Contragitive decoupling defense method for intelligent identification of communication signal modulation mode
Through the combination of empirical modal decomposition and generative adversarial network, the adversarial perturbation is detected and processed, and the adversarial training process is solved, and the problem of weak perturbation detection is improved, which is improved by improving the robustness of the intelligent modulation recognition model. and recognition accuracy.
Patent Information
- Application Number
- CN202411150491.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-21
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-08-21
AI Technical Summary
The existing defense methods have poor anti-detection performance under weak perturbations, making it difficult to accurately detect adversarial perturbations, and there are still residues when filtering out adversarial perturbations, which limits the performance improvement of the intelligent modulation recognition model and is difficult to ensure the original. The accuracy of the identification of the sample.
The adversariality of the detection signal is decomposed by empirical modality, and the signal waveform of the generated adversarial network is randomly shaped, decoupling the adversarial training process into the original branch and the adversarial branch, and the modulation method of the predicted signal is fused based on the adversarial detection results.
The detection effect of weak perturbations is improved, the adversarial perturbation residual problem is overcome, the robustness and recognition accuracy of the intelligent modulation recognition model are improved, and the recognition accuracy of the original samples is ensured.
Smart Images

Figure CN119094291A_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present application relate to the technical field of communication signal modulation identification, and in particular to an adversarial decoupling defense method for intelligent identification of communication signal modulation modes. Background Art
[0002] With the development of wireless communication networks, the amount of data and data patterns generated by communication devices have increased dramatically. Artificial intelligence has endowed wireless communication systems with the ability to automatically and intelligently process communication data, greatly improving the efficiency of data processing and communication. Automatic Modulation Recognition (AMR) is a key technology in cognitive radio and non-cooperative communication. This technology identifies the modulation mode of the signal based on features or likelihood, providing important information for subsequent steps such as demodulation. However, traditional AMR technology relies on manually extracted features or prior knowledge of the channel, and the recognition efficiency is not high. In recent years, many researchers have tried to apply deep learning (DL) to recognition tasks, using deep neural networks (DNN) to extract and classify the intrinsic features of the signal, greatly improving the recognition speed and accuracy. In communication scenarios such as cognitive radio, edge computing, interference identification, and radio monitoring, intelligent recognition models based on deep learning have shown great advantages. For AMR systems with complex and changeable channels, DL can automatically extract features within the signal and accurately identify the modulation mode of the signal by utilizing its powerful nonlinear mapping capabilities.
[0003] However, DNN-based intelligent recognition models have been proven to be vulnerable to adversarial attacks that are difficult for humans to detect. Attackers can use the broadcast nature of wireless channels to inject carefully designed tiny adversarial perturbations into the receiver, which seriously affects the reliability of intelligent recognition models. Depending on the attack phase, common attacks include poisoning attacks that pollute training data in the training phase and evasion attacks that mislead model predictions in the inference phase. Since poisoning attacks require knowledge of the training set information of the target model, which is usually difficult to achieve due to data privacy, evasion attacks pose a greater threat to intelligent modulation recognition models. Depending on the degree of understanding of the target model, common evasion attacks usually include white-box attacks against known models and black-box attacks against unknown models. Lin et al. introduced attack methods such as Fast Gradient Sign Method (FGSM), Basic Iterative Method (BIM), and Momentum Iterative Method (MIM) from the image field to the communication field, proving the vulnerability of DL-based modulation recognition models. Liu et al. proposed a dynamic iterative method (DIM) to attack the white-box modulation classifier, which improved the success rate of the attack. In practice, attackers do not know the structure and parameters of the model, and often use the transferability of adversarial samples to carry out black-box attacks. Hu et al. proposed a substitute meta-learning black-box attack method, which combines meta-learning with the training of substitute models to improve training efficiency and attack performance. Dong et al. generated adversarial samples on the reconstructed substitute model and directly transferred the adversarial samples to the inaccessible black-box model to carry out the attack.
[0004] In response to the security risks brought by adversarial attacks to intelligent recognition models, researchers have conducted research on adversarial sample detection to reject the input model after detecting adversarial samples. In order to protect the intelligent modulation recognition model from attacks, Xu et al. realized the detection of adversarial signals by fusing multiple features of radio signals. However, in many application scenarios, it is difficult to simply detect whether the input is an adversarial sample, and adversarial defense methods are also needed to identify the true category of the input. Zhang et al. used homomorphic filtering to attenuate adversarial perturbations with high-frequency features in the signal to improve the robustness of the AMC model. Chen et al. used the principle of distillation learning to perform multiple knowledge extraction through adversarial training (AT) and normal training to improve the robustness of the AMR model to attacks. AT is a simple and effective defense method that uses adversarial sample knowledge to train the classifier and reduce the vulnerability of the classifier to adversarial samples. However, AT is only highly robust to specific attacks and is difficult to adapt to new types of attacks with stronger aggressiveness. Therefore, Kim et al. proposed a Gaussian Smoothing (GS) method, which uses Gaussian noise to enhance training data and improves the robustness of the modulation classifier against unknown attacks. However, although GS weakens the adversarial nature of the sample, it reduces the recognition accuracy of the classifier for the original sample.
[0005] With the development of generative artificial intelligence, generative adversarial networks (GANs) have emerged. Some researchers have used autoencoders and GANs to eliminate adversarial nature in input data. Autoencoders can learn data representation through their own encoders and decoders, while GANs learn data distribution through their own generators and discriminators to achieve more advanced data generation tasks. Sahay et al. developed a denoising autoencoder that improved the reliability of the power allocation model based on deep learning by learning the mapping relationship between potential adversarial samples and corresponding original samples. Dong et al. built a defensive GAN-based end-to-end communication system through triple training, and used GAN to enhance the robustness of the communication system against adversarial attacks. Traditional GANs use random noise as input, and there is a problem that the training process is prone to mode collapse. Unlike using noise that follows a priori distribution as input, Wang et al. used adversarial signals as input to the generator in GAN, trained and applied the generator to eliminate adversarial perturbations, and weakened the impact of attacks on the modulation classifier. However, some perturbations still remain in the signal processed by GAN, which makes it difficult to further improve the recognition accuracy of the model.
