Radiation source individual identification method, device, electronic device and storage medium

Abstract

The present application relates to a method, apparatus, electronic device, and storage medium for identifying individual radiation sources. The method includes the following steps: acquiring an electromagnetic signal to be identified; preprocessing the electromagnetic signal to be identified to obtain target input data; inputting the target input data into a trained radiation source information encoder for feature extraction to obtain target radiation source information features; the radiation source information encoder is obtained through generative adversarial training for feature decoupling; feature decoupling is performed to decouple radiation source information features inherent to the radiation source itself from signal information features associated with the electromagnetic signal itself in the original collected electromagnetic signal; and performing radiation source individual identification on the target radiation source information features to obtain radiation source individual identification results. The method can decouple radiation source information features from signal information features in the electromagnetic signal, thereby improving the accuracy and robustness of radiation source individual identification.

Description

Technical Field

[0001] The present application relates to the field of electromagnetic signal processing, and in particular to a method, device, electronic device and storage medium for identifying individual radiation sources. Background Art

[0002] Individual radiator identification is a technology that extracts characteristic information that characterizes the identity of the signal-transmitting device, thereby identifying individual devices. Traditional methods for individual radiator identification typically manually extract radiator characteristics during transient states (e.g., unstable signal states caused by powering on and off, or switching between operating states). These characteristics include instantaneous frequency, instantaneous phase, instantaneous amplitude, signal envelope, and transient signal spectrum characteristics; or radiator signal characteristics during stable operating states, such as reference frequency, modulation constellation, and signal preamble. Feature engineering and identification models are then used to achieve individual identification. With the increasing complexity and sophistication of electronic devices and the continuous evolution of modulation technologies, the electromagnetic environment is becoming increasingly complex, resulting in a greater diversity of electromagnetic signals. Consequently, these traditional methods are increasingly unable to meet the accuracy requirements for individual radiator identification.

[0003] As a result, artificial intelligence technology, particularly neural networks, has begun to be applied to the identification of individual radiation sources. Currently, algorithms using neural networks for the task of identifying individual radiation sources typically consist of a feature extraction module and a classifier module. Using an end-to-end training framework, these algorithms take received electromagnetic signals as input and output individual identification results. These algorithms rely on training data and may achieve good results for test data similar to the training data, but are less effective for classifying and identifying unknown radiation source data. Furthermore, because the model closely matches the training data, even for radiation sources similar to the training data, test results can be significantly affected by variations in the content and modulation of the emitted signals. Therefore, current related technologies still face challenges in identifying individual radiation sources, with low accuracy and robustness that need to be addressed. Summary of the Invention

[0004] In this embodiment, a method, device, electronic device, and storage medium for identifying individual radiation sources are provided to solve the problem of low recognition accuracy and robustness in the related art of identifying individual radiation sources.

[0005] In a first aspect, this embodiment provides a method for identifying an individual radiation source, including:

[0006] Acquiring an electromagnetic signal to be identified;

[0007] Preprocessing the electromagnetic signal to be identified to obtain target input data;

[0008] Inputting the target input data into a trained radiation source information encoder for feature extraction to obtain target radiation source information features; the radiation source information encoder is obtained through generative adversarial training for feature decoupling; the feature decoupling is to decouple the radiation source information features inherent in the original collected electromagnetic signal and associated with the radiation source itself from the signal information features associated with the electromagnetic signal itself;

[0009] Perform radiation source individual identification on the target radiation source information feature to obtain a radiation source individual identification result.

[0010] In some embodiments, the radiation source information encoder is used as a component of a pre-built radiation source individual identification system to perform the generative adversarial training; the radiation source individual identification system further includes: a data preprocessor, a signal information encoder, a signal generator, a radiation source identifier, a signal discriminator, and a decoupling discriminator; wherein during the generative adversarial training:

[0011] The data preprocessor is used to perform data preprocessing tasks on the received signal;

[0012] The signal information encoder is used to perform a signal information feature extraction task;

[0013] The radiation source information encoder is used to perform the radiation source information feature extraction task;

[0014] The signal generator is used to perform a signal reconstruction task according to the signal information characteristics and the radiation source information characteristics;

[0015] The signal discriminator is used to determine whether the reconstructed signal generated by the signal generator is true;

[0016] The radiation source identifier is used to perform a radiation source individual identification task based on radiation source information features;

[0017] The decoupling discriminator is used to determine the category to which the radiation source information feature belongs.

[0018] In some embodiments, preprocessing the electromagnetic signal to be identified to obtain target input data includes:

[0019] Based on the data preprocessor, the electromagnetic signal to be identified is matrix processed to obtain the target input data.

[0020] In some embodiments, the training process of generating adversarial training includes:

[0021] Based on unlabeled data in the sample data set, adversarial training is performed on the feature extraction capability of the signal information encoder, the feature extraction capability of the radiation source information encoder, and the signal reconstruction capability of the signal generator;

[0022] After completing the training based on the unlabeled data, the feature extraction capability of the radiation source information encoder is supervisedly trained based on the labeled data in the sample data set, so that the radiation source information features extracted by the radiation source information encoder meet the preset radiation source individual identification conditions.

[0023] In some embodiments, adversarial training is performed on the feature extraction capability of the signal information encoder, the feature extraction capability of the radiation source information encoder, and the signal reconstruction capability of the signal generator based on unlabeled data in the sample dataset, including:

[0024] Based on the unlabeled data, the components of the radiation source individual recognition system are iteratively trained according to the first information decoupling loss, the first information reconstruction loss, and the first generative adversarial loss until a preset first training termination condition is reached; wherein,

[0025] The first information decoupling loss is determined based on output data of the radiation source information encoder;

[0026] The first information reconstruction loss is determined based on output data of the signal generator;

[0027] The first generative adversarial loss is determined based on output data of the radiation source information encoder.

