Construct an m-dimensional Hadamard matrix with a column width of m = col2+1, where 2col2 is the first segment vector The column width is . Then the m-dimensional Hadamard matrix It can be expressed as:

[0013] ;

[0014] Where, The matrix is ​​a square matrix, and the elements of the matrix can only take two values: 1 and -1. .

[0015] Furthermore, step S6 further includes constructing the second system of linear equations based on the determined first parameters, including: traversing a starting position function according to the correct code length and the correct constraint length, and then segmenting the binary code stream data to form a second segment vector P of a second mapping vector. Based on the second segment vector P, a corresponding Hadamard matrix is ​​constructed, and then a Walsh-Hadamard transform is performed, and a correct starting position and a check matrix are determined based on a Walsh spectrum peak.

[0016] In a second aspect, the present application further discloses a computer-readable storage medium storing computer-executable instructions. When the computer-executable instructions are loaded and executed by a processor, the above-mentioned method for blind identification of non-systematic convolutional code parameters is implemented.

[0017] In a third aspect, a computer program product is also disclosed. When the computer program product is run on a terminal, the terminal executes the above-mentioned method for blind identification of non-systematic convolutional code parameters.

[0018] The beneficial effects of the present invention are:

[0019] This method can blindly identify the parameters of non-systematic convolutional codes without knowing any parameters. In non-cooperative communications, the encoded identification parameters can be used to decode the information and recover the original information. Furthermore, a segmented process is used to decompose the high-dimensional equations solved using the Walsh-Hadamard transform into two lower-dimensional equations, reducing the computational complexity from exponential to exponential. This reduced complexity also makes this method easy to implement and fast. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 Flowchart of a method for blind identification of non-systematic convolutional code parameters according to some embodiments of the present application;

[0021] Figure 2 is a schematic diagram of a first mapping vector and a first segmentation vector according to some embodiments of the present application;

[0022] Figure 3 A simulation principle block diagram of a computer program product according to some embodiments of the present application. DETAILED DESCRIPTION

[0023] The following will clearly and completely describe the technical solutions of the present invention in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work shall fall within the scope of protection of the present invention.

[0024] See Figure 1-Figure 3 , understand the non-systematic convolutional code parameter blind identification method, medium and program product of the embodiments of the present application.

[0025] Before explaining, some professional terms are explained: Walsh-Hadamard transform is an orthogonal transform, and Walsh spectrum is the spectral representation of the signal under the Walsh function basis.

[0026] First, a method for blindly identifying parameters of a non-systematic convolutional code according to an embodiment of the present application is described, comprising the following steps:

[0027] S1. Obtain binary code stream data (binary code stream), initial code length, and initial constraint length of the non-systematic convolutional code (scrambling code) to be identified.

[0028] Specifically, the binary code stream data of the non-systematic convolutional code is converted into 2 n Segmentation is performed, where n is an integer between 1 and 12, the initial code length range is 2 to 8, and traversal is performed within the initial code length range.

[0029] In some examples, the code stream sequence of the binary code stream data is y (j), where j ranges from 1, 2, 3…M, and M is the buffer length of the code stream sequence, and its value is 2 n An integer multiple of .

[0030] S2. Construct a binary linear equation system based on the code stream data, the initial code length, and the initial constraint length. In this embodiment, it is defined as the first linear equation system. The row coefficients of the first linear equation system are mapped to decimal data to obtain a first mapping vector V , whose length is , L is a positive integer.

[0031] Specifically, according to the length of the binary bit stream segment (i.e., the truncation length formed by traversing within the initial code length), the first linear equation system constructed is a binary linear equation system, and the row coefficients of the binary linear equation system are mapped to decimal data to obtain the first mapping vector VAmong them, the encoding process of non-systematic convolutional code has a linear characteristic, that is, the output codeword satisfies the linear relationship with the input information and the encoder state. Therefore, based on its code stream data, initial code length, and initial constraint length, a binary linear equation system can be constructed to obtain its corresponding first mapping vector V .

[0032] Exemplarily, the process of mapping the row coefficients of the first linear equation system (which is a binary linear equation system) into decimal data is as follows:

[0033] Assume that the coefficient vector corresponding to the i-th row of the coefficient matrix of the binary linear equation system is converted to a decimal number , whose vector length is , where n and K are both positive integers, n is the code length, and K is the constraint length. Then use The construction length is The coefficient vector of the new two-variable linear equation system The construction method is to convert the vector The position in is set to m ( m is the number of times this decimal number appears, which is also the number of times this code word appears), and the rest are set to 0.

[0034] For example =10, then Set it to 1, and the rest to 0.

[0035] but ;

[0036] Assume that there is N Two-variable linear equations, then the coefficient matrix of the first system of equations corresponds to the first mapping vector V for:

[0037] .

