A multi-level LDPC decoding method based on CSK modulation

Abstract

The invention discloses a multi-level LDPC decoding method based on CSK modulation. The method comprises the following steps: performing radio frequency front-end processing on a received navigation radio frequency signal to obtain a digital intermediate frequency signal, performing carrier stripping on the signal, performing CSK demodulation through a matched filter module, transmitting a correlation value after CSK demodulation and corresponding bit data to a decoding module for performing LDPC decoding operation to obtain decoded telegram data, improving multi-level LDPC decoding performance by utilizing mutually orthogonal symbols modulated by CSK and soft decision information output by CSK demodulation, initializing a confidence vector of a received symbol by utilizing a CSK demodulation output result during LDPC decoding, avoiding recalculation of information to be decoded during the initialization process, thereby having smaller calculation complexity and storage complexity, and solving the problem of balancing decoding performance and complexity in the existing multi-level LDPC decoding method.

Description

Technical Field

[0001] The invention relates to the technical field of communication and navigation, and in particular to a multi-level LDPC decoding method based on CSK modulation. Background Art

[0002] In modern satellite navigation systems, LDPC codes are widely used in channel coding of navigation messages. For example, the modernized signal L1C in the GPS system uses binary LDPC coding, and the signals B1C and B2a in the Beidou satellite navigation system use 64-base LDPC coding, which can ensure the reliability of message data transmission. LDPC (Low-Density Parity Check) codes are a type of error-correcting codes with performance close to the Shannon limit. In the past two decades, binary LDPC codes have attracted considerable attention and developed rapidly due to their excellent performance. Multi-base LDPC codes can achieve greater performance gains than binary LDPC codes, but at the cost of higher computational complexity and storage resources, which has hindered the practical application and development of multi-base LDPC codes.

[0003] On the other hand, in order to increase the broadcast rate of navigation information without changing the spread spectrum code rate and spread spectrum gain, the latest satellite navigation system introduces code shift keying (CSK) modulation signals based on traditional spread spectrum signals, such as the L6 signal of the QZSS system. CSK modulation is a new type of multi-level orthogonal modulation method, which represents bit information by transmitting a variety of different pseudo code waveform signals, each of which is obtained by cyclic shifting a basic pseudo code, which can greatly improve the information transmission efficiency compared with the traditional PSK modulation method.

[0004] Compared with the modulation and broadcasting of CSK signals, the demodulation processing of CSK signals at the receiving end is relatively complicated, and its computational complexity increases exponentially with the number of bits (N) of the modulated data (q = 2 N ). Secondly, as for the multi-level LDPC decoding after CSK demodulation, the most commonly used decoding algorithms are the back propagation (BP) algorithm based on the finite field GF(q) and the logarithmic domain belief propagation (LLR-BP) algorithm. These algorithms are very complex compared to the binary LDPC decoding algorithm. Therefore, from the perspective of signal reception, the combination of CSK modulation and multi-level LDPC coding will bring great challenges to the design of receivers for new satellite navigation signals, and it is of great significance to design a demodulation and decoding method with low computational complexity and no performance loss. Summary of the invention

[0005] The technical problem to be solved by the present invention is that the computational complexity of the prior art is high. The purpose is to provide a multi-level LDPC decoding method based on CSK modulation, which has high performance and low complexity and can ensure real-time processing of high-rate CSK modulated signal demodulation and high-reliability decoding of multi-level LDPC codes.

[0006] The present invention is achieved through the following technical solutions:

[0007] A multi-level LDPC decoding method based on CSK modulation comprises the following steps:

[0008] S1, receiving a navigation radio frequency signal, performing radio frequency front-end processing on the navigation radio frequency signal, and obtaining a digital intermediate frequency signal;

[0009] S2. Generate a local carrier, wherein the local carrier includes an in-phase carrier signal and an orthogonal carrier signal, and use the in-phase carrier signal and the orthogonal carrier signal to perform carrier stripping on the digital intermediate frequency signal to obtain a carrier stripped signal;

[0010] S3, generating a multi-channel local pseudo-random code sequence according to the code phase determined by the CSK modulation mode, performing matching correlation calculation on the generated multi-channel local pseudo-random code sequence and the signal after the carrier stripping, obtaining a multi-channel correlation value and corresponding bit data, and performing CSK demodulation on the multi-channel correlation value and the corresponding bit data;

[0011] S4, performing a multi-base LDPC decoding operation on the correlation value and the corresponding bit data after CSK demodulation to obtain decoded telegram data.

