Radio frequency fingerprint extraction method and system using logarithmic difference characteristics of channels at different moments

Abstract

The invention discloses a radio frequency fingerprint extraction method and system using logarithmic difference characteristics of channels at different moments, and the method and system are used for extracting radio frequency fingerprints of wireless equipment, and the method comprises the following steps: carrying out the collection and preprocessing of a wireless signal, obtaining a time domain receiving signal, and carrying out the frequency domain transformation of the time domain receiving signal, and obtaining a frequency domain receiving signal; selecting two sections of receiving symbols which pass through different channels and do not generate deep fading to perform differential operation and logarithmic differential operation, and further performing division operation to obtain radio frequency fingerprints without channels; and superposing and averaging multiple groups of results after division operation to eliminate the influence of noise, and finally obtaining radio frequency fingerprints which do not contain channels and noise. According to the radio frequency fingerprint extraction method, the difference and logarithmic difference characteristics of channels at different moments are utilized, the influence of time-varying channels on radio frequency fingerprint extraction is effectively eliminated, meanwhile, noise is restrained, the stability of radio frequency fingerprint extraction under the channel variable environment is improved, and the radio frequency fingerprint extraction method is effective and universal for various wireless signals.

Description

Technical Field

[0001] The present invention belongs to the field of physical layer security technology, and in particular relates to a radio frequency fingerprint extraction method and system using the logarithmic differential characteristics of channels at different times. Background Art

[0002] With the development of mobile communication technology, especially the maturity of 5G and 6G technology, the current society has entered the era of the Internet of Everything and the Intelligent Connection of Everything. However, in wireless communication systems, the identity information of users and the transmission content between devices are easily tampered with and hijacked, which will pose a serious threat to personal and property safety. Traditional cryptography-based authentication technology usually uses encryption algorithms or authentication servers to achieve information confidentiality, but it is difficult to resist the brute force attack of modern quantum computers. On the contrary, the Radio Frequency Fingerprint (RFF) technology, an identity authentication mechanism based on physical layer security, is more reliable. The radio frequency fingerprint is naturally random and unique to the device. The hardware characteristics of each device are unique, difficult to clone, and are not affected by changes in the channel or position. Therefore, it can be used as an identity identifier of the device. When the device transmits a wireless signal, the difference in its hardware causes different devices to produce different radio frequency response parameters, and the radio frequency fingerprint characteristics are parasitic in these transmitted signals. In addition, the radio frequency fingerprint authentication technology does not need to change the architecture of the original communication system, and can achieve non-perceptual authentication.

[0003] However, in mobile communication systems, wireless terminals are often in a mobile environment with time-varying channels, which causes certain interference to the extraction of stable RF fingerprint features from the received signal, thereby reducing the authentication performance. In recent years, various RF fingerprint extraction methods have been proposed for signals of different communication protocols, but they still cannot guarantee robustness in time-varying channel environments, and the extracted RF fingerprint features are still mixed with channel and noise components. Therefore, it is urgent to propose a universal and unified RF fingerprint extraction method for all types of wireless signals, which can not only eliminate the influence of channels and noise on RF fingerprint extraction, but also retain stable RF fingerprint features for identity authentication. Summary of the invention

[0004] The purpose of the present invention is to provide a radio frequency fingerprint extraction method and system using the logarithmic differential characteristics of channels at different times, by selecting two segments of symbols at different times to perform differential operations and logarithmic differential operations in the frequency domain, thereby eliminating the influence of time-varying channels on radio frequency fingerprint extraction, and by superimposing and averaging the operation results of multiple groups of symbols to suppress noise, ultimately obtaining radio frequency fingerprint features that do not contain channels and noise, so as to solve the technical problems mentioned in the background technology.

[0005] To solve the above technical problems, the present invention provides a radio frequency fingerprint extraction method and system using the logarithmic differential characteristics of channels at different times, eliminating the influence of time-varying channels and noise on radio frequency fingerprint extraction, so as to obtain stable radio frequency fingerprint features.