[0006] Through the above analysis, it is found that the currently proposed defense methods still have the following problems: First, the currently proposed defense methods have poor detection effect on adversarial samples under weak perturbation power, and it is difficult to accurately detect the adversarial nature hidden in the samples; Second, when the currently proposed defense methods filter out the adversarial perturbations in the samples, some perturbations still remain, which limits the further improvement of the performance of the intelligent modulation recognition model; Third, the currently proposed defense methods mainly focus on improving the robustness of the model to adversarial samples, and it is difficult to ensure the recognition accuracy of the model for the original clean samples. Summary of the invention
[0007] In view of this, the embodiments of the present application propose an adversarial decoupling defense method for intelligent identification of communication signal modulation modes. The adversarial nature of the detection signal is detected based on empirical mode decomposition, which overcomes the problem of poor adversarial detection performance under weak disturbances. The signal waveform is reconstructed by a generative adversarial network and randomly shaped, which overcomes the problem of residual adversarial perturbations in adversarial samples. The adversarial training process is decoupled into an original branch and an adversarial branch, and the modulation mode of the signal is predicted based on the adversarial detection results. While ensuring the recognition accuracy of the original sample, the robustness of the intelligent modulation recognition model to attacks is effectively improved, thereby effectively improving the recognition accuracy and robustness to attacks.
[0008] To achieve the above-mentioned purpose, an embodiment of the present application proposes an adversarial decoupling defense method for intelligent identification of communication signal modulation modes, comprising the following steps: performing empirical mode decomposition (EMD) on an input signal, decomposing the input signal into a combination of several intrinsic mode function (IMF) components and a residual component, and performing high-frequency feature enhancement on the first IMF component, inputting the input signal after high-frequency feature enhancement into an adversarial detector of a pre-trained recognition model, and determining the adversarial suspicion of the input signal; using a pre-trained generative adversarial network to reconstruct the waveform of the input signal to weaken hidden adversarial features; inputting the input signal into an original branch of a pre-trained recognition model to obtain a prediction vector of the original branch, inputting the input signal after waveform reconstruction into an adversarial branch of a pre-trained recognition model to obtain a prediction vector of the adversarial branch, fusing the prediction vector of the original branch and the prediction vector of the adversarial branch based on the adversarial suspicion, and determining the modulation mode of the input signal based on the final prediction vector obtained by fusion.
[0009] To achieve the above-mentioned purpose, an embodiment of the present application also proposes an adversarial decoupling defense system for intelligent identification of communication signal modulation modes, including: an adversarial detection module, a waveform reconstruction module and an adversarial training decoupling prediction module; the adversarial detection module is used to perform empirical mode decomposition on the input signal, decompose the input signal into a combination of several IMF components and residual components, and enhance the high-frequency features of the first IMF component, and input the input signal after high-frequency feature enhancement into the adversarial detector of the pre-trained recognition model to determine the adversarial suspicion of the input signal; the waveform reconstruction module is used to use the pre-trained generative adversarial network to reconstruct the waveform of the input signal to weaken the hidden adversarial features; the adversarial training decoupling prediction module is used to input the input signal into the original branch of the pre-trained recognition model to obtain the prediction vector of the original branch, input the waveform-reconstructed input signal into the adversarial branch of the pre-trained recognition model to obtain the prediction vector of the adversarial branch, fuse the prediction vector of the original branch and the prediction vector of the adversarial branch based on the adversarial suspicion, and determine the modulation mode of the input signal based on the final prediction vector obtained by fusion.
[0010] To achieve the above-mentioned objectives, an embodiment of the present application also proposes an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the adversarial decoupling defense method for intelligent identification of communication signal modulation modes as described above.
[0011] To achieve the above-mentioned objectives, an embodiment of the present application also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it can implement the adversarial decoupling defense method for intelligent identification of communication signal modulation modes as described above.
[0012] The adversarial decoupling defense method for intelligent identification of communication signal modulation modes proposed in the embodiment of the present application uses empirical mode decomposition to enhance high-frequency adversarial features in the input signal, and designs an adversarial detector to detect hidden threats in the input signal, overcoming the problem of poor adversarial detection performance under weak disturbances, improving the detection effect of weak disturbances, and providing necessary information for adversarial defense. The waveform of the input signal is reconstructed using a generative adversarial network, and the waveform is flexibly randomly shaped using the reconstructed disturbance, overcoming the problem of residual adversarial disturbances, better restoring the original modulation features, and improving the recognition performance of the adversarial branch. The traditional adversarial training process is decoupled into the original branch and the adversarial branch, and the prediction results of the two branches are fused according to the adversarial detection results, overcoming the problem that the intelligent modulation recognition model is difficult to balance the recognition accuracy of the original sample and the adversarial sample, and improving the robustness of the intelligent modulation recognition model to attacks while ensuring the recognition accuracy of the original sample, thereby effectively improving the recognition accuracy and robustness to attacks.
