Radio signal forgery system for resisting deep learning

Abstract

The invention discloses a radio signal forgery system for resisting deep learning, which belongs to the technical field of data processing, and comprises a deep detection end, the deep detection end is connected with an environment early warning end, and the environment early warning end is connected with a shift grading end; the deep detection end is used for collecting signals of open shared information in real time and carrying out deep learning monitoring and deep confrontation signal progression in real time, and the environment early warning end is used for judging the surrounding environment of the radio signal anti-counterfeiting system in real time by detecting the multipath effect and noise of the radio signal anti-counterfeiting system in real time. And the shift grading end is used for tracking the radio signal anti-counterfeiting position at the current moment in real time and adjusting the radio signal anti-counterfeiting position and the anti-counterfeiting angle at regular time. According to the radio signal counterfeiting system for confronting deep learning, the security of radio signal anti-counterfeiting in the method for confronting deep learning is improved, whether the future radio signal anti-counterfeiting environment is abnormal or not is predicted in real time, and the radio signal anti-counterfeiting intensity is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a radio signal forgery system for countering deep learning. Background Art

[0002] Radio signals are a communication method that uses wireless electromagnetic waves to transmit information in space. They can transmit signals such as sound, text, data, and images. The generation and propagation of radio signals are the basis of modern communication technology, and their principles and applications have widely affected people's lives and industrial development. A radio signal forgery system for countering deep learning is an advanced electronic countermeasure technology that aims to deceive and reduce the classification accuracy of deep learning models by generating forged signals highly similar to the original radio signals. This technology is particularly important in the field of military communication security.

[0003] Application No. CN202110005855.3 discloses a radio signal forgery method for countering deep learning, which solves the technical problem that it is difficult for signal classifiers based on deep learning in the fields of artificial intelligence and electronic countermeasures to counter. The forged signals of the present invention have extremely high similarity to the original radio signals, effectively counter signal classifiers based on deep learning, and reduce the classification accuracy of radio signal modulation types. The forged signals of the present invention basically do not affect the understanding of radio signal content when the interference signals are unknown.

[0004] After retrieving the above patents, it is found that there are some deficiencies in the radio signal anti-counterfeiting system for countering deep learning: 1. The openness and sharing of wireless communication make wireless signals vulnerable to various security threats, such as illegal intrusion, malicious attacks, etc. Although the methods for countering deep learning can enhance the security of signals, they also introduce new security vulnerabilities; 2. The actual radio signal transmission environment is complex and changeable, including factors such as multipath effects, noise, and interference. These factors may weaken the effect of countering signals; 3. Shift invariance reduces the attacker's requirements for model knowledge and synchronization, which makes the attack easier to implement. However, this also means that the defense mechanism needs to have higher flexibility and adaptability to detect and respond to various shift-varying attacks.

[0005] Therefore, a radio signal forgery system for countering deep learning is proposed to solve the above problems. Summary of the Invention

[0006] The main purpose of the present invention is to provide a radio signal forgery system for countering deep learning to solve the problems raised in the above background.

[0007] To achieve the above purpose, the technical solution adopted by the present invention is: a radio signal forgery system for countering deep learning, including a depth detection end, an environment warning end, and a shift grading end; The depth detection end is used to collect signals of open - shared information in real - time, conduct deep - learning monitoring and progressive deep - adversarial signals in real - time, and increase the anti - counterfeiting strength of electrical signals against deep learning in real - time; The environmental warning end is used to detect the multipath effect and noise of the radio signal forgery system in real - time, judge the surrounding environment of the radio signal anti - counterfeiting system in real - time, predict whether the future radio signal anti - counterfeiting environment is abnormal in real - time, and perform corrective compensation in real - time when the environment is abnormal; The shifting and grading end is used to track the anti - counterfeiting position of the radio signal at the current moment in real - time, adjust the anti - counterfeiting position and anti - counterfeiting angle of the radio signal regularly, self - track the anti - counterfeiting position and anti - counterfeiting angle of the radio signal after shifting in real - time, and update the defense signal in time according to the tracking information.