[0006] A radio frequency fingerprint extraction method using the logarithmic differential characteristics of channels at different times comprises the following steps:

[0007] Step 1: collect and preprocess the wireless signal to obtain a preprocessed time domain received signal y(n), and further transform the time domain received signal y(n) into a frequency domain to obtain a frequency domain received signal Y(k);

[0008] Step 2: Select the frequency domain received signal Y of two segments of symbols that have experienced different channels and do not produce deep fading 1 (k) and Y 2 (k) performing a difference operation and a logarithmic difference operation, and further performing a division operation to obtain a radio frequency fingerprint F(k) without a channel;

[0009] Step 3: Superimpose and average the radio frequency fingerprints F(k) obtained by step 2 for multiple groups of symbols to eliminate the influence of noise and obtain the final radio frequency fingerprint without channel and noise.

[0010] Furthermore, in step 1, the wireless signal acquisition includes the receiver sampling the wireless signal to obtain an initial received signal; the wireless signal preprocessing includes selecting symbols for RF fingerprint extraction from the initial received signal, completing time synchronization and carrier frequency offset estimation and compensation; the preprocessed time domain received signal y(n) is expressed as

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

[0012] Where n=0, 1, ..., N-1, N is the number of sampling points of the selected symbol, x(n) is the local ideal signal corresponding to the selected symbol, r(n) is the radio frequency fingerprint, h(n) is the wireless channel, and "*" represents the convolution operation; the frequency domain transformation includes performing a frequency domain transformation on the time domain received signal y(n) to obtain a frequency domain received signal Y(k), which is expressed as

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

[0014] Where k=0, 1, ..., N-1, X(k), R(k), and H(k) are the frequency domain expressions of the local ideal signal, the radio frequency fingerprint, and the wireless channel, respectively.

[0015] Further, in step 2, selecting two segments of symbols that experience different channels and do not produce deep fading includes selecting two segments of received symbols with a time interval of τ, where i>Δ, Δ is the minimum incoherence time of the channel, and the channel H experienced by the two segments of symbols at this time 1 (k) and H 2 (k) have no correlation, so they have different channel characteristics; select the frequency domain received signal Y of the two segments of symbols that do not produce deep fading 1 (k) and Y 2 (k) includes selecting two segments of received symbols whose frequency domain channel is flat and whose frequency domain channel amplitude fluctuates within the range of ±TH, where TH is the maximum value of the frequency domain channel amplitude fluctuation under the premise that deep channel fading does not occur, wherein the frequency domain channel is written in the form of H(k)=1+ΔH(k), that is, |ΔH(k)|<TH, |·| represents absolute value operation; the frequency domain received signal Y of the two segments of symbols 1 (k) and Y 2 (k) After the preprocessing and frequency domain transformation in step 1, the two symbols can be two symbols of the same length in the same frame, or two symbols of the same length belonging to different frames. The corresponding local ideal signals are X 1 (k) and X 2 (k).

[0016] Furthermore, when the local ideal signals corresponding to the two selected symbol segments are the same, that is, both are X(k), in step 2, the two frequency domain received signals Y 1 (k) and Y 2 (k) performs a differential operation, which is specifically expressed as

[0017] Y 1 (k)-Y 2 (k) = X(k)R(k)[H 1 (k)-H 2 (k)]

[0018] For two frequency domain received signals Y 1 (k) and Y 2 (k) Perform logarithmic difference operation, which is specifically expressed as

[0019]

[0020] Where ln[·] represents the logarithmic operation;

[0021] ln[1+ΔH(k)] is approximately equivalent to ΔH(k), so the above formula can be expressed as

[0022] LqCy 1 (k)-lnH 2 (k) = ln[1 + ΔH 1(k)]-ln[1+ΔH 2 (k)]≈ΔH 1 (k)-ΔH 2 (k)

[0023] =H 1 (k)-H 2 (k)

[0024] The two frequency domain received signals Y 1 (k) and Y 2 The difference operation result of (k) is divided by the two-segment frequency domain received signal Y 1 (k) and Y 2 The logarithmic difference operation result of (k) is obtained to obtain the RF fingerprint F(k) without the channel, which is specifically expressed as

[0025]

[0026] When the local ideal signals corresponding to the two selected symbols are different, the local signal components are first removed from the received signal in the frequency domain, and then the difference operation and logarithmic difference operation are performed. 1 (k) By Instead, Y 2 (k) By Instead, the obtained channel-free radio frequency fingerprint F(k) is expressed as F(k)=R(k).