[0013] Optionally, performing empirical mode decomposition on the input signal to decompose the input signal into a form composed of a plurality of IMF components and a residual component includes:
[0014] The local maximum and minimum points of the signal are connected by cubic spline curves to obtain the upper envelope and the lower envelope, and the average values m1(t) of the upper envelope and the lower envelope are calculated. The average value m1(t) is expressed as:
[0015]
[0016] Among them, e max (t) represents the upper envelope, e min (t) represents the lower envelope;
[0017] The intermediate signal C is obtained by subtracting the input signal from the average value m1(t). 1,1 (t), intermediate signal C 1,1 (t) is expressed as:
[0018] C 1,1 (t) = x(t) - m1(t)
[0019] Where x(t) represents the input signal;
[0020] Determine the intermediate signal C 1,1 Is the absolute value of the difference between the number of extreme points and the number of zero crossing points of (t) less than 2? If the intermediate signal C 1,1 If the absolute value of the difference between the number of extreme points of (t) and the number of zero crossing points is less than 2, then the intermediate signal C 1,1(t) is taken as the first IMF component IMF1, and the residual component r1(t) is obtained, r1(t) = x(t) - IMF1. If the intermediate signal C 1,1 If the absolute value of the difference between the number of extreme points and the number of zero-crossing points of (t) is greater than or equal to 2, continue to decompose;
[0021] After several iterations, the remaining component r is obtained. n (t), r n (t) = r n-1 (t)-IMF n , when the residual component r n When (t) is monotonic, the empirical mode decomposition ends and the residual component R(t) is obtained, R(t) = r n (t), the input signal x(t) is decomposed into a combination of several IMF components and residual components R(t), expressed as:
[0022]
[0023] Among them, IMF i represents the i-th IMF component, n is the total number of IMF components;
[0024] The high-frequency feature enhancement of the first IMF component is achieved by the following formula:
[0025]
[0026] Among them, λ is the preset high-frequency feature enhancement factor, Represents the input signal after high-frequency features are enhanced.
[0027] Optionally, the step of inputting the input signal after high-frequency feature enhancement into an adversarial detector of a pre-trained recognition model to determine the adversarial suspiciousness of the input signal comprises: inputting the input signal after high-frequency feature enhancement into an adversarial detector of a pre-trained recognition model to determine the adversarial suspiciousness of the input signal Input into the pre-trained recognition model; the output layer of the recognition model outputs the logits value and will Input to the binary adversarial detector consisting of only three fully connected layers to obtain the detection probability output by the binary adversarial detector The detection probability As the adversarial suspiciousness μ of the input signal x(t).
[0028] Optionally, the pre-trained generative adversarial network consists of a generator G and a discriminator D. When training the generator G and the discriminator D, the original sample x and its corresponding adversarial sample x are trained. * As input;
[0029] Define the loss function L for training the generator GG (x,x * ) and the loss function L for training the discriminator D D (x,x * ) are:
[0030]
[0031] Among them, M is the number of samples, N is the sample length, and β is the preset identification loss coefficient;
[0032] The training process of the generative adversarial network is to minimize L G (x,x * ) and L D (x,x * ) Continuously update the network parameters to enhance the reconstruction ability of the generator G and the identification ability of the discriminator D;
[0033] After training, the test sample is input into the generator G, which reconstructs a modulated signal waveform that is similar to the test sample and can mislead the discriminator D to identify it as true.
[0034] Optionally, the using a pre-trained generative adversarial network to reconstruct a waveform of an input signal includes:
[0035] Input the input signal x(t) into the pre-trained generative adversarial network, and record the difference between the input signal x(t) and the output G[x(t)] of the generator G of the generative adversarial network as the reconstructed perturbation Δx(t), Δx(t) = x(t) - G[x(t)];
[0036] Replace the reconstructed disturbance Δx(t) with random noise Δn with the same power as the reconstructed disturbance Δx(t) to obtain the input signal x after waveform reconstruction r (t), the input signal x after waveform reconstruction r (t) is expressed as:
[0037] x r (t) = G[x(t)] + Δn
[0038]
[0039] Where φ is a random signal of length N that follows a standard normal distribution.
[0040] Optionally, the original branch is trained using only original samples, and the adversarial branch is trained using only adversarial samples;
[0041] The Projection Gradient Descent (PGD) method is used to generate adversarial samples to train the adversarial branch. The adversarial samples are randomly initialized before iteration. The iterative process gradually enhances the adversarial nature of the adversarial samples in multiple directions. The iterative process is expressed as:
[0042]
[0043] Among them, S represents the introduced random perturbation, y is the unique hot encoding of the true label, represents the adversarial sample after the cth iteration, α is the preset proportional coefficient, and sign(·) represents the sign function;
[0044] Using the pre-trained generative adversarial network, the final adversarial sample x * Reconstruct the waveform to obtain the reconstructed adversarial sample Then reconstruct the adversarial sample Perform Gaussian enhancement to obtain enhanced adversarial samples Enhanced adversarial examples It is expressed as:
[0045]
[0046] Where ng represents Gaussian noise with standard deviation σ, and s represents the number of samples added with noise.
[0047] Optionally, the prediction vector of the original branch and the prediction vector of the adversarial branch are fused based on the adversarial suspicion, and the modulation mode of the input signal is determined based on the final prediction vector obtained by fusion, which is implemented by the following formula:
[0048] f[x(t)]=(1-μ)f ori [x(t)]+μf adv [x r (t)]
[0049]
[0050] Among them, f ori [x(t)] represents the prediction vector of the original branch, f adv [x r (t)] represents the prediction vector of the adversarial branch, f[x(t)] represents the final prediction vector, K represents the total number of modulation modes, and f[y k |x(t)] represents the predicted probability of the kth modulation mode. BRIEF DESCRIPTION OF THE DRAWINGS
[0051] Figure 1It is a flowchart of an adversarial decoupling defense method for intelligent identification of communication signal modulation mode provided in one embodiment of the present application;
[0052] Figure 2 It is a schematic diagram of implementing an adversarial decoupling defense for intelligent identification of communication signal modulation modes provided in one embodiment of the present application;
[0053] Figure 3 It is a structural schematic diagram of an adversarial decoupling defense system for intelligent identification of communication signal modulation modes provided in another embodiment of the present application;
[0054] Figure 4 It is a comparison result diagram of an adversarial decoupling defense method for intelligent identification of communication signal modulation mode and an Inspection-Net method provided in another embodiment of the present application;
[0055] Figure 5 It is a comparison result diagram of an adversarial decoupling defense method and a Gaussian smoothing defense method for intelligent identification of communication signal modulation modes provided in another embodiment of the present application;
[0056] Figure 6 It is a structural schematic diagram of an electronic device provided in another embodiment of the present application. DETAILED DESCRIPTION
[0057] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the embodiments of the present application will be described in detail below in conjunction with the accompanying drawings. However, it will be appreciated by those skilled in the art that in the present application, many technical details are proposed in order to enable the reader to better understand the present application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical scheme claimed in the present application can also be implemented. The division of the following embodiments is for the convenience of description, and the specific implementation of the present application should not be construed as any limitation, and the various embodiments can be combined and referenced with each other under the premise of no contradiction.