[0008] Preferably, the depth detection end includes an open - sharing module, an information acquisition module, and an adversarial progression module; The open - sharing module includes a signal acquisition unit and an open - sharing unit; The signal acquisition unit is used to generate interference signals for radio modulation signals by combining the differential evolution algorithm with visual constraints, add the interference signals to the radio modulation signals, forge and send adversarial signals, and receive the forged adversarial signals in real - time through a data collector; The open - sharing unit is used to achieve interconnection of open - shared data for deep learning through big data and Internet interconnection.

[0009] Preferably, the information acquisition module includes a depth monitoring unit and a depth adversarial unit; The depth monitoring unit is used to detect and track the anti - counterfeiting data of radio signals for deep learning in real - time through a data tracker and a data detector to obtain the basic information of the anti - counterfeiting data of radio signals; The depth adversarial unit is used to detect the adversarial data of radio signal forgery at the current moment in real - time through a data detector to obtain the adversarial information of the anti - counterfeiting data of radio signals.

[0010] Preferably, the adversarial progression module includes an adversarial progression unit, an adversarial tracking unit, and an adversarial adjustment unit; The adversarial progression unit is used to receive open - shared data in real - time, conduct adversarial progression on the adversarial information of the anti - counterfeiting data of radio signals in real - time, and calculate the adversarial correction value of the anti - counterfeiting data of radio signals in real - time. The formula is as follows: ; Where, A represents the relationship between the adversarial data of the anti - counterfeiting data of radio signals and the deep - learning resistance data, specifically as follows: Represents the basic value for the first radio signal anti-counterfeiting data against data correction, Represents the basic value for the first radio signal anti-counterfeiting data against data reaching the deep learning resistance data, Represents the basic value for the second radio signal anti-counterfeiting data against data correction, Represents the basic value for the second radio signal anti-counterfeiting data against data reaching the deep learning resistance data, Represents the basic value of the deep learning resistance data at the current moment, which represents the correction values of different time points and different deep learning resistance data; The anti-counterfeiting tracking unit is used to track the correction value between the radio signal anti-counterfeiting data and the deep learning resistance data at the current moment through a data tracker in real time; The anti-counterfeiting adjustment unit is used to adjust the anti-counterfeiting intensity in real time according to the correction value.

[0011] Preferably, the environmental warning terminal includes an environmental monitoring module, a warning analysis module, and a warning tracking module; The environmental monitoring module is used to detect the multipath effect, noise, and interference signals during radio signal anti-counterfeiting through an environmental detector in real time.

[0012] Preferably, the warning analysis module includes a multipath effect unit, a noise detection unit, and an interference analysis unit; The multipath effect unit is used to calculate the multipath effect value of the radio signal at the current moment in real time, and the calculation formula is as follows: ; Wherein, is the received signal multipath effect value, is the number of resolvable multipaths, is the time-varying attenuation of the signal amplitude of the nth path, determined by path loss and shadow fading, is the signal path transmission delay value of different paths, is the phase change amount of the nth path multipath signal; The noise detection unit is used to calculate the noise of the radio signal at the current moment in real time, and the calculation formula is as follows: ; Wherein, represents the total output noise at the current moment, KT 0 represents the thermal noise power at the current moment, G i represents the gains of each stage, N Ai represents the internal noise of each stage of radio signals; The interference analysis unit is used to perform environmental prediction on the environmental parameters at the current moment in real time. The prediction method is as follows: Step Ⅰ: Set the environmental standard parameters for the radio signal anti-counterfeiting and anti-interference data corresponding to the deep learning channels, and use the radio signal anti-counterfeiting data environment detection device to detect the environmental parameters corresponding to the deep learning channels in real time; Step Ⅱ: Calculate the difference between the environmental parameters corresponding to the deep learning channels and the environmental standard parameters. If the difference = ±0.2, it is determined that the environmental parameters tend to the safe value. If the difference > 0.2 or < -0.2, it is determined that the environmental parameters exceed the safe value, and an environmental warning is issued. Taking three days as a cycle, predict the environmental parameters of the corresponding deep learning channels during radio signal anti-counterfeiting and anti-interference based on the environmental parameters of three cycles. If the average values of the environmental parameters of the three cycles all exceed the environmental standard parameters, it is determined that the radio signal anti-counterfeiting and anti-interference environment of the corresponding deep learning channel is abnormal.

