Polarization code blind identification method, device and system

Abstract

The invention discloses a polarization code blind identification method, device and system, and belongs to the technical field of communication. When the code length of the polarization code is identified, a key code word index set of which the information freezing attribute is easily influenced by a code rate under different code lengths in a preset code length set is summarized in advance, the frozen bit proportion information of a demodulation signal under different code lengths is extracted based on the key code word index set, and the size relation of the frozen bit proportion information is compared to identify the code length of the polarization code. Identifying the real polarization code length of the signal sent by the sending end; on the basis of identifying the code length, key bit attributes of a code word sequence under the code length are identified and judged, and an information bit set and a code rate parameter are identified. In the process of traversing different code lengths of the preset code length set, all the polar code codeword indexes do not need to be completely traversed, the key codeword index in the polar code codeword indexes is recognized in advance, high accuracy can be achieved with low calculation complexity, and the polar code blind recognition method is efficient and accurate.

Description

Technical Field

[0001] The present invention belongs to the field of communication technologies, and more specifically, relates to a method, apparatus, and system for blind identification of polar codes. Background Art

[0002] The rapid development of mobile communication technologies has put forward higher requirements for transmission reliability and efficiency. As an encoding scheme for 5G control channels, polar codes have become a research hotspot due to their excellent performance that can reach the Shannon limit. In digital communication, channel coding resists channel interference by introducing redundant data, and polar codes have been strictly proven to reach the Shannon limit in binary discrete memoryless channels, having important theoretical value and application prospects. In practical scenarios such as non-cooperative communication and spectrum monitoring, signal blind identification technology has become a key research direction. This technology can, without knowing the coding parameters, identify the coding type and parameters by analyzing the characteristics of the received signal, providing core parameter support for the channel decoding system.

[0003] With the rapid development of mobile communication technologies and the increasing diversification of application scenarios, the encoding and decoding technologies of polar codes have moved from theoretical research to practical applications. Against this background, combining signal blind identification technology with polar code theory provides a new solution to solve signal analysis problems in scenarios such as non-cooperative communication and spectrum monitoring.

[0004] Existing methods for blind identification of polar codes mainly include two key traversal steps: First, it is necessary to traverse different candidate polar code length parameters and segment the received sequence according to the code length. Second, for each candidate polar code length, perform a traversal-based verification feature decision on the bits in the segmented sequence to obtain the bit indices with information bit characteristics at this code length, and then estimate the code rate. By comparing the magnitudes of the estimated code rates under different candidate polar code lengths, select the polar code length corresponding to the minimum code rate as the recognition result of the true code length, and further determine the corresponding code rate parameter and information bit index as the recognition true values according to the recognition result of the true code length. This method requires two complete traversal processes. During the traversal process, complex verification feature operations and threshold decision operations are required, which results in a high computational complexity. These problems have hindered the wide application of this technology to a certain extent, and it is impossible to achieve high accuracy with a low computational complexity, and thus cannot meet the requirements of anti-interference and real-time processing. Therefore, it is of great significance to study efficient and reliable polar code signal blind identification technology. Summary of the Invention

[0005] Aiming at the above defects or improvement requirements of the prior art, the present invention provides a method, apparatus, and system for blind identification of polar codes to solve the technical problem that the prior art cannot achieve high accuracy with a low computational complexity.

[0006] In order to achieve the above object, in a first aspect, the present invention provides a polarization code blind recognition method, comprising:

[0007] S1. Based on a key codeword index set under different code lengths in a preset code length set, frozen bit ratio information of a demodulated signal under a corresponding code length is extracted, and a real polar code length of a signal sent by a transmitter is identified by comparing the size relationship of frozen bit ratio information of the demodulated signal under different code lengths; the key codeword index set includes: a polar code codeword index whose frozen attribute is affected by a code rate in a polar code codeword index under a corresponding code length; the demodulated signal is a signal after demodulating a received signal; the received signal is a signal after a transmitted signal is transmitted through a channel; and the transmitted signal is a signal sent by a transmitter and encoded and modulated by a polar code;

[0008] S2. Divide the received signal into subsequences based on the actual polar code length of the transmitted signal, perform feature recognition and threshold judgment on the bits in each subsequence, obtain the information bit index in the subsequence, and obtain the code rate by statistics.

[0009] Further preferably, the above S1 includes:

[0010] S11, set l=1;

[0011] S12, when l = 1, based on N l The corresponding second key codeword index set calculates the demodulated signal in code length N l The second frozen position ratio information When l = L, based on N l The corresponding first key codeword index set calculates the demodulated signal in code length N l The first frozen position ratio information When l=2,3,…,L-1, based on N l The corresponding second key codeword index set calculates the demodulated signal in code length N l The second frozen position ratio information Based on N l The corresponding first key codeword index set calculates the demodulated signal in code length N l The first frozen position ratio information N l is the lth code length in the preset code length set; the preset code length set includes L code lengths of different sizes, and the code lengths are all powers of 2, and the larger code length of two adjacent code lengths is twice the smaller code length; L≥2;

[0012] in, Based on N l The corresponding a-th key codeword index set The calculation results are:

[0013] Divide the demodulated signal into multiple subsequences of length N l and construct a codeword matrix M with the subsequences as row vectors l ;

[0014] For each key codeword index in , take it as the column index, and obtain the corresponding column vector from the polarization kernel matrix F l used when performing polarization code encoding with code length N l . Then, perform the frozen bit check feature operation between the obtained column vector and each row vector of M l and take the average to obtain the frozen bit feature of this key codeword index, and further obtain the proportion information of the key codeword indexes that are frozen bits in

[0015] which is mapped based on ; l' = 1, 2, …, L - 1; including: polarization codeword indexes 1, 2, …, N l'+1 where the frozen bits of the polarization codeword indexes fluctuate under different code rates in the preset code rate set;

[0016] S13. Determine whether 2 ≤ l ≤ L is satisfied. If yes, go to S14; otherwise, set l = l + 1 and go to S12;

[0017] S14. Determine whether or l = L is satisfied. If yes, go to S15; otherwise, set l = l + 1 and go to S12;

[0018] S15. When , take N l-1 as the true polarization code length of the signal sent by the transmitter; when l = L, take N l as the true polarization code length of the signal sent by the transmitter.