[0058] In order to solve the technical problem that the currently proposed defense methods have poor defense effects, an embodiment of the present application proposes an adversarial decoupling defense method for intelligent identification of communication signal modulation modes, which is applied to electronic devices, wherein the electronic device can be a terminal or a server. This embodiment and the following embodiments are all described by taking the server as an example. The following is a specific description of the implementation details of the adversarial decoupling defense method for intelligent identification of communication signal modulation modes proposed in this embodiment. The following content is only the implementation details provided for the convenience of understanding and is not necessary for the implementation of this solution.
[0059] The specific process of the adversarial decoupling defense method for intelligent identification of communication signal modulation modes proposed in this embodiment can be as follows: Figure 1 As shown, including:
[0060] Step 101, perform empirical mode decomposition on the input signal, decompose the input signal into a combination of several IMF components and residual components, enhance the high-frequency features of the first IMF component, input the input signal after high-frequency feature enhancement into the adversarial detector of the pre-trained recognition model, and determine the adversarial suspicion of the input signal.
[0061] In the specific implementation, after obtaining the input signal, the server first needs to perform empirical mode decomposition on the input signal, decomposing the input signal into a combination of several IMF components and residual components, and then enhance the high-frequency features of the first IMF component. The input signal with high-frequency feature enhancement is then input into the adversarial detector of the pre-trained recognition model to determine the adversarial suspicion of the input signal.
[0062] Most of the defense methods proposed so far are based on DNN. Due to the high-dimensional characteristics of DNN, the slight difference in the input signal will become larger and larger during the propagation between network layers, and will eventually mislead the output of the network. This is also the reason why adversarial samples are aggressive. Therefore, the output of the last feature extraction layer of the recognition model can be used to detect the adversarial nature of the sample and distinguish the original sample from the adversarial sample. Using the logit output of the recognition model to train a simple binary detector as an adversarial detector, better detection results can be achieved without changing the network structure. The output of the binary detector is usually a probability, which can be regarded as the threat level of the sample to the recognition model. However, when the perturbation power of the adversarial sample is very small, the detection effect of this detection method is not ideal. To this end, the server uses empirical mode decomposition to detect adversarial samples and improve the accuracy of adversarial detection under weak perturbations.
[0063] Compared with the original samples, adversarial samples often have higher power at high frequencies, so the high-frequency characteristics of adversarial samples can be used to detect the suspiciousness of samples. EMD is an adaptive signal time-frequency processing method that decomposes signals based on the time scale characteristics of the data itself without presetting any basis functions. It can separate signals into IMF components and residual functions of different frequencies, and is particularly suitable for the analysis and processing of nonlinear and non-stationary signals.
[0064] In one example, when the server performs empirical mode decomposition on the input signal and decomposes the input signal into a combination of several IMF components and residual components, the server first connects the local maximum points and local minimum points of the signal through a cubic spline curve to obtain the upper envelope and the lower envelope, and calculates the average value m1(t) of the upper envelope and the lower envelope. The average value m1(t) is expressed as:
[0065]
[0066] Among them, e max (t) represents the upper envelope, e min (t) represents the lower envelope.
[0067] Then, the server needs to use the input signal and the average value m1(t) to make a difference to get the intermediate signal C 1,1 (t), intermediate signal C 1,1 (t) is expressed as:
[0068] C 1,1 (t) = x(t) - m1(t)
[0069] Where x(t) represents the input signal.
[0070] After that, the server needs to determine the intermediate signal C 1,1 Whether the absolute value of the difference between the number of extreme points of (t) and the number of zero crossing points is less than 2, that is, to judge whether C 1,1 Whether the number of extreme points of (t) is equal to the number of zero-crossing points or differs by at most one, and whether the upper envelope and the lower envelope at any time are locally symmetrical relative to the time axis.
[0071] If the intermediate signal C 1,1 The absolute value of the difference between the number of extreme points of (t) and the number of zero-crossing points is less than 2, and the upper envelope and the lower envelope at any time are locally symmetrical with respect to the time axis, then the server converts the intermediate signal C 1,1 (t) is taken as the first IMF component IMF1, and the residual component r1(t) is obtained, r1(t)=x(t)-IMF1.
[0072] If the intermediate signal C 1,1 If the absolute value of the difference between the number of extreme points and the number of zero-crossing points of (t) is greater than or equal to 2, or the upper envelope and the lower envelope at any time are not locally symmetrical with respect to the time axis, the server needs to continue decomposing. After several iterations (n iterations), the server can obtain the residual component r n (t), r n (t) = r n-1 (t)-IMF n , when the residual component rn When (t) is monotonic, the empirical mode decomposition ends and the residual component R(t) is obtained, R(t) = r n (t).
[0073] At this point, the input signal x(t) is decomposed into a combination of several IMF components and residual components R(t):
[0074]
[0075] Among them, IMF i represents the i-th IMF component, and n is the total number of IMF components.