[0013] Preferably, the warning tracking module includes a signal capture unit and an interference compensation unit; The signal capture unit is used to track and receive the environmental values and abnormal signals of radio signals in real time through a data tracker and a data collector; The interference compensation unit is used to perform real-time compensation and adjustment on the multipath effect of radio signals according to the calculated correction value.

[0014] Preferably, the shift and grading end includes a position monitoring module, a shift timing module, and a shift tracking module; The position monitoring module is used to detect the real-time position of radio signal anti-counterfeiting and anti-interference in real time through a data tracker and a locator, and obtain the real-time position of radio signal anti-counterfeiting and anti-interference at the current moment.

[0015] Preferably, the shift timing module includes a positioning adjustment unit, and the timing adjustment unit is used to perform timing adjustment on the position and angle of radio signal anti-counterfeiting and anti-interference through a timer and a data sensor, so as to realize the timing and angle adjustment of the position and angle of radio signal anti-counterfeiting and anti-interference.

[0016] Preferably, the shift tracking module includes a shift self-tracking unit and a defense update unit; The shift self-tracking unit is used to perform real-time tracking on the anti-counterfeiting position and anti-counterfeiting angle of radio signals through real-time tracking of a data tracker and a data sensor, and obtain the anti-counterfeiting position and anti-counterfeiting angle of radio signal anti-counterfeiting and anti-interference at the current moment; The defense update unit is used to perform real-time detection on the deep learning resistance data according to the shifted anti-counterfeiting position and anti-counterfeiting angle of radio signals, and update the defense strength of radio signal anti-counterfeiting and anti-interference data in real time.

[0017] The present invention has the following beneficial effects: (1. In the present invention, by setting a depth detection end, when used to counter the radio signal forgery operation of deep learning, through real-time deep learning monitoring and the progression of deep adversarial signals, the intensity of countering deep learning can be detected in real time, and by increasing the anti-counterfeiting intensity of the electrical signal for countering deep learning in real time, the wireless signal is less threatened by security, further increasing the security of radio signal anti-counterfeiting in the method of countering deep learning.

[0018] (2. In the present invention, by setting an environment warning end, when used to counter the radio signal forgery operation of deep learning, through real-time detection of the multipath effect and noise of the radio signal forgery system, and real-time judgment of the surrounding environment of the radio signal anti-counterfeiting system, the environmental data of the radio signal at the current moment can be tracked in real time, and whether the future radio signal anti-counterfeiting environment is abnormal can be predicted in real time through the environmental data at the current moment, enhancing the anti-counterfeiting effect of the radio signal in countering deep learning and ensuring the anti-counterfeiting accuracy of the radio signal.

[0019] (3. In the present invention, by setting a shift and grading end, when used to counter the radio signal forgery operation of deep learning, by regularly adjusting the anti-counterfeiting position and anti-counterfeiting angle of the radio signal, when anti-counterfeiting the radio signal, the attacker's self-adaptability to model knowledge and synchronization can be improved, and by real-time self-tracking of the anti-counterfeiting position and anti-counterfeiting angle of the shifted radio signal, the flexibility and adaptive ability of the defense mechanism are enhanced, so as to better detect and respond to various attack behaviors with shift changes, and the anti-counterfeiting intensity of the radio signal increases. Brief Description of the Drawings

[0020] Figure 1 It is a schematic diagram of the overall system architecture of a radio signal forgery system for countering deep learning according to the present invention; Figure 2 It is a schematic diagram of the architecture of the depth detection end of a radio signal forgery system for countering deep learning according to the present invention; Figure 3 It is a schematic diagram of the architecture of the environment warning end of a radio signal forgery system for countering deep learning according to the present invention; Figure 4 It is a schematic diagram of the architecture of the shift and grading end of a radio signal forgery system for countering deep learning according to the present invention. Detailed Embodiments

[0021] To make the technical means, creative features, achieved purposes and effects of the present invention easy to understand, the present invention will be further described below in conjunction with specific embodiments.

