A network authentication method based on radio frequency signal gene analysis

Abstract

The present invention discloses a network authentication method based on radio frequency signal gene analysis, including: S01: Collecting the radio frequency baseband signals of the communication radiation source individuals to be identified; S02: Segmenting the collected radio frequency baseband signals and respectively converting them into equipotential star maps; S03: Using a pre-trained neural network to respectively identify the equipotential star maps from the same radio frequency baseband signal, and obtaining the identification results of the communication radiation source individuals based on the characteristics of the equipotential star maps; S04: Judging whether the identification results are consistent. If they are consistent, using the identification results as the authentication results. If they are inconsistent, proceeding to the next step; S05: Converting the whole radio frequency baseband signal into a differential equipotential star map, and using a pre-trained neural network for identification, obtaining the identification results of the communication radiation source individuals based on the characteristics of the differential equipotential star map, and using the identification results as the authentication results. By converting the signals into different star maps, the present invention has a relatively high success rate in identifying communication radiation source individuals.

Description

Technical Field

[0001] The present invention relates to the field of identification and authentication, and particularly to a network authentication method based on radio frequency signal gene analysis. Background Art

[0002] The electromagnetic field is the material basis for the transmission of electromagnetic waves. When modulating the current at the excitation source and attaching the information to be transmitted, that is, the intentional modulation information, to the signal, the unintentional modulation information reflecting the characteristics of the radiation source itself will also be carried by the signal. After these information-carrying electromagnetic waves depart from the excitation source and propagate step by step outward, they are continuously affected by the transmission medium. Although retaining the characteristics of the electromagnetic field at the excitation source, their characteristics may change to a great extent during the propagation process. It is necessary to carry out theoretical research and technical processing to extract the intentional modulation information in the electromagnetic wave signal. There are a large number of studies on the identification of modulation methods for reference. For the unintentional modulation information among them, affected by the signal transmission environment and the individual characteristics of the radiation source, there may be a characteristic that can reflect the inherent and essential characteristics of the radiation source, serving as the gene of the radiation source, similar to the genetic principle of biological genes. Although going through vicissitudes, through this long historical book of genes, clues can still be found by following the clues.

[0003] With the continuous emergence of information security problems brought by wireless communication networks, especially problems such as user identity impersonation, replay attacks, and device cloning. In recent years, attack incidents that have occurred in various countries (such as causing load shedding, line overload disconnection, and cascading failures in the power grid through intelligent data attacks) have gradually exposed various hidden dangers in the information security of the power grid. How to accurately identify and authenticate Internet of Things objects is the primary problem faced by the power Internet of Things and also the basis for the application of the power Internet of Things. Data attacks by illegally connected devices in the power Internet of Things will cause serious interference and threats to the entire network. It is difficult to ensure the security of the communication system only relying on traditional application layer password authentication methods. Therefore, it is of great significance to design an effective physical layer authentication system. Summary of the Invention

[0004] Aiming at the problem that the prior art lacks effective physical layer authentication means, the present invention provides a network authentication method based on radio frequency signal gene analysis, which ensures the accuracy of the identification result through two-level physical layer authentication and identification, and can complete accurate authentication.

[0005] The following is the technical solution of the present invention.

[0006] A network authentication method based on radio frequency signal gene analysis includes the following steps:

[0007] S01: Collect the radio frequency baseband signal of the communication radiation source individual to be identified;

[0008] S02: Segment the collected RF baseband signals and convert them into equipotential star charts respectively;

[0009] S03: Use a pre-trained neural network to identify the equipotential star charts from the same RF baseband signal respectively, and obtain the individual identification results of communication radiation sources based on the characteristics of the equipotential star charts;

[0010] S04: Determine whether the identification results are consistent. If they are consistent, use the identification results as the authentication results. If they are inconsistent, proceed to the next step;

[0011] S05: Convert the entire RF baseband signal into a differential equipotential star chart, and use a pre-trained neural network for identification to obtain the individual identification results of communication radiation sources based on the characteristics of the differential equipotential star chart, and use this identification result as the authentication result.