[0028] In some embodiments, supervised training is performed on the feature extraction capability of the radiation source information encoder based on the labeled data in the sample data set, so that the radiation source information features extracted by the radiation source information encoder meet the preset radiation source individual identification conditions, including:

[0029] Based on the labeled data, iteratively training the components of the radiation source individual identification system according to a second information decoupling loss, a second information reconstruction loss, a second generative adversarial loss, and a radiation source classification loss until a preset second training termination condition is reached;

[0030] The second information decoupling loss is determined based on output data of the radiation source information encoder;

[0031] The second information reconstruction loss is determined based on output data of the signal generator;

[0032] The second generative adversarial loss is determined based on output data of the radiation source information encoder;

[0033] The radiation source classification loss is determined based on output data of the radiation source identifier.

[0034] In some embodiments, the method further comprises:

[0035] Collect sample electromagnetic signals using several radio stations as sample radiation sources;

[0036] splicing the in-phase component and the orthogonal component in the sample electromagnetic signal to obtain a spliced ​​signal; segmenting the spliced ​​signal to obtain a segmented signal; and performing convolution processing on the segmented signal to obtain a sample electromagnetic signal matrix;

[0037] The sample electromagnetic signal matrix is ​​used as input for the generative adversarial training.

[0038] In a second aspect, this embodiment provides a device for identifying individual radiation sources, including: an acquisition module, a preprocessing module, a feature extraction module, and an identification module; wherein:

[0039] The acquisition module is used to acquire the electromagnetic signal to be identified;

[0040] The preprocessing module is used to preprocess the electromagnetic signal to be identified to obtain target input data;

[0041] The feature extraction module is configured to input the target input data into a trained radiation source information encoder for feature extraction to obtain target radiation source information features; the radiation source information encoder is obtained through generative adversarial training for feature decoupling; the feature decoupling is the decoupling of radiation source information features inherent to the radiation source itself in the originally collected electromagnetic signal from signal information features associated with the electromagnetic signal itself;

[0042] The identification module is used to perform radiation source individual identification on the target radiation source information characteristics to obtain a radiation source individual identification result.

[0043] In a third aspect, an electronic device is provided in this embodiment, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the radiation source individual identification method described in the first aspect is implemented.

[0044] In a fourth aspect, a storage medium is provided in this embodiment, on which a computer program is stored. When the program is executed by a processor, the method for individual identification of radiation sources described in the first aspect is implemented.

[0045] Compared to related technologies, this embodiment provides a method, apparatus, electronic device, and storage medium for identifying individual radiation sources. The method first acquires an electromagnetic signal to be identified; then, preprocesses the electromagnetic signal to be identified to obtain target input data; then, inputs the target input data into a trained radiation source information encoder for feature extraction to obtain target radiation source information features. The radiation source information encoder is obtained through generative adversarial training for feature decoupling. Feature decoupling involves decoupling the radiation source information features inherent to the original acquired electromagnetic signal from the signal information features associated with the electromagnetic signal itself. Finally, the target radiation source information features are used to perform radiation source individual identification, obtaining a radiation source individual identification result. This method decouples the radiation source information features from the signal information features in the electromagnetic signal, enabling radiation source individual identification based on the radiation source information features of the radiation source itself, thereby improving the accuracy and robustness of radiation source individual identification.

[0046] The details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0048] Figure 1 This is a hardware structure block diagram of a terminal of the radiation source individual identification method according to an embodiment of the present application;

[0049] Figure 2 is a flow chart of a method for identifying individual radiation sources according to an embodiment of the present application;

[0050] Figure 3 This is a data flow diagram of feature extraction in an embodiment of the present application;

[0051] Figure 4 This is a schematic diagram of a data preprocessing process according to an embodiment of the present application;

[0052] Figure 5 This is a schematic diagram of data flow for generative adversarial training in some embodiments of the present application;

[0053] Figure 6 is a flow chart of a method for identifying individual radiation sources in some embodiments of the present application;

[0054] Figure 7 This is a structural block diagram of the radiation source individual identification device according to an embodiment of the present application. DETAILED DESCRIPTION

[0055] In order to more clearly understand the purpose, technical solutions and advantages of the present application, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments.

[0056] Unless otherwise defined, technical or scientific terms used in this application shall have the ordinary meanings as understood by persons of ordinary skill in the art to which this application belongs. The terms "a," "an," "the," "these," and similar expressions in this application do not denote limitations on quantity and may be singular or plural. The terms "comprise," "include," "have," and any variations thereof, as used in this application, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or device comprising a series of steps or modules (units) is not limited to the listed steps or modules (units) but may include unlisted steps or modules (units) or other steps or modules (units) inherent to the process, method, product, or device. The terms "connected," "connected," "coupled," and similar expressions used in this application are not limited to physical or mechanical connections but may include electrical connections, whether direct or indirect. As used in this application, "plurality" means two or more. "And / or" describes an association between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can mean: A exists alone; A and B exist simultaneously; or B exists alone. Generally, the character " / " indicates that the objects in the preceding and following relationship are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.

[0057] The method embodiment provided in this embodiment can be executed in a terminal, a computer or a similar computing device. For example, running on a terminal, Figure 1 FIG. 1 is a block diagram of the hardware structure of the terminal of the radiation source individual identification method of this embodiment. Figure 1 As shown, the terminal may include one or more ( Figure 1 The processor 102 (only one is shown) and a memory 104 for storing data, wherein the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA. The terminal may also include a transmission device 106 for communication functions and an input / output device 108. It will be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above terminal. Figure 1 More or fewer components than shown, or with Figure 1 Different configurations shown.