[0022] Embodiment 1 Please refer to Figures 1 to 2Shown: A radio signal forgery system for countering deep learning, including a depth detection end, an environmental warning end, and a shift and grading end; The depth detection end is used to collect signals of open and shared information in real time, monitor deep learning in real time, and progress the deep confrontation signals, and increase the anti-counterfeiting intensity of the electrical signals against deep learning in real time; The environmental warning end is used to detect the multipath effect and noise of the radio signal forgery system in real time, judge the surrounding environment of the radio signal anti-counterfeiting system in real time, predict whether the future radio signal anti-counterfeiting environment is abnormal in real time, and perform corrective compensation in real time when the environment is abnormal; The shift and grading end is used to track the anti-counterfeiting position of the radio signal at the current moment in real time, adjust the anti-counterfeiting position and anti-counterfeiting angle of the radio signal regularly, and perform real-time self-tracking on the anti-counterfeiting position and anti-counterfeiting angle of the shifted radio signal, and update the defense signal in time according to the tracking information.

[0023] In some embodiments, the depth detection end includes an open and shared module, an information acquisition module, and an anti-counterfeiting progression module; The open and shared module includes a signal acquisition unit and an open and shared unit; The signal acquisition unit is used to generate interference signals for radio modulation signals by combining the differential evolution algorithm with visual constraints, add the interference signals to the radio modulation signals, forge and send confrontation signals, and receive the forged confrontation signals in real time through a data collector; The open and shared unit is used to achieve interconnection of the open and shared data of deep learning through big data and Internet interconnection.

[0024] In some embodiments, the information acquisition module includes a depth monitoring unit and a deep confrontation unit; The depth monitoring unit is used to detect and track the anti-counterfeiting data of the radio signal of deep learning in real time through a data tracker and a data detector to obtain the basic information of the anti-counterfeiting data of the radio signal; The deep confrontation unit is used to detect the confrontation data of the radio signal forgery at the current moment in real time through a data detector to obtain the confrontation information of the anti-counterfeiting data of the radio signal.

[0025] In some embodiments, the anti-counterfeiting progression module includes an anti-counterfeiting progression unit, an anti-counterfeiting tracking unit, and an anti-counterfeiting adjustment unit; The anti-counterfeiting progression unit is used to receive the open and shared data in real time, progress the confrontation information of the anti-counterfeiting data of the radio signal in real time, and calculate the confrontation correction value of the anti-counterfeiting data of the radio signal in real time. The formula is as follows: ; Wherein, AIndicates the relationship between the anti-counterfeiting data of radio signals, the countermeasure data, and the data for resisting deep learning, as follows: Indicates the basic value corrected by the first anti-counterfeiting data of radio signals against countermeasure data Indicates the basic value for the first anti-counterfeiting data of radio signals against countermeasure data to reach the data for resisting deep learning Indicates the basic value corrected by the second anti-counterfeiting data of radio signals against countermeasure data Indicates the basic value for the second anti-counterfeiting data of radio signals against countermeasure data to reach the data for resisting deep learning Indicates the basic value of the data for resisting deep learning at the current moment, where the corrected values of different time points and different data for resisting deep learning are represented The countermeasure tracking unit is used to track in real time the correction value between the anti-counterfeiting countermeasure data of radio signals and the data for resisting deep learning at the current moment through a data tracker The countermeasure adjustment unit is used to adjust the countermeasure intensity in real time according to the correction value