[0038] Where vector V The length is .

[0039] S3, the first mapping vector V Perform segmentation processing to obtain the first segment vector , based on the first segment vector Construct the Hadamard matrix. Specifically, for the first mapping vector V Segment and construct the first segment vector , the first segment vector The column width is col2, the row width is row2, and the Hadamard matrix is ​​constructed based on the column width col2.

[0040] For a better understanding, let's further explain S3 by combining the examples listed in S2. Figure 2 As shown, the first mapping vector V The length is A row vector of The first shot vector V according to The first segment vector formed after segmentation is a A matrix of dimension, where , l Is a positive integer. The Hadamard matrix (i.e., Hadamard matrix) refers to the first mapping vector corresponding to the coefficients of the first set of equations V Segment into the first segment vector Ψ , then follow the first segment vector The column width constructs an m-dimensional Hadamard matrix, where m = col2+1, then the m-dimensional Hadamard matrix It can be expressed as:

[0041] ;

[0042] Where, The matrix is ​​a square matrix, and the elements of the matrix can only take two values: 1 and -1. .

[0043] S4, for the constructed Hadamard matrix and the segmented vector, that is, the first segmented vector Ψ A Walsh-Hadamard transform is performed to obtain a first spectral distribution of a Walsh spectrum (ie, a Walsh spectrum).

[0044] In detail, assuming is the first segment vector Ψ Walsh spectrum expression of the first piecewise vector Ψ The corresponding Hadamard matrix is , then the first segment vector Ψ The Walsh-Hadamard transform can be expressed as: , where the meanings of the parameters are the same as those in the above content.

[0045] Based on this, the embodiment of the present application performs the first mapping vector V By segmenting, the problem of solving high-dimensional equations by Walsh-Hadamard transform is decomposed into the problem of solving two lower-dimensional equations. N The Walsh transform of a linear equation system is essentially a matrix multiplication, and its direct computational complexity is , is an exponential complexity. If calculated directly, N When it is large, it often exceeds the computer memory. Considering that the form of Walsh transform is similar to Fourier transform, a fast Walsh transform algorithm can be used by using the principle of butterfly operation to reduce the computational complexity to .

[0046] The details are as follows:

[0047] Assume that the first mapping vector V， Divided into , Two sections, then there are , the corresponding Hadamard is of order N, so the vector V Walsh-Hadamard variation for:

[0048] ;

[0049] Rearranging the two parts of the above formula into two rows gives:

[0050] ;

[0051] Further:

[0052] ;

[0053] In the formula is the N-1 order Hadamard matrix corresponding to the first segment vector. From the above inference, it can be seen that after the mapping vector is divided into two segments, the Hadamard matrix changes from N order to N-1 order.

[0054] In this way, we can recursively calculate the first vector V The Walsh transform is performed after segmenting the data, thus avoiding the extensive computation required by the Hadamard matrix. The fast Walsh-Hadamard transform significantly reduces computational complexity, saves storage space and registers, facilitates practical engineering applications, and significantly improves computational efficiency.

[0055] S5. Obtain the correct code length and the correct constraint length based on the position of the Walsh spectrum peak. Specifically, detect the peak of the non-zero point in the first spectrum distribution, including: searching for a peak of the first spectrum distribution outside the zero point position as a valid peak, so as to determine the first parameter based on the peak position of the first spectrum distribution, and then execute S6; otherwise, increase the initial code length by 1, and then re-execute S2 and subsequent steps.

[0056] Specifically, the correct code length and the correct constraint length can be identified by the peak search method, and then the first segment vector Ψ The Walsh spectrum will have a peak other than at zero. The value at this peak is converted into a binary vector and used as the check vector. If both the code length and constraint length are incorrect, the Walsh spectrum will have a peak only at zero and no peaks elsewhere. In this case, increase the initial code length by 1 and continue with S2 and subsequent steps. If the code length exceeds 8, exit the loop and output the corresponding prompt message.

[0057] The peak of the Walsh spectrum may be found according to a threshold value.

[0058] S6. Construct a new binary linear equation system based on the determined first parameter, which is defined as the second linear equation system here and obtain a second mapping vector, perform a Walsh-Hadamard transform on the second segment vector of the second mapping vector, and obtain a correct starting position and a correct check matrix.

[0059] Specifically, the starting position function can be traversed according to the correct code length and the correct constraint length, and then the binary code stream data can be segmented. Based on the code stream data, code length, and constraint length, a second linear equation group is constructed, and the row coefficients of the second linear equation group are mapped to decimal data to obtain the second mapping vector 。 Then, the second mapping vector is segmented to form a new second segment vector P; and a Hadamard matrix is ​​constructed according to the second segment vector P, and then a Walsh-Hadamard transform is performed, and the correct starting position and check matrix are determined according to the Walsh spectrum peak.