[0012] The present invention obtains a digital intermediate frequency signal by front-end processing the received navigation radio frequency signal, and performs carrier stripping on the signal. CSK demodulation is performed through a matched filter module, and the correlation value and corresponding bit data after CSK demodulation are transmitted to a decoding module for LDPC decoding operation to obtain decoded telegram data. The CSK modulation is mutually orthogonal symbols and the CSK demodulation output is soft decision information, so as to improve the multi-level LDPC decoding performance. During LDPC decoding, the CSK demodulation output result is directly used to initialize the confidence vector of the received symbol, avoiding the recalculation of the information to be decoded during the initialization process, thereby having a smaller calculation complexity and storage complexity, solving the balance problem between decoding performance and complexity in the existing multi-level LDPC decoding method, and also obtaining the benefit of improving information transmission efficiency due to the use of CSK modulation, and being able to ensure real-time processing of high-speed CSK modulation signal demodulation and high-reliability multi-level LDPC code decoding, thereby effectively reducing the implementation complexity.

[0013] As a further limitation of the present invention, the RF front-end processing includes: filtering, down-converting and analog-to-digital converting the navigation RF signal.

[0014] As a further limitation of the present invention, the carrier stripping specifically includes: multiplying the in-phase carrier signal and the digital intermediate frequency signal to obtain an in-phase branch signal, and multiplying the orthogonal carrier signal and the digital intermediate frequency signal to obtain an orthogonal branch signal.

[0015] As a further limitation of the present invention, the CSK demodulation method specifically comprises the following steps:

[0016] S31, generating multiple local pseudo-random code sequences according to a preset CSK modulation mode;

[0017] S32, performing matching correlation calculation on the signal after carrier stripping and multiple local pseudo-random code sequences respectively to obtain multi-channel correlation values;

[0018] S33, sorting the multi-path correlation values ​​according to their magnitude, and converting the corresponding code phases into bit data respectively;

[0019] S34, intercepting and sorting some correlation values ​​and corresponding bit data, performing CSK demodulation on the intercepted and sorted some correlation values ​​and corresponding bit data, and using them as input to the LDPC decoding module.

[0020] As a further limitation of the present invention, the local pseudo code sequence includes a pseudo code sequence with multiple code phases orthogonal to each other, the multiple code phases correspond to original bit data, and the matching correlation value obtained by the matching correlation calculation represents soft decision information of the received original bit data.

[0021] As a further limitation of the present invention, the number of symbol bits of the CSK modulation is consistent with the number of symbol bits of the multi-level LDPC encoding, and the truncated length of the CSK demodulation output vector is consistent with the truncated length of the confidence vector of the multi-level LDPC decoding operation.

[0022] As a further limitation of the present invention, the method for multi-ary LDPC decoding specifically comprises the following steps:

[0023] S41, setting the maximum number of iterations, and initializing the current number of iterations, initializing the confidence vector of the received symbol and the confidence vector passed from the variable node to the check node according to the CSK demodulation output correlation value and bit data;

[0024] S42, for each variable node, calculating the decision symbol and the confidence vector transmitted from the variable node to the check node according to the variable node update rule;

[0025] S43, using the check matrix to calculate the checksum, if the checksum is zero, then the decoding is declared successful, the decision symbol sequence is used as the decoding output and the decoding is terminated, otherwise the next step is executed;

[0026] S44, for each check node, calculating the confidence vector transferred from the check node to the variable node according to the check node update rule;

[0027] S45. The current number of iterations is increased by one. If the current number of iterations reaches the maximum number of iterations, the decoding is terminated and a failure is declared. Otherwise, the process goes to step S42.

[0028] As a further limitation of the present invention, the variable node update rule of the multi-ary LDPC decoding method is:

[0029] When iterative decoding is initialized, the r bits of data and the corresponding correlation values ​​output after CSK demodulation are taken to form the confidence vector L of the truncated received information. j,r =(x r ,LLR(x r )), where x r is r finite field elements, LLR(x r ) is the corresponding log-likelihood ratio (LLR) value;

[0030] Set the variable node VN j Passed to the verification node CN f The location vector V2C j→i Initialize to L j ' ,r :V2C j→i =L′ j,r =L j,r ·h i,j =(x r ·h i,j ,LLR(x r )), where h i,j is an element in the check matrix H;

[0031] In the iterative decoding process, let C2V f→j It is the connected check node CN f Passed to variable node VN j The length of r is the degree vector, using VN j All received confidence vectors C2V f→j (f≠i), calculate VN j Pass to CN i The confidence vector The finite field elements Yes f,j The inverse element of rThe operation represents sorting the elements in the confidence vector and intercepting the first r elements.

