Radio frequency signal transient section detection method and device based on energy spectrum coefficient entropy

Abstract

The invention relates to the field of physical layer security, in particular to a radio frequency signal transient section detection method and device based on energy spectrum coefficient entropy, and the method comprises the steps: obtaining a radio frequency signal, and preprocessing the radio frequency signal; segmenting the preprocessed radio-frequency signal by using a sliding window, and calculating an energy spectrum coefficient corresponding to the radio-frequency signal of each window so as to form an energy spectrum coefficient track; the entropy of each element in the energy spectrum coefficient trajectory is calculated in sequence, so that an energy spectrum coefficient entropy trajectory is obtained; calculating a posterior probability density corresponding to the energy spectrum coefficient entropy trajectory; and carrying out peak detection on the posterior probability density curve to obtain starting and ending moments of the transient section of the radio frequency signal. According to the method, the change behavior of the signal power envelope is simulated by using the energy spectrum coefficient entropy trajectory, the posterior probability density curve is constructed to compress the edge of the transient section into the pulse, the starting moment and the ending moment of the transient section are extracted through peak detection, and the method has the characteristics that a preset threshold value is not needed, and calculation is moderate.

Description

Technical Field

[0001] The present invention relates to the field of physical layer security, and in particular to a method and device for detecting transient sections of radio frequency signals based on energy spectrum coefficient entropy. Background Art

[0002] In the information age, driven by the rapid development of electronic information technology, the gradual maturity of wireless communication technology has led to a surge in the number of wireless devices, bringing unprecedented changes to people's lives. However, the development of information technology is a double-edged sword: while bringing convenience to people, it can also pose certain risks to information privacy. To improve the security of wireless networks, radio frequency fingerprinting (RFFI) technology has become a research hotspot in wireless network security. Each wireless device has a unique RF fingerprint, which can effectively distinguish and identify devices. Based on the composition of the RF signal, RF fingerprint features can be divided into features based on transient signals and features based on steady-state signals. Transient signals are generated when a transmitter switches on and off. If the transient state is not correctly captured, the subsequent features obtained will not be the actual characteristics of the transmitter. Therefore, capturing the transient segment is the basis for extracting the transient RF fingerprint and is an extremely critical and challenging part of the RF fingerprint extraction process.

[0003] There are five traditional transient detection methods: Variance Fractal Dimension Threshold Detection (VFDTD), Bayesian Step Change Detection (BSCD), Mean Change Point Detection (MCPD), Energy Criterion (EC-α), and Sliding Window Variance Trajectory-based Posterior Probability Density Detection (SWVT-P2D2). These transient detection methods can detect the starting point of transient segments in different ways, but there is still room for improvement. For example, they have high computational complexity, require pre-set thresholds, and are unable to detect the end point of transient segments. Summary of the Invention

[0004] In light of the technical issues of the aforementioned traditional transient detection methods, such as high computational complexity, reliance on preset thresholds, and inability to accurately detect the endpoints of transient segments, a method and apparatus for detecting transient segments of radio frequency signals based on energy spectrum coefficient entropy is provided. This invention primarily utilizes sliding window segmentation technology combined with Hilbert transform and discrete Fourier transform to extract energy spectrum coefficients. By constructing an energy spectrum coefficient entropy trajectory and calculating its posterior probability density, it ultimately achieves synchronous detection of the start and end times of the transient segment, thereby reducing computational complexity, avoiding manual threshold intervention, and improving detection accuracy.

[0005] The technical means adopted in the present invention are as follows:

[0006] A method for detecting transient sections of radio frequency signals based on energy spectrum coefficient entropy, comprising the following steps:

[0007] Acquire a radio frequency signal, and pre-process the radio frequency signal;

[0008] Use a sliding window to segment the preprocessed RF signal, calculate the energy spectrum coefficient corresponding to the RF signal in each window, and thus form an energy spectrum coefficient trajectory;

[0009] Calculating the entropy of each element in the energy spectrum coefficient trajectory in sequence, thereby obtaining the energy spectrum coefficient entropy trajectory;

[0010] Calculating the posterior probability density corresponding to the energy spectrum coefficient entropy trajectory;

[0011] Peak detection is performed on the posterior probability density curve to obtain the start time and end time of the transient section of the radio frequency signal at one time.