[0058] The memory 104 can be used to store computer programs, such as application software programs and modules, such as the computer program corresponding to the radiation source individual identification method in this embodiment. The processor 102 executes the computer program stored in the memory 104 to execute various functional applications and data processing, thereby implementing the above-mentioned method. The memory 104 may include high-speed random access memory (RAM) and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include memory located remotely from the processor 102, and such remote memory may be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0059] Transmission device 106 is used to receive or transmit data via a network. This network may include a wireless network provided by the terminal's communications provider. In one embodiment, transmission device 106 includes a network interface controller (NIC), which can be connected to other network devices via a base station to enable communication with the Internet. In another embodiment, transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0060] In this embodiment, a method for identifying individual radiation sources is provided. Figure 2 is a flow chart of the radiation source individual identification method of this embodiment, as shown in FIG. Figure 2 As shown, the process includes the following steps:

[0061] Step S210: Acquire the electromagnetic signal to be identified.

[0062] The electromagnetic signal to be identified can be a currently received electromagnetic signal emitted by a radiation source of unknown device identity. Feature extraction and individual radiation source identification are performed on the electromagnetic signal to determine the device identity of the corresponding radiation source. This can be used, for example, in radar model identification, wireless device authentication, and vehicle identity authentication scenarios.

[0063] Step S220: pre-process the electromagnetic signal to be identified to obtain target input data.

[0064] The collected electromagnetic signal to be identified can be adjusted in data format to form target input data compatible with feature extraction, meeting the input requirements for individual radiation source identification. In some embodiments, the collected electromagnetic signal to be identified is a quadrature carrier modulated (IQ modulated) electromagnetic signal. The in-phase component (I component) and the quadrature component (Q component) of the IQ modulated electromagnetic signal can be concatenated and then segmented into signal segments. The signal segments are then flattened to a predetermined size through convolution to form a signal matrix as the target input data.

[0065] In step S230, the target input data is input into the trained radiation source information encoder for feature extraction to obtain target radiation source information features; the radiation source information encoder is obtained through generative adversarial training for feature decoupling; feature decoupling is to decouple the radiation source information features inherent in the original collected electromagnetic signal associated with the radiation source itself from the signal information features associated with the electromagnetic signal itself.

[0066] The radiation source information encoder can be specifically obtained through generative adversarial training. In the generative adversarial training process, feature decoupling is mainly performed on the radiation source information features and signal information features in the original collected electromagnetic signal, so that the trained radiation source information encoder can accurately extract the radiation source information features from the target input data. Among them, the radiation source information feature is the information feature related to the radiation source in the electromagnetic signal, which can express the characteristics of the radiation source itself and focus on the inherent information of the individual radiation source. In this embodiment, the radiation source information feature is a feature automatically learned using deep learning. Therefore, the radiation source information feature extracted here is not required to be consistent with the physical characteristics of the radiation source or to be directly understood and interpreted. The signal information feature can be an information feature related to the electromagnetic signal, such as a signal feature related to changes in parameters such as signal modulation mode and transmission frequency.

[0067] Generative adversarial training can be performed based on a pre-built generative adversarial network structure to decouple the emitter information features from the signal information features in the received signal. During the training process, electromagnetic signal samples can be collected to construct a sample dataset. Subsequently, a pre-defined iterative optimization method is used to iteratively train each component of the generative adversarial network structure (including the emitter information encoder), updating the model parameters involved in each component. Finally, the trained emitter information encoder is used in the actual emitter information feature extraction stage of the individual emitter identification process.

[0068] The aforementioned radiation source information encoder can be a feature extraction architecture implemented using a convolutional neural network. For example, a residual network (ResNet) can be used as the backbone of the radiation source information encoder to adapt to multi-scale data input, followed by a feature concatenation module and a feature fusion module. It is understood that those skilled in the art may also select other architectures for the radiation source information encoder based on the needs of actual application scenarios, and this embodiment does not impose any limitations thereto.

[0069] In this step, a generative adversarial mechanism is used to decouple the characteristics of the radiation source itself from the signal characteristics of the electromagnetic signal, thereby successfully extracting the radiation source information characteristics from the electromagnetic signal to perform individual radiation source task identification.

[0070] Step S240 , performing radiation source individual identification on the target radiation source information features to obtain a radiation source individual identification result.

[0071] A pre-built recognition model can be used to identify the target radiation source information features and determine the individual radiation source identification result. Specifically, the target radiation source information features can be input into a radiation source identifier trained synchronously with the radiation source information encoder to obtain the aforementioned individual radiation source identification result. The radiation source identifier can utilize two stacked multi-layer perceptrons to implement the recognition function. It is understood that those skilled in the art may also use other structures with feature processing and recognition capabilities to perform individual radiation source identification.

[0072] Compared with related technologies, this embodiment enhances the focus on the inherent characteristics of the radiation source itself, utilizes a generative adversarial mechanism to decouple the individual inherent characteristics of the radiation source from the signal characteristics in the received signal, and utilizes the inherent characteristics to perform identification tasks. It is universally applicable to different forms of electromagnetic signals and eliminates interference with the identification task caused by signal content, modulation methods, etc. For example, it is applicable to electromagnetic signals of different frequencies and modulation methods. It can effectively extract high-quality feature representations, thereby improving the accuracy and robustness of individual radiation source identification.

[0073] Therefore, this embodiment, through the above steps S210 to S240, first obtains the electromagnetic signal to be identified; then, preprocesses the electromagnetic signal to be identified to obtain target input data; next, inputs the target input data into a trained radiation source information encoder for feature extraction to obtain target radiation source information features; the radiation source information encoder is obtained through generative adversarial training for feature decoupling; feature decoupling is the decoupling of radiation source information features inherent to the radiation source itself in the original collected electromagnetic signal from signal information features associated with the electromagnetic signal itself; finally, individual radiation source identification is performed on the target radiation source information features to obtain a radiation source individual identification result. This can achieve the decoupling of radiation source information features and signal information features in the electromagnetic signal, and realize individual radiation source identification based on the radiation source information features of the radiation source itself, thereby improving the accuracy and robustness of individual radiation source identification.