[0032] As a further limitation of the present invention, in each iterative decoding process of the multi-ary LDPC decoding method, the variable node makes a decision in each iteration, and selects the finite field element with the largest confidence in the confidence vector as the decision value, and the decision formula is:

[0033]

[0034] As a further limitation of the present invention, the check node update rule of the multi-ary LDPC decoding method is:

[0035] For each check node CN i , receiving the confidence vector V2C transmitted by all variable nodes connected to it j→i , calculate the confidence vector C2V i→j :

[0036] Each addition operation is defined as the basic calculation of the check node: input two confidence vectors of length r, add the finite field elements from different confidence vectors to obtain candidate elements and calculate the corresponding LLR values, sort the LLR values ​​of all candidate elements, and intercept the first r LLR values ​​and the corresponding finite field elements to form the output confidence vector.

[0037] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0038] 1. Selecting the same number of symbol bits for CSK modulation as that for multi-base LDPC coding ensures that N bits of information are treated as one symbol unit throughout the demodulation and decoding process, which can effectively avoid information loss caused by information-bit conversion.

[0039] 2. The output of CSK demodulation is the sorted correlation value and the corresponding bit information. Compared with the conventional CSK demodulation which only outputs the hard decision bit information corresponding to the correlation peak, it retains a more complete amount of soft decision information. The output result of this CSK demodulation can be directly input into the LDPC decoding module for initialization of the confidence vector. Since the symbols modulated by CSK are mutually orthogonal, it can also ensure that the input symbols of the multi-level LDPC decoding are orthogonal, thereby accelerating the convergence speed of the decoding iteration and effectively improving the performance loss of the existing multi-level LDPC decoding algorithm in the waterfall domain.

[0040] 3. The LDPC decoding module directly uses the CSK demodulation output result to initialize the confidence vector of the received symbol, avoiding the recalculation of the information to be decoded during the initialization process; the length of the CSK demodulation output vector and all confidence vectors in the LDPC decoding process are truncated to r, r≤q, so the technical solution has a smaller computational complexity and storage complexity, solves the balance problem between decoding performance and complexity in the existing multi-base LDPC decoding method, and also obtains the benefit of improving information transmission efficiency due to the use of CSK modulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:

[0042] Figure 1 It is a flowchart of a multi-ary LDPC decoding method based on CSK modulation in an embodiment of the present invention;

[0043] Figure 2 A block diagram of a matched filter module in an embodiment of the present invention for implementing CSK demodulation;

[0044] Figure 3 The present invention is a block diagram of a multi-ary LDPC decoding module in an embodiment of the present invention. DETAILED DESCRIPTION

[0045] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0046] Example 1

[0047] like Figure 1 As shown, this embodiment provides a multi-level LDPC decoding method based on CSK modulation, comprising the following steps:

[0048] S1, receiving the navigation radio frequency signal, performing radio frequency front-end processing on the navigation radio frequency signal, and obtaining a digital intermediate frequency signal;

[0049] S2. Generate a local carrier, where the local carrier includes an in-phase carrier signal and an orthogonal carrier signal, and use the in-phase carrier signal and the orthogonal carrier signal to perform carrier stripping on the digital intermediate frequency signal to obtain a carrier stripped signal;

[0050] S3, generating a multi-channel local pseudo-random code sequence according to the code phase determined by the CSK modulation mode, performing matching correlation calculation on the generated multi-channel local pseudo-random code sequence and the signal after the carrier stripping, obtaining a multi-channel correlation value and corresponding bit data, and performing CSK demodulation on the multi-channel correlation value and the corresponding bit data;

[0051] S4, performing a multi-base LDPC decoding operation on the correlation value and the corresponding bit data after CSK demodulation to obtain decoded telegram data.

