A data-independent anti-multipath radio frequency fingerprint extraction method and system

Abstract

The present application relates to a kind of data-independent anti-multipath radio frequency fingerprint extraction method and system, for extracting the radio frequency fingerprint of wireless device.The method implementation steps include: sampling and preprocessing to wireless signal, obtain received signal;The time-domain signal of received signal is transformed in frequency domain and logarithm operation to obtain cepstrum domain signal;Select two frames of signals experiencing channel incoherence to do cross-correlation operation and further remove the correlation caused by local signal, as radio frequency fingerprint expression;The cross-correlation value between multiple symbols is superimposed and averaged to enhance fingerprint, while eliminating noise.The method of the present application effectively decouples the radio frequency fingerprint component in wireless signal from wireless multipath channel component, varying transmission data component, while enhancing the representation of radio frequency fingerprint, improving the stability of fingerprint, and suppressing noise, which is an effective data-independent anti-multipath radio frequency fingerprint extraction method for any wireless transmission signal.

Description

Technical Field

[0001] This invention relates to the fields of wireless communication and physical layer security, and particularly to a data-independent method and system for multipath-resistant radio frequency fingerprint extraction. Background Technology

[0002] With the development of wireless communication, especially the widespread adoption of 4G and 5G technologies, the secure transmission of information in wireless communication systems is more vulnerable to eavesdropping, forgery, or tampering. Compared to traditional cryptographic authentication technologies, physical layer security-based authentication mechanisms have lower complexity and network overhead, and can achieve seamless authentication without affecting normal communication or relying on third-party plugins. Currently, Radio Frequency Fingerprint (RFF) technology has become an important technology in the field of physical layer security. RFF fingerprints are inherent hardware information carried by wireless devices, mainly generated by factors such as the tolerances of electronic components in the circuit, filter characteristics, and nonlinearities caused by power amplifiers and mixers. Since the hardware characteristics of each device are unique and cloning is difficult, RFF-based authentication technology fully utilizes the hardware differences between different devices to identify them and resist identity attacks and deception.

[0003] Every wireless signal is transmitted according to a prescribed frame format, which includes both fixed-structure signals such as preambles and random-sequence signals such as data symbols. The radio frequency fingerprint of each terminal is implicit in these transmitted signals.

[0004] In modern 4G and 5G wireless communication systems, wireless terminals often move at high speeds, and the wireless channels that signals traverse also change rapidly. When the amount of data transmitted is large, the signal typically has a larger bandwidth, making it more susceptible to the effects of multipath channels. However, existing methods have not yet proposed an effective RF fingerprint extraction method for broadband signals in variable channel environments. Furthermore, most existing methods are designed for extracting RF fingerprints from preamble symbols with fixed structures, which limits their application to randomly changing data signals in real-world scenarios.

[0005] Therefore, there is an urgent need to propose a data-independent, multipath-resistant radio frequency fingerprint extraction method that is applicable to any wireless signal. This method needs to overcome the influence of wireless multipath channels while extracting stable and unchanging radio frequency fingerprints from randomly changing data. Summary of the Invention

[0006] Objective of the Invention: To address the problems existing in the prior art, this invention provides a data-independent, multipath-resistant radio frequency fingerprint extraction method and system for extracting radio frequency fingerprints from wireless devices. This technical solution selects two frames of signals that have experienced incoherent channels. By performing cross-correlation operations in the cepstral domain, the influence of wireless multipath channels and varying data portions on radio frequency fingerprint extraction is eliminated. Furthermore, by averaging the cross-correlation values ​​between multiple symbols, the fingerprint is enhanced and noise is suppressed, thereby achieving robust radio frequency fingerprint extraction.