[0019] Further preferably, it is obtained through the following method:

[0020] Construct a polarization codeword index vector V l” = [1, 2, …, N l” ; Estimate the channel reliability of the polarization channel corresponding to each codeword index in vector V l” to obtain the corresponding polarization channel reliability value;

[0021] Multiply vector V l” by the permutation matrix B l” to rearrange vector V l” and obtain vector V' l”; permutation matrix B l” is the permutation matrix used for polar code encoding with code length N l” ;

[0022] For each code rate R in the preset code rate set: Mark the N l” codeword indices with the smallest polar channel reliability values in the vector V' l” (1 - R) as frozen, and the remaining codeword indices as unfrozen; When the i-th codeword index in the vector V' l” is different from the (i + N l” / 2)-th codeword index, take the i-th codeword index in the vector V' l” as the key codeword index corresponding to N l” at the code rate R, or, take both the i-th codeword index and the (i + N l” / 2)-th codeword index in the vector V' l” as the key codeword index corresponding to N l” at the code rate R; i = 1, 2, …, N l” / 2;

[0023] Add all the key codeword indices corresponding to N l” at different code rates R in the preset code rate set to , and remove duplicates from ;

[0024] where a is 1 or 2; l” = 2, 3, …, L.

[0025] Further preferably, based on mapping to obtain:

[0026] Obtain the position indices of the elements belonging to l’+1 in the vector V” as the associated position indices; Obtain the key codeword indices of the vector V' l' at each associated position index and add them to ;

[0027] where the vector V” l’+1 is the first half of the vector V' l'+1 ; The vector V' l'+1 is obtained by multiplying the vector V l'+1 = [1, 2, …, N l'+1 by the permutation matrix B l'+1 ; The vector V' l' is obtained by multiplying the vector V l' = [1, 2, …, N l' by the permutation matrix B l' ;

[0028] Further preferably, the above demodulation signal is an LLR signal, which is the signal after soft demodulation of the received signal.

[0029] Further preferably, the above received signal is the signal received from the sending end and encoded by a polar code and modulated by BPSK;

[0030] The above demodulation signal is the signal after BPSK soft demodulation of the received signal.

[0031] Further preferably, wherein, wherein, is the number of key codeword indices in ; λ k is the frozen bit feature of the k-th key codeword index in , specifically: J l is the number of rows of matrix M l ; f represents the frozen bit check feature operation; M l (j,:) is the row vector corresponding to the j-th row of matrix M l ; F(:,k) is the column vector obtained by taking the k-th key codeword index as the column index of matrix F; c is the correction parameter.

[0032] Further preferably, the above feature recognition and threshold decision on the bits in the subsequence include: performing selective feature recognition and threshold decision on the bits in the subsequence based on the general partial order rule of the polar code, including:

[0033] Dividing the bits in the subsequence according to the corresponding indices based on the general partial order rule of the polar code to obtain a plurality of bit sets; for each bit set, sorting the bits therein in descending order of channel reliability corresponding to the bit indices;

[0034] For each bit set, performing feature calculation and threshold decision on the bits therein in order. When a certain bit is determined to be a frozen bit, all subsequent bits in the bit set are frozen bits, and thus the information bit decision results of each bit in the subsequence are obtained.

[0035] In a second aspect, the present invention provides a communication method, which is applied to a receiving end and includes:

[0036] Receiving the signal after the transmitted signal is transmitted through the channel and demodulating it to obtain a demodulation signal; wherein, the transmitted signal is the signal transmitted by the transmitting end and encoded by a polar code and modulated;

[0037] For the demodulated signal, perform the polarization code blind recognition method provided in the first aspect of the present invention to obtain the true polarization code length, information bit index, and code rate of the transmitted signal;

[0038] Based on the true polarization code length, information bit index, and code rate of the transmitted signal obtained by recognition, perform polarization code decoding on the demodulated signal to obtain the information sequence generated by the transmitting end.

[0039] In a third aspect, the present invention provides a communication device, including: a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, it executes the communication method provided in the second aspect of the present invention.

[0040] In a fourth aspect, the present invention provides a communication system, including: a transmitting end and a receiving end;

[0041] The transmitting end is used to perform polarization code encoding and modulation on the information sequence to be transmitted and then transmit it to the receiving end;

[0042] The receiving end is used to execute the communication method provided in the second aspect of the present invention.

[0043] In a fifth aspect, the present invention further provides a computer-readable storage medium, the computer-readable storage medium includes a stored computer program, wherein when the computer program is run by a processor, it controls the device where the storage medium is located to execute the method provided in the first aspect or the second aspect of the present invention.

[0044] In a sixth aspect, the invention further provides a computer program product, including computer program / instructions, and when the computer program / instructions are executed by a processor, the method provided in the first aspect or the second aspect of the present invention is implemented.