[0052] The RF front-end processing includes: filtering, down-converting and analog-to-digital conversion of the navigation RF signal; obtaining a digital intermediate frequency signal by performing front-end processing on the received navigation RF signal, and stripping the carrier thereof. The carrier stripping specifically includes: multiplying the in-phase carrier signal and the digital intermediate frequency signal to obtain the in-phase branch signal, and multiplying the orthogonal carrier signal and the digital intermediate frequency signal to obtain the orthogonal branch signal. The carrier stripped signal is CSK demodulated by a matched filter module, and the correlation value and the corresponding bit data after CSK demodulation are transmitted for LDPC decoding operation to obtain the decoded telegram data. During LDPC decoding, the CSK demodulation output result is directly used to initialize the confidence vector of the received symbol, avoiding the recalculation of the information to be decoded during the initialization process, so that it has a smaller computational complexity and storage complexity, solves the balance problem between decoding performance and complexity in the existing multi-level LDPC decoding method, and also obtains the benefit of improving information transmission efficiency due to the use of CSK modulation, which can ensure real-time processing of high-speed CSK modulation signal demodulation and high-reliability multi-level LDPC code decoding.

[0053] In some possible embodiments, the number of bits N contained in each symbol of the CSK modulated signal contained in the navigation radio frequency signal matches the number of bits contained in each symbol in the multi-level LDPC coding based on the finite field GF(q), so that the N bits in the information are treated as a symbol unit in the entire demodulation and decoding process, which can effectively avoid the information loss introduced by the information-bit conversion and maximize the advantages of CSK modulation. The CSK demodulation outputs the sorted correlation value and the corresponding bit information, which retains a more complete soft decision information compared with the conventional CSK demodulation that only outputs the hard decision bit information corresponding to the correlation peak. The output result of this CSK demodulation can be directly input into the LDPC decoding module for initialization of the confidence vector. Since the symbols of CSK modulation are mutually orthogonal, it can also ensure that the input symbols of the multi-level LDPC decoding are orthogonal, thereby accelerating the decoding iteration convergence speed and effectively improving the performance loss of the existing multi-level LDPC decoding algorithm in the waterfall domain. The LDPC decoding module directly uses the CSK demodulation output result to initialize the confidence vector of the received symbol, avoiding the recalculation of the information to be decoded during the initialization process; the length of the CSK demodulation output vector and all confidence vectors in the LDPC decoding process are truncated to r, so the technical solution has a smaller computational complexity and storage complexity, solves the balance problem between decoding performance and complexity in the existing multi-base LDPC decoding method, and also obtains the benefit of improving information transmission efficiency due to the use of CSK modulation.

[0054] like Figure 2 As shown, the signal after carrier stripping generates a multi-channel local pseudo-random code sequence according to the code phase determined by the CSK modulation method, and performs matching correlation calculation with the signal after carrier stripping to obtain multi-channel correlation values ​​and corresponding bit data, and complete CSK demodulation. The CSK demodulation method specifically includes the following steps:

[0055] S31, generating multiple local pseudo-random code sequences according to a preset CSK modulation mode;

[0056] S32, performing matching correlation calculation on the signal after carrier stripping and multiple local pseudo-random code sequences respectively to obtain multi-channel correlation values;

[0057] S33, sorting the multi-path correlation values ​​according to their magnitude, and converting the corresponding code phases into bit data respectively;

[0058] S34, intercepting and sorting some correlation values ​​and corresponding bit data, performing CSK demodulation on the intercepted and sorted some correlation values ​​and corresponding bit data, and using them as input to the LDPC decoding module.

[0059] The local pseudo code sequence includes a pseudo code sequence with multiple code phases orthogonal to each other, the multiple code phases correspond to the original bit data, the matching correlation value obtained by the matching correlation calculation represents the soft decision information of the received original bit data, the number of symbol bits of CSK modulation corresponds to the number of symbol bits of multi-base LDPC coding, the truncated length of the CSK demodulation output vector corresponds to the truncated length of the confidence vector of the multi-base LDPC decoding operation, and the number of symbol bits of CSK modulation is selected to be the same as the number of symbol bits of multi-base LDPC coding, which can ensure that N-bit information is treated as a symbol unit in the entire demodulation and decoding process, and effectively avoid the information loss introduced by information-bit conversion.