[0007] Technical Solution: To achieve the above technical objectives, the present invention provides a data-independent, multipath-resistant radio frequency fingerprint extraction method comprising the following steps:

[0008] (1) The receiver samples the radio signal, then preprocesses each symbol or sequence segment to complete time synchronization and carrier frequency offset estimation and compensation, obtaining the preprocessed time-domain received signal y(n), which is expressed as:

[0009] y(n) = x(n) * h T (n)*h(n)*h R (n)+z(n)

[0010] Where n = 0, 1, ..., M-1, M is the number of sampling points for a symbol or a sequence, x(n) is the local ideal signal, and h T (n) is the equivalent filter at the transmitting end, h R Let h(n) be the receiver's equivalent filter, h(n) be the wireless channel, and z(n) be the additive noise. The combined equivalent filters at the transmitter and receiver can be approximated as the radio frequency fingerprint rff(n), i.e., rff(n) = h(n). T (n)*h R (n). When the signal-to-noise ratio is high enough, the noise part can be ignored, and the above equation can be simplified to:

[0011] y(n) = x(n) * rff(n) * h(n)

[0012] (2) The frequency domain signal Y(k) is obtained by performing a frequency domain transformation on the time-domain received signal y(n), which is expressed as follows:

[0013] Y(k)=X(k)RFF(k)H(k)

[0014] Where k = 0, 1, ..., M-1; X(k), RFF(k), and H(k) are the frequency domain expressions of the local ideal signal, the radio frequency fingerprint, and the wireless channel, respectively. Further, the frequency domain signal Y(k) is transformed into the cepstral domain signal C(k) by performing a logarithmic operation, and the transformation formula is:

[0015] C(k) = ln Y(k) = ln|X(k)RFF(k)H(k)| + j[argX(k)RFF(k)H(k) + 2lπ], l = 0, ±1, ±2, … where ln[·] represents the logarithmic operation, |·| represents the absolute value operation, and arg[·] represents the complex argument operation. Further expansion of the real and imaginary parts of the above cepstral domain signal C(k) transforms the multiplicative relationship between the channel and the fingerprint into an additive relationship, specifically expressed as:

[0016] C (k) = ln |

[0017] (3) Select two subframes with a time interval of τ, where τ > Δ, and Δ is the minimum incoherence time of the channel. At this time, there is no correlation between the channels experienced by the two frames of signals. Perform cross-correlation operation on the cepstral domain signals C1(k) and C2(k) obtained by the above steps (1)-(2) for the two frames of signals. The specific formula is as follows:

[0018]

[0019] Where t = 0, 1, ..., M-1; corr[·] represents complex conjugate correlation operation; C1(k) and C2(k) represent the cepstral domain signals corresponding to the two selected frames of signals, respectively; Y1(k) and Y2(k) correspond to the frequency domain received signals obtained by the logarithmic operation of C1(k) and C2(k); m and t are index values, used to obtain the values ​​of the m-th and mt-th points in Y1(k) and Y2(k). Further, it can be expanded as follows:

[0020] R(t)=corr[lnX1(k),lnX2(k)]+corr[lnX1(k),lnRFF(k)]+corr[lnX1(k),lnH2(k)]

[0021] +corr[lnRFF(k),lnX2(k)]+corr[lnRFF(k),lnRFF(k)]

[0022] +corr[lnRFF(k),lnH2(k)]+corr[lnH1(k),lnX2(k)]

[0023] +corr[lnH1(k),lnRFF(k)]+corr[lnH1(k),lnH2(k)]

[0024] Since the two frames are incoherent, the correlation between the channels, corr[lnH1(k),lnH2(k)], is approximately 0. Furthermore, there is no correlation between channels H1(k), H2(k) and fingerprint RFF(k), channels H1(k), H2(k) and data X1(k), X2(k), or fingerprint RFF(k) and data X1(k), X2(k). Therefore, the above equation can be approximately expressed as:

[0025] R(t)≈corr[lnX1(k),lnX2(k)]+corr[lnRFF(k),lnRFF(k)]

[0026] (4) Based on the content of the local signal, the above cross-correlation operation results are further processed in the following two cases to remove the correlation between the local signals X1(k) and X2(k) to obtain the radio frequency fingerprint expression F(k).

[0027] (4.1) When X1(k) and X2(k) transmit the same data or a fixed sequence such as a preamble symbol, corr[lnX1(k),lnX2(k)] is a fixed value. Therefore, the result of subtracting this fixed value from R(t) is used as the RF fingerprint expression.