[0045] Generally speaking, through the above technical solutions conceived by the present invention, the following beneficial effects can be achieved:

[0046] 1. The present invention provides a blind recognition method for polar codes, which improves the execution order of the existing hierarchical recognition strategy for polar code parameters. First, it quickly recognizes the true code length of the polar code of the signal transmitted by the transmitting end, and then proceeds with the progressive process of recognizing the code rate and the information bit index set, optimizing the feature extraction efficiency and reducing the computational complexity. Among them, when recognizing the true code length of the polar code of the signal transmitted by the transmitting end, the key codeword index sets in which the information freezing attribute is susceptible to the influence of the code rate under different code lengths in the preset code length set are pre-inducted. Based on the key codeword index sets, the freezing bit ratio information of the demodulated signal under different code lengths is extracted, and the true code length of the signal transmitted by the transmitting end is recognized by comparing the magnitude relationship of the freezing bit ratio information under different code lengths. In the process of traversing different code lengths in the preset code length set in the present invention, it is not necessary to completely traverse all polar codeword indexes, but the key codeword indexes among them are pre-recognized, and high accuracy can be achieved with relatively low computational complexity. It is an efficient and accurate blind recognition method for polar codes.

[0047] 2. Further, the blind recognition method for polar codes provided by the present invention performs selective feature recognition and threshold decision on the bits in each subsequence based on the general partial order rule of polar codes. First, the codeword indexes in the subsequence are sorted in descending order of channel reliability based on the general partial order rule of polar codes; then, the feature calculation and threshold decision of information bits and freezing bits are performed on each bit in order. When a certain codeword is judged as a freezing bit, the subsequent codewords are all freezing bits. By differentiating and processing different bit positions through the diversity sorting of the general partial order rule of polar codes, the algorithm complexity is further reduced while ensuring the recognition accuracy.

[0048] 3. Further, the blind recognition method for polar codes provided by the present invention uses to calculate the feature representing the freezing bit ratio to screen the code length corresponding to the maximum freezing bit ratio information; in the calculation process of the feature representing the freezing bit ratio, since different code lengths in the preset code length set will cause different degrees of error accumulation (usually, the larger the code length, the more significant the error accumulation, making the feature value λ k show a systematic downward bias), the present invention proposes a correction method based on statistics: first, calculate the mean and the standard deviation of the feature, and then compensate and correct the feature value through the formula . This correction method effectively cancels the error accumulation effect caused by the increase in code length by introducing the weighted term of the standard deviation , thereby obtaining a more robust feature value estimation and being able to further improve the recognition accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a flowchart of the communication method based on blind recognition of polar codes provided by the embodiment of the present invention;

[0050] Figure 2 Schematic diagram of a communication system based on blind recognition of polar codes provided by an embodiment of the present invention;

[0051] Figure 3 Schematic diagram of the performance of the polar code length parameter recognized by the blind recognition method provided by the present invention under different signal-to-noise ratio conditions;

[0052] Figure 4 Schematic diagram of the performance of the polar code rate parameter and the information bit set recognized by the blind recognition method provided by the present invention under different signal-to-noise ratio conditions. Detailed implementation manners

[0053] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0054] To achieve the above objectives, in a first aspect, the present invention provides a method for blind recognition of polar codes, including:

[0055] S1. Based on the key codeword index sets corresponding to different code lengths in a preset code length set, extract the frozen bit ratio information of the demodulated signal at the corresponding code length, and identify the true polar code length of the signal transmitted by the transmitting end by comparing the size relationships of the frozen bit ratio information of the demodulated signal at different code lengths; the key codeword index sets include: the polar codeword indices in the polar codeword indices corresponding to the code length where the freezing attribute is affected by the code rate; the demodulated signal is the signal after demodulating the received signal; the received signal is the signal after the transmitted signal passes through the channel; the transmitted signal is the signal that has been polar-coded and modulated and transmitted by the transmitting end;

[0056] In an optional implementation manner, the above S1 includes:

[0057] S11. Let l = 1;

[0058] S12. When l = 1, calculate the second frozen bit ratio information of the demodulated signal at the code length N l based on the corresponding second key codeword index set l When l = L, calculate the first frozen bit ratio information of the demodulated signal at the code length N based on the corresponding first key codeword index set l When l = 2, 3,..., L - 1, calculate the frozen bit ratio information of the demodulated signal at the code length N l based on the corresponding first key codeword index set When l = 2, 3,..., L - 1, based on N lThe corresponding second key codeword index set calculates the proportion information of the second frozen bits of the demodulated signal with code length N l under the second frozen bit proportion information Based on N l The corresponding first key codeword index set calculates the proportion information of the first frozen bits of the demodulated signal with code length N l under the first frozen bit proportion information

[0059] N l is the l-th code length in the preset code length set; the preset code length set includes L code lengths of different sizes, and the code lengths are all powers of 2, and the larger code length of two adjacent sizes is twice the smaller code length; L≥2. In an alternative embodiment, the preset code length set is {16, 32, 64, 128, 512, 1024}.

[0060] It should be noted that the above method for demodulating the received signal can be a soft demodulation method or can be demodulated by a hard decision method, which is not limited here. Preferably, a soft demodulation method is adopted. At this time, the demodulated signal is an LLR signal.

[0061] Preferably, in an alternative embodiment, the above received signal is a signal received from the sending end that has been polar code encoded and BPSK modulated; the above demodulated signal is a signal obtained by performing BPSK soft demodulation on the received signal.

[0062] Based on N l The corresponding a-th key codeword index set is calculated as follows:

[0063] The demodulated signal is divided into multiple subsequences of length N l and a codeword matrix M with the subsequences as row vectors is constructed l ;

[0064] Taking each key codeword index in as the column index, from the polarization kernel matrix F l used when performing polarization code encoding with code length N l the corresponding column vector is obtained, and after performing the frozen bit check feature operation with each row vector of M l and taking the average, the frozen bit feature of this key codeword index is obtained, and then the proportion information of the key codeword indexes that are frozen bits in is obtained as

[0065] It should be noted that there are various methods for calculating here, which are not limited.