[0060] In some possible embodiments, the method for multi-ary LDPC decoding specifically includes the following steps:

[0061] S41, setting the maximum number of iterations, and initializing the current number of iterations, initializing the confidence vector of the received symbol and the confidence vector passed from the variable node to the check node according to the CSK demodulation output correlation value and bit data;

[0062] S42, for each variable node, calculating the decision symbol and the confidence vector transmitted from the variable node to the check node according to the variable node update rule;

[0063] S43, using the check matrix to calculate the checksum, if the checksum is zero, then the decoding is declared successful, the decision symbol sequence is used as the decoding output and the decoding is terminated, otherwise the next step is executed;

[0064] S44, for each check node, calculating the confidence vector transferred from the check node to the variable node according to the check node update rule;

[0065] S45. The current number of iterations is increased by one. If the current number of iterations reaches the maximum number of iterations, the decoding is terminated and a failure is declared. Otherwise, the process goes to step S42.

[0066] Example 2

[0067] like Figure 3 As shown, the multi-level LDPC code is composed of the check matrix H under the finite field GF(q) m×n Definition: Each row of the H matrix corresponds to a check node CN i , each column corresponds to a variable node VN j , the specific implementation is as follows:

[0068] Set the maximum number of iterations itr max , the current iteration number itr = 0. Using the r correlation values ​​and bit data of each symbol output by CSK demodulation, the confidence vector L of the received symbol is initialized j(0≤j<n), and each variable node VN j (0≤j<n) is passed to each check node CN i The confidence vector V2C of (0≤i<m) j→i , where the number of finite field elements contained in the transmitted belief vector is r, r≤q;

[0069] For each variable node VN j , calculate the decision symbol according to the variable node update rule and confidence vector V2C j→i ;

[0070] Using the check matrix H, calculate the checksum If s = 0, the decoding is declared successful and the symbol sequence is determined As the decoding output and terminate the decoding; otherwise, execute the next step;

[0071] For each check node CN i , according to the check node update rule, calculate from the check node CN i Passed to variable node VN j The confidence vector C2V i→j ;

[0072] Let itr = itr + 1. If itr = itr max , the decoding is terminated and a failure is declared; otherwise, the decoding is turned to each variable node VN j , calculate the decision symbol according to the variable node update rule and confidence vector V2C j→i ;

[0073] The specific update rules for the above variable nodes are:

[0074] If the current iteration number itr = 0, take the r bits of data and the corresponding correlation values ​​output after CSK demodulation to form the truncated confidence vector L j,r =(x r ,LLR(x r )), where x r is a finite field element, LLR(x r ) is the corresponding log-likelihood ratio (LLR) value. V2C j→i Initialized to L′ j,r :V2C j→i =L′ j,r =L j,r ·h i,j =(x r ·h i,j ,LLR(x r )), where h i,j is an element in the check matrix H;

[0075] If the current number of iterations itr>0, set C2V f→j It is the connected check node CN f Passed to variable node VN j The length of r is the degree vector, using VN j All received confidence vectors C2V f→j (f≠i), calculate VN j Pass to CN i The confidence vector V2C j→i : The finite field elements Yes f,j The inverse element of r The operation represents sorting the elements in the confidence vector and intercepting the first r elements.

[0076] In addition, each variable node makes a decision every time it is calculated, and selects the finite field element with the largest confidence in the confidence vector as the decision value. The decision formula is:

[0077]

[0078] The specific update rules for the above check nodes are:

[0079] For each check node CN i , receiving the confidence vector V2C transmitted by all variable nodes connected to it j→i , calculate the confidence vector C2V i→j : Each addition operation is defined as the basic calculation of the check node: input two confidence vectors of length r, add the finite field elements from different confidence vectors to obtain candidate elements and calculate the corresponding LLR values, sort the LLR values ​​of all candidate elements, and intercept the first r LLR values ​​and the corresponding finite field elements to form the output confidence vector.