[0027] Furthermore, in step 3, the radio frequency fingerprints F(k) obtained by step 2 for multiple groups of symbols are superimposed and averaged, including calculating the F(k) obtained by the steps in step 2 for the L groups of symbols. 1 (k), F 2 (k), ..., F L (k), and then the results are superimposed and averaged to obtain the final RF fingerprint without channel and noise Its specific expression is

[0028]

[0029] where i = 1, 2, ..., L, F i (k) represents the result of the division operation in step 2 on the i-th group of symbols.

[0030] The present invention also discloses a radio frequency fingerprint extraction system using the logarithmic differential characteristics of channels at different times, the system comprising:

[0031] A signal acquisition module is used to sample the wireless signal to obtain an initial received signal;

[0032] The signal preprocessing module is used to preprocess the received symbols for extracting RF fingerprints, specifically including time synchronization and carrier frequency offset estimation and compensation, to obtain the preprocessed time domain received signal y(n), and further transform the frequency domain to obtain the frequency domain received signal Y(k);

[0033] The RF fingerprint extraction module is used to extract the RF fingerprint from the frequency domain received signal of the received symbol, and extract the RF fingerprint from the frequency domain received signal Y of the selected two segments of symbols that have experienced different channels and do not produce deep fading. 1 (k) and Y 2 (k) performing a difference operation and a logarithmic difference operation, and further performing a division operation to obtain a radio frequency fingerprint F(k) without a channel;

[0034] The RF fingerprint denoising module is used to superimpose and average the RF fingerprints F(k) obtained by the RF fingerprint extraction module for multiple groups of symbols to eliminate the influence of noise and obtain the final RF fingerprint without channel and noise.

[0035] The radio frequency fingerprint extraction method and system of the present invention using the logarithmic differential characteristics of channels at different times have the following advantages:

[0036] The method of the present invention utilizes the logarithmic differential characteristics of the channel at different times, performs differential operations and logarithmic differential operations on the frequency domain received signal, and further performs a division operation, so as to effectively eliminate the wireless variable channel component in the received signal, and superimposes and averages the results of the division operation on multiple groups of symbols to eliminate the influence of noise, and finally obtains a radio frequency fingerprint without channel and noise. The present invention has a great improvement in the stability of radio frequency fingerprint extraction in a channel variable environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 Schematic diagram of the flow of a radio frequency fingerprint extraction method using logarithmic differential characteristics of channels at different times in an embodiment of the present invention;

[0038] Figure 2 Schematic diagram of the PSBCH synchronization subframe structure in the LTE-V2X vehicle networking communication standard in an embodiment of the present invention;

[0039] Figure 3 The radio frequency fingerprints extracted by four different LTE-V2X communication devices for the PSSS symbols of two PSBCH synchronization subframes in the embodiment of the present invention are as follows: Waveform comparison chart;

[0040] Figure 4 This is a schematic diagram of the recognition results of 12 LTE-V2X communication devices in different environments based on the method of the present invention in an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solution and advantages of the present invention application more clearly presented, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the scope of the present invention patent application.

[0042] In actual mobile communication environments, due to the influence of time-varying channels, it becomes difficult to extract a stable RF fingerprint from a received signal doped with a wireless channel. Therefore, the present invention provides a RF fingerprint extraction method and system applicable to any wireless signal in a channel-variable environment, which utilizes the differences and logarithmic differential characteristics of channels at different times to effectively eliminate channel components and ultimately extract a stable RF fingerprint without channels and noise.

[0043] The radio frequency fingerprint extraction method using the logarithmic differential characteristics of the channel at different times provided in the embodiment of the present invention can be applied to wireless communication systems that adopt any communication protocol, such as Wi-Fi, LTE, etc. In this embodiment, the method of the present invention is explained and illustrated in detail by taking the LTE-V2X vehicle network as an example. Vehicle network devices are usually in a time-varying channel environment. The receiver in the vehicle network system is configured with the radio frequency fingerprint extraction system provided in this embodiment, which is used to receive the wireless radio frequency signal transmitted by the transmitter, and use the radio frequency fingerprint extraction method provided in this embodiment to extract the radio frequency fingerprint of the corresponding device from the received signal for identity authentication and other purposes.