[0026] Carry out deep learning monitoring and the progression of deep countermeasure signals in real time, so that while achieving openness and sharing in countering deep learning, the intensity of countering deep learning can be detected in real time, and the countermeasure intensity can be adjusted in a timely manner according to the deep learning intensity. By increasing the anti-counterfeiting intensity of the electrical signal for countering deep learning in real time, the wireless signal is less threatened by security, and further increases the security of radio signal anti-counterfeiting in the method of countering deep learning

[0027] Embodiment 2 Please refer to Figure 3 As shown: Based on Embodiment 1, in some embodiments, the environmental warning end includes an environmental monitoring module, a warning analysis module, and a warning tracking module The environmental monitoring module is used to detect the multipath effect, noise, and interference signals during the anti-counterfeiting countermeasure of radio signals in real time through an environmental detector

[0028] In some embodiments, the warning analysis module includes a multipath effect unit, a noise detection unit, and an interference analysis unit The multipath effect unit is used to calculate the multipath effect value of the radio signal at the current moment in real time, and the calculation formula is as follows: ; Wherein, is the multipath effect value of the received signal is the number of resolvable multipaths is the time-varying attenuation of the signal amplitude of the nth path, determined by path loss and shadow fading is the signal path transmission delay value of different paths is the phase change amount of the nth path multipath signal; The noise detection unit is used to calculate the noise of the radio signal at the current moment in real time, and the calculation formula is as follows: ; Wherein, represents the total output noise at the current moment, KT 0 represents the thermal noise power at the current moment, G i represents the gains of each stage, N Ai represents the internal noise of the radio signals at each stage; The interference analysis unit is used to perform environment prediction on the environmental parameters at the current moment in real time, and the prediction method is as follows: Step Ⅰ: Set the environmental standard parameters of the radio signal anti-counterfeiting and anti-interference data corresponding to the deep learning channel, and use the radio signal anti-counterfeiting data environment detection device to detect the environmental parameters of the corresponding deep learning channel in real time; Step Ⅱ: Calculate the difference between the environmental parameters of the corresponding deep learning channel and the environmental standard parameters. If the difference = ±0.2, it is determined that the environmental parameters tend to the safe value. If the difference > 0.2 or < -0.2, it is determined that the environmental parameters exceed the safe value, and an environmental warning is issued. Taking three days as a cycle, predict the environmental parameters of the corresponding deep learning channel during radio signal anti-counterfeiting and anti-interference based on the environmental parameters of three cycles. If the average values of the environmental parameters of the three cycles all exceed the environmental standard parameters, it is determined that the radio signal anti-counterfeiting and anti-interference environment of the corresponding deep learning channel is abnormal.

[0029] In some embodiments, the warning tracking module includes a signal capture unit and an interference compensation unit; The signal capture unit is used to track and receive the environmental values and abnormal signals of the radio signal in real time through a data tracker and a data collector; The interference compensation unit is used to perform real-time compensation and adjustment on the multipath effect of the radio signal according to the calculated correction value.

[0030] By tracking the radio signal anti-counterfeiting position at the current moment in real time and regularly adjusting the radio signal anti-counterfeiting position and anti-counterfeiting angle, when radio signal anti-counterfeiting is performed, the anti-counterfeiting position and anti-counterfeiting angle can be regularly adjusted, improving the attacker's self-adaptability to model knowledge and synchronization, increasing the attack defense effect, and by performing real-time self-tracking on the radio signal anti-counterfeiting position and anti-counterfeiting angle after displacement in real time and updating the defense signal in time according to the tracking information, the defense mechanism can be updated in real time, enhancing the flexibility and self-adaptive ability of the defense mechanism, so as to better detect and respond to various attack behaviors with displacement changes.

[0031] Embodiment III Please refer toFigure 4 As shown: Based on the basis of Embodiment 1, in some embodiments, the shift and grading end includes a position monitoring module, a shift timing module, and a shift tracking module; The position monitoring module is used to detect the real-time position of the radio signal anti-counterfeiting and countermeasure in real time through a data tracker and a locator, and obtain the real-time position of the radio signal anti-counterfeiting and countermeasure at the current moment.