[0060] The specific transformation process is basically the same as in the above example, and the specific process will not be repeated here.

[0061] S7. Combining the first parameter and the correct starting position, solving a generator matrix based on the coding structure of the non-systematic convolutional code of the corresponding code rate, and outputting parameter results of the non-systematic convolutional code.

[0062] In detail, the coding rate of the non-systematic convolutional code can be calculated based on the length of the correct information bit and the code length identified in the above steps. Different generating matrix recognition algorithms are selected according to different coding rates. There are three coding rates: 1 / 2 code rate, 3 / 4 code rate, and 7 / 8 code rate.

[0063] Furthermore, different generator matrix recognition algorithms are selected according to different coding rates, which means that when the coding rate is 1 / 2, the method of solving the greatest common divisor by using Euclidean method; when the coding rate is 3 / 4 and 7 / 8, the coding structure generator matrix recognition method of 3 / 4 and 7 / 8 is used.

[0064] Specifically, the selection of a 1 / 2 bit rate is achieved by using the Euclidean method to solve the greatest common divisor. First, a binary bit stream vector is reconstructed based on the original binary bit stream data and coding parameters. The binary bit stream vector refers to a bit stream vector Y containing two complete code lengths. Then, by solving the Euclidean greatest common divisor Q of the two bit stream vectors, the coding sequence Y=Q*G can be used to calculate the generator matrix G=Y / Q, thereby completing the identification of the generator matrix.

[0065] The method for identifying a coding structure generator matrix for coding rates of 3 / 4 rate and 7 / 8 rate by utilizing the coding structure generator matrix of 3 / 4 rate and 7 / 8 rate refers to first constructing a generator polynomial and coefficient matrix of 1 / 2 rate based on the code length and constraint length identified in step S6, and then deleting the related sequence to obtain the generator matrix of 3 / 4 rate and 7 / 8 rate.

[0066] Among them, the parameter identification results of the non-systematic convolutional code mainly include information bits, starting positions, constraint lengths, check matrices, code lengths and information bit lengths, and generator matrices.

[0067] Therefore, the method of the embodiment of the present application can realize blind identification of non-systematic convolutional code parameters when any parameters are unknown. In non-cooperative communication, information can be decoded based on the coding identification parameters, thereby restoring the original information.

[0068] According to other embodiments of the present application, a computer-readable storage medium is also disclosed, in which computer-executable instructions are stored. When the computer-executable instructions are loaded and executed by a processor, any of the above-mentioned non-systematic convolutional code parameter blind identification methods is implemented.

[0069] According to other embodiments of the present application, a computer program product comprising instructions is further disclosed. When the computer program product is run on a terminal, the terminal executes any of the above-mentioned methods for blind identification of non-systematic convolutional code parameters.

[0070] Specific combination Figure 3 Understanding, which includes:

[0071] The data acquisition module is used to acquire the code stream data, initial code length, and initial constraint length of the non-systematic convolutional code to be identified.

[0072] A binary equation group construction module is used to execute S2 in the non-systematic convolutional code parameter blind identification method and S6 to construct the second linear equation group based on the determined first parameter and obtain the second mapping vector.

[0073] The Walsh-Hadamard transformation module is used to perform Walsh-Hadamard transformation on the second mapping vector in S3, S4, S5 and S6 in the non-systematic convolutional code parameter blind identification method to obtain the correct starting position and the correct check matrix.

[0074] The non-systematic convolutional code parameter identification module is used to execute S7 in the non-systematic convolutional code parameter blind identification method.

[0075] The foregoing description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein and should not be construed as excluding other embodiments. Rather, the present invention can be used in various other combinations, modifications, and environments and can be modified within the scope of the concept described herein through the above teachings or techniques or knowledge in the relevant field. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention are intended to be protected by the appended claims.

Claims

1. A method for blind identification of non-systematic convolutional code parameters, characterized in that: include: Obtaining code stream data, an initial code length, and an initial constraint length of a non-systematic convolutional code to be identified to construct a first linear equation group, and performing decimal data mapping on row coefficients of the first linear equation group to obtain a first mapping vector; Segmentally processing the first mapping vector to obtain a first segmented vector, constructing a first transformation matrix based on the first segmented vector, then performing an orthogonal transformation on the first transformation matrix to obtain a first spectrum distribution, and determining first parameters based on a peak value of the first spectrum distribution; wherein the first parameters include a correct code length, a correct constraint length, and correct information bits; constructing a second linear equation system based on the determined first parameter, mapping row coefficients of the second linear equation system into decimal data to obtain a second mapping vector, segmenting the second mapping vector to form a second segmented vector of the second mapping vector, constructing a second transformation matrix based on the second segmented vector, performing an orthogonal transformation on the second transformation matrix constructed from the second segmented vector of the second mapping vector to obtain a second spectral distribution, and determining a correct starting position based on a peak value of the second spectral distribution; After generating a generator matrix based on the first parameter and the correct starting position, a parameter result of the non-systematic convolutional code is output.