[0074] In one embodiment, a radiation source information encoder is used as a component of a pre-built radiation source individual identification system for generative adversarial training; the radiation source individual identification system further comprises: a data preprocessor, a signal information encoder, a signal generator, a radiation source identifier, a signal discriminator, and a decoupling discriminator; in the generative adversarial training process:

[0075] The data preprocessor is used to perform data preprocessing tasks on the received signal; the signal information encoder is used to perform signal information feature extraction tasks; the radiation source information encoder is used to perform radiation source information feature extraction tasks; the signal generator is used to perform signal reconstruction tasks based on signal information features and radiation source information features; the signal discriminator is used to determine whether the reconstructed signal generated by the signal generator is true; the radiation source identifier is used to perform radiation source individual identification tasks based on radiation source information features; and the decoupling discriminator is used to determine the category to which the radiation source information features belong.

[0076] In this embodiment, a radiation source individual identification system is implemented based on the structure of a generative adversarial network. The radiation source individual identification system may include a data preprocessor, a signal information encoder, a radiation source information encoder, a signal generator, a radiation source identifier, a signal discriminator, and a decoupling discriminator. The data preprocessor is used to preprocess the signal data input to the radiation source individual identification system so that the data format and size of the signal data meet the input requirements of the radiation source individual identification system. The signal information encoder, as a feature extraction structure, is used to extract signal-related information features of the electromagnetic signal, such as signal features related to changes in parameters such as signal modulation mode and transmission frequency. The radiation source information encoder is used to obtain radiation source information features related to the electromagnetic signal and the inherent information of the radiation source itself. The signal generator is used to combine the signal information features output by the signal information encoder with the radiation source information features output by the radiation source information encoder to perform signal reconstruction to obtain a reconstructed signal. The radiation source identifier is used to use the radiation source information features output by the radiation source information encoder to perform radiation source individual identification. The signal discriminator is used to determine whether the reconstructed signal generated by the signal generator is sufficiently realistic. The decoupling discriminator is used to determine the category to which the feature extracted by the radiation source information encoder belongs, for example, to determine whether the feature extracted by the radiation source information encoder is a radiation source information feature or a signal information feature.

[0077] The signal information encoder and the radiation source information encoder can adopt the same structure and have a similar feature extraction process. For example, the signal information encoder and the radiation source information encoder can both include a backbone network, a feature splicing module, and a feature fusion module. The backbone network can be ResNet. Figure 3 This is a data flow diagram for feature extraction in this embodiment. Figure 3 The feature extraction process can adapt the feature extraction of the signal information encoder and the feature extraction of the radiation source information encoder. Figure 3 As shown, the signal after the data preprocessor is in matrix form ( Figure 3 Matrix 1, Matrix 2, Matrix 3, and Matrix 4 in the image are input to the backbone network. After processing, the backbone network outputs the corresponding features (Feature 1, Feature 2, Feature 3, and Feature 4) for each input matrix. Matrices 1 through 4 represent different scales of the same signal segment to capture multi-scale information. For example, the signal is divided into segment a and then pieced together to form Matrix 1, the signal is divided into segment b and then pieced together to form Matrix 2, and so on. Different signal segmentation and splicing can result in different coverage within the receptive field during convolution calculations, thereby capturing the connections between different time-series segments of the signal.

[0078] Then, the features output by the backbone network are input into the feature splicing module for feature splicing. Direct splicing can be used to obtain splicing features. The features output by the splicing feature module are then passed through the feature fusion module of the fully connected layer to fuse multi-scale feature information and obtain the final output signal information features or radiation source information features.

[0079] For example, the signal generator may employ a deconvolutional neural network with a residual connection structure. The radiation source identifier may employ two stacked multi-layer perceptrons. The signal discriminator and decoupling discriminator may employ a convolutional network with a fully connected layer.

[0080] It is understandable that those skilled in the art may also select other network structures as the above-mentioned signal generator, signal discriminator, signal discriminator, decoupling discriminator and radiation source identifier according to the requirements of the actual application scenario, and this embodiment does not make specific limitations on this.

[0081] In one embodiment, preprocessing the electromagnetic signal to be identified to obtain target input data may specifically include: performing matrix processing on the electromagnetic signal to be identified based on a data preprocessor to obtain the target input data. The electromagnetic signal to be identified can be concatenated, segmented, and convolved to form a signal matrix as the target input data, thereby adapting it to the input of a subsequent signal information encoder and a radiation source information encoder.

[0082] Additionally, in one embodiment, the training process of the above-mentioned generative adversarial training may include:

[0083] Based on the unlabeled data in the sample data set, adversarial training is performed on the feature extraction ability of the signal information encoder, the feature extraction ability of the radiation source information encoder, and the signal reconstruction ability of the signal generator; after completing the training based on the unlabeled data, supervised training is performed on the feature extraction ability of the radiation source information encoder based on the labeled data in the sample data set, so that the radiation source information features extracted by the radiation source information encoder meet the preset radiation source individual identification conditions.

[0084] In this embodiment, generative adversarial training is divided into two parts. The first part is pre-training based on unlabeled data. This training mainly utilizes the decoupling and recombination of data features, using a generative adversarial network to train the signal information encoder, emitter information encoder, and signal generator. The second part of the training is supervised fine-tuning based on the first part of the training using labeled data.

[0085] The first part of the training primarily utilizes a signal information encoder and a source information encoder to extract features from the input signal. The signal information features are then decoupled from the source information features. A decoupling discriminator is then used to determine the category of features extracted by the source information encoder (either source information features or signal information features). This constrains the generation of source information features to meet the requirements of feature decoupling while preserving the original signal information, enabling the signal generator to reconstruct the original signal. The signal discriminator then determines whether the reconstructed signal from the signal generator is sufficiently realistic (this can be determined by comparing its similarity to the real signal). This approach improves the performance of the signal generator, source information encoder, and signal information encoder through adversarial generation. Ultimately, training improves the feature extraction capabilities of the signal information encoder and source information encoder, resulting in decoupled signal information features and source information features.