[0028] F(k)=R(t)-corr[lnX1(k),lnX2(k)]

[0029] (4.2) When X1(k) and X2(k) are transmitted random data, the correlation between the random data is very low, corr[lnX1(k),lnX2(k)] is approximately 0, and R(t) can be approximated as the RF fingerprint expression.

[0030] F(k)=R(t)≈corr[lnRFF(k),lnRFF(k)]

[0031] (5) The cross-correlation operation performed in steps (3)-(4) above is performed on two symbols. When a frame of signal has L symbols, L cross-correlation operations can be performed between them. 2 For secondary correlation, the fingerprint after superposition and averaging is used. To further enhance fingerprint characterization, improve the stability of radio frequency fingerprints, and eliminate the influence of noise, it is specifically expressed as follows:

[0032]

[0033] Where i = 1, 2, ..., L 2 F i (k) represents the radio frequency fingerprint obtained by performing the i-th correlation.

[0034] The present invention also provides a data-independent anti-multipath radio frequency fingerprint extraction system, including...

[0035] The data frame acquisition module is used to receive wireless signals and sample them to obtain time-domain baseband signals.

[0036] The signal preprocessing module is used to perform time synchronization and carrier frequency offset estimation and compensation on each symbol or sequence in the time-domain baseband signal to obtain the preprocessed time-domain received signal y(n);

[0037] The radio frequency fingerprint extraction module is used to perform frequency domain transformation and logarithmic operation on the time domain received signal y(n) to obtain the cepstral domain signal C(k). Then, two frames of signals that have experienced channel incoherence are selected for cross-correlation operation and the correlation caused by local signals is further removed, which is used as the radio frequency fingerprint expression F(k).

[0038] The fingerprint enhancement module is used to superimpose and average multiple symbols of radio frequency fingerprints to enhance the fingerprint while eliminating noise.

[0039] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages:

[0040] The method of this invention effectively decouples the radio frequency fingerprint component in the wireless signal from the variable wireless channel components and changing data components by performing cross-correlation operations in the cepstral domain. This eliminates the influence of channel and data on radio frequency fingerprint extraction, achieving the purpose of data independence and multipath resistance in radio frequency fingerprints. Furthermore, by superimposing and averaging the cross-correlation values ​​between multiple symbols, the stability of the radio frequency fingerprint is enhanced, and noise is suppressed. This method greatly improves the radio frequency fingerprint recognition performance of wireless devices under conditions of variable channel and data. Attached Figure Description

[0041] Figure 1 This is a flowchart illustrating a data-independent, multipath-resistant radio frequency fingerprint extraction method in one embodiment.

[0042] Figure 2 This is a schematic diagram of the PSBCH synchronization subframe structure of the LTE-V2X vehicle-to-everything (V2X) communication standard in one embodiment;

[0043] Figure 3 This is an example of an RF fingerprint extracted from DMRS symbol packets in the PSBCH synchronization subframe signal by four different LTE-V2X communication devices. Waveform comparison chart. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and are not intended to limit the scope of this patent application.

[0045] Considering that in actual wireless communication environments, channels are complex and variable, and the data transmitted by wireless devices in actual communication is also constantly changing, it is difficult to extract a stable and unchanging radio frequency fingerprint from the mixed wireless multipath channels and changing data signals. The extracted radio frequency fingerprint often contains a mixture of channel information and actual communication data information.

[0046] Therefore, the present invention provides a data-independent anti-multipath radio frequency fingerprint extraction method applicable to any wireless signal, which can decouple the radio frequency fingerprint component in the user's transmitted signal from the wireless multipath channel component and the changing data component, and extract the radio frequency fingerprint that does not contain channel information and data information.

[0047] The data-independent, multipath-resistant radio frequency fingerprint extraction method provided in this application can be applied to wireless network communication devices with any frame structure, such as Wi-Fi and 4G LTE. In this embodiment, taking a vehicle-to-everything (V2X) communication device as an example, the device includes a transmitter and a receiver. The receiver receives the radio frequency signal transmitted by the transmitter and uses the radio frequency fingerprint extraction method provided in this embodiment to extract the corresponding device's radio frequency fingerprint from the received radio frequency signal. This radio frequency fingerprint can be used for identity recognition or access authentication, etc.