[0066] In an optional implementation manner, whether the key codeword index is a frozen bit is determined based on the frozen bit feature of the key codeword index, and then statistics are The proportion of key codeword indexes in the frozen bits is directly used as

[0067] In another optional embodiment, a method capable of characterizing statistics The characteristic of the proportion of key codeword indexes in the frozen bits is There are many ways to represent, which are not limited here. For example, in an optional implementation, in, for The number of key codeword indexes in ; k for The frozen bit features of the kth key codeword index in are as follows: J l is the matrix M l The number of rows; f represents the frozen bit check feature operation; Ml(j,:) is the matrix M l The row vector corresponding to the jth row of ; F(:, k) is the column vector obtained by taking the kth key codeword index as the column index of the matrix F. In another optional implementation mode 2, considering that due to the noise interference in the transmission process, the method has an error accumulation effect in the feature operation process, resulting in a decrease in recognition accuracy, in, Wherein, c is a correction parameter determined according to the error distribution characteristic, c∈R. Preferably, another optional implementation mode 2 is adopted.

[0068] It should be noted that there are many methods for calculating the frozen bit check feature, which are not limited here. In an optional implementation, Among them, m k For code length N l The generator matrix G used in polar code encoding l The index set of the non-zero elements in the kth column of y; j,m is the matrix M l The bit indexed by m in the row vector corresponding to the jth row of ; σ is the standard deviation of the channel noise; the channel in this embodiment is an AWGN channel. l For code length N l The permutation matrix B used in polar code encoding l With the polarization kernel matrix F l The product of .

[0069] based on The mapping obtains the following:

[0070] Obtain vector V". l’+1 Obtain the position indices of the elements belonging to as the associated position indices; obtain vector V'. l' Obtain the key word indices at each associated position index and add them to ;

[0071] Among them, vector V" l’+1 is the first half of vector V'. l'+1 Vector V' l'+1 is vector V l'+1 =[1, 2, …, N l'+1 multiplied by the permutation matrix B l'+1 ; vector V' l' is vector V l' =[1, 2, …, N l' multiplied by the permutation matrix B l' .

[0072] In an alternative embodiment, it is obtained in the following manner:

[0073] Construct the polarization code word index vector V l” =[1, 2, …, N l” ; perform channel reliability estimation on the polarization channels corresponding to each code word index in vector V l” to obtain the corresponding polarization channel reliability values;

[0074] Multiply vector V l” by the permutation matrix used during the polarization code encoding at the transmitting end to rearrange vector V l” and obtain vector V'. l” ;

[0075] For each code rate R in the preset code rate set: Mark the N l” code word indices with the smallest polarization channel reliability values among the code word indices in vector V' l” (1 - R) as frozen, and mark the remaining code word indices as non - frozen; when the identification of the i - th code word index in vector V' l” is different from that of the (i + N l” / 2) - th code word index, use the i - th code word index in vector V' l” as the key word index corresponding to N l” at code rate R, or use the identifications of both the i - th code word index and the (i + N l” / 2) - th code word index in vector V' l” as the key word index corresponding to N l”The corresponding key codeword index; i = 1, 2, …, N l” / 2; In an alternative embodiment, the above preset code rate set is

[0076] For different code rates R in the preset code rate set, N l” The corresponding key codeword indexes are all added to and perform deduplication on ;

[0077] where a is 1 or 2; l” = 2, 3, …, L.

[0078] It should be noted that there are various methods for channel reliability estimation, such as the Beta - expansion theory estimation method, the Bhattacharyya parameter method, the Gaussian approximation method, etc., which are not limited here. Preferably, the Beta - expansion theory estimation method is adopted to estimate the polarization channel weight PW corresponding to each codeword index in the vector V l” as the corresponding polarization channel reliability value.

[0079] For any codeword index idx in the vector V l” where 1 ≤ idx ≤ N l” , N l” = 2 n ; The polarization channel weight corresponding to the codeword index idx is:

[0080]

[0081] where i r is the r - th bit from the low - order to the high - order direction after converting idx - 1 to binary; β is a preset weight.

[0082] After obtaining the polarization channel weights of all codeword indexes in the vector V l” , the larger the polarization channel weight, the higher the reliability of the corresponding polarization channel.

[0083] In an alternative embodiment, Based on mapping to obtain including:

[0084] Obtain the position indexes of the elements belonging to l ” +1 in the vector V' as the associated position indexes; Obtain the key codeword indexes of the vector V' l' under each associated position index and add them to ;

[0085] where the vector V' l ” +1 is the vector V'l'+1 The first half in; vector V' l'+1 is vector V l'+1 = 1, 2, …, N l'+1 and permutation matrix B l'+1 multiplied to obtain; vector V' l' is vector V l' = [1, 2, …, N l' and permutation matrix B l' multiplied to obtain.

[0086] S13. Determine whether 2 ≤ l ≤ L is satisfied. If so, go to S14; otherwise, let l = l + 1 and go to S12;

[0087] S14. Determine whether or l = L is satisfied. If so, go to S15; otherwise, let l = l + 1 and go to S12;

[0088] S15. When , take N l-1 as the true polarization code length of the signal sent by the transmitter; when l = L, take N l as the true polarization code length of the signal sent by the transmitter.

[0089] S2. Based on the true polarization code length of the transmitted signal, perform subsequence partitioning on the received signal, perform feature recognition and threshold decision on the bits in each subsequence, obtain the information bit indices in the subsequence, and statistically obtain the code rate.

[0090] It should be noted that there are various methods for performing feature recognition and threshold decision on the bits in the subsequence. Feature extraction can adopt methods such as the average likelihood ratio method, the average likelihood difference method, the average cosine ratio method, etc.; threshold decision can adopt methods such as the maximum-minimum error probability criterion method, the zero-error probability derivative balance method, etc., which are not limited here.