[0076] Among several IMF components, IMF1 has the highest frequency. The main information of adversarial perturbation is hidden in IMF1. Under weak perturbation, the high-frequency features of adversarial samples and original samples are not significant, which limits the effect of adversarial sample detection. Therefore, the server needs to increase the power of IMF1 to enhance the adversarial hidden in the sample and provide better feature differences for subsequent networks to facilitate detection. The server enhances the high-frequency features of the first IMF component, which can be achieved through the following formula:
[0077]
[0078] Among them, λ is the preset high-frequency feature enhancement factor, Represents the input signal after high-frequency features are enhanced.
[0079] In one example, the server obtains the input signal after high-frequency feature enhancement After that, the input signal with enhanced high-frequency features can be Input into the pre-trained recognition model, the output layer of the recognition model outputs the logits value and will Input to the binary adversarial detector consisting of only three fully connected layers to obtain the detection probability output by the binary adversarial detector Finally, the recognition model will detect the probability As the adversarial suspiciousness μ of the input signal x(t), the adversarial suspiciousness μ of the input signal x(t) can be expressed as:
[0080]
[0081] Among them, τ is the preset decision threshold.
[0082] Step 102, using the pre-trained generative adversarial network, reconstructs the waveform of the input signal to weaken the hidden adversarial features.
[0083] In the specific implementation, after obtaining the input signal, the server also needs to use the pre-trained generative adversarial network to reconstruct the waveform of the input signal to weaken the hidden adversarial features.
[0084] It is worth noting that step 102 and step 101 can be performed synchronously.
[0085] In an example, the pre-trained generative adversarial network consists of a generator G and a discriminator D. When training the generator G and the discriminator D, the server can send the original sample x and its corresponding adversarial sample x * as input.
[0086] When training the generator G, the waveform reconstructed by the generator G must be close to the original signal waveform in the time domain and be able to deceive the discriminator D to make an incorrect prediction. When training the discriminator D, it is necessary to reduce the probability of the discriminator D predicting the reconstructed sample while increasing the probability of the discriminator D predicting the original sample. Therefore, the server defines the loss function L for training the generator G G (x,x * ) and the loss function L for training the discriminator D D (x,x * ) are:
[0087]
[0088] Among them, M is the number of samples, N is the sample length, and β is the preset identification loss coefficient.
[0089] The training process of the generative adversarial network is to minimize L G (x,x * ) and L D (x,x * ) Continuously update the network parameters to enhance the reconstruction ability of the generator G and the identification ability of the discriminator D.
[0090] After training, the test sample is input into the generator G, which reconstructs a modulated signal waveform that is similar to the test sample and can mislead the discriminator D to identify it as true.
[0091] The server inputs the input signal x(t) into the pre-trained generative adversarial network, and records the difference between the input signal x(T) and the output G[x(t)] of the generator G of the generative adversarial network as the reconstructed perturbation Δx(t), Δx(t) = x(t) - G[x(t)].
[0092] Subsequently, the server replaces the reconstructed disturbance Δx(t) with a random noise Δn having the same power as the reconstructed disturbance Δx(t) to obtain the input signal x after waveform reconstruction r (t), the input signal x after waveform reconstructionr (t) can be expressed as:
[0093] x r (t) = G[x(t)] + Δn
[0094]
[0095] Where φ is a random signal of length N that follows a standard normal distribution.
[0096] For the generative adversarial network, when the input is an adversarial sample, Δx(t) contains adversarial information. After replacing it with random noise, this part of the adversarial information will be destroyed. When the input is the original sample, Δx(t) only contains random noise. Replacing it with random noise has little effect on the prediction of the recognition model. Compared with the traditional method of directly adding random noise to the input sample, the setting of the generative adversarial network can adaptively adjust the noise power according to the size of the reconstructed disturbance, avoiding the influence of the modulation characteristics of the sample due to unreasonable noise power setting.
[0097] Step 103: input the input signal into the original branch of the pre-trained recognition model to obtain a prediction vector of the original branch; input the waveform-reconstructed input signal into the adversarial branch of the pre-trained recognition model to obtain a prediction vector of the adversarial branch; fuse the prediction vector of the original branch and the prediction vector of the adversarial branch based on the adversarial suspicion; and determine the modulation method of the input signal based on the final prediction vector obtained by fusion.
[0098] In the specific implementation, considering that in the training process of traditional adversarial training, adversarial samples need to be added to the training set to train the recognition model together with the original samples, which easily causes overfitting in the training process and reduces the recognition accuracy of the recognition model for the original samples. In order to solve this problem, the server chooses to decouple the adversarial training into the original branch and the adversarial branch. The server inputs the input signal into the original branch to obtain the prediction vector of the original branch, inputs the input signal after waveform reconstruction into the adversarial branch to obtain the prediction vector of the adversarial branch, and finally fuses the prediction vector of the original branch and the prediction vector of the adversarial branch based on the adversarial suspicion, and determines the modulation mode of the input signal based on the final prediction vector obtained by fusion.
[0099] In one example, the original branch is trained only with original samples, and the adversarial branch is trained only with adversarial samples. The server generates adversarial samples using the projected gradient descent method to train the adversarial branch. The adversarial samples are randomly initialized before iteration. The iterative process gradually enhances the adversarial nature of the adversarial samples in multiple directions. The iterative process is expressed as:
[0100]
[0101] Among them, S represents the introduced random perturbation, y is the unique hot encoding of the true label, represents the adversarial sample after the cth iteration, α is the preset proportional coefficient, and sign(·) represents the sign function.
[0102] In order to further improve the generalization of the adversarial branch to unknown attacks, the server can perform Gaussian enhancement on the reconstructed adversarial sample. Gaussian enhancement uses the training set of the Gaussian noise enhancement model to improve the robustness of the modulation recognition model to possible disturbances in multiple directions. Specifically, the server uses the pre-trained generative adversarial network to generate a Gaussian noise enhancement on the final adversarial sample x. * Reconstruct the waveform to obtain the reconstructed adversarial sample Then reconstruct the adversarial sample Perform Gaussian enhancement to obtain enhanced adversarial samples Enhanced adversarial examples It can be expressed as:
[0103]
[0104] Where ng represents Gaussian noise with standard deviation σ, and s represents the number of samples added with noise.