[0048] In one embodiment, such as Figure 1 As shown, a data-independent, multipath-resistant radio frequency fingerprint extraction method is provided, comprising the following steps:

[0049] (1) The wireless signal is sampled and preprocessed to obtain the received signal.

[0050] In this embodiment, as Figure 2 As shown, taking the synchronization subframe structure of the Physical Sidelink Broadcast Channel (PSBCH) in the LTE-V2X communication standard for vehicle-to-everything (V2X) as an example, the subframe structure adopts the resource block approach, which contains 7 fixed sequence symbols, representing the primary sidelink synchronization signal (PSSS), secondary sidelink synchronization signal (SSSS), demodulation reference signal (DMRS), etc., and also contains 6 PSBCH symbols with randomly varying data.

[0051] In this embodiment, the receiver samples the baseband signal of the PSBCH synchronization subframe at a sampling rate of 30.72 Mbps. Since one PSBCH synchronization subframe is 1 ms long, the entire subframe corresponds to 30720 sampling points. Subsequently, preprocessing is performed on each of the 13 symbols (7 fixed sequence symbols and 6 PSBCH symbols) to complete time synchronization, carrier frequency offset estimation, and compensation, resulting in the preprocessed time-domain received signal y(n), which is expressed as...

[0052] y(n) = x(n) * h T (n)*h(n)*h R (n)+z(n)

[0053] Where n = 0, 1, ..., M-1, M is the number of sampling points for one symbol, x(n) is the local ideal signal, and h T (n) is the equivalent filter at the transmitting end, h R y(n) is the equivalent filter at the receiver, h(n) is the wireless channel, and z(n) is the additive noise. It should be noted that the value of M is 62 when processing PSSS and SSSS symbols, and 72 when processing DMRS and PSBCH symbols. During the process of the radio frequency signal being transmitted from the transmitter of the vehicle-to-everything (V2X) communication device through the wireless multipath channel and then received by the receiver, the transmitted resource block data x(n) and the wireless multipath channel h(n) are both included in y(n). The equivalent filters at the transceiver end can be approximated as the radio frequency fingerprint rff(n), i.e., rff(n) = h(n). T (n)*h R (n). Therefore, the received signal y(n) of each symbol in the PSBCH synchronization subframe contains the device's data information, channel information, and RF fingerprint information. In this example, the signal-to-noise ratio of the received signal is high, exceeding 20dB, and the noise component is negligible. The above formula can be simplified to...

[0054] y(n) = x(n) * rff(n) * h(n)

[0055] (2) The cepstral domain signal is obtained by performing frequency domain transformation and logarithmic operation on the time domain signal y(n) to transform the multiplicative relationship between the channel and the fingerprint into an additive relationship.

[0056] In this embodiment, for each symbol of the PSBCH synchronization subframe, the simplified time-domain signal y(n) is transformed in the frequency domain to obtain the frequency-domain signal Y(k), which is represented as follows:

[0057] Y(k)=X(k)RFF(k)H(k)

[0058] Where k = 0, 1, ..., M-1. When transforming the PSSS and SSSS symbols, the value of M is 62; when transforming the DMRS and PSBCH symbols, the value of M is 72. X(k), RFF(k), and H(k) are the frequency domain expressions of the local ideal signal, RF fingerprint, and wireless channel, respectively. Further, the frequency domain signal Y(k) is transformed into the cepstral domain signal C(k) by performing a logarithmic operation, and the transformation formula is:

[0059] C(k)=ln Y(k)=ln|X(k)RFF(k)H(k)|+j[argX(k)RFF(k)H(k)+2lπ], l=0,±1,±2,…

[0060] Where ln[·] represents the logarithmic operation, |·| represents the absolute value operation, and arg[·] represents the complex argument operation; the real and imaginary parts of the above cepstral domain signal C(k) are further expanded to transform the multiplicative relationship between the channel and the fingerprint into an additive relationship, specifically expressed as follows:

[0061] C (k) = ln |

[0062] (3) Select two frames of signals that have experienced uncorrelated channels, perform cross-correlation operation, and further remove the correlation caused by local signals to obtain the radio frequency fingerprint expression.