[0012] Furthermore, preprocessing the radio frequency signal includes:

[0013] Performing filtering on the radio frequency signal using a bandpass filter;

[0014] The filtered RF signal is normalized.

[0015] Furthermore, the method of using a sliding window to segment the preprocessed RF signal and calculating the energy spectrum coefficient corresponding to the RF signal in each window to form an energy spectrum coefficient trajectory includes:

[0016] The preprocessed RF signal s(n) is segmented using a sliding window with a window length of W. When the starting point of the window moves to the position with index n, a real sequence truncated signal is obtained.

[0017] Perform Hilbert transform on each segment of the real number sequence truncated signal to achieve the goal of transforming the real number sequence truncated signal Transform the truncated signal into a complex sequence

[0018] Perform discrete Fourier transform on the complex sequence truncated signal in each window and calculate the energy spectrum coefficient vector corresponding to each window:

[0019]

[0020] in, is the energy spectrum coefficient vector corresponding to each window, W is the window length, is the complex sequence truncated signal, k is the frequency index, n is the time index within the window, and j is the imaginary unit;

[0021] The energy spectrum coefficient vector corresponding to each window Take the average value and convert the energy spectrum coefficient vector corresponding to each window into The energy spectrum coefficient scalar corresponding to each window is converted into the energy spectrum coefficient scalar corresponding to each window, and the energy spectrum coefficient scalar corresponding to each window is arranged in order to obtain the energy spectrum coefficient trajectory C(n) corresponding to the radio frequency signal of each window.

[0022] Furthermore, the step of sequentially calculating the entropy of each element in the energy spectrum coefficient trajectory to obtain the energy spectrum coefficient entropy trajectory includes:

[0023] The energy spectrum coefficient trajectory is segmented using a sliding window method to construct a statistical sample of the energy spectrum coefficient trajectory, and the proportion probability of each element in the energy spectrum coefficient trajectory is calculated:

[0024]

[0025] Where W is the window length, p i is the proportion probability of each element in the energy spectrum coefficient trajectory, C(i) is the energy spectrum coefficient value at the i-th time point, C(k) is the energy spectrum coefficient value at the k-th time point, i is the current time index, and k is the summation index variable;

[0026] According to the proportion probability of each element, the entropy of each element in the energy spectrum coefficient trajectory is calculated in turn, thereby constructing the energy spectrum coefficient entropy trajectory:

[0027] H(n)=-P n log2P n

[0028] Among them, H(n) is the entropy trajectory of the energy spectrum coefficient, P n is the probability value corresponding to time point n.

[0029] Furthermore, the calculation of the posterior probability density corresponding to the energy spectrum coefficient entropy trajectory includes:

[0030] Bayesian estimation is used to detect the step changes on both sides of the entropy trajectory, and the posterior probability density curve is obtained:

[0031]

[0032] Among them, H(n) is the energy spectrum coefficient entropy trajectory, N is the sequence length, m is the index variable, and p({m}|H) is the posterior probability density.

[0033] Furthermore, the peak detection of the posterior probability density curve to obtain the start time and the end time of the transient section of the radio frequency signal at one time includes:

[0034] The posterior probability density curve presents two peak points, and the index numbers corresponding to the peak points of the curve are the starting time and the ending time of the transient segment respectively.

[0035] A radio frequency signal transient section detection device based on energy spectrum coefficient entropy, used to implement the above radio frequency signal transient section detection method based on energy spectrum coefficient entropy, comprises:

[0036] Signal acquisition module: used to acquire radio frequency signals;

[0037] Preprocessing module: used to preprocess the acquired RF signal;

[0038] The first calculation module is used to calculate the entropy of the energy spectrum coefficient of the truncated signal after the sliding window segmentation and construct the entropy trajectory;

[0039] The second calculation module is used to calculate the posterior probability density of the energy spectrum coefficient entropy trajectory;

[0040] Peak detection module: used to detect the peak of the posterior probability density curve and obtain the starting and ending times of the transient section of the RF signal.

[0041] Compared with the prior art, the present invention has the following advantages:

[0042] This method accurately detects the start of transient segments and, compared to existing transient detection algorithms, successfully detects the end of transient signals for the first time. Furthermore, while some methods typically require extensive experimