Lightweight channel coding blind identification method and system based on multi-modal feature fusion

Abstract

The invention belongs to the technical field of wireless communication, and particularly discloses a lightweight channel coding blind identification method and system based on multi-modal feature fusion. The method comprises the following steps: firstly, detecting a signal-to-noise ratio of a collected signal, and if the signal-to-noise ratio is smaller than a signal-to-noise ratio threshold value, respectively carrying out LDPC code identification, Polar code identification, Turbo code identification, convolutional code identification and linear block code identification on the collected signal based on a mathematical statistics theory; and otherwise, inputting the acquired signal into a channel coding identification model for blind identification of channel coding. According to the method, the signal statistical features and the signal depth features are adaptively selected based on the signal-to-noise ratio of the collected signals for blind recognition, high efficiency of mathematical statistical recognition and high precision of deep learning recognition are integrated, and dynamic balance of signal blind recognition calculation efficiency and recognition precision is achieved.

Description

Technical Field

[0001] This application belongs to the field of wireless communication technology, and more specifically, relates to a lightweight channel coding blind recognition method and system based on multi-modal feature fusion. Background Art

[0002] Channel coding blind recognition refers to the process of inferring the channel coding type and parameters (such as code length, code rate, generating polynomial, etc.) from the received signal without knowing the coding parameters. It is of great significance in non-cooperative communication and signal analysis.

[0003] Currently, channel coding blind recognition technologies are mainly divided into two categories. One is the channel coding recognition method based on mathematical statistics theory. This method analyzes the periodic components of the received signal and realizes the recognition of channel coding by artificially extracting coding features. Traditional channel recognition algorithms usually can only recognize specific channel codings and are difficult to recognize the channel coding type in a complex electromagnetic environment, and the algorithm complexity is relatively high. The other is the deep learning method based on data accumulation. This method annotates known data, usually using the original bit stream as the input, and extracts high-dimensional features from the original bit stream through a convolutional neural network to distinguish different channel coding types. However, the problem with it is that the number of model parameters is large and it is difficult to be deployed to edge devices for real-time recognition. Summary of the Invention

[0004] In view of the above defects or improvement requirements of the prior art, this application provides a lightweight channel coding blind recognition method and system based on multi-modal feature fusion, aiming to solve the technical problem that the existing channel coding blind recognition is too complex.

[0005] To achieve the above object, in a first aspect, this application provides a lightweight channel coding blind recognition method based on multi-modal feature fusion, including: Detect the signal-to-noise ratio of the collected signal. If the signal-to-noise ratio is less than the signal-to-noise ratio threshold, then perform LDPC code recognition, Polar code recognition, Turbo code recognition, convolutional code recognition, and linear block code recognition on the collected signal based on signal statistical features; otherwise, input the collected signal into a channel coding recognition model for blind recognition of channel coding. The channel coding recognition model is trained using the following steps: Generate a data set including LDPC code signals, Polar code signals, Turbo code signals, convolutional code signals, and linear block code signals through simulation; Add noise to each type of simulated signal in the data set and adaptively adjust the amplitude scaling factor of each type of simulated signal according to the signal power; Construct a channel coding recognition model based on a convolutional neural network and use the data set to train the channel coding recognition model.

[0006] Preferably, specifically: Pre-construct a candidate parity-check matrix of LDPC code; If the average a posteriori log-likelihood ratio of the parity-check matrix and the acquired signal is higher than a preset first threshold, it is determined that the acquired signal is an LDPC code, and the LDPC code parameters are identified based on the candidate parity-check matrix.

[0007] Preferably, the average a posteriori log-likelihood ratio of the parity-check matrix and the acquired signal is specifically:

[0008] Among them, is the candidate parity-check matrix, is the acquired signal, is the th bit in the acquired signal bit, is the acquired signal the total number of bits in; is and the a posteriori log-likelihood ratio of.

[0009] Preferably, the Polar code identification for the acquired signal is specifically: Use a sliding window to intercept the acquired signal to obtain multiple groups of codewords; Construct a matrix from multiple groups of codewords, and determine whether holds. If so, it is determined that the acquired signal is a Polar code, and the Polar code parameters are identified; where represents that the matrix is a row column complex matrix, is the number of codewords, is the candidate code length, represents finding the rank of the matrix of.

[0010] Preferably, the Turbo code identification for the acquired signal is specifically: Construct a Hankel matrix based on the acquired signal; Traverse different candidate values of the code length , if it can satisfy , then it is determined that the acquired signal is a Turbo code, and is the code length of the Turbo code.

[0011] Preferably, the convolutional code identification for the acquired signal is specifically: The collected signal is detected using the cyclic spectrum correlation function. If a cyclic frequency can be detected, it is determined that the collected signal is a convolutional code, and the convolutional code parameters are identified based on the cyclic frequency.

[0012] Preferably, the collected signal is detected using the cyclic spectrum correlation function, specifically:

[0013] where is the collected signal, is the time delay, is the time length of the analysis window, is the base of the natural logarithm, is pi, is the cyclic frequency, is the imaginary unit.

[0014] Preferably, the linear block code of the collected signal is identified, specifically: Candidate codewords of the linear block code are pre-constructed; The code pattern likelihood of the collected signal is calculated using the following formula :

[0015] where is the th candidate codeword, is the th bit in the collected signal, is the code length of the collected signal; If there exists a code pattern likelihood greater than a preset second threshold, it is determined that the collected signal is a linear block code, and the linear block code parameters are identified based on the code pattern likelihood .

[0016] Preferably, the amplitude scaling factor of each type of simulation signal is adaptively adjusted according to the signal power, specifically:

[0017] is the scaled signal, is the original signal, is 's mean value, is the noise variance, is 's variance, is the square root of the signal-to-noise ratio.

[0018] In a second aspect, the present application provides a lightweight channel coding blind recognition system based on multi-modal feature fusion, including: The signal-to-noise ratio detection module is used to detect the signal-to-noise ratio of the acquired signal. If the signal-to-noise ratio is less than the signal-to-noise ratio threshold, the signal statistical feature recognition module is called; otherwise, the model recognition module is called. The signal statistical feature recognition module is used to perform LDPC code recognition, Polar code recognition, Turbo code recognition, convolutional code recognition, and linear block code recognition on the acquired signal based on signal statistical features. The model recognition module is used to input the acquired signal into the channel coding recognition model for blind recognition of channel coding. The model training module is used to generate a data set including LDPC code signals, Polar code signals, Turbo code signals, convolutional code signals, and linear block code signals through simulation; add noise to each type of simulated signal in the data set, and adaptively adjust the amplitude scaling factor of each type of simulated signal according to the signal power; construct a channel coding recognition model based on a convolutional neural network, and use the data set to train the channel coding recognition model.

[0019] Generally speaking, compared with the prior art, the above technical solution conceived by this application has the following beneficial effects: (1) The blind channel coding recognition method of this application adaptively selects the statistical features and deep features of the signal based on the signal-to-noise ratio of the acquired signal for blind recognition. The mathematical statistics recognition algorithm is used for signal recognition under low signal-to-noise ratio conditions to reduce the computational complexity, and the deep learning recognition method is used under high signal-to-noise ratio conditions to improve the recognition accuracy, thereby realizing the dynamic balance between the computational efficiency and recognition accuracy of signal blind recognition.

[0020] (2) The channel coding recognition model constructed by deep learning in this application has a noise addition training strategy during the training process, ensuring that the recognition accuracy rate is greater than 90% under high signal-to-noise ratio conditions.

[0021] (3) The coding blind recognition method based on signal statistical features proposed in this application is more suitable for low signal-to-noise ratio conditions, and has lower computational complexity compared with traditional coding blind recognition methods. Description of the Drawings

[0022] Figure 1 is a flowchart of a lightweight blind channel coding recognition method based on multi-modal feature fusion provided by an embodiment of this application.

[0023] Figure 2 is a schematic diagram of the change trend of the loss rate during the training process of the channel coding recognition model provided by an embodiment of this application.

[0024] Figure 3 is a schematic diagram of the change trend of the recognition accuracy rate during the training process of the channel coding recognition model provided by an embodiment of this application.

[0025] Figure 4 This is a schematic diagram of the composition structure of a lightweight channel coding blind recognition system based on multi-modal feature fusion provided by an embodiment of the present application. Detailed implementation manners

[0026] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application 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 application and are not used to limit the present application.