[0086] For the second part of the training, the radiation source information features after decoupling the original signal need to meet the requirements for individual radiation source identification. For example, the accuracy of individual radiation source identification based on the extracted radiation source information features must meet the preset accuracy requirements. During the second part of the training process, the radiation source identifier can be used to perform the radiation source individual identification task. Therefore, the classification loss can be added to the loss constructed in the first part of the training to form the total loss of the second part, further constraining the radiation source information feature extraction, so that the radiation source information features extracted by the radiation source information encoder can ultimately be used for the radiation source individual identification task.

[0087] Existing algorithms that use neural network technology to identify individual emitters require human intervention during the collection, labeling, and processing of relevant data. This limits the size of the dataset, especially the labeled data, and thus the model's recognition performance. Furthermore, because the model is highly fitted to the training data, even for emitters similar to the training data, test results can be significantly affected by variations in the emitted signal content, modulation, and other factors.

[0088] During the first training phase, this embodiment fully utilizes unlabeled data, addressing the challenges of data labeling and limited model performance in related techniques, thereby enhancing the model's generalization capabilities. Consequently, compared to related techniques, this approach reduces reliance on labeled data and eliminates limitations on individual radiation source recognition performance. During the second training phase, supervised fine-tuning based on labeled data utilizes a small amount of labeled data to adaptively adjust model parameters, further enhancing model recognition accuracy.

[0089] In one embodiment, based on the unlabeled data in the sample data set, adversarial training is performed on the feature extraction capability of the signal information encoder, the feature extraction capability of the radiation source information encoder, and the signal reconstruction capability of the signal generator. Specifically, the training may include:

[0090] Based on unlabeled data, the components of the radiation source individual identification system are iteratively trained according to the first information decoupling loss, the first information reconstruction loss and the first generative adversarial loss until the preset first training termination condition is reached; wherein, the first information decoupling loss is determined based on the output data of the radiation source information encoder; the first information reconstruction loss is determined based on the output data of the signal generator; and the first generative adversarial loss is determined based on the output data of the radiation source information encoder.

[0091] In each iteration, the signal data pair processed by the data preprocessor in the unlabeled data is input into the signal information encoder for signal information feature extraction to obtain a first output feature pair, and is also input into the radiation source information encoder for radiation source information feature extraction to obtain a second output feature pair; the first output feature pair and the second output feature are cross-combined and input into the signal generator for signal reconstruction to obtain a first output signal pair; the first output signal pair is input into the signal information encoder for signal information feature extraction to obtain a third output feature pair, and is also input into the radiation source information encoder for radiation source information feature extraction to obtain a fourth output feature pair; the third output feature pair and the fourth output feature pair are input into the signal generator for signal reconstruction to obtain a second output signal pair. Finally, based on the radiation source information features in the first output feature pair and the second output feature pair, the first information decoupling loss and the generative adversarial loss are determined, and based on the second output signal pair and the signal pair processed by the data preprocessor, the signal reconstruction loss is determined.

[0092] The training process of the first part is as follows: First, considering that the first part does not involve the task of identifying individual radiation sources, the parameters of the radiation source identifier are frozen to reduce the amount of calculation and speed up the model inference. Then, training signal data pairs are randomly extracted from the training set of the sample data set and processed into matrix form by the data preprocessor, for example ,Will Input to the signal information encoder respectively Radiation source information encoder Perform feature extraction to obtain signal information features 、 (first output feature pair) and radiation source information features 、 (Second output feature pair); cross-combine two pairs of feature pairs and 、 Combination 、 Combine and input to the signal generator to obtain the reconstructed signal and (the first output signal pair), and then and Input signal information encoder Radiation source information encoder , get the signal information characteristics 、 (Third output feature pair) and radiation source information feature 、 (the fourth output feature pair); 、 and 、 Input signal generator to get reconstructed signal and (Second output signal pair).

[0093] By calculating each loss and gradient, back propagation and parameter update are completed, thus completing the first part of the training. In this embodiment, the loss of the first part of the training is The calculation of includes the first information decoupling loss, the first information reconstruction loss and the first generative adversarial loss.

[0094] The calculation method of the first information decoupling loss is:

[0095]

[0096] ;

[0097] The signal reconstruction loss is calculated as:

[0098] ;

[0099] The generative adversarial loss is calculated as:

[0100]

[0101] ;

[0102] Finally, the total loss of the first part is:

[0103] .

[0104] The “#” above indicates the connection of formula segments.

[0105] In the first part of the training, the signal information encoder is used to extract signal information features, and the radiation source information encoder is used to extract radiation source information features. By using a cross-combination method and combining it with a generative adversarial network, the validity of the features can be guaranteed, and the relevant loss function is used to constrain the above feature extraction during the training process, thereby ultimately achieving the purpose of feature decoupling. is the output of the decoupling discriminator, is the output of the signal discriminator, and the specific forms of the decoupling discriminator and the signal discriminator are not limited.

[0106] Additionally, in one embodiment, supervised training is performed on the feature extraction capability of the radiation source information encoder based on the labeled data in the sample data set, so that the radiation source information features extracted by the radiation source information encoder meet the preset radiation source individual identification conditions, which may specifically include:

[0107] Based on the labeled data, the components of the radiation source individual identification system are iteratively trained according to the second information decoupling loss, the second information reconstruction loss, the second generative adversarial loss and the radiation source classification loss until the preset second training termination condition is reached; the second information decoupling loss is determined based on the output data of the radiation source information encoder; the second information reconstruction loss is determined based on the output data of the signal generator; the second generative adversarial loss is determined based on the output data of the radiation source information encoder; and the radiation source classification loss is determined based on the output data of the radiation source identifier.