[0044] like Figure 1 As shown, a method for extracting radio frequency fingerprints using the logarithmic differential characteristics of channels at different times is provided, comprising the following steps:

[0045] (1) The wireless signal is collected and preprocessed to obtain a preprocessed time domain received signal y(n), which is further transformed into a frequency domain to obtain a frequency domain received signal Y(k).

[0046] like Figure 2As shown, the wireless signal takes the physical sidelink broadcast channel (PSBCH, Physical Sidelink Broadcast Channel) synchronization subframe in the LTE-V2X vehicle networking communication standard as an example. The subframe contains 7 fixed sequence symbols, specifically including 2 identical primary synchronization signals (PSSS, Primary Sidelink Synchronization Signal), 2 identical secondary synchronization signals (SSSS, Secondary Sidelink Synchronization Signal) and 3 demodulation reference signals (DMRS, Demodulation Reference Signal), of which 2 of the 3 DMRS symbols are also identical symbols. In this embodiment, the receiver in the LTE-V2X vehicle networking system samples the wireless signal of the PSBCH synchronization subframe at a sampling rate of 30.72Mbps. The length of a PSBCH synchronization subframe is 1ms, and a total of 30720 sampling points are obtained, which are used as the initial received signal, and the initial received signal contains radio frequency fingerprint information as well as channel and noise. Then, 7 fixed sequence symbols are selected as the receiving symbols to be extracted for RF fingerprint preprocessing, and time synchronization, carrier frequency offset estimation and compensation are completed to obtain the preprocessed time domain received signal y(n), which is expressed as

[0047] y(n)=x(n)*r(n)*h(n)

[0048] Where n = 0, 1, ..., N-1, N is the number of sampling points of the selected symbol, x(n) is the local ideal signal corresponding to the selected symbol, r(n) is the RF fingerprint, h(n) is the wireless channel, and "*" represents the convolution operation. Specifically, when processing PSSS and SSSS symbols, the value of N is 62; when processing DMRS symbols, the value of N is 72.

[0049] For each of the 7 selected received symbols, the time domain received signal y(n) is transformed into the frequency domain to obtain the frequency domain received signal Y(k), which is expressed as

[0050] Y(k)=X(k)R(k)H(k)

[0051] Where k = 0, 1, ..., N-1, X(k), R(k), H(k) are the frequency domain expressions of the local ideal signal, the RF fingerprint, and the wireless channel, respectively. Specifically, when the PSSS and SSSS symbols are transformed, the value of N is 62; when the DMRS symbols are transformed, the value of N is 72.

[0052] (2) Select the frequency domain received signal Y of two segments of symbols that have experienced different channels and do not produce deep fading 1(k) and Y 2 (k) performs a difference operation and a logarithmic difference operation, and further performs a division operation to obtain a radio frequency fingerprint F(k) without a channel.

[0053] Select two PSBCH synchronization subframes with a time interval of more than 2s, which is much longer than the minimum incoherence time of the channel. Select a segment of the same symbol as the local ideal signal from these two subframes, and their frequency domain received signals are Y 1 (k) and Y 2 (k), the corresponding local ideal signal is X(k), and the channel H experienced by the two symbols is 1 (k) and H 2 (k) are not correlated with each other and therefore have different channel characteristics. The PSBCH synchronization subframe includes three types of fixed sequence symbols, and any sequence with the same local signal selected from two frames can constitute a group of such symbols. In this embodiment, when the Internet of Vehicles device used is in motion and the channel experienced is time-varying, two sequences with the same local signal corresponding to the same frame can also be regarded as a group.

[0054] In addition, the channels experienced by the two segments of the selected set of symbols cannot produce deep fading, so as to ensure that the frequency domain channel is flat and the frequency domain channel amplitude fluctuates within the range of ±TH, where TH is the maximum value of the frequency domain channel amplitude fluctuation without deep channel fading, where the frequency domain channel can be written in the form of H(k)=1+ΔH(k), that is, |ΔH(k)|<TH, and |·| represents the absolute value operation. In this example, to ensure that the frequency domain channel meets the above conditions, TH is set to 0.2.