[0027] The terms "first" and "second" in the description and claims of this article are used to distinguish different objects, rather than to describe a specific order of the objects. For example, the first codeword and the second codeword are used to distinguish different codewords, rather than to describe the specific order of the codewords.

[0028] In the embodiments of the present application, words such as "exemplary" or "for example" are used to represent examples, illustrations or explanations. Any embodiment or design solution described as "exemplary" or "for example" in the embodiments of the present application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Exactly speaking, using words such as "exemplary" or "for example" aims to present relevant concepts in a specific manner.

[0029] In the description of the embodiments of the present application, unless otherwise specified, the meaning of "a plurality of" refers to two or more. For example, a plurality of codewords refers to two or more codewords, etc.

[0030] First, the technical terms involved in the embodiments of the present application will be introduced.

[0031] LDPC code (Low-Density Parity-Check Code) is a linear block error-correcting code with excellent performance, having an error-correcting ability close to the Shannon limit and an efficient decoding process.

[0032] Polar code (Polar code) is a linear block code based on the channel polarization phenomenon. Through the channel polarization process, the Polar code can synthesize multiple parallel independent channels into some perfect channels with a capacity close to 1 and pure noise channels with a capacity close to 0, and select to directly transmit information on the perfect channels to approach the channel capacity.

[0033] Turbo code (Turbo Code) is a parallel concatenated convolutional code, and its core idea is to improve the randomness and error-correcting ability of coding by connecting multiple convolutional codes in parallel and inserting an interleaver between them.

[0034] The Hankel matrix is a special form of matrix, characterized in that the elements on each diagonal are the same.

[0035] The embodiments of the present application will be described below with reference to the accompanying drawings in the embodiments of the present application.

[0036] The embodiments of the present application disclose a lightweight channel coding blind recognition method based on multi-modal feature fusion, as Figure 1 shown, which specifically includes the following steps: Step 1: Collect signals.

[0037] Step 2: Detect the signal-to-noise ratio of the collected signals:

[0038] Wherein, is the signal-to-noise ratio of the collected signals; is the voltage value measured at the output end of the receiver when there is a signal input; is the noise voltage value measured at the output end of the receiver when there is no signal input.

[0039] Step 3: Adaptively select signal statistical features or signal depth features according to the signal-to-noise ratio for signal blind recognition: Step 31: Compare the magnitude relationship between the signal-to-noise ratio and the signal-to-noise ratio threshold. If the signal-to-noise ratio is greater than the signal-to-noise ratio threshold, go to Step 32; otherwise, go to Step 33.

[0040] Step 32: Based on the theory of mathematical statistics, perform LDPC code recognition, Polar code recognition, Turbo code recognition, convolutional code recognition, and linear block code recognition on the collected signals respectively: Step 321: Perform LDPC code recognition on the collected signals: Pre-construct a candidate parity-check matrix of the LDPC code ; Derive the parity-check matrix and the average posterior log-likelihood ratio of the collected signal :

[0041] Wherein, is the candidate parity-check matrix, is the collected signal, is the th bit in the collected signal is the total number of bits in the collected signal is and posterior log-likelihood ratio.

[0042] If the average a posteriori log-likelihood ratio of the parity-check matrix and the acquired signal is higher than a preset first threshold, it is determined that the acquired signal is an LDPC code, and the LDPC code parameters are identified based on the candidate parity-check matrix.

[0043] Taking the Gaussian channel as an example:

[0044] is the natural exponential function, is the power of the channel noise, is the th acquired signal, is the actual value of the symbol; represents proportionality.

[0045] Step 322: Identify the Polar code for the acquired signal: The code length of the Polar code is usually a power of 2. Perform a sliding window intercept analysis on the acquired signal. If it is a Polar code, the code length corresponding to the peak of the autocorrelation of the acquired signal satisfies .

[0046] Use a sliding window to intercept the acquired signal to obtain multiple groups of codewords; Construct a matrix from multiple groups of codewords, and determine whether holds. If so, it is determined that the acquired signal is a Polar code, and the Polar code parameters are identified; where represents that the matrix is a row column complex matrix, is the number of codewords, is the candidate code length, represents finding the rank of the matrix .