[0080] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A multi-level LDPC decoding method based on CSK modulation, characterized in that: The following steps are involved: S1, receiving a navigation radio frequency signal, performing radio frequency front-end processing on the navigation radio frequency signal, and obtaining a digital intermediate frequency signal; S2. Generate a local carrier, wherein the local carrier includes an in-phase carrier signal and an orthogonal carrier signal, and use the in-phase carrier signal and the orthogonal carrier signal to perform carrier stripping on the digital intermediate frequency signal to obtain a carrier stripped signal; S3, generating a multi-channel local pseudo-random code sequence according to the code phase determined by the CSK modulation mode, performing matching correlation calculation on the generated multi-channel local pseudo-random code sequence and the signal after the carrier stripping, obtaining a multi-channel correlation value and corresponding bit data, and performing CSK demodulation on the multi-channel correlation value and the corresponding bit data; The method for CSK demodulation specifically comprises the following steps: S31, generating multiple local pseudo-random code sequences according to a preset CSK modulation mode; S32, performing matching correlation calculation on the signal after carrier stripping and multiple local pseudo-random code sequences respectively to obtain multi-channel correlation values; S33, sorting the multi-path correlation values ​​according to their magnitude, and converting the corresponding code phases into bit data respectively; S34, intercepting and sorting partial correlation values ​​and corresponding bit data, performing CSK demodulation on the intercepted and sorted partial correlation values ​​and corresponding bit data, and using them as input to the LDPC decoding module; The number of symbol bits of the CSK modulation is consistent with the number of symbol bits of the multi-ary LDPC coding, and the truncated length of the CSK demodulation output vector is consistent with the truncated length of the confidence vector of the multi-ary LDPC decoding operation; S4, performing a multi-base LDPC decoding operation on the correlation value and the corresponding bit data after CSK demodulation to obtain decoded telegram data; The multi-ary LDPC decoding method specifically comprises the following steps: S41, setting the maximum number of iterations, and initializing the current number of iterations, initializing the confidence vector of the received symbol and the confidence vector passed from the variable node to the check node according to the CSK demodulation output correlation value and bit data; S42, for each variable node, calculating the decision symbol and the confidence vector transmitted from the variable node to the check node according to the variable node update rule; S43, using the check matrix to calculate the checksum, if the checksum is zero, then the decoding is declared successful, the decision symbol sequence is used as the decoding output and the decoding is terminated, otherwise the next step is executed; S44, for each check node, calculating the confidence vector transferred from the check node to the variable node according to the check node update rule; S45. The current number of iterations is increased by one. If the current number of iterations reaches the maximum number of iterations, the decoding is terminated and a failure is declared. Otherwise, the process goes to step S42.

2. The multi-level LDPC decoding method based on CSK modulation according to claim 1, characterized in that: The radio frequency front-end processing includes: filtering, down-converting and analog-to-digital conversion of the navigation radio frequency signal.

3. The multi-level LDPC decoding method based on CSK modulation according to claim 1, characterized in that: The carrier stripping specifically includes: multiplying the in-phase carrier signal with the digital intermediate frequency signal to obtain an in-phase branch signal, and multiplying the quadrature carrier signal with the digital intermediate frequency signal to obtain a quadrature branch signal.

4. The multi-level LDPC decoding method based on CSK modulation according to claim 1, characterized in that: The local pseudo-random code sequence includes a pseudo-code sequence with multiple code phases orthogonal to each other, the multiple code phases correspond to original bit data, and the matching correlation value obtained by the matching correlation calculation represents soft decision information of the received original bit data.

5. The multi-level LDPC decoding method based on CSK modulation according to claim 1, characterized in that: The variable node update rule of the multi-ary LDPC decoding method is: When iterative decoding is initialized, the r bits of data and the corresponding correlation values ​​output after CSK demodulation are taken to form the confidence vector of the truncated received information. ,in x r is r finite field elements, LLR ( x r ) is the corresponding log-likelihood ratio (LLR) value; The variable node Passed to the check node The confidence vector Initialize to : ,in is an element in the check matrix H; In the iterative decoding process, let Is a connected check node Pass to variable node The confidence vector of length r is used All confidence vectors received ,calculate Pass to The confidence vector : , where the finite field elements yes The inverse of (•) r The operation represents sorting the elements in the confidence vector and intercepting the first r elements, r≤q.

6. The multi-level LDPC decoding method based on CSK modulation according to claim 1, characterized in that: In each iterative decoding process of the multi-level LDPC decoding method, the variable node makes a decision in each iteration, and selects the finite field element with the largest confidence in the confidence vector as the decision value. The decision formula is: 。 7. The multi-level LDPC decoding method based on CSK modulation according to claim 1, characterized in that: The check node update rule of the multi-ary LDPC decoding method is: For each check node , receives the confidence vectors transmitted by all variable nodes connected to it , calculate the confidence vector : ; in Each addition operation is defined as the basic calculation of the check node: input two confidence vectors of length r, add the finite field elements from different confidence vectors to obtain candidate elements and calculate the corresponding LLR values, sort the LLR values ​​of all candidate elements, and intercept the first r LLR values ​​and the corresponding finite field elements to form the output confidence vector.

Images