2. The method for blind identification of non-systematic convolutional code parameters according to claim 1, wherein: Also includes: After obtaining the first spectrum distribution, detecting a non-zero peak in the spectrum distribution as a valid peak, and determining the first parameter according to a position of the peak; Otherwise, that is, the code length and constraint length are both incorrect, and the spectrum distribution has a peak only at the zero point position, and no peaks at other positions, then the initial code length is incrementally adjusted and the following steps are re-executed: constructing the first linear equation group and determining the first mapping vector and subsequent steps.

3. The method for blind identification of non-systematic convolutional code parameters according to claim 2, wherein: The specific steps include: S1. Obtain code stream data, initial code length, and initial constraint length of the non-systematic convolutional code to be identified; S2: construct the first linear equations based on the code stream data, the initial code length, and the initial constraint length, and perform decimal data mapping on the row coefficients of the first linear equations to obtain the first mapping vector V ; S3. Segment processing is performed on the first mapping vector to obtain a first segment vector Ψ , based on the first segment vector Ψ Construct the Hadamard matrix; S4, constructing the first segment vector of the Hadamard matrix Ψ Performing a Walsh-Hadamard transform to obtain the first spectral distribution of the Walsh spectrum; S5, detecting the peak at a non-zero position in the first spectrum distribution, including: searching for the peak of the first spectrum distribution as a valid peak outside the zero position, so as to determine the first parameter according to the peak position of the first spectrum distribution, and then executing S6; Otherwise, that is, the code length and constraint length are both incorrect, and the spectrum distribution has a peak only at the zero point position, and no peaks at other positions, then the initial code length is increased by 1, and then S2 and subsequent steps are re-executed; S6. Constructing a second linear equation system based on the determined first parameters, mapping row coefficients of the second linear equation system into decimal data to obtain a second mapping vector, segmenting the second mapping vector to obtain a second segmented vector, and performing a Walsh-Hadamard transform on the second segmented vector of the second mapping vector to obtain the correct starting position and the correct check matrix; S7. In combination with the first parameter and the correct starting position, the generator matrix is ​​solved based on the coding structure of the non-systematic convolutional code of the corresponding code rate, and the recognition result is output.

4. The method for blind identification of non-systematic convolutional code parameters according to claim 3, wherein: Solving the generator matrix based on the coding structure of the non-systematic convolutional code of the corresponding code rate includes: The coding rate of the non-systematic convolutional code is calculated according to the length of the correct information bits and the code length, and then different generator matrix recognition algorithms are selected according to the coding rate.

5. The method for blind identification of non-systematic convolutional code parameters according to claim 4, wherein: The coding rates include: 1 / 2 code rate, 3 / 4 code rate, and 7 / 8 code rate.

6. The method for blind identification of non-systematic convolutional code parameters according to claim 3, wherein: Said S1 comprises: The binary code stream data of the non-systematic convolutional code is converted into binary code according to 2 n Segmentation is performed, where n is an integer between 1 and 12, the code length range is 2 to 8, and traversal is performed within the code length range to form a truncated sequence.

7. The method for blind identification of non-systematic convolutional code parameters according to claim 3, wherein: In the S3: According to the first segment vector Ψ Construct an m-dimensional Hadamard matrix with a column width of m = col2+1, where col2 is the first segment vector Ψ The column width of the m-dimensional Hadamard matrix Expressed as: ； Where, Hadamard matrix The matrix is ​​a square matrix, and the elements of the matrix have only two values ​​1 and -1. 。 8. The method for blind identification of non-systematic convolutional code parameters according to claim 3, wherein: The constructing the second linear equation system based on the determined first parameter includes: The starting position function is traversed according to the correct code length and the correct constraint length, and then the binary code stream data is segmented. A second linear equation group is constructed based on the code stream data, code length, and constraint length. The row coefficients of the second linear equation group are mapped to decimal data to obtain a second mapping vector, and then the second mapping vector is segmented to form a new second segmentation vector P.

9. A computer-readable storage medium, characterized in that Computer executable instructions are stored, and when the computer executable instructions are loaded and executed by a processor, the non-systematic convolutional code parameter blind identification method according to any one of claims 1 to 8 is implemented.

10. A computer program product, characterized in that When the computer program product is run on a terminal, the terminal executes the method for blind identification of non-systematic convolutional code parameters according to any one of claims 1 to 8.

Images