[0105] In an example, the server fuses the prediction vector of the original branch and the prediction vector of the adversarial branch based on the adversarial suspicion, and determines the modulation mode of the input signal based on the final prediction vector obtained by fusion, which can be implemented by the following formula:
[0106] f[x(t)]=(1-μ)f ori [x(t)]+μf adv [x r (t)]
[0107]
[0108] Among them, f ori [x(t)] represents the prediction vector of the original branch, f adv [x r (t)] represents the prediction vector of the adversarial branch, f[x(t)] represents the final prediction vector, K represents the total number of modulation modes, and f[y k|x(t)] represents the predicted probability of the kth modulation mode. If μ = 0, it means that the input signal is completely treated as the original sample. At this time, the recognition accuracy of the recognition model is consistent with the original branch, avoiding the decrease in the recognition accuracy of the original sample by the recognition model. If 0 < μ < 1, it means that the adversarial detector cannot completely determine the nature of the input signal, and needs to combine the results of the original branch and the adversarial branch for recognition. If μ = 1, it means that the input signal is an adversarial sample, and the recognition result of the recognition model is consistent with the adversarial branch.
[0109] At this point, the adversarial decoupling defense for intelligent identification of communication signal modulation methods has been completed. The entire implementation process of the adversarial decoupling defense for intelligent identification of communication signal modulation methods can be described as follows: Figure 2 shown.
[0110] In this embodiment, empirical mode decomposition is used to enhance the high-frequency adversarial features in the input signal, and an adversarial detector is designed to detect hidden threats in the input signal, thereby overcoming the problem of poor adversarial detection performance under weak perturbations, improving the detection effect of weak perturbations, and providing necessary information for adversarial defense. The waveform of the input signal is reconstructed using a generative adversarial network, and the waveform is flexibly randomly shaped using the reconstructed perturbation, thereby overcoming the problem of residual adversarial perturbations, better restoring the original modulation features, and improving the recognition performance of the adversarial branch. The traditional adversarial training process is decoupled into the original branch and the adversarial branch, and the prediction results of the two branches are fused according to the adversarial detection results, thereby overcoming the problem that the intelligent modulation recognition model is difficult to balance the recognition accuracy of the original sample and the adversarial sample, and improving the robustness of the intelligent modulation recognition model to attacks while ensuring the recognition accuracy of the original sample, thereby effectively improving the recognition accuracy and robustness to attacks.
[0111] Another embodiment of the present application proposes an adversarial decoupling defense system for intelligent identification of communication signal modulation modes. The details of the adversarial decoupling defense system for intelligent identification of communication signal modulation modes proposed in this embodiment are described in detail below. The following content is only the implementation details provided for the convenience of understanding and is not necessary for the implementation of this example. Figure 3 It is a schematic diagram of the adversarial decoupling defense system for intelligent identification of communication signal modulation modes proposed in this embodiment, including: an adversarial detection module 201, a waveform reconstruction module 202 and an adversarial training decoupling prediction module 203.
[0112] The adversarial detection module 201 is used to perform empirical mode decomposition on the input signal, decompose the input signal into a combination of several IMF components and residual components, and enhance the high-frequency features of the first IMF component. The input signal after high-frequency feature enhancement is input into the adversarial detector of the pre-trained recognition model to determine the adversarial suspicion of the input signal.
[0113] The waveform reconstruction module 202 is used to reconstruct the waveform of the input signal using a pre-trained generative adversarial network to weaken hidden adversarial features.
[0114] The adversarial training decoupling prediction module 203 is used to input the input signal into the original branch of the pre-trained recognition model to obtain the prediction vector of the original branch, input the input signal after waveform reconstruction into the adversarial branch of the pre-trained recognition model to obtain the prediction vector of the adversarial branch, fuse the prediction vector of the original branch and the prediction vector of the adversarial branch based on the adversarial suspicion, and determine the modulation method of the input signal based on the final prediction vector obtained by fusion.
[0115] It is not difficult to find that this embodiment is a system embodiment corresponding to the above method embodiment, and this embodiment can be implemented in conjunction with the above method embodiment. The relevant technical details and technical effects mentioned in the above embodiments are still valid in this embodiment, and in order to reduce repetition, they are not repeated here. Accordingly, the relevant technical details mentioned in this embodiment can also be applied in the above embodiments.
[0116] It is worth mentioning that all modules involved in this embodiment are logic modules. In practical applications, a logic unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. In addition, in order to highlight the innovative part of this application, this embodiment does not introduce units that are not closely related to solving the technical problems proposed by this application, but this does not mean that there are no other units in this embodiment.
[0117] The adversarial decoupling defense method and system for intelligent identification of communication signal modulation modes proposed in this application can be used to defend against adversarial attacks on intelligent modulation identification of communication signals, including white-box attacks, black-box attacks, reliability attacks, and targeted attacks. It can also be used to defend against adversarial attacks in other fields such as image recognition based on deep learning.