[0091] Preferably, in an alternative implementation, the above-mentioned feature recognition and threshold decision on the bits in the subsequence include: performing selective feature recognition and threshold decision on the bits in the subsequence based on the general partial order rule (UPO rule) of the polarization code, including:

[0092] Partition the bits in the subsequence according to the corresponding indices based on the general partial order rule of the polarization code to obtain multiple bit sets; for each bit set, sort the bits in it in descending order according to the channel reliability corresponding to the bit indices;

[0093] For each bit set, the bits therein are sequentially subjected to feature calculation and threshold decision. When a certain bit is determined to be a frozen bit, all subsequent bits in the bit set are frozen bits, and thus the information bit decision results of each bit in the subsequence are obtained.

[0094] The present invention designs a differential decision strategy and proposes the above-mentioned method for identifying the code rate and information bit set of a general partial order assisted polar code: First, according to the UPO rule of the polar code, the polarization channel indexes under the true polarization code length of the transmitted signal at the transmitting end are subjected to diversity (set partitioning) and sorting, and then the subsequence of the received signal is subjected to set partitioning and sorting. Then, the feature calculation and threshold decision of information bits and frozen bits are sequentially performed on the bits in each subset of the subsequence, and it is judged whether some bits of the subset can directly determine the bit attribute according to the UPO sorting. Finally, the information bit set is quickly identified and the code rate parameter is counted. This method distinguishes and processes different bits through UPO diversity sorting, further reducing the computational complexity while ensuring the recognition accuracy.

[0095] To further illustrate the polar code blind recognition method provided by the present invention, it is described in detail below in conjunction with a specific embodiment:

[0096] In this embodiment, the polar code blind recognition method is applied to a communication method, such as Figure 1 shown, and the entire communication method includes:

[0097] Transmitting end:

[0098] A1. Signal generation and transmission: The transmitting end encodes and modulates the information sequence and then transmits it;

[0099] Receiving end:

[0100] A2. Signal preprocessing: Perform signal preprocessing steps such as synchronization, filtering, and demodulation on the received signal to obtain a valid data frame;

[0101] A3. Polar code length parameter identification: According to the received sequence, through fast sorting of polarization weights, analyze the difference characteristics of the frozen bit distribution of different candidate code lengths, and identify the code length parameter;

[0102] A4. Polar code rate and information bit set identification: According to the general partial order rule, perform selective feature identification and decision on the bits of the polar code, obtain the complete information bit parameters, and further count the code rate information;

[0103] A5. Channel decoding: Perform decoding on the LLR values of the demodulated codeword sequence according to the identified polar code encoding parameters to obtain the information sequence;

[0104] Specifically, in operation A1, for signal generation and transmission, at the transmitting end, any combination of coding parameters is selected for polar code encoding, and then the sequence is modulated using BPSK. The obtained modulated signal is transmitted through an additive white Gaussian noise channel (AWGN).

[0105] Specifically, in step A2, for signal preprocessing, for the transmitted signal received at the receiving end, signal processing steps such as matched filtering, downsampling, frame synchronization, channel estimation, and demodulation are performed to obtain the soft decision LLR sequence of BPSK demodulation.

[0106] Specifically, in operation A3, for polar code length parameter identification, based on the received sequence, the candidate code lengths are traversed. The polarization weights are calculated for the bit indices and quickly sorted based on the candidate code lengths to obtain the set of frozen feature detection bits under different candidate code lengths. The parity feature is calculated using the demodulated LLR sequence, and further the difference feature of the frozen bit distributions of adjacent candidate code lengths is obtained to identify the code length parameter, specifically including:

[0107] A31. The process of traversing the candidate code lengths and calculating the polarization weights for quick sorting: Traverse the candidate code lengths N from small to large, and estimate and sort the polarization channel weights corresponding to the bit indices under the code length N through the Beta expansion theory.

[0108] In a polar code of length N = 2 n take any bit index 1 ≤ idx ≤ N, and after converting idx - 1 to binary, obtain i n i n-1 …i1; where i n is the highest bit; the polarization channel weight corresponding to the codeword index idx is:

[0109]

[0110] where i r is the r-th bit from the low-order to high-order direction after converting idx - 1 to binary; β is a preset weight.

[0111] After obtaining the polarization channel weights of all codeword indices, the larger the polarization channel weight, the higher the reliability of the corresponding polarization channel of the codeword index. Sort the codeword indices according to the polarization channel weight values from large to small to obtain the channel reliability index sorting.

[0112] A32. Process of obtaining the frozen bit distribution difference feature between adjacent candidate code lengths in the preset code length set and identifying the code length parameter: When traversing the code lengths to segment and combine the received sequence to obtain sequence samples of different lengths, and forming a codeword matrix with the sequence samples as row vectors, when traversing to the candidate code length N, calculate the proportion information of the key codeword indexes that are frozen bits in the first key codeword index set Ω1 corresponding to the code length N as the frozen bit difference feature η1. At the same time, convert Ω1 into the second key codeword index set Ω2 at the code length N / 2 according to the permutation matrix in the encoding process, and similarly calculate the frozen bit difference feature η2 at the code length N / 2 according to the second key codeword index set Ω2. When the frozen bit difference feature η1 at the candidate code length N is greater than the frozen bit difference feature η2 at the candidate code length N / 2, and the frozen bit difference feature at the candidate code length 2N is less than the frozen bit difference feature at the code length N, identify that the true polar code length is N.