[0012] In the present invention, after segmenting the collected signals, identification is first performed by means of equipotential star charts. If the results are inconsistent, further identification is performed by means of differential equipotential star charts. Through two-level identification guarantees, the accuracy of identification is basically ensured, so it can be used for device authentication.

[0013] Preferably, the conversion process of the equipotential star chart includes:

[0014] Segment the RF baseband signal and represent it as a two-dimensional constellation diagram;

[0015] According to the density distribution of the vector endpoints in the two-dimensional constellation diagram, distinguish different density regions by color to obtain the equipotential star chart.

[0016] Preferably, the process of representing as a two-dimensional constellation diagram includes: on a two-dimensional coordinate with I as the horizontal axis and Q as the vertical axis, represent the vector endpoints of the RF baseband signal under the projection of a specified basis vector to obtain a vector diagram. After all vector endpoints are represented, the two-dimensional constellation diagram corresponding to the RF baseband signal is obtained; where I is the in-phase signal of the RF baseband signal, which is the cos component of the RF baseband signal, and Q is the quadrature signal of the RF baseband signal, which is the sin component of the RF baseband signal.

[0017] Preferably, the neural network is a deep convolutional neural network, and the training process is: select a part from the equipotential star charts as training samples, and the rest as identification samples. Set the parameter structure of the deep convolutional neural network, and then import the training samples for training to obtain the trained deep convolutional neural network.

[0018] Preferably, the conversion process of the differential equipotential star chart includes:

[0019] Perform differential and normalization processing on the RF baseband signal to obtain a differential constellation diagram;

[0020] According to the point density distribution in the differential constellation diagram, different density regions are color - differentiated to obtain a differential equipotential planet diagram.

[0021] Preferably, the differential processing process includes:

[0022] For the radio frequency baseband signal where t is the sampling point position, x(t) is the transmitter baseband signal, and f t is the transmitter carrier frequency;

[0023] The differential processing is expressed as:

[0024]

[0025] where: d(t) is the signal after differential processing, y * is the conjugate value of y; n is 1; e -j2πθn is the phase rotation factor; r(t) is the signal received by the receiver and r(t)=s(t), f r is the receiver carrier frequency, is the phase error when the receiver receives the signal.

[0026] Preferably, the process of step S01 includes: using a spectrum analyzer to collect the radio frequency baseband signals of the communication radiation source individuals to be identified respectively at a preset sampling frequency and sampling time.

[0027] Preferably, the color - differentiation of different density regions includes: performing gradient coloring according to the density difference of vector endpoints, and then adjusting the brightness of the coloring according to the density difference, where the brightness decreases as the density increases.

[0028] The substantial effects of the present invention include: by converting the signal into different planet diagrams, using a neural network for feature extraction and recognition, and combining two - level authentication recognition, the success rate of identifying communication radiation source individuals is relatively high, and it can be used for physical - layer authentication of devices. Brief Description of the Drawings

[0029] Figure 1 is a schematic diagram of the two - dimensional constellation diagram generation process of an embodiment of the present invention;

[0030] Figure 2 is an equipotential planet diagram of an embodiment of the present invention;

[0031] Figure 3 is a test result confusion matrix diagram based on the characteristics of the differential constellation trajectory diagram of an embodiment of the present invention;

[0032] Figure 4 is a test result confusion matrix diagram based on the characteristics of the equipotential planet diagram of an embodiment of the present invention;

[0033] Figure 5 It is a schematic diagram of the generation process of the differential equipotential planet map of the embodiment of the present invention. Detailed implementation manners

[0034] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in conjunction with the embodiments. Obviously, the described embodiments are only a part rather than all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0035] It should be understood that in various embodiments of the present invention, the magnitudes of the sequence numbers of the various processes do not mean the order of execution, and the order of execution of the various processes should be determined according to their functions and internal logics, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0036] It should be understood that in the present invention, "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device including a series of steps or units does not necessarily have to be limited to those clearly listed steps or units, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0037] The following uses specific embodiments to elaborate on the technical solutions of the present invention in detail. The embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0038] Embodiment:

[0039] A network authentication method based on radio frequency signal gene analysis includes the following steps:

[0040] S01: Collect the radio frequency baseband signal of the communication radiation source individual to be identified;