[0063] In this embodiment, two PSBCH synchronization subframes with a time interval exceeding 2 seconds are selected. This time interval far exceeds the minimum incoherence time of the channel. Moreover, the vehicle-to-everything (V2X) communication device used in this embodiment is in motion, and the two selected signal frames are located in different positions and environments. Therefore, the channels experienced by the two signal frames are not correlated. Cross-correlation is performed on the cepstral domain signals C1(k) and C2(k) obtained from the two signal frames according to steps (1)-(2) above. The specific formula is as follows:

[0064]

[0065] Where t = 0, 1, ..., M-1. When cross-correlation is performed between PSSS and SSSS symbols, the value of M is 62; when cross-correlation is performed between DMRS and PSBCH symbols, the value of M is 72. corr[·] represents complex conjugate correlation operation; C1(k) and C2(k) represent the cepstral domain signals corresponding to the two selected frames of signals, respectively; Y1(k) and Y2(k) correspond to the frequency domain received signals obtained by the logarithmic operation of C1(k) and C2(k); m and t are index values, used to obtain the values ​​of the m-th and mt-th points in Y1(k) and Y2(k). Further, it can be expanded as follows:

[0066] R(t)=corr[lnX1(k),lnX2(k)]+corr[lnX1(k),lnRFF(k)]+corr[lnX1(k),lnH2(k)]

[0067] +corr[lnRFF(k),lnX2(k)]+corr[lnRFF(k),lnRFF(k)]

[0068] +corr[lnRFF(k),lnH2(k)]+corr[lnH1(k),lnX2(k)]

[0069] +corr[lnH1(k),lnRFF(k)]+corr[lnH1(k),lnH2(k)]

[0070] Since the two frames are incoherent, the correlation between the channels, corr[lnH1(k),lnH2(k)], is approximately 0. Furthermore, there is no correlation between channels H1(k), H2(k) and fingerprint RFF(k), channels H1(k), H2(k) and data X1(k), X2(k), or fingerprint RFF(k) and data X1(k), X2(k). Therefore, the above equation can be approximately expressed as:

[0071] R(t)≈corr[lnX1(k),lnX2(k)]+corr[lnRFF(k),lnRFF(k)]

[0072] Furthermore, in this embodiment, the cross-correlation operation results will be further processed according to the content of the local signal x(n) of each symbol in the PSBCH synchronization subframe in the following two cases to remove the correlation between the local signals X1(k) and X2(k) and thus obtain the radio frequency fingerprint expression F(k).

[0073] (3.1) When X1(k) and X2(k) are fixed sequence symbols of the three types, PSSS, SSSS, and DMRS, corr[lnX1(k),lnX2(k)] is a fixed value. Therefore, the result of subtracting this fixed value from R(t) is used as the radio frequency fingerprint expression.

[0074] F(k)=R(t)-corr[lnX1(k),lnX2(k)]

[0075] (3.2) When X1(k) and X2(k) are PSBCH symbols with random data variation, the correlation between random data is very low, corr[lnX1(k),lnX2(k)] is approximately 0, and R(t) can be approximated as the RF fingerprint expression.

[0076] F(k)=R(t)≈corr[lnRFF(k),lnRFF(k)]

[0077] (4) Average the cross-correlation values ​​between multiple symbols to enhance the fingerprint while eliminating noise.

[0078] In this embodiment, the cross-correlation operation performed in step (3) above is performed on two symbols. When a frame of signal has L symbols, L cross-correlation operations can be performed between them. 2 Secondary correlation, using fingerprints after superposition and averaging. This can further enhance fingerprint characterization, improve the stability of RF fingerprints, and eliminate the influence of noise. Specifically, a PSBCH synchronization subframe contains 13 symbols, of which 7 are fixed sequence symbols and 6 are random data PSBCH symbols. Within the fixed sequence, two PSSS symbols can be correlated 4 times, two SSSS symbols can be correlated 4 times, and three DMRS symbols can be correlated 9 times; the six random data PSBCH symbols can be correlated 36 times. That is, for a PSBCH synchronization subframe with multiple sets of different symbols in a single signal frame, cross-correlation operations can be performed between each set of symbols to enhance the fingerprint. For each set of L symbols, L... 2 The specific expression for the sub-cross-correlation is:

[0079]

[0080] Where i = 1, 2, ..., L 2 F i (k) represents the RF fingerprint obtained from the i-th correlation. When performing cross-correlation on PSSS and SSSS symbols, L is 2; when performing cross-correlation on DMRS symbols, L is 3; and when performing cross-correlation on PSBCH symbols, L is 6. The calculated value for each group of symbols... Combined, they yield the final stable and enhanced RF fingerprint of the PSBCH synchronization subframe.

[0081] Figure 3 This embodiment demonstrates the radio frequency fingerprints extracted from DMRS symbol packets in the PSBCH synchronization subframe signal by four different LTE-V2X communication devices. Waveform comparison chart. Radio frequency fingerprints of 4 devices. By observing and comparing the waveforms side by side, it can be seen that each actual device has different sequences. Each point k in the RF fingerprint is unique, and this difference can be used to distinguish different devices, enabling identity recognition and access authentication. Using the RF fingerprint extraction method provided in this embodiment, four different LTE-V2X communication devices were identified and authenticated at a signal-to-noise ratio of 20dB. In a mobile environment with changing channels, the accuracy rate for identifying legitimate devices reached 99.9%, and the rejection rate for illegitimate devices reached 94.5%.

[0082] Therefore, the data-independent anti-multipath radio frequency fingerprint extraction method provided in this embodiment of the invention can effectively eliminate the influence of wireless multipath channels and randomly changing transmitted data on radio frequency fingerprint extraction. It also suppresses noise and enhances fingerprint characterization by superimposing and averaging the cross-correlation values ​​between multiple symbols, thereby achieving robust data-independent anti-multipath radio frequency fingerprint extraction, which is more conducive to use for identity recognition or access authentication.

[0083] The present invention also provides a data-independent anti-multipath radio frequency fingerprint extraction system, including...

[0084] The data frame acquisition module is used to receive wireless signals and sample them to obtain time-domain baseband signals.

[0085] The signal preprocessing module is used to perform time synchronization and carrier frequency offset estimation and compensation on each symbol or sequence in the time-domain baseband signal to obtain the preprocessed time-domain received signal y(n);

[0086] The radio frequency fingerprint extraction module is used to perform frequency domain transformation and logarithmic operation on the time domain received signal y(n) to obtain the cepstral domain signal C(k). Then, two frames of signals that have experienced channel incoherence are selected for cross-correlation operation and the correlation caused by local signals is further removed, which is used as the radio frequency fingerprint expression F(k).

[0087] The fingerprint enhancement module is used to superimpose and average multiple symbols of radio frequency fingerprints to enhance the fingerprint while eliminating noise.

[0088] It should be understood that although the steps in the flowcharts described in the embodiments above are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise expressly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Executing the steps in the flowchart in different orders is within the scope of this specification.

[0089] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0090] The embodiments described above are merely one implementation method of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this invention should be determined by the appended claims.