[0113] Specifically, the process of obtaining the set of frozen feature detection bits at different candidate code lengths: For a sequence with code length N According to the channel reliability index sorting, when the candidate code rate changes, Some bit attributes in the sequence switch between information bits and frozen bits, and the bit indexes of this part are summarized into the set of frozen feature detection bits. In this way, the set of frozen feature detection bits in the case of code length N is obtained, and the subsequent steps of calculating the bit check feature are all selected from the indexes of this set;

[0114] When N is the minimum code length in the preset code length set, the corresponding set of frozen feature detection bits is only the second key codeword index set corresponding to N, and calculate the second frozen bit proportion information of the demodulated signal at the code length N based on the second key codeword index set corresponding to N; when N is the maximum code length in the preset code length set, the corresponding set of frozen feature detection bits is only the first key codeword index set corresponding to N, and calculate the first frozen bit proportion information of the demodulated signal at the code length N based on the first key codeword index set corresponding to N; when N is other code lengths in the preset code length set except the minimum code length and the maximum code length, the corresponding set of frozen feature detection bits includes both the first key codeword index set and the second key codeword index set; calculate the second frozen bit proportion information of the demodulated signal at the code length N based on the second key codeword index set corresponding to N; calculate the first frozen bit proportion information of the demodulated signal at the code length N based on the first key codeword index set corresponding to N.

[0115] The second key codeword index set corresponding to the code length N / 2 is mapped based on the first key codeword index set corresponding to the code length N; the first key codeword index set corresponding to the code length N includes: the polar codeword indexes 1, 2,..., N that have fluctuations in frozen bits at different code rates in the preset code rate set.

[0116] During the process of calculating the check feature in the frozen bit difference feature process: According to the polarization code encoding theory, the encoding sequence x and the polarization kernel matrix F N Obtain the permutation sequence As shown in the following formula:

[0117]

[0118] Among them, the polarization kernel matrix satisfies F N F N =I. When the index k in the permutation sequence is a frozen bit, the corresponding bit v k is 0. There is the following check relationship formula (frozen bit check feature operation formula):

[0119]

[0120] Among them, f j,k represents the element in the j-th row and k-th column of the polarization kernel matrix F N

[0121] The frozen bit difference feature is specifically calculated using the LLR sequence. The process is as follows: In the AWGN channel, under BPSK soft demodulation, the log-likelihood ratio LLR X|Y (x|y) of the transmitted sequence x corresponding to the received sequence y is expressed as:

[0122]

[0123] Furthermore, the check relationship formula is deduced and simplified into the following form:

[0124]

[0125] Among them, the demodulated signal is divided into multiple subsequences of length N, and a codeword matrix M with the subsequences as row vectors is constructed; m k is the index set of the non-zero elements in the k-th column of the generating matrix G used for polar code encoding with a code length of N; y j,m is the bit with index m in the row vector corresponding to the j-th row of the codeword matrix M; σ is the standard deviation of the AWGN channel noise; the generating matrix G is the product of the permutation matrix B and the polarization kernel matrix F used for polar code encoding with a code length of N. sign is the sign operation, and min is the minimum value operation.

[0126] The frozen bit feature of the k-th key codeword index in a certain key codeword index set is: J l is the number of rows of the matrix M l ; f represents the frozen bit check feature operation; M​l (j, :) is the row vector corresponding to the j-th row of matrix M l ; F(:, k) is the column vector obtained by using the k-th key codeword index as the column index of matrix F. The corresponding frozen bit difference feature where K is the number of key codeword indices in the corresponding key codeword index set; λ k is the frozen bit feature of the k-th key codeword index in the corresponding key codeword index set.

[0127] In operation A4, first, the general partial order rule UPO of the polar code is used to identify the code rate and frozen bits of the polar code. Based on the code length parameter, the bit indices are diversified and sorted according to the general partial order. The information freezing attribute features of the bits are calculated using the LLR sequence according to the diversity order, and the bit decision is completed through the feature threshold calculation. Finally, the complete information bit index set and the code rate parameter are statistically obtained, specifically including:

[0128] Diversify the bit indices obtained according to the identified code length parameter according to the general partial order (Universal Partial Order, UPO) rule: For bits with a code length of N, the reliability of the signals of some indices is predicted by the addition rule and the left exchange rule for the channel reliability. According to the UPO rule, the bit indices are diversified and sorted. The channel reliability of each diversity as a whole shows a gradually increasing trend, and the reliability comparison of each sub-channel is limited within each diversity and cannot be directly compared across diversities. According to this diversity and sorting, the diversities are detected in the order of decreasing overall reliability between diversities, and the information freezing attributes of the bits are detected in the order of decreasing channel reliability within each diversity.

[0129] Among them, the calculation process of the information freezing attribute feature of the bit using the LLR sequence: The original bit sequence obtained by mixing information bits and frozen bits can be obtained through the following formula:

[0130] x × G N = x × G N × G N = x

[0131] where, if the bit u in the sequence i is a frozen bit, it can be obtained by the operation of the sequence x i and the i-th column vector g N in the generator matrix G i , as shown in the following check relation formula:

[0132]

[0133] The verification relation formula is expressed by the posterior LLR approximation formula as follows:

[0134]

[0135] The size of matrix X is M×N, where M represents the number of sequence samples, and M groups of eigenvalues λ are obtained. j For the average value of this group of characteristic quantities, we have:

[0136]

[0137] The process of calculating the characteristic threshold and making a decision is as follows: In BPSK modulation, the bit error probability is P, and the probability calculation of u j can be expressed as:

[0138]

[0139] where w represents the code weight of the column vector g of the generating matrix j of P e is the bit error rate calculated according to the noise variance under BPSK modulation and is expressed as:

[0140]

[0141] where exfc is the complementary error function and is expressed as:

[0142]

[0143] Finally, the decision threshold is calculated as:

[0144]

[0145] The complete information bit set and the code rate parameter are obtained through statistics, specifically as follows: When the index bit with higher channel reliability in the UPO sorting is determined to be a frozen bit according to the feature calculation, then all the index bits with lower channel reliability and later UPO sorting in this subset are directly determined to be frozen bits, thereby efficiently identifying and obtaining the code rate and the information bit set. The above steps are executed in sequence for each bit in the subset until the feature recognition of all the index bits in the subset is completed, and the information bit set is obtained. The code rate parameter is obtained through statistics.

[0146] In operation A5, the channel decoding specifically includes: Based on the identified code length, code rate, and information bit index parameters, polar code belief propagation (BP) is performed for polar code decoding to obtain information bits.

[0147] As Figure 2 shown, the present invention also provides a communication system based on blind recognition of polar codes, including:

[0148] RF module: It is used to receive the signal to be recognized with an unknown modulation and coding method and perform RF front-end processing;

[0149] Signal preprocessing module: It synchronizes, filters and demodulates the received signal to extract the valid data frame;

[0150] Polarization code coding parameter identification module: Based on the demodulated LLR sequence, through segmented reshaping, polarization kernel matrix and generating matrix feature extraction, it realizes the blind identification of polarization code coding parameters, including code length, code rate and information bit position;

[0151] Polarization code decoding module: According to the identified coding parameters, it decodes the demodulated LLR sequence to restore the original information sequence.

[0152] To verify the effectiveness of the present invention, further explanation will be made below in combination with experimental data:

[0153] The simulation experiment is implemented on the USR-2954R and LabVIEW platforms. The modulated and coded signal in the embodiment of this case is generated by MATLAB. According to Figure 1 The experiment is completed according to the steps A1 to A5 shown, and the code length, code rate, and information bit index set (or frozen bit index set) of the polarization code identification are obtained respectively.

[0154] Figure 3 It shows the performance of the blind identification method for the polarization code code length parameter in the embodiment of the present invention under different signal-to-noise ratio conditions. Figure 4 It shows the performance of the blind identification method for the polarization code rate parameter and information bit set in the embodiment of the present invention under different signal-to-noise ratio conditions.

[0155] In the second aspect, the present invention provides a communication method, which is applied to the receiving end and includes:

[0156] Receiving the signal after the transmitted signal passes through the channel transmission and demodulating it to obtain the demodulated signal; wherein, the transmitted signal is the signal that has been polarization-coded and modulated by the transmitting end;

[0157] Performing the polarization code blind identification method provided in the first aspect of the present invention on the demodulated signal to obtain the true polarization code length, information bit index and code rate of the transmitted signal;

[0158] Based on the true polarization code length, information bit index and code rate of the transmitted signal obtained by identification, performing polarization code decoding on the demodulated signal to obtain the information sequence generated by the transmitting end.

[0159] The related technical solutions are the same as the polarization code blind identification method provided in the first aspect of the present invention, and will not be limited here.

[0160] In a third aspect, the present invention provides a communication device, comprising: a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, it executes the communication method provided in the second aspect of the present invention.

[0161] The related technical solutions are the same as the communication method provided in the second aspect of the present invention and are not limited herein.

[0162] In a fourth aspect, the present invention provides a communication system, comprising: a sending end and a receiving end;

[0163] The sending end is configured to perform polar code encoding and modulation on the information sequence to be transmitted and then send it to the receiving end;

[0164] The receiving end is configured to execute the communication method provided in the second aspect of the present invention.

[0165] The related technical solutions are the same as the communication method provided in the second aspect of the present invention and are not limited herein.

[0166] In a fifth aspect, the present invention further provides a computer-readable storage medium, where the computer-readable storage medium includes a stored computer program, and when the computer program is run by a processor, it controls the device where the storage medium is located to execute the method provided in the first aspect or the second aspect of the present invention.

[0167] The related technical solutions are the same as the method provided in the first aspect or the second aspect of the present invention and are not limited herein.

[0168] In a sixth aspect, the invention further provides a computer program product, comprising a computer program / instructions, and when the computer program / instructions are executed by a processor, they implement the method provided in the first aspect or the second aspect of the present invention.

[0169] The related technical solutions are the same as the method provided in the first aspect or the second aspect of the present invention and are not limited herein.

[0170] Those skilled in the art can easily understand that the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A blind recognition method for polar codes, characterized in that, It includes: S1. Based on the key codeword index sets under different code lengths in the preset code length set, extract the frozen bit ratio information of the demodulated signal at the corresponding code length, and identify the true polarization code length of the transmitted signal by comparing the magnitude relationships of the frozen bit ratio information of the demodulated signal at different code lengths; The key codeword index set includes: the polarization codeword indexes whose frozen attributes are affected by the code rate among the polarization codeword indexes at the corresponding code length; the demodulated signal is the signal obtained by demodulating the received signal; The received signal is the signal after the transmitted signal passes through the channel; the transmitted signal is the signal that is encoded and modulated by the polarization code and then transmitted by the transmitting end; S2. Based on the true polarization code length of the transmitted signal, divide the received signal into subsequences, perform feature recognition and threshold decision on the bits in each subsequence, obtain the information bit indexes in the subsequence, and statistically obtain the code rate.

2. The blind recognition method of polar codes according to claim 1, wherein, The S1 includes: S11. Let l = 1; S12. When l = 1, based on N l Calculate the proportion information of the second frozen bits of the demodulated signal at code length N based on the corresponding second set of critical codeword indices l When l = L, based on N Calculate the proportion information of the first frozen bits of the demodulated signal at code length N based on the corresponding first set of critical codeword indices l When l = 2, 3, …, L - 1, based on N l Calculate the proportion information of the second frozen bits of the demodulated signal at code length N based on the corresponding second set of critical codeword indices Calculate the proportion information of the first frozen bits of the demodulated signal at code length N based on the corresponding first set of critical codeword indices l N l is the l-th code length in the preset code length set; the preset code length set includes L code lengths of different sizes, and the code lengths are all powers of 2, and the larger code length of two adjacent sizes is twice the smaller code length; L ≥ 2; Based on N l Calculate the proportion information of the first frozen bits of the demodulated signal at code length N based on the corresponding first set of critical codeword indices l When l = 2, 3, …, L - 1, based on N Calculate the proportion information of the second frozen bits of the demodulated signal at code length N based on the corresponding second set of critical codeword indices l N Based on N l The corresponding a-th key codeword index set Calculated as: Divide the demodulated signal into multiple subsequences of length N l and construct a codeword matrix M with the subsequences as row vectors l ; Taking each key codeword index in as the column index, obtaining the corresponding column vector from the polarization kernel matrix F l used when performing polarization code encoding with code length N, and respectively performing the frozen bit check feature operation with each row vector of M l and then taking the average to obtain the frozen bit feature of this key codeword index, and further obtaining l the proportion information of the key codeword indexes that are frozen bits in as Based on mapping to obtain; l' = 1, 2, …, L - 1; including: polarization codeword indexes 1, 2, …, N l'+1 in the polarization codeword indexes where the frozen bits fluctuate at different code rates in the preset code rate set; S13. Judge whether 2 ≤ l ≤ L is satisfied. If so, go to S14; otherwise, let l = l + 1 and go to S12; S14. Determine whether the following is satisfied or l = L. If so, go to S15; otherwise, set l = l + 1 and go to S12; S15. When , take N l-1 as the true code length of the polar code for the signal sent by the transmitting end; when l = L, take N l as the true code length of the polar code for the signal sent by the transmitting end.

3. The polarization code blind recognition method according to claim 2, wherein Obtained by the following method: Construct the polarization codeword index vector V l” = [1, 2, …, N l” ; Estimate the channel reliability for each polarization channel corresponding to the codeword index in vector V l” to obtain the corresponding polarization channel reliability value; Multiply the vector V l” by the permutation matrix B l” to rearrange the vector V l” and obtain the vector V' l” ; the permutation matrix B l” is the permutation matrix used in polar code encoding with a code length of N l” ; For each code rate R in the preset code rate set: The vector V' l” in which the N l” (1 - R) codeword indices with the smallest polarization channel reliability values are marked as frozen, and the remaining codeword indices are marked as unfrozen; when the i-th codeword index in the vector V' l” is different from the identification of the (i + N l” / 2)-th codeword index, the i-th codeword index in the vector V' l” is used as the key codeword index corresponding to N l” at the code rate R, or, the identifications of both the i-th codeword index and the (i + N l” / 2)-th codeword index in the vector V' l” are used as the key codeword index corresponding to N l” at the code rate R; i = 1, 2, …, N l” / 2; Add all the key codeword indices corresponding to N at different code rates R in the preset code rate set to l” and remove duplicates from ; ​ where a is 1 or 2; l” = 2, 3, …, L.

4. The blind identification method of polar code according to claim 2, characterized in that Based on Mapped to include: Obtain vector V' l ” +1 Obtain the position index of the elements belonging to as the associated position index; Obtain vector V' l' The key codeword index under each associated position index, and add it to ; Among them, the vector V' l ” +1 is the first half of the vector V' l'+1 ; the vector V' l'+1 is the vector V l'+1 = [1, 2, …, N l'+1 multiplied by the permutation matrix B l'+1 ; the vector V' l' is the vector V l' = [1, 2, …, N l' multiplied by the permutation matrix B l' .

5. The blind recognition method of polar codes according to claim 2, characterized in that, The received signal is the signal received from the transmitting end that is encoded by the polarization code and modulated by BPSK; The demodulated signal is the signal obtained by performing BPSK soft demodulation on the received signal.

6. The blind recognition method of polar code according to claim 5, characterized in that Among them, is the number of key codeword indexes in; λ k is the frozen bit feature of the k-th key codeword index in, specifically: J l is the number of rows of matrix M l ; f represents the frozen bit check feature operation; M l (j,:) is the row vector corresponding to the j-th row of matrix M l ; F(:,k) is the column vector obtained by taking the k-th key codeword index as the column index of matrix F; c is the correction parameter.

7. The blind identification method of polar codes according to any one of claims 1-6, characterized in that, Performing feature recognition and threshold decision on the bits in the subsequence includes: selectively performing feature recognition and threshold decision on the bits in the subsequence based on the general partial order rule of the polarization code, including: Dividing the bits in the subsequence according to the corresponding indexes based on the general partial order rule of the polarization code to obtain multiple bit sets; for each bit set, sort the bits in it in descending order of the channel reliability corresponding to the bit indexes; For each bit set, perform feature calculation and threshold decision on the bits in it in order. When a certain bit is determined to be a frozen bit, all subsequent bits in the bit set are frozen bits, and thus the information bit decision results of each bit in the subsequence are obtained.

8. A communication method, characterized in that, It includes: Receive the signal after the transmitted signal passes through the channel and demodulate it to obtain the demodulated signal; the transmitted signal is the signal that is encoded by the polarization code and modulated and then transmitted by the transmitting end; For the demodulated signal, execute the polarization code blind recognition method described in any one of claims 1 - 7 to obtain the true polarization code length, information bit indexes and code rate of the transmitted signal; Based on the identified true polarization code length, information bit indexes and code rate of the transmitted signal, perform polarization code decoding on the demodulated signal to obtain the information sequence generated by the transmitting end.

9. A communication device, characterized in that, It includes: A memory and a processor, the memory stores a computer program, and when the processor executes the computer program, it executes the communication method described in claim 8.

10. A communication system, characterized in that, It includes: a transmitting end and a receiving end; The transmitting end is used to encode and modulate the information sequence to be transmitted and then transmit it to the receiving end; The receiving end is used to execute the communication method described in claim 8.

Images