[0108] The second part of the training uses labeled data for supervised fine-tuning. In each iteration:

[0109] Randomly extract the training signal data from the training set and input it into the aforementioned radiation source individual recognition system, and then process it with the data preprocessor to obtain ; Signal Information Encoder Radiation source information encoder , get the signal information characteristics 、 and radiation source information characteristics 、 ; Cross-combine two pairs of features and 、 and 、 Input signal generator to get reconstructed signal and , and then and Input signal information encoder Radiation source information encoder , get the signal information characteristics 、 and radiation source information characteristics 、 ;Will 、 and 、 Input signal generator to get reconstructed signal and ; The radiation source information characteristics 、 、 、 Input radiation source identifier Get the classification judgment p1, p2 and 、 .

[0110] The calculation of the loss in the second part of training includes the second information decoupling loss, the second information reconstruction loss, the second generative adversarial loss, and the radiation source classification loss. The calculation process of the second information decoupling loss is:

[0111]

[0112] ;

[0113] The calculation process of the second signal reconstruction loss is:

[0114] ;

[0115] The calculation process of the second generation adversarial loss is:

[0116]

[0117] ;

[0118] The calculation process of radiation source classification loss is:

[0119] ;

[0120] Finally, the total loss of the second part of training is:

[0121] .

[0122] Additionally, in one embodiment, the above-mentioned radiation source individual identification method may further include:

[0123] Sample electromagnetic signals are collected using several radio stations as sample radiation sources; the in-phase components and orthogonal components in the sample electromagnetic signals are spliced ​​to obtain a spliced ​​signal; the spliced ​​signal is segmented to obtain a segmented signal; the segmented signal is convolved to obtain a sample electromagnetic signal matrix; and the sample electromagnetic signal matrix is ​​used as input for generative adversarial training.

[0124] When collecting sample electromagnetic signals, the acquisition conditions can be set to direct radiation or diffraction, and various parameters such as frequency band and modulation mode can be configured during acquisition. The collected signals can be IQ modulated electromagnetic signals. The sample dataset can be divided into a training set, a validation set, and a test set, for example, in a ratio of 3:1:1. If a total of 15,000 data points are collected, the training set, validation set, and test set will contain 9,000, 3,000, and 3,000 data points, respectively.

[0125] In the data preprocessor P, the acquired IQ modulated electromagnetic signal is preprocessed: for an IQ signal with a length of L, the signal 、 Components are spliced ​​to obtain a spliced ​​signal :

[0126] ;

[0127] in Quantity and The size of the components is (1, L). The signal after splicing The size is (2, L), then the signal Divide into [K1, K2, K3, K4] segments, each segment size is (2, N i ), satisfying the relationship:

[0128] ;

[0129] By convolution, the signal segments are flattened into the size (1, N i ), the signal fragments are reassembled into the input matrix As the input to the radiation source information encoder and the signal information encoder, this data preprocessing can mine the correlation between the I / Q components of the same time segment and between different time segments, and realize multi-scale data fusion.

[0130] Figure 4 FIG. 1 is a schematic diagram of a data preprocessing process of this embodiment. Figure 4 As shown, for the input signal a, by splicing the I component and Q component, a spliced ​​signal a is obtained, the size of which is (2, L). After that, the spliced ​​signal a is divided into several signal segments, such as Figure 4 K1 to K4 in the i ), i takes the value of 1, 2, 3, or 4. Then, the size of each signal segment is flattened, and the size of each signal is (1, N i ). Each signal fragment is combined into a matrix to obtain matrix 1, matrix 2, matrix 3, and matrix 4.

[0131] Figure 5 Schematic diagram of data flow for generating adversarial training in some embodiments, such as Figure 5 As shown, input signal 1 and input signal 2 are input into the data preprocessor as a pair of signals. After data preprocessing, the processed signal 1 is obtained ( Figure 5 The solid arrows in the data preprocessor point to the signal information encoder and the radiation source information encoder) and the processed signal 2 ( Figure 5 The data preprocessor in the figure points to the signal information encoder and the radiation source information encoder (dashed arrow). The processed signal 1 and the processed signal 2 are respectively input into the signal information encoder and the radiation source information encoder for feature extraction. The signal information of signal 1 (solid arrow from the upper signal information encoder pointing to the upper signal generator) and the signal information of signal 2 (dashed arrow from the upper signal information encoder pointing to the upper signal generator) output by the signal information encoder are cross-combined together with the radiation source information of signal 1 (solid arrow from the upper radiation source information encoder pointing to the upper signal generator and the upper decoupling discriminator) and the radiation source information of signal 2 (dashed arrow from the upper radiation source information encoder pointing to the upper signal generator and the upper decoupling discriminator) output by the radiation source information encoder. The cross-combination is then input into the signal generator for signal reconstruction, resulting in signal 1' (dashed arrow from the upper signal generator pointing to the lower signal information encoder and the lower radiation source information encoder) and signal 2' (solid arrow from the upper signal generator pointing to the lower signal information encoder and the lower radiation source information encoder). The radiation source information of signal 1 and signal 2 output by the radiation source information encoder are input into the decoupling discriminator to determine the category to which the feature belongs. Signal 1' and signal 2' are input into the signal discriminator to verify the authenticity of the signal. In addition, signal 1' and signal 2' are respectively input into the signal information encoder and the radiation source information encoder for feature extraction, and the obtained radiation source information of signal 1' (the dotted arrow of the lower-layer radiation source information encoder points to the lower-layer signal generator and the lower-layer decoupling discriminator) and the radiation source information of signal 2' (the solid arrow of the lower-layer radiation source information encoder points to the lower-layer signal generator and the lower-layer decoupling discriminator) are input into the decoupling discriminator to determine the category to which the feature belongs. Afterwards, the signal information of signal 1' (the dotted arrow of the lower-layer signal information encoder points to the lower-layer signal generator), the signal information of signal 2' (the solid arrow of the lower-layer signal information encoder points to the lower-layer signal generator), the radiation source information of signal 1', and the radiation source information of signal 2' are cross-combined and input into the signal generator for signal reconstruction to obtain the reconstructed signal , reconstruct the signal . Reconstruct the signal , reconstruct the signal The signal is input into the signal discriminator to verify its authenticity. In addition, based on the radiation source identifier, the radiation source information of signal 1, signal 2, signal 1', and signal 2' output by the radiation source information encoder are individually identified as radiation sources to obtain a radiation source identification result. Based on the radiation source identification result and the corresponding radiation source identification label, the accuracy of the identification result is verified.