[0055] Performing differential operation and logarithmic differential operation on the frequency domain received signals of the two selected symbol segments, and further performing division operation, specifically including:

[0056] For two frequency domain received signals Y 1 (k) and Y 2 (k) Perform differential operation. The specific formula is:

[0057] Y 1 (k)-Y 2 (k) = X(k)R(k)[H 1 (k)-H 2 (k)]

[0058] For two frequency domain received signals Y 1 (k) and Y 2 (k) Perform logarithmic difference operation. The specific formula is:

[0059]

[0060] Where ln[·] represents the logarithmic operation. Since the amplitude of ΔH(k) is small, ln[1+ΔH(k)] is approximately equivalent to ΔH(k), so the above formula can be further expressed as

[0061] LqCy 1 (k)-lnH 2 (k) = ln[1 + ΔH 1 (k)]-ln[1+ΔH 2 (k)]≈ΔH 1 (k)-ΔH 2 (k)

[0062] =H 1 (k)-H 2 (k)

[0063] The two frequency domain received signals Y 1 (k) and Y 2 The difference operation result of (k) is divided by the two-segment frequency domain received signal Y 1 (k) and Y 2 The logarithmic difference operation result of (k) eliminates the channel information and obtains the RF fingerprint F(k) without the channel, which is specifically expressed as

[0064]

[0065] (3) The RF fingerprints F(k) obtained by step (2) for multiple groups of symbols are superimposed and averaged to eliminate the influence of noise and obtain the final RF fingerprint without channel and noise.

[0066] Since the two selected symbols with the same local ideal signal can be two sequences of the same local signal corresponding to different PSBCH subframes, or two sequences of the same local signal corresponding to the same PSBCH subframe, multiple groups of symbols that meet the requirements of step (2) can be obtained in the end. The total of L groups of symbols are subjected to differential and logarithmic differential operations as described in step (2), and further divided to obtain F 1 (k), F 2 (k), ..., F L (k) Then, the above results are superimposed and averaged to eliminate the influence of noise and obtain the final RF fingerprint without channel and noise. Its specific expression is

[0067]

[0068] where i = 1, 2, ..., L, F i (k) represents the result of the division operation on the i-th group of symbols.

[0069] Figure 3 The RF fingerprints extracted by four different LTE-V2X communication devices for the PSSS symbols of two PSBCH synchronization subframes in this embodiment are shown. Waveform comparison chart. The RF fingerprint is obtained by performing differential and logarithmic differential operations on two PSSS sequences with the same local signal in two PSBCH frames at different times, and further performing division operations and superposition average operations. The RF fingerprints of the four devices are The waveforms are plotted on the same graph and compared to show the RF fingerprint of each device. There are significant differences between the waveforms, which can be used to distinguish different devices and achieve identity authentication.

[0070] Figure 4 The schematic diagram of the recognition results of 12 LTE-V2X communication devices in different environments based on the method of the present invention in this embodiment is shown. The device communication environment includes a combination of indoor line-of-sight scenes, corridor non-line-of-sight scenes, outdoor scenes, static scenes, mobile scenes, etc. The scenes are rich and basically include various common scenes of daily communication. In the above-mentioned various communication scenarios, the RF fingerprint extraction method provided in this embodiment is used to identify 12 LTE-V2X communication devices, and the recognition accuracy rate is above 95%, and the accuracy rate of some mobile scenes can reach 100%. In indoor line-of-sight scenes and corridor non-line-of-sight scenes, the accuracy rate when moving is higher than the accuracy rate when stationary, which verifies that the method of the present invention eliminates the interference of wireless variable channels by utilizing the difference characteristics of channels at different times.

[0071] It can be seen that the RF fingerprint extraction method using the logarithmic differential characteristics of the channels at different times provided by the embodiment of the present invention can utilize the differences and logarithmic differential characteristics of the channels at different times to effectively eliminate the influence of time-varying channels on RF fingerprint extraction in various complex environments, and finally obtain stable RF fingerprint features.