[0118] In another embodiment, in order to evaluate the performance of the adversarial decoupling defense method and system (hereinafter referred to as EMD-AD) for intelligent identification of communication signal modulation modes proposed in this application, technicians conducted a series of simulation experiments. In the simulation experiment, the ResNet network was selected as the intelligent modulation recognition model to be protected, and three fully connected layers were used to form an adversarial detector for detecting adversarial samples, and Dropout was used to prevent overfitting. The types of modulation signals to be identified include 8 digital signals: 8PSK, QPSK, BPSK, GFSK, CPFSK, PAM4, QAM16 and QAM64, and two analog signals: WBFM and AM-DSB. The simulation parameters for studying the influence of disturbance level on defense performance are set as follows: the modulation signal-to-noise ratio is SNR=10dB, the disturbance level ε∈[0, 0.003] and the interval is 0.0003, and the decision threshold for detecting adversarial samples is set to 0.5. In the simulation experiment, 1000 samples are taken each time to verify the attack performance of the adversarial samples generated for the test signal. When verifying the adversarial sample detection performance, the EMD-AD proposed in this application is compared with the more mature Inspection-Net, and the two detection performance indicators of Accuracy and F1-Score are compared. The simulation results are as follows: Figure 4 When verifying the defense performance, the adversarial decoupling defense (hereinafter referred to as ADD) proposed in this application is compared with the more mature Gaussian Smoothing (GS) defense method to compare the defense effects of the two on FGSM, MIM, and DIM attacks. The simulation results are shown in Figure 5 As shown. Figure 4 It can be seen that the adversarial detection performance of EMD-AD is better than that of Inspection-Net, and it has better detection effect under weak perturbations. At the same time, it can be seen from Figure 5 that compared with GS, the ADD method not only ensures the recognition accuracy of the recognition model for the original samples, but also enhances the robustness of the model against attacks.
[0119] Another embodiment of the present application provides an electronic device, the structure of which is as follows: Figure 6 As shown, it includes: at least one processor 301; and a memory 302 that is communicatively connected to the at least one processor 301; wherein the memory 302 stores instructions that can be executed by the at least one processor 301, and the instructions are executed by the at least one processor 301 so that the at least one processor 301 can execute the adversarial decoupling defense method for intelligent identification of communication signal modulation methods described in the above-mentioned method embodiments.
[0120] Among them, the memory and the processor are connected in a bus manner, and the bus may include any number of interconnected buses and bridges, and the bus connects various circuits of one or more processors and memories together. The bus can also connect various other circuits such as peripherals, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be one element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices on a transmission medium. The data processed by the processor is transmitted on a wireless medium via an antenna, and further, the antenna also receives data and transmits the data to the processor.
[0121] The processor is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory can be used to store data used by the processor when performing operations.
[0122] Another embodiment of the present application proposes a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it can implement the adversarial decoupling defense method for intelligent identification of communication signal modulation modes described in the above-mentioned method embodiments.
[0123] That is, those skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a program, and the program is stored in a storage medium, including a number of instructions to enable a device (which can be a single-chip microcomputer, chip, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, ROM (Read-Only Memory), RAM (Random Access Memory), disk or optical disk and other media that can store program codes.
[0124] Those skilled in the art will appreciate that the above embodiments are specific embodiments for implementing the present application, and in actual applications, various changes may be made thereto in form and detail without departing from the spirit and scope of the present application.
Claims
1. An adversarial decoupling defense method for intelligent identification of communication signal modulation mode, characterized in that: include: Perform empirical mode decomposition on the input signal to decompose it into a combination of several IMF components and residual components, and enhance the high-frequency features of the first IMF component. The input signal after high-frequency feature enhancement is input into the adversarial detector of the pre-trained recognition model to determine the adversarial suspicion of the input signal. Using the pre-trained generative adversarial network, the waveform of the input signal is reconstructed to weaken the hidden adversarial features; The input signal is input into the original branch of the pre-trained recognition model to obtain the prediction vector of the original branch, the input signal after waveform reconstruction is input into the adversarial branch of the pre-trained recognition model to obtain the prediction vector of the adversarial branch, the prediction vector of the original branch and the prediction vector of the adversarial branch are fused based on the adversarial suspicion, and the modulation method of the input signal is determined based on the final prediction vector obtained by fusion.
2. The adversarial decoupling defense method for intelligent identification of communication signal modulation modes according to claim 1 is characterized in that: The step of performing empirical mode decomposition on the input signal to decompose the input signal into a combination of several IMF components and residual components includes: The local maximum and minimum points of the signal are connected by cubic spline curves to obtain the upper envelope and the lower envelope, and the average values m1(t) of the upper envelope and the lower envelope are calculated. The average value m1(t) is expressed as: Among them, e max (t) represents the upper envelope, e min (t) represents the lower envelope; The intermediate signal C is obtained by subtracting the input signal from the average value m1(t). 1,1 (t), intermediate signal C 1,1 (t) is expressed as: C 1,1 (t)=x(t)-m1(t) Where x(t) represents the input signal; Determine the intermediate signal C 1,1 (t) whether the absolute value of the difference between the number of extreme points and the number of zero-crossing points is less than 2, and whether the upper envelope and the lower envelope at any time are locally symmetrical with respect to the time axis; If the intermediate signal C 1,1 The absolute value of the difference between the number of extreme points of (t) and the number of zero crossing points is less than 2, and the upper envelope and the lower envelope at any time are locally symmetrical with respect to the time axis, then the intermediate signal C 1,1 (t) is taken as the first IMF component IMF1, and the residual component r1(t) is obtained, r1(t)=x(t)-IMF1; If the intermediate signal C 1,1 If the absolute value of the difference between the number of extreme points and the number of zero-crossing points of (t) is greater than or equal to 2, or the upper envelope and the lower envelope at any time are not locally symmetrical with respect to the time axis, continue to decompose; After several iterations, the remaining component r is obtained. n (t), r n (t) = r n-1 (t)-IMF n , when the residual component r n When (t) is monotonic, the empirical mode decomposition ends and the residual component R(t) is obtained, R(t) = r n (t), the input signal x(t) is decomposed into a combination of several IMF components and residual components R(t), expressed as: Among them, IMF i represents the i-th IMF component, n is the total number of IMF components; The high-frequency feature enhancement of the first IMF component is achieved by the following formula: Among them, λ is the preset high-frequency feature enhancement factor, Represents the input signal after high-frequency features are enhanced.
3. The adversarial decoupling defense method for intelligent identification of communication signal modulation modes according to claim 2 is characterized in that: The step of inputting the input signal after high-frequency feature enhancement into the adversarial detector of the pre-trained recognition model to determine the adversarial suspiciousness of the input signal includes: The input signal after high-frequency features are enhanced Input into the pre-trained recognition model; The output layer of the recognition model outputs the logits value and will Input to the binary adversarial detector consisting of only three fully connected layers to obtain the detection probability output by the binary adversarial detector The detection probability As the adversarial suspiciousness μ of the input signal x(t).
4. The adversarial decoupling defense method for intelligent identification of communication signal modulation modes according to claim 3 is characterized in that: The pre-trained generative adversarial network consists of a generator G and a discriminator D. When training the generator G and the discriminator D, the original sample x and its corresponding adversarial sample x are * As input; Define the loss function L for training the generator G G (x,x * ) and the loss function L for training the discriminator D D (x,x * ) are: Among them, M is the number of samples, N is the sample length, and β is the preset identification loss coefficient; The training process of the generative adversarial network is to minimize L G (x,x * ) and L D (x,x * ) Continuously update the network parameters to enhance the reconstruction ability of the generator G and the identification ability of the discriminator D; After training, the test sample is input into the generator G, which reconstructs a modulated signal waveform that is similar to the test sample and can mislead the discriminator D to identify it as true.
5. The adversarial decoupling defense method for intelligent identification of communication signal modulation modes according to claim 4 is characterized in that: The method of using a pre-trained generative adversarial network to reconstruct a waveform of an input signal includes: Input the input signal x(t) into the pre-trained generative adversarial network, and record the difference between the input signal x(t) and the output G[x(t)] of the generator G of the generative adversarial network as the reconstructed perturbation Δx(t), Δx(t) = x(t) - G[x(t)]; Replace the reconstructed disturbance Δx(t) with random noise Δn with the same power as the reconstructed disturbance Δx(t) to obtain the input signal x after waveform reconstruction r (t), the input signal x after waveform reconstruction r (t) is expressed as: x r (t)=G[x(t)]+Δn Where φ is a random signal of length N that follows a standard normal distribution.
6. The adversarial decoupling defense method for intelligent identification of communication signal modulation modes according to claim 5 is characterized in that: The original branch is trained only with original samples, and the adversarial branch is trained only with adversarial samples; The projected gradient descent method is used to generate adversarial samples to train the adversarial branch. The adversarial samples are randomly initialized before iteration. The iterative process gradually enhances the adversarial nature of the adversarial samples in multiple directions. The iterative process is expressed as: Among them, S represents the introduced random perturbation, y is the unique hot encoding of the true label, represents the adversarial sample after the cth iteration, α is the preset proportional coefficient, and sign(·) represents the sign function; Using the pre-trained generative adversarial network, the final adversarial sample x * Reconstruct the waveform to obtain the reconstructed adversarial sample Then reconstruct the adversarial sample Perform Gaussian enhancement to obtain enhanced adversarial samples Enhanced adversarial examples It is expressed as: Where ng represents Gaussian noise with standard deviation σ, and s represents the number of samples added with noise.
7. The adversarial decoupling defense method for intelligent identification of communication signal modulation modes according to claim 6 is characterized in that: The prediction vector of the original branch and the prediction vector of the adversarial branch are fused based on the adversarial suspicion, and the modulation mode of the input signal is determined based on the final prediction vector obtained by fusion, which is implemented by the following formula: f[x(t)]=(1-μ)f ori [x(t)]+μf adv [x r (t)] Among them, f ori [x(t)] represents the prediction vector of the original branch, f adv [x r (t)] represents the prediction vector of the adversarial branch, f[x(t)] represents the final prediction vector, K represents the total number of modulation modes, and f[y k |x(t)] represents the predicted probability of the kth modulation mode.
8. An adversarial decoupling defense system for intelligent identification of communication signal modulation modes, characterized in that: include: Adversarial detection module, waveform reconstruction module and adversarial training decoupled prediction module; The adversarial detection module is used to perform empirical mode decomposition on the input signal, decompose the input signal into a combination of several IMF components and residual components, and enhance the high-frequency features of the first IMF component. The input signal after high-frequency feature enhancement is input into the adversarial detector of the pre-trained recognition model to determine the adversarial suspicion of the input signal; The waveform reconstruction module is used to reconstruct the waveform of the input signal using the pre-trained generative adversarial network to weaken the hidden adversarial features; The adversarial training decoupling prediction module is used to input the input signal into the original branch of the pre-trained recognition model to obtain the prediction vector of the original branch, input the input signal after waveform reconstruction into the adversarial branch of the pre-trained recognition model to obtain the prediction vector of the adversarial branch, fuse the prediction vector of the original branch and the prediction vector of the adversarial branch based on the adversarial suspicion, and determine the modulation method of the input signal based on the final prediction vector obtained by fusion.
9. An electronic device, characterized in that: include: at least one processor; and, a memory communicatively coupled to the at least one processor; In which, the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the adversarial decoupling defense method for intelligent identification of communication signal modulation modes as described in any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it is able to implement the adversarial decoupling defense method for intelligent identification of communication signal modulation modes as described in any one of claims 1 to 7.
Citation Information
Patent Citations
Signal modulation type recognition attack defense method based on generative adversarial network
CN113378644A
Communication signal noise reduction method and system based on relative average generative adversarial network
CN115442191A
Multi-feature fusion generative adversarial network ultra-short-term wind power prediction method
CN115526376A
Rotary machinery fault diagnosis method and system based on improved EEMD and generative adversarial network
CN118395364A
Training of a machine learning model for predictive maintenance tasks
WO2024012735A1
Cited By
Dual cooperative defense method for communication signal modulation intelligent identification confrontation
CN119691559A