[0132] Figure 6 is a flow chart of a method for identifying an individual radiation source in some embodiments, such as Figure 6 As shown, the radiation source individual identification method includes the following steps:

[0133] Step S601 , collecting samples to construct an electromagnetic signal dataset; wherein, electromagnetic signals are collected using a plurality of ultrashort wave broadcasting stations as radiation sources to construct the electromagnetic signal dataset.

[0134] Step S602: constructing a radiation source individual identification system; the structure of the radiation source individual identification system may refer to the above embodiment.

[0135] Step S603: Based on the generative adversarial mechanism, the radiation source individual recognition system based on feature decoupling is trained. The specific training process can be referred to the above embodiment and will not be described in detail here.

[0136] Step S604: After the training is completed, a data preprocessor is used to preprocess the electromagnetic signal to be identified to obtain target input data.

[0137] Step S605: input the target input data into the trained radiation source information encoder for feature extraction to obtain target radiation source information features.

[0138] Step S606: Perform radiation source individual identification on the target radiation source information characteristics to obtain a radiation source individual identification result.

[0139] This embodiment also provides a device for individually identifying radiation sources, which is used to implement the above-mentioned embodiments and preferred embodiments. Details already described will not be repeated here. The terms "module," "unit," "subunit," etc., used below, may refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented using software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0140] Figure 7 is a structural block diagram of the radiation source individual identification device 70 of this embodiment, as shown in FIG. Figure 7 As shown, the radiation source individual identification device 70 includes: an acquisition module 72, a pre-processing module 74, a feature extraction module 76 and an identification module 78; wherein:

[0141] An acquisition module 72 is used to acquire an electromagnetic signal to be identified;

[0142] The preprocessing module 74 is used to preprocess the electromagnetic signal to be identified to obtain target input data; the feature extraction module 76 is used to input the target input data into the trained radiation source information encoder for feature extraction to obtain the target radiation source information characteristics; the radiation source information encoder is obtained through generative adversarial training for feature decoupling; feature decoupling is to decouple the radiation source information characteristics inherent in the original collected electromagnetic signal associated with the radiation source itself from the signal information characteristics associated with the electromagnetic signal itself; the identification module 78 is used to perform radiation source individual identification on the target radiation source information characteristics to obtain the radiation source individual identification result.

[0143] It should be noted that the above modules can be functional modules or program modules, and can be implemented through software or hardware. For modules implemented through hardware, the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0144] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and optional implementation modes, which will not be described in detail in this embodiment.

[0145] This embodiment further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0146] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0147] Optionally, in this embodiment, the processor may be configured to execute the following steps through a computer program:

[0148] S1, obtaining the electromagnetic signal to be identified;

[0149] S2, preprocessing the electromagnetic signal to be identified to obtain target input data;

[0150] S3, inputting the target input data into a trained radiation source information encoder for feature extraction to obtain target radiation source information features; the radiation source information encoder is obtained through generative adversarial training for feature decoupling; feature decoupling is the process of decoupling the radiation source information features inherent in the original collected electromagnetic signal and associated with the radiation source itself from the signal information features associated with the electromagnetic signal itself;

[0151] S4, performing radiation source individual identification on the target radiation source information characteristics to obtain a radiation source individual identification result.

[0152] Accordingly, the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementation modes, and will not be repeated in this embodiment.

[0153] In addition, in conjunction with the radiation source individual identification method provided in the above embodiments, a storage medium may also be provided in this embodiment to implement the method. The storage medium stores a computer program; when the computer program is executed by a processor, any of the radiation source individual identification methods in the above embodiments is implemented.

[0154] It should be understood that the specific embodiments described herein are only used to explain this application and are not used to limit it. Based on the embodiments provided in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0155] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0156] Obviously, the accompanying drawings are merely examples or embodiments of the present application. A person skilled in the art can also apply the present application to other similar situations based on these drawings without inventive effort. Furthermore, it is understandable that, although the work involved in this development process may be complex and lengthy, certain design, manufacturing, or production changes based on the technical content disclosed in this application are merely routine technical means for a person skilled in the art and should not be considered to constitute a deficiency in the disclosure of the present application.

[0157] The term "embodiment" as used in this application refers to specific features, structures, or characteristics described in conjunction with the embodiment that can be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily mean that the embodiment is the same, nor does it mean that it is mutually exclusive with other embodiments and is independent or optional. It is understood, either explicitly or implicitly, by those skilled in the art that the embodiments described in this application can be combined with other embodiments when there is no conflict.