[0072] The embodiment of the present invention also includes a radio frequency fingerprint extraction system using the logarithmic differential characteristics of the channel at different times, which can also be applied to a wireless communication system using any communication protocol. Taking the radio frequency fingerprint extraction for the LTE-V2X vehicle networking system as an example, a radio frequency fingerprint extraction system using the logarithmic differential characteristics of the channel at different times is provided, including:

[0073] The signal acquisition module is used to sample the wireless signal sent by the LTE-V2X vehicle networking equipment to obtain the initial received signal;

[0074] The signal preprocessing module is used to preprocess the received symbols for extracting RF fingerprints, such as PSSS, SSSS, and DMRS symbols in the PSBCH synchronization subframe, including time synchronization and carrier frequency offset estimation and compensation, to obtain the preprocessed time domain received signal y(n), and further transform the frequency domain to obtain the frequency domain received signal Y(k);

[0075] The RF fingerprint extraction module is used to extract the RF fingerprint from the frequency domain received signal of the received symbol, and extract the RF fingerprint from the frequency domain received signal Y of the selected two segments of symbols that have experienced different channels and do not produce deep fading. 1 (k) and Y 2 (k) performing a difference operation and a logarithmic difference operation, and further performing a division operation, to obtain a radio frequency fingerprint F(k) without a channel, wherein the selected two segments of symbols are two segments of symbols with the same local ideal signal, which may be two sequences with the same local signal corresponding to different PSBCH subframes, or may be two sequences with the same local signal corresponding to the same PSBCH subframe;

[0076] The RF fingerprint denoising module is used to superimpose and average the RF fingerprints F(k) obtained by the RF fingerprint extraction module for multiple groups of symbols to eliminate the influence of noise and obtain the final RF fingerprint without channel and noise.

[0077] The technical features and steps of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features and steps in the flow charts in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features and implementation steps, they should be considered to be within the scope of this specification.

[0078] It should be noted that the above-mentioned embodiment only expresses one implementation method of the present invention, and its description is relatively specific and detailed, but it cannot be understood as limiting the scope of the present invention. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A radio frequency fingerprint extraction method using the logarithmic differential characteristics of channels at different times, characterized in that: The method comprises the following steps: Step 1: collect and preprocess the wireless signal to obtain a preprocessed time domain received signal y(n), and further transform the time domain received signal y(n) into a frequency domain to obtain a frequency domain received signal Y(k); Step 2: Select the frequency domain received signals Y1(k) and Y2(k) of two segments of symbols that have experienced different channels and do not produce deep fading, perform differential operations and logarithmic differential operations, and further perform division operations to obtain a radio frequency fingerprint F(k) without a channel; Step 3: Superimpose and average the radio frequency fingerprints F(k) obtained by step 2 for multiple groups of symbols to eliminate the influence of noise and obtain the final radio frequency fingerprint without channel and noise.

2. The method for extracting radio frequency fingerprints using logarithmic differential characteristics of channels at different times according to claim 1, characterized in that: In step 1, the wireless signal acquisition includes sampling the wireless signal by the receiver to obtain an initial received signal; the wireless signal preprocessing includes selecting symbols for RF fingerprint extraction from the initial received signal, completing time synchronization and carrier frequency offset estimation and compensation for the initial received signal; the preprocessed time domain received signal y(n) is expressed as y(n)=x(n)*r(n)*h(n) Where n=0, 1, ..., N-1, N is the number of sampling points of the selected symbol, x(n) is the local ideal signal corresponding to the selected symbol, r(n) is the radio frequency fingerprint, h(n) is the wireless channel, and "*" represents the convolution operation; the frequency domain transformation includes performing a frequency domain transformation on the time domain received signal y(n) to obtain a frequency domain received signal Y(k), which is expressed as Y(k)=X(k)R(k)H(k) Where k=0, 1, ..., N-1, X(k), R(k), and H(k) are the frequency domain expressions of the local ideal signal, the radio frequency fingerprint, and the wireless channel, respectively.