[0041] S02: Segment the collected radio frequency baseband signal and convert it into an equipotential planet map respectively;

[0042] S03: Use the pre-trained neural network to identify the equipotential planet maps from the same radio frequency baseband signal respectively, and obtain the identification result of the communication radiation source individual based on the equipotential planet map features;

[0043] S04: Judge whether the identification results are consistent. If they are consistent, use the identification result as the authentication result. If they are inconsistent, execute the next step;

[0044] S05: Convert the entire RF baseband signal into a differential equipotential star map, and use a pre-trained neural network for recognition to obtain the individual recognition result of the communication radiation source based on the characteristics of the differential equipotential star map. This recognition result is used as the authentication result.

[0045] In this embodiment, after the collected signal is segmented, it is first recognized by means of an equipotential star map. If the results are inconsistent, the differential equipotential star map is further used for recognition. Through the two-level recognition guarantee, the accuracy of recognition is basically ensured, so it can be used for device authentication.

[0046] Among them, the generation processes of the two star maps are slightly different. Taking 20 Wi-Fi network card devices of the same manufacturer, model, and batch as an example. The baseband signal acquisition device is an FSW26 type spectrum analyzer, and the acquisition environment is an indoor laboratory scene. A total of 20 Wi-Fi network card devices are collected, and 50 samples are collected for each device; the signal sampling frequency is 80 MHz, and each collection is 1.75 ms, that is, the number of sample points per sample is 140,000 (taking a single channel as an example). Among them, the number of points in the effective data transmission segment after removing the signal noise segment by the variance trajectory change point detection algorithm is 80,000 (all are steady-state signals). Then, the effective data transmission segment is sliced (with 10,000 points as a new sample) for processing. Then, 8 effective data transmission segment fragments are cut out for each sample. Taking each fragment as a sample, each device becomes a total of 50×8 = 400 samples. At this time, there are a total of 20×400 = 8000 samples (after generating the equipotential star map, 6400 samples are randomly selected to generate for the training of the deep convolutional neural network, and the remaining 1600 samples are used for recognition testing. Among them, for each wireless device, the number of training samples is 320, and the number of testing samples is 80).

[0047] Among them, the equipotential star map conversion and recognition process includes:

[0048] Such as Figure 1 As shown, represent the sample of the RF baseband signal as a two-dimensional constellation map; on the two-dimensional coordinate with I as the horizontal axis and Q as the vertical axis, represent the vector endpoints of the RF baseband signal under the specified basis vector projection to obtain a vector map. After all vector endpoints are represented, the two-dimensional constellation map corresponding to the RF baseband signal is obtained. I is the in-phase signal of the RF baseband signal, which is the cos component of the RF baseband signal, and Q is the quadrature signal of the RF baseband signal, which is the sin component of the RF baseband signal.

[0049] According to the density distribution of the vector endpoints in the two-dimensional constellation map, distinguish different density regions by color to obtain an equipotential star map, such as Figure 2The equipotential planet diagram of one of the devices is shown below. Different density regions are color-coded, including: gradient coloring according to the density difference at the vector endpoints, and then adjusting the brightness of the coloring according to the density difference, where the brightness decreases as the density increases. The colored image is more distinguishable than the uncolored one. However, to further improve the distinguishability of its features, the form of adjusting brightness is adopted here, making the density difference more obvious, adding differences in brightness on the basis of different colors, and highlighting the features of each image.

[0050] Use a pre-trained neural network to identify the equipotential planet diagram and obtain the individual identification result of the communication radiation source based on the features of the equipotential planet diagram.

[0051] The neural network in this embodiment is a deep convolutional neural network. After generating the equipotential planet diagram, 6,400 samples are randomly selected for training the deep convolutional neural network, and the remaining 1,600 samples are used for identification testing. For each wireless device, the number of training samples is 320, and the number of testing samples is 80.

[0052] The designed deep convolutional neural network structure is shown in Table 1.