[0032] The shift timing module includes a positioning adjustment unit, and the timing adjustment unit is used to adjust the position and angle of the radio signal anti-counterfeiting and countermeasure at regular intervals through a timer and a data sensor, so as to realize the adjustment of the position and angle of the radio signal anti-counterfeiting and countermeasure at regular intervals and at a fixed angle.

[0033] In some embodiments, the shift tracking module includes a shift self-tracking unit and a defense update unit; The shift self-tracking unit is used to track the anti-counterfeiting position and anti-counterfeiting angle of the radio signal in real time through the real-time tracking of the data tracker and the data sensor, and obtain the anti-counterfeiting position and anti-counterfeiting angle of the radio signal anti-counterfeiting and countermeasure at the current moment; The defense update unit is used to detect the deep learning resistance data in real time according to the shifted anti-counterfeiting position and anti-counterfeiting angle of the radio signal, and update the defense strength of the radio signal anti-counterfeiting and countermeasure data in real time.

[0034] By tracking the anti-counterfeiting position of the radio signal at the current moment in real time, and adjusting the anti-counterfeiting position and anti-counterfeiting angle of the radio signal at regular intervals, when anti-counterfeiting the radio signal, it is possible to adjust the anti-counterfeiting position and anti-counterfeiting angle at regular intervals, improve the attacker's self-adaptability to the model knowledge and synchronization, increase the attack defense effect, and by tracking the shifted anti-counterfeiting position and anti-counterfeiting angle of the radio signal in real time and updating the defense signal in time according to the tracking information in real time, the defense mechanism can be updated in real time, so that the flexibility and self-adaptive ability of the defense mechanism are enhanced, so as to better detect and respond to various attack behaviors with shift changes.

[0035] In the present invention, there is provided a radio signal forgery system for countering deep learning. When the system operates, it first generates interference signals for radio modulation signals by combining a differential evolution algorithm with visual constraints, adds the interference signals to the radio modulation signals, forges and transmits countermeasure signals. During this process, it collects signals of open and shared information in real time, monitors deep learning in real time, and progresses deep countermeasure signals, increasing the anti-counterfeiting strength of electrical signals against deep learning in real time. This enables real-time detection of the strength of countering deep learning while achieving openness and sharing, and timely adjustment of the countermeasure strength according to the deep learning strength. By increasing the anti-counterfeiting strength of electrical signals against deep learning in real time, the wireless signals are less threatened by security, further enhancing the security of radio signal anti-counterfeiting in the method of countering deep learning. By detecting the multipath effect and noise of the radio signal forgery system in real time, judging the surrounding environment of the radio signal anti-counterfeiting system in real time, predicting whether the future radio signal anti-counterfeiting environment is abnormal in real time, and performing corrective compensation in real time when the environment is abnormal, it can track the environmental data of the radio signal at the current moment in real time, predict whether the future radio signal anti-counterfeiting environment is abnormal based on the environmental data at the current moment in real time, and perform corrective compensation in real time when the environment is abnormal, enhancing the anti-counterfeiting effect of radio signals in countering deep learning and ensuring the anti-counterfeiting accuracy of radio signals. By tracking the anti-counterfeiting position of the radio signal at the current moment in real time, regularly adjusting the anti-counterfeiting position and angle of the radio signal, and performing real-time self-tracking on the shifted anti-counterfeiting position and angle of the radio signal, and timely updating the defense signal according to the tracking information in real time, when anti-counterfeiting the radio signal, it can regularly adjust the anti-counterfeiting position and angle, improving the attacker's self-adaptability to model knowledge and synchronization, increasing the attack defense effect. And by performing real-time self-tracking on the shifted anti-counterfeiting position and angle of the radio signal in real time and timely updating the defense signal according to the tracking information in real time, the defense mechanism can be updated in real time, enhancing the flexibility and self-adaptive ability of the defense mechanism, so as to better detect and respond to various attack behaviors with shift changes, and increasing the anti-counterfeiting strength of radio signals.