[0047] Step 323: Identify the Turbo code for the acquired signal: Construct a Hankel matrix based on the acquired signal;

[0048] The Hankel matrix is a special structure matrix, and its definition requires that all elements on the anti-diagonal are equal, that is, it satisfies the relationship:

[0049] Assume that the sequence of the acquired signal is , determine the number of rows and the number of columns and meet the conditions and fill it according to the following rules:

[0050]

[0051]

[0052] Traverse different candidate values of code length If it can meet then it is determined that the collected signal is a Turbo code, and is the code length of the Turbo code.

[0053] Traverse different candidate values of code length If it can meet then it is determined that the collected signal is a Turbo code, and is the code length of the Turbo code.

[0054] Step 324: Identify the convolutional code for the collected signal: Use the cyclic spectral correlation function to detect the collected signal. If the cyclic frequency can be detected, it is determined that the collected signal is a convolutional code, and the convolutional code parameters are identified based on the cyclic frequency.

[0055] The cyclic spectral correlation function is specifically:

[0056] where is the collected signal, is the time delay, is the time length of the analysis window, is the base of the natural logarithm, is pi, is the cyclic frequency, is the imaginary symbol.

[0057] Step 325: Identify the linear block code for the collected signal: Pre-construct the candidate codewords of the linear block code; Calculate the code pattern likelihood of the collected signal using the following formula :

[0058] where is the th candidate codeword, is the th bit in the collected signal, is the code length of the collected signal; is the log-likelihood ratio; If there is a code pattern likelihood greater than a preset second threshold, it is determined that the collected signal is a linear block code, and based on the code pattern likelihood the linear block code parameters are identified.

[0059] Step 33: Input the collected signal into a channel coding recognition model for blind recognition of channel coding.

[0060] In the embodiments of the present application, the channel coding recognition model is trained by the following steps: Step 1: Generate a data set including LDPC code signals, Polar code signals, Turbo code signals, convolutional code signals, and linear block code signals through simulation; Step 11: Generate LDPC code signals through simulation, specifically: Derive the generator matrix of the LDPC code from the equation ; where is the matrix transpose; represents taking the modulus; is the parity-check matrix of the LDPC code, and a sparse matrix is randomly generated specifically by the random construction method where the number of non-zero elements in each row and each column is fixed, satisfying the low-density characteristic; Then generate the LDPC code from the equation ; where is the generated LDPC code, with a length of including information bits and parity bits, both are positive integers; is the original information bit sequence; Step 12: Generate Polar code signals through simulation, specifically: The generation formula of the Polar code is as follows:

[0061] where is the encoded Polar code sequence, is the input information bit sequence, is the sequence length, is the generator matrix of the Polar code, and its construction formula is as follows:

[0062] where is the basic kernel matrix; represents for times of Kronecker product recursive operation, from the formula Deduced ; is a bit-reversed permutation matrix used to adjust the order of the input sequence; Step 13: Generate a Turbo code signal through simulation, specifically: The Turbo code signal is ; where is the original information sequence; is the first parity-check sequence; is the second parity-check sequence; For the input original information sequence , the RSC encoder will output the original information as and the parity bit , the parity-check sequence The calculation process is shown in the following formula:

[0063] where, is the constraint length; represents the exclusive OR operation; is the binary number of the feedback polynomial, is the binary number of the feedforward polynomial; represents taking the modulus; is the th original information; Step 14: Generate a convolutional code signal through simulation, specifically: For a convolutional code with a code rate of and a constraint length of , its encoding process is defined by generating polynomials, and each generating polynomial corresponds to an output bit:

[0064] where, is the th output bit at time ; is the input bit at time ; is the generating polynomial coefficient, indicating whether the th output is connected to the value of the -stage shift register; represents taking the modulus; and are positive integers; Step 15: Generate a linear block code signal through simulation, specifically: Select the generating matrix of the linear block code; input the information vector and the matrix Perform matrix multiplication to obtain channel-coded data ; verify through the parity-check matrix ; where, ; where, and are positive integers, is the transpose of the matrix.

[0065] Step 2: Add high-frequency noise and low-frequency noise to each type of simulation signal in the dataset according to a ratio, and adaptively adjust the amplitude scaling factor of each type of simulation signal according to the signal power. The scaling formula is as follows;

[0066] is the scaled signal, is the original signal, is the mean value of, is the noise variance, is the variance of, is the square root of the signal-to-noise ratio.