[0053] Network layer Parameter structure Input layer 227×227×3 Convolutional layer 55×55×96 Pooling layer 27×27×96 Normalization layer 27×27×96 Convolutional layer 27×27×256 Pooling layer 13×13×256 Normalization layer 13×13×256 Convolutional layer 13×13×384 Convolutional layer 13×13×384 Convolutional layer 13×13×256 Pooling layer 6×6×256 Normalization layer 6×6×256 Fully connected layer 9 216 Fully connected layer 4 096 Fully connected layer 4 096 Output layer 20

[0054] Table 1 Deep convolutional neural network structure

[0055] Finally, through the identification and authentication of the deep convolutional neural network, the individual identification results of the communication radiation source based on the features of the differential constellation trajectory diagram and the individual identification results of the communication radiation source based on the features of the equipotential planet diagram are obtained respectively. The test results are as shown in Figure 3 and Figure 4 It can be seen from the figure that the success rate of individual identification of the communication radiation source based on the features of the differential constellation trajectory diagram is 88.6%, while the success rate of individual identification of the communication radiation source based on the features of the equipotential planet diagram is 90.4%. This shows that under the same deep convolutional neural network model architecture, compared with the traditional statistical graph domain method based on the constellation diagram, this embodiment can improve the identification accuracy to a certain extent without reducing the calculation efficiency (when using a laptop with a 4.0 GHz dual-processor, the average calculation time for each identification of the deep convolutional neural network model does not exceed 20 ms).

[0056] The differential equipotential planet diagram conversion and identification process includes:

[0057] Assume that the radio frequency signal emitted by an individual communication radiation source where t is the sampling point position, x(t) is the transmitter baseband signal, f tis the carrier frequency of the transmitter. If the RF circuit of an individual communication radiation source is ideal and the channel is also ideal, the signal r(t) received by the receiver is equal to s(t).

[0058] The receiver down-converts the signal to obtain the baseband signal where f r is the carrier frequency of the receiver, and is the phase error when the receiver receives the signal.

[0059] When f r ≠ f t , the baseband signal obtained by the receiver's down-conversion is where θ = f r - f t . Since the demodulated signal contains a residual frequency deviation θ, there is a phase rotation factor e j2πθt for each sampling point of the baseband signal. Since this phase rotation factor varies with the position t of the sampling point, it will cause the overall rotation of the constellation trajectory diagram.

[0060] In most coherent demodulation communication systems, estimating the frequency deviation and phase deviation can obtain the estimated frequency deviation and phase deviation The receiver uses the estimated results to compensate the received signal for frequency deviation and phase deviation, thereby obtaining a stable constellation diagram. In the method of extracting RF fingerprints based on the constellation diagram, since the purpose of the receiver is not to accurately demodulate each received signal modulation symbol, the received signal can be differentially processed at a certain interval n to obtain a more stable constellation diagram. The method of differential processing is

[0061]

[0062] where: d(t) is the signal after differential processing; y * is the conjugate value of y; n is 1. Although the signal d(t) after differential processing still contains a phase rotation factor e -j2πθn ; however, this phase rotation factor is a constant value and does not change with the change of the sampling point position. Therefore, the new I and Q two-channel signals after differential processing only contain a phase rotation factor with a constant value, and a relatively stable constellation diagram can be obtained without estimating and compensating the carrier frequency deviation and phase deviation of the receiver. Figure 5 In, according to the different densities of the two-dimensional differential constellation diagram points, different colors are assigned to different regions, and the one-dimensional signal is converted into a two-dimensional color image (like an ultra-high-definition X-ray film), which more comprehensively describes the subtle features of the signal.

[0063] Still retrain using a depth convolutional neural network with the same structure for recognition and authentication, and respectively obtain the individual recognition results of communication radiation sources based on the equipotential planet map features without differential processing, the individual recognition results of communication radiation sources based on differential constellation trajectory map features, and the individual recognition results of communication radiation sources based on the method proposed in this paper. The recognition success rate obtained is 98.6%.

[0064] From the description of the above embodiments, those skilled in the art can understand that for the convenience and conciseness of description, only the division of the above functional modules is used as an example. In actual applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the specific device is divided into different functional modules to complete all or part of the functions described above.

[0065] In the embodiments provided in this application, it should be understood that the disclosed structures and methods can be implemented in other ways. For example, the embodiments of the structure described above are only illustrative. For example, the division of modules or units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another structure, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of structures or units can be in electrical, mechanical or other forms.