[0036] The above are only the preferred embodiments of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications, and environments, and can be changed within the scope of the concept described herein through the above teachings or the techniques or knowledge in related fields. And any changes and modifications made by those skilled in the art without departing from the spirit and scope of the present invention shall fall within the protection scope of the appended claims of the present invention.

Claims

1. A radio signal forgery system for countering deep learning, characterized in that: It includes a depth detection end, the depth detection end is connected to the environment warning end, and the environment warning end is connected to the displacement step-by-step end; The deep detection end is used to collect signals of open and shared information in real time, monitor deep learning and advance deep confrontation signals in real time, and increase the anti-counterfeiting strength of electrical signals against deep learning in real time; The environmental warning terminal is used to detect the multipath effect and noise of the radio signal counterfeiting system in real time, judge the surrounding environment of the radio signal anti-counterfeiting system in real time, and predict in real time whether the future radio signal anti-counterfeiting environment is abnormal, and make corrections and compensations in real time when the environment is abnormal; The shifting step end is used to track the anti-counterfeiting position of the radio signal at the current moment in real time, and adjust the anti-counterfeiting position and anti-counterfeiting angle of the radio signal at regular intervals, as well as to self-track the anti-counterfeiting position and anti-counterfeiting angle of the radio signal after the shift in real time, and update the defense signal in real time according to the tracking information.

2. The radio signal forgery system for countering deep learning according to claim 1, characterized in that: The depth detection end includes an open sharing module, the open sharing module is connected to the information acquisition module, and the information acquisition module is connected to the confrontation progressive module; The open sharing module includes a signal acquisition unit and an open sharing unit; The signal acquisition unit is used to generate an interference signal for the radio modulation signal in combination with the visually restricted differential evolution algorithm, and add the interference signal to the radio modulation signal, forge and send the adversarial signal, and receive the forged adversarial signal in real time through the data acquisition instrument; The open sharing unit is used to realize the intercommunication of open sharing data of deep learning through big data and the Internet.

3. The radio signal counterfeiting system for countering deep learning according to claim 2, characterized in that: The information acquisition module includes a depth monitoring unit and a depth confrontation unit; The deep monitoring unit is used to detect and track the radio signal anti-counterfeiting data learned in depth in real time through a data tracker and a data detector to obtain basic information of the radio signal anti-counterfeiting data; The deep adversarial unit is used to detect the adversarial data of the radio signal forgery at the current moment in real time through a data detector, and obtain the adversarial information of the radio signal anti-counterfeiting data.

4. The radio signal counterfeiting system for countering deep learning according to claim 3, characterized in that: The adversarial progressive module includes an adversarial progressive unit, an adversarial tracking unit and an adversarial adjustment unit; The countermeasure progressive unit is used to receive the open shared data in real time, and countermeasure progressive the countermeasure information of the radio signal anti-counterfeiting data in real time, and calculate the countermeasure correction value of the radio signal anti-counterfeiting data in real time. The formula is as follows: ； in, A The relationship between radio signal anti-counterfeiting data adversarial data and deep learning resistance data is as follows: Indicates the basic value of the first radio signal anti-counterfeiting data against data correction, Indicates the basic value of the first radio signal anti-counterfeiting data adversarial data reaching the deep learning resistance data, Indicates the basic value of the second radio signal anti-counterfeiting data against data correction, Indicates the basic value of the second radio signal anti-counterfeiting data adversarial data reaching the deep learning resistance data, Indicates the basic value of deep learning resistance data at the current moment, and here indicates the corrected value of different time points and different deep learning resistance data; The countermeasure tracking unit is used to track the correction value between the radio signal anti-counterfeiting countermeasure data and the deep learning resistance data at the current moment in real time through a data tracker; The confrontation adjustment unit is used to adjust the confrontation strength in real time according to the correction value.

5. The radio signal counterfeiting system for countering deep learning according to claim 1, characterized in that: The environmental early warning terminal includes an environmental monitoring module, the environmental monitoring module is connected to the early warning analysis module, and the early warning analysis module is connected to the early warning tracking module; The environment monitoring module is used to detect the multipath effect, noise and interference signal in real time during the anti-counterfeiting confrontation of radio signals through an environment detector.

6. The radio signal counterfeiting system for countering deep learning according to claim 5, characterized in that: The early warning analysis module includes a multipath effect unit, a noise detection unit and an interference analysis unit; The multipath effect unit is used to calculate the multipath effect value of the radio signal at the current moment in real time, and the calculation formula is as follows: ； in, is the multipath effect value of the received signal, is the number of resolvable multipaths, is the time-varying attenuation of the signal amplitude on the nth path, determined by path loss and shadow fading, is the signal path transmission delay value of different paths, is the phase change of the n-th multipath signal; The noise detection unit is used to calculate the noise of the radio signal at the current moment in real time, and the calculation formula is as follows: ； in, represents the total output noise at the current moment, KT 0 represents the thermal noise power at the current moment, G i represents the gain of each level, N Ai Represents the internal noise of radio signals at all levels; The interference analysis unit is used to perform environmental prediction on the environmental parameters at the current moment in real time, and the prediction method is as follows: Step I: Set the environmental standard parameters of the radio signal anti-counterfeiting countermeasure data corresponding to the deep learning channel, and use the radio signal anti-counterfeiting data environmental detection equipment to detect the environmental parameters of the corresponding deep learning channel in real time; Step II: Calculate the difference between the environmental parameters of the corresponding deep learning channel and the environmental standard parameters. If the difference = ±0.2, it is judged that the environmental parameters are approaching the safe value. If the difference > 0.2 or < -0.2, it is judged that the environmental parameters exceed the safe value, and an environmental warning is issued. Three days are counted as a cycle. The environmental parameters of the corresponding deep learning channel during radio signal anti-counterfeiting confrontation are predicted based on the environmental parameters of the three cycles. If the average values ​​of the environmental parameters of the three cycles exceed the environmental standard parameters, it is judged that the radio signal anti-counterfeiting confrontation environment of the corresponding deep learning channel is abnormal.

7. The radio signal counterfeiting system for countering deep learning according to claim 6, characterized in that: The early warning tracking module includes a signal capture unit and an interference compensation unit; The signal capture unit is used to track and receive the environmental values ​​and abnormal signals of the radio signal in real time through a data tracker and a data collector; The interference compensation unit is used to perform real-time compensation adjustment on the multipath effect of the radio signal according to the calculated correction value.

8. The radio signal counterfeiting system for countering deep learning according to claim 1, characterized in that: The shift step end includes a position monitoring module, the position monitoring module is connected to the shift timing module, and the shift timing module is connected to the shift tracking module; The position monitoring module is used to detect the real-time position of the radio signal anti-counterfeiting confrontation in real time through a data tracker and a locator, and obtain the real-time position of the radio signal anti-counterfeiting confrontation at the current moment.

9. The radio signal counterfeiting system for countering deep learning according to claim 8, characterized in that: The shift timing module includes a positioning adjustment unit, which is used to perform timing adjustment on the radio signal anti-counterfeiting countermeasure position and angle through a timer and a data sensor, thereby achieving timing and angle adjustment on the radio signal anti-counterfeiting countermeasure position and angle.

10. The radio signal forgery system for countering deep learning according to claim 9, characterized in that: The displacement tracking module includes a displacement self-tracking unit and a defense update unit; The displacement self-tracking unit is used to track the anti-counterfeiting position and anti-counterfeiting angle of the radio signal in real time through real-time tracking of the data tracker and the data sensor, and obtain the anti-counterfeiting countermeasure position and anti-counterfeiting countermeasure angle of the radio signal at the current moment; The defense updating unit is used to perform real-time detection on the deep learning resistance data according to the shifted radio signal anti-counterfeiting confrontation position and anti-counterfeiting confrontation angle, and to update the defense strength of the radio signal anti-counterfeiting confrontation data in real time.

Images