[0158] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of patent protection. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for identifying individual radiation sources, characterized in that: include: Acquiring an electromagnetic signal to be identified; Preprocessing the electromagnetic signal to be identified to obtain target input data; Inputting the target input data into a trained radiation source information encoder for feature extraction to obtain target radiation source information features; The radiation source information encoder is obtained by generative adversarial training for feature decoupling; The generative adversarial training includes: based on unlabeled data in the sample data set, conducting adversarial training on the feature extraction capability of the signal information encoder, the feature extraction capability of the radiation source information encoder, and the signal reconstruction capability of the signal generator; after inputting the data output by the data preprocessor into the signal information encoder and the radiation source information encoder respectively, cross-combining the data output by the signal information encoder and the radiation source information encoder respectively, and then inputting the data into the signal generator; After completing the training based on the unlabeled data, supervised training is performed on the feature extraction capability of the radiation source information encoder based on the labeled data in the sample data set, and a radiation source identifier is used to perform the radiation source individual identification task. Based on the loss constructed by the adversarial training, a classification loss is added to form the total loss of the supervised training, and the radiation source information features are further constrained so that the radiation source information features extracted by the radiation source information encoder meet the preset radiation source individual identification conditions; The feature decoupling is to decouple the radiation source information features inherent to the radiation source itself in the originally collected electromagnetic signal from the signal information features associated with the electromagnetic signal itself; Perform radiation source individual identification on the target radiation source information feature to obtain a radiation source individual identification result.

2. The method for identifying individual radiation sources according to claim 1, wherein: The radiation source information encoder is a component of a pre-built radiation source individual identification system, and performs the generative adversarial training; The radiation source individual identification system further includes: a data preprocessor, a signal information encoder, a signal generator, a radiation source identifier, a signal discriminator, and a decoupling discriminator; wherein in the generative adversarial training process: The data preprocessor is used to perform data preprocessing tasks on the received signal; The signal information encoder is used to perform a signal information feature extraction task; The radiation source information encoder is used to perform the radiation source information feature extraction task; The signal generator is used to perform a signal reconstruction task according to the signal information characteristics and the radiation source information characteristics; The signal discriminator is used to determine whether the reconstructed signal generated by the signal generator is true; The radiation source identifier is used to perform a radiation source individual identification task based on radiation source information features; The decoupling discriminator is used to determine the category to which the radiation source information feature belongs.

3. The method for identifying individual radiation sources according to claim 2, wherein: Preprocessing the electromagnetic signal to be identified to obtain target input data includes: Based on the data preprocessor, the electromagnetic signal to be identified is matrix processed to obtain the target input data.

4. The method for identifying individual radiation sources according to claim 2, wherein: Based on unlabeled data in the sample data set, adversarial training is performed on the feature extraction capability of the signal information encoder, the feature extraction capability of the radiation source information encoder, and the signal reconstruction capability of the signal generator, including: Based on the unlabeled data, the components of the radiation source individual recognition system are iteratively trained according to the first information decoupling loss, the first information reconstruction loss, and the first generative adversarial loss until a preset first training termination condition is reached; wherein, The first information decoupling loss is determined based on output data of the radiation source information encoder; The first information reconstruction loss is determined based on output data of the signal generator; The first generative adversarial loss is determined based on output data of the radiation source information encoder.

5. The method for identifying individual radiation sources according to claim 2, wherein: Based on the labeled data in the sample data set, supervised training is performed on the feature extraction capability of the radiation source information encoder so that the radiation source information features extracted by the radiation source information encoder meet the preset radiation source individual identification conditions, including: Based on the labeled data, iteratively training the components of the radiation source individual identification system according to a second information decoupling loss, a second information reconstruction loss, a second generative adversarial loss, and a radiation source classification loss until a preset second training termination condition is reached; The second information decoupling loss is determined based on output data of the radiation source information encoder; The second information reconstruction loss is determined based on output data of the signal generator; The second generative adversarial loss is determined based on output data of the radiation source information encoder; The radiation source classification loss is determined based on output data of the radiation source identifier.

6. The method for identifying an individual radiation source according to any one of claims 1 to 5, wherein: The method further comprises: Collect sample electromagnetic signals using several radio stations as sample radiation sources; splicing the in-phase component and the orthogonal component in the sample electromagnetic signal to obtain a spliced ​​signal; segmenting the spliced ​​signal to obtain a segmented signal; and performing convolution processing on the segmented signal to obtain a sample electromagnetic signal matrix; The sample electromagnetic signal matrix is ​​used as input for the generative adversarial training.

7. A radiation source individual identification device, characterized in that: include: Acquisition module, preprocessing module, feature extraction module and recognition module; wherein: The acquisition module is used to acquire the electromagnetic signal to be identified; The preprocessing module is used to preprocess the electromagnetic signal to be identified to obtain target input data; The feature extraction module is used to input the target input data into the trained radiation source information encoder for feature extraction to obtain target radiation source information features; the radiation source information encoder is obtained through generative adversarial training for feature decoupling; the generative adversarial training includes: based on unlabeled data in the sample data set, adversarial training is performed on the feature extraction capability of the signal information encoder, the feature extraction capability of the radiation source information encoder, and the signal reconstruction capability of the signal generator; after the data output by the data preprocessor is respectively input to the signal information encoder and the radiation source information encoder, the data output by the signal information encoder and the radiation source information encoder are cross-combined and input to the signal generator; After completing the training based on the unlabeled data, supervised training is performed on the feature extraction capability of the radiation source information encoder based on the labeled data in the sample data set, and a radiation source identifier is used to perform the radiation source individual identification task. Based on the loss constructed by the adversarial training, a classification loss is added to form the total loss of the supervised training, and the radiation source information features are further constrained so that the radiation source information features extracted by the radiation source information encoder meet the preset radiation source individual identification conditions; the feature decoupling is to decouple the radiation source information features inherent to the radiation source itself in the originally collected electromagnetic signal from the signal information features associated with the electromagnetic signal itself; The identification module is used to perform radiation source individual identification on the target radiation source information characteristics to obtain a radiation source individual identification result.

8. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to perform the radiation source individual identification method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the radiation source individual identification method according to any one of claims 1 to 6 are implemented.

Images