3. The method for extracting radio frequency fingerprints using logarithmic differential characteristics of channels at different times according to claim 2, characterized in that: In the step 2, selecting two segments of symbols that experience different channels and do not produce deep fading includes selecting two segments of received symbols with a time interval of τ, wherein τ>Δ, Δ is the minimum incoherent time of the channel, and at this time, the channels H1(k) and H2(k) experienced by the two segments of symbols are not correlated, and therefore have different channel characteristics; selecting the frequency domain received signals Y1(k) and Y2(k) of the two segments of symbols that do not produce deep fading includes selecting two segments of received symbols whose frequency domain channels are flat and whose frequency domain channel amplitudes fluctuate within the range of ±TH, TH being the maximum value of the frequency domain channel amplitude fluctuation that can be reached without deep channel fading, wherein the frequency domain channel is written in the form of H(k)=1+ΔH(k), that is, |ΔH(k)|<TH, |·| represents absolute value operation; the frequency domain received signals Y1(k) and Y2(k) of the two segments of symbols are obtained after preprocessing and frequency domain transformation in step 1; The two-segment symbols are two symbols with the same length in the same frame, or two symbols with the same length belonging to different frames, and their corresponding local ideal signals are X1(k) and X2(k) respectively.

4. The method for extracting radio frequency fingerprints using logarithmic differential characteristics of channels at different times according to claim 3, characterized in that: When the local ideal signals corresponding to the two selected symbol segments are the same, that is, both are X(k), in step 2, a differential operation is performed on the two frequency domain received signals Y1(k) and Y2(k), which is specifically expressed as Y1(k)-Y2(k)=X(k)R(k)[H1(k)-H2(k)] Perform logarithmic difference operation on the two frequency domain received signals Y1(k) and Y2(k), which can be expressed as Where ln[·] represents the logarithmic operation; ln[1+ΔH(k)] is approximately equivalent to ΔH(k), so the above formula can be expressed as ln H1(k)-ln H2(k)=ln[1+ΔH1(k)]-ln[1+ΔH2(k)]≈ΔH1(k)-ΔH2(k) =H1(k)-H2(k) The difference operation result of the two frequency domain received signals Y1(k) and Y2(k) is divided by the logarithmic difference operation result of the two frequency domain received signals Y1(k) and Y2(k) to obtain the RF fingerprint F(k) without the channel, which is specifically expressed as When the local ideal signals corresponding to the two selected symbols are different, the local signal components are first removed from the received signal in the frequency domain, and then the difference operation and logarithmic difference operation are performed, that is, Y1(k) in the above formula is replaced by Instead, Y2(k) is given by Instead, the obtained channel-free radio frequency fingerprint F(k) is expressed as F(k)=R(k).

5. The method for extracting radio frequency fingerprints using logarithmic differential characteristics of channels at different times according to claim 4, characterized in that: In step 3, the radio frequency fingerprints F(k) obtained by step 2 for multiple groups of symbols are superimposed and averaged, including calculating the L groups of symbols according to the steps in step 2 to obtain F1(k), F2(k), ..., F L (k), and then the results are superimposed and averaged to obtain the final RF fingerprint without channel and noise Its specific expression is where i = 1, 2, ..., L, F i (k) represents the result of the division operation in step 2 on the i-th group of symbols.

6. A radio frequency fingerprint extraction system using the logarithmic differential characteristics of channels at different times, characterized in that: The radio frequency fingerprint extraction method using the logarithmic differential characteristics of the channel at different times as described in any one of claims 1 to 5 is used, and the system includes: A signal acquisition module is used to sample the wireless signal to obtain an initial received signal; The signal preprocessing module is used to preprocess the received symbols for extracting RF fingerprints, specifically including time synchronization and carrier frequency offset estimation and compensation, to obtain the preprocessed time domain received signal y(n), and further transform the frequency domain to obtain the frequency domain received signal Y(k); The radio frequency fingerprint extraction module is used to extract the radio frequency fingerprint from the frequency domain received signal of the received symbol, perform differential operation and logarithmic differential operation on the frequency domain received signals Y1(k) and Y2(k) of the selected two segments of symbols that have experienced different channels and do not produce deep fading, and further perform division operation to obtain the radio frequency fingerprint F(k) without the channel; The RF fingerprint denoising module is used to superimpose and average the RF fingerprints F(k) obtained by the RF fingerprint extraction module for multiple groups of symbols to eliminate the influence of noise and obtain the final RF fingerprint without channel and noise.

Images