Claims

1. A data-independent method for multipath-resistant radio frequency fingerprint extraction, characterized in that, The method includes the following steps: Step 1: Sample and preprocess the wireless signal to obtain the time-domain received signal; Step 2: Perform frequency domain transformation and logarithmic operation on the received time domain signal to obtain the cepstral domain signal, so as to transform the multiplicative relationship between the channel and the fingerprint into an additive relationship; Step 3: Select two frames of signals that have experienced uncorrelated channels, perform cross-correlation calculations, and further remove the correlation caused by local signals to obtain the RF fingerprint expression; Step 4: Average the cross-correlation values ​​between multiple symbols in a frame of signal to enhance the fingerprint while eliminating noise; In step 1, after sampling the wireless signal, the receiver preprocesses each symbol or sequence segment. The preprocessing steps include time synchronization and carrier frequency offset estimation and compensation. The preprocessed time-domain received signal... Represented as ； in M is the number of sampling points for one symbol. For ideal local signal, For the equivalent filter at the transmitting end, For the receiver's equivalent filter, For wireless channels, This is additive noise; where the equivalent filters at the transceiver ends together approximate the RF fingerprint. ,Right now When the signal-to-noise ratio is high, the above expression for the received signal simplifies to: 。 2. The data-independent, multipath-resistant radio frequency fingerprint extraction method according to claim 1, characterized in that, In step 2, the frequency domain transformation is performed on the received signal in the time domain. Frequency domain transformation is performed to obtain the frequency domain signal. , which is represented as ； in ; These are the frequency domain expressions for the local ideal signal, the radio frequency fingerprint, and the wireless channel, respectively; the logarithmic operation is performed on the frequency domain signal. Perform logarithmic operations to transform it into a cepstral domain signal. Its transformation formula is ； in, This indicates the logarithmic operation. This indicates taking the absolute value. This represents the operation of taking the complex argument; for the above cepstral domain signal The real and imaginary parts are further expanded to transform the multiplicative relationship between the channel and the fingerprint into an additive relationship, specifically represented as follows: 。 3. The data-independent, multipath-resistant radio frequency fingerprint extraction method according to claim 2, characterized in that, In step 3, the two selected signal frames have a time interval of [missing information]. The two subframes, in which , This is the minimum incoherence time of the channel, at which point the two frames of signals experience no correlation between the channels; the cross-correlation operation is performed on the cepstral domain signals of the selected two frames of signals. and The specific formula for cross-correlation is as follows: ； in ; Represents the conjugate operation of complex numbers; and These represent the cepstral domain signals corresponding to the two selected signal frames, respectively. , Corresponding to and The frequency domain received signal from which the logarithmic operation is performed, where m and t are index values, is used to obtain... , The cross-correlation formula for the values ​​of the m-th and mt-th points is approximately expressed as follows: 。 4. The data-independent, multipath-resistant radio frequency fingerprint extraction method according to claim 3, characterized in that, In step 3, further removing the correlation caused by local signals involves processing the cross-correlation results in two ways, depending on the content of the local signals, to remove the local signals. The correlation is used to obtain the radio frequency fingerprint expression. The specific operation is as follows: (4.1) When When the transmitted data is the same or a preamble symbol, For a fixed value, Subtracting this fixed value results in the radio frequency fingerprint expression. ； (4.2) When When the data is random for transmission, Approximation as an expression for radio frequency fingerprints 。 5. The data-independent, multipath-resistant radio frequency fingerprint extraction method according to claim 1, characterized in that, In step 4, for a signal frame with L symbols, L is performed between the multiple symbols. 2 Secondary correlation, using fingerprints after superposition and averaging. The specific formula for averaging the cross-correlation values ​​among multiple symbols is as follows: ； in , Indicates doing the first The radio frequency fingerprint obtained by secondary correlation.

6. A data-independent, multipath-resistant radio frequency fingerprint extraction system for implementing the method of any one of claims 1-5, characterized in that, include The data frame acquisition module is used to receive wireless signals and sample them to obtain time-domain baseband signals. The signal preprocessing module is used to perform time synchronization and carrier frequency offset estimation and compensation on each symbol or sequence in the time-domain baseband signal to obtain the preprocessed time-domain received signal. ; The radio frequency fingerprint extraction module is used to analyze the received signal in the time domain. Frequency domain transformation and logarithmic operation are performed to obtain the cepstral domain signal. Subsequently, two frames of signals that have experienced channel incoherence are selected, cross-correlation is performed, and the correlation caused by local signals is further removed to obtain the radio frequency fingerprint expression. ; The fingerprint enhancement module is used to superimpose and average multiple symbols of radio frequency fingerprints to enhance the fingerprint while eliminating noise.

Images