[0067] Step 3: Intercept 1024 consecutive data for each type of simulation signal and divide them into a training set and a test set according to a ratio of 8:2; Step 4: Construct a channel coding recognition model based on a convolutional neural network and set the model training parameters as follows: Training batch: 200; Batch size: 64; Learning rate: 0.01; Loss function: Cross-entropy loss function; Optimizer: Stochastic gradient descent.

[0068] Input all the training sets into the channel coding recognition model for model training. During the training process, as Figure 2 shown, after 200 iterations, the loss value converges, and the model training is completed. As Figure 3 shown, during the training process, after 200 iterations, the recognition accuracy of the model approaches the stable optimum.

[0069] Finally, input all the test sets into the model after training convergence for testing, and finally obtain the trained channel coding recognition model.

[0070] This application embodiment also discloses a lightweight channel coding blind recognition system based on multi-modal feature fusion, as Figure 4 shown, specifically including: The signal-to-noise ratio detection module is used to detect the signal-to-noise ratio of the acquired signal. If the signal-to-noise ratio is less than the signal-to-noise ratio threshold, the signal statistical feature recognition module is called; otherwise, the model recognition module is called. The signal statistical feature recognition module is used to perform LDPC code recognition, Polar code recognition, Turbo code recognition, convolutional code recognition, and linear block code recognition on the acquired signal based on signal statistical features. The model recognition module is used to input the acquired signal into a channel coding recognition model for blind recognition of channel coding. The model training module is used to generate a data set including LDPC code signals, Polar code signals, Turbo code signals, convolutional code signals, and linear block code signals through simulation; add noise to each type of simulated signal in the data set, and adaptively adjust the amplitude scaling factor of each type of simulated signal according to the signal power; construct a channel coding recognition model based on a convolutional neural network, and use the data set to train the channel coding recognition model.

[0071] It should be understood that the above system is used to execute the method in the above embodiment. The corresponding program modules in the system have similar implementation principles and technical effects to those described in the above method. The working process of this system can refer to the corresponding process in the above method and will not be elaborated here.

[0072] Based on the method in the above embodiment, an electronic device is provided in an embodiment of the present application. The electronic device includes: a processor (Processor), a communication interface (Communications Interface), a memory (Memory), and a communication bus. Among them, the processor, the communication interface, and the memory complete communication with each other through the communication bus. The processor can call the logical instructions in the memory to execute the method in the above embodiment.

[0073] In addition, when the logical instructions in the above memory are implemented in the form of a software functional unit and sold or used as an independent product, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application.

[0074] Based on the method in the above embodiment, a computer-readable storage medium is provided in an embodiment of the present application. The computer-readable storage medium stores a computer program. When the computer program runs on a processor, the processor is caused to execute the method in the above embodiment.

[0075] Based on the method in the above embodiments, an embodiment of the present application provides a computer program product. When the computer program product runs on a processor, it causes the processor to execute the method in the above embodiments.

[0076] It can be understood that the processor in the embodiment of the present application may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. The general-purpose processor may be a microprocessor or any conventional processor.

[0077] The method steps in the embodiment of the present application may be implemented in a hardware manner or by a processor executing software instructions. The software instructions may be composed of corresponding software modules, and the software modules may be stored in a random access memory (RAM), flash memory, read-only memory (ROM), programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), registers, hard disks, removable hard disks, CD-ROMs, or any other form of storage medium well known in the art. An exemplary storage medium is coupled to the processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium may also be a component of the processor. The processor and the storage medium may be located in an ASIC.

[0078] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted through the computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more integrated available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)), etc.

[0079] It can be understood that the various numerical numbers involved in the embodiments of the present application are only for the convenience of description and are not used to limit the scope of the embodiments of the present application.

[0080] Those skilled in the art can easily understand that the above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A lightweight channel coding blind recognition method based on multi-modal feature fusion, characterized in that, Including: Detect the signal-to-noise ratio of the acquired signal. If the signal-to-noise ratio is less than the signal-to-noise ratio threshold, perform LDPC code recognition, Polar code recognition, Turbo code recognition, convolutional code recognition, and linear block code recognition on the acquired signal respectively based on signal statistical characteristics; otherwise, input the acquired signal into a channel coding recognition model for blind recognition of channel coding. The channel coding recognition model is trained by the following steps: Generate a data set including LDPC code signals, Polar code signals, Turbo code signals, convolutional code signals, and linear block code signals through simulation; Add noise to each type of simulated signal in the data set, and adaptively adjust the amplitude scaling coefficient of each type of simulated signal according to the signal power; Construct a channel coding recognition model based on a convolutional neural network, and use the data set to train the channel coding recognition model.

2. The lightweight channel coding blind recognition method according to claim 1, wherein Perform LDPC code recognition on the acquired signal, specifically: Pre-construct a candidate parity-check matrix of the LDPC code. If the average posterior log-likelihood ratio of the parity-check matrix and the acquired signal is higher than a preset first threshold, determine that the acquired signal is an LDPC code, and identify the LDPC code parameters based on the candidate parity-check matrix.

3. The lightweight channel coding blind recognition method according to claim 2, characterized in that, The average posterior log-likelihood ratio of the parity-check matrix and the acquired signal is specifically: Among them, is a candidate check matrix, is the acquired signal, is the acquired signal the -th bit in, is the acquired signal the total number of bits in; is and the posterior log-likelihood ratio of.

4. The lightweight channel coding blind recognition method according to claim 1, characterized in that Perform Polar code recognition on the acquired signal, specifically: Use a sliding window to intercept the acquired signal to obtain multiple groups of codewords. Construct a matrix from multiple groups of codewords , and determine Whether it holds. If so, determine that the collected signal is a Polar code and identify the Polar code parameters; where represents the matrix is a row column complex matrix, is the number of codewords, is the candidate code length, represents finding the rank of the matrix .

5. The lightweight channel coding blind recognition method according to claim 1, characterized in that Perform Turbo code recognition on the acquired signal, specifically: Construct a Hankel matrix based on the collected signals ; Traverse different candidate code lengths , if it can satisfy , then it is determined that the collected signal is a Turbo code, and is the code length of the Turbo code.

6. The lightweight channel coding blind recognition method according to claim 1, characterized in that Perform convolutional code recognition on the acquired signal, specifically: Use a cyclic spectral correlation function to detect the acquired signal. If a cyclic frequency can be detected, determine that the acquired signal is a convolutional code, and identify the convolutional code parameters based on the cyclic frequency.

7. The lightweight channel coding blind recognition method according to claim 6, wherein Using a cyclic spectral correlation function to detect the acquired signal is specifically: Among them, is the acquired signal, is the time delay, is the time length of the analysis window, is the base of the natural logarithm, is the pi, is the cyclic frequency, is the imaginary symbol.

8. The lightweight channel coding blind recognition method according to claim 1, characterized in that Perform linear block code recognition on the acquired signal, specifically: Pre-construct candidate codewords of the linear block code. Calculate the pattern likelihood of the acquired signal using the following formula :[[]]END]] Among them, is the th candidate codeword, is the th bit in the acquired signal, is the code length of the acquired signal; If there exists a code pattern likelihood greater than a preset second threshold, it is determined that the acquired signal is a linear block code, and based on the code pattern likelihood the linear block code parameters are identified.

9. The lightweight channel coding blind recognition method according to claim 1, wherein And adaptively adjust the amplitude scaling coefficient of each type of simulated signal according to the signal power, specifically: is the scaled signal, is the original signal, is the mean value of, is the noise variance, is the variance of, is the square root of the signal-to-noise ratio.

10. A lightweight channel coding blind recognition system based on multi-modal feature fusion, characterized in that, Including: A signal-to-noise ratio detection module, used to detect the signal-to-noise ratio of the acquired signal. If the signal-to-noise ratio is less than the signal-to-noise ratio threshold, call the signal statistical feature recognition module, otherwise call the model recognition module. A signal statistical feature recognition module, used to perform LDPC code recognition, Polar code recognition, Turbo code recognition, convolutional code recognition, and linear block code recognition on the acquired signal respectively based on signal statistical characteristics. A model recognition module, used to input the acquired signal into a channel coding recognition model for blind recognition of channel coding. A model training module, used to generate a data set including LDPC code signals, Polar code signals, Turbo code signals, convolutional code signals, and linear block code signals through simulation; Add noise to each type of simulated signal in the data set, and adaptively adjust the amplitude scaling coefficient of each type of simulated signal according to the signal power; Construct a channel coding recognition model based on a convolutional neural network, and use the data set to train the channel coding recognition model.

Images