[0066] The units described as separate components may or may not be physically separated. The components displayed as units may be one physical unit or multiple physical units, that is, they can be located in one place, or they can be distributed to multiple different places. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0067] In addition, each functional unit in the embodiments of this application can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0068] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solution of the embodiments of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The software product is stored in a storage medium and includes several instructions for causing a device (which can be a single-chip microcomputer, a chip, etc.) or a processor to execute all or part of the steps of the methods of the embodiments of the present application. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks, or optical discs.

[0069] The above content is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered by the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A network authentication method based on radio frequency signal gene analysis, characterized in that Including the following steps: S01: Collect the radio frequency baseband signals of the communication radiation source individuals to be identified; S02: Segment the collected radio frequency baseband signals and convert them into equipotential star maps respectively; S03: Use the pre-trained neural network to identify the equipotential star maps from the same radio frequency baseband signal respectively, and obtain the identification results of the communication radiation source individuals based on the characteristics of the equipotential star maps; S04: Judge whether the identification results are consistent. If they are consistent, use the identification results as the authentication results. If they are not consistent, execute the next step; S05: Convert the whole radio frequency baseband signal into a differential equipotential star map, and use the pre-trained neural network for identification to obtain the identification results of the communication radiation source individuals based on the characteristics of the differential equipotential star map, and use this identification result as the authentication result.

2. The network authentication method based on radio frequency signal gene analysis according to claim 1, wherein, The conversion process of the equipotential star map includes: Segment the radio frequency baseband signal and represent it as a two-dimensional constellation map; According to the density distribution of the vector endpoints in the two-dimensional constellation map, distinguish the colors of different density regions to obtain the equipotential star map.

3. The network authentication method based on radio frequency signal gene analysis according to claim 2, wherein The process of representing as a two-dimensional constellation map includes: On the two-dimensional coordinate with I as the horizontal axis and Q as the vertical axis, represent the vector endpoints of the radio frequency baseband signal under the projection of the specified basis vector to obtain a vector diagram. After all the vector endpoints are represented, obtain the two-dimensional constellation map corresponding to the radio frequency baseband signal; where I is the in-phase signal of the radio frequency baseband signal, which is the cos component of the radio frequency baseband signal, and Q is the quadrature signal of the radio frequency baseband signal, which is the sin component of the radio frequency baseband signal.

4. A network authentication method based on radio frequency signal gene analysis according to claim 1, characterized in that, The neural network is a deep convolutional neural network, and the training process is: Select a part from the equipotential star maps as training samples, and the rest as identification samples. Set the parameter structure of the deep convolutional neural network, and then import the training samples for training to obtain the trained deep convolutional neural network.

5. The network authentication method based on radio frequency signal gene analysis according to claim 1 or 2, characterized in that The conversion process of the differential equipotential star map includes: Perform differential and normalization processing on the radio frequency baseband signal to obtain a differential constellation map; According to the point density distribution in the differential constellation map, distinguish the colors of different density regions to obtain the differential equipotential star map.

6. The network authentication method based on radio frequency signal gene analysis according to claim 5, wherein The process of differential processing includes: For the radio frequency baseband signal where t is the sampling point position, x(t) is the transmitter baseband signal, and f t is the transmitter carrier frequency; The differential processing is expressed as: Where: d(t) is the signal after differential processing, y * is the conjugate value of y; n is 1; e -j2πθn is the phase rotation factor; r(t) is the signal received by the receiver and r(t) = s(t), f r is the carrier frequency of the receiver, is the phase error when the receiver receives the signal.

7. The network authentication method based on radio frequency signal gene analysis according to claim 1, wherein The process of step S01 includes: Use a spectrum analyzer to collect the radio frequency baseband signals of the communication radiation source individuals to be identified at a preset sampling frequency and sampling time respectively.

8. The network authentication method based on radio frequency signal gene analysis according to claim 2 or 5, characterized in that, The distinguishing of the colors of different density regions includes: Gradually color according to the density difference of the vector endpoints, and then adjust the brightness of the coloring according to the density difference, where the brightness decreases as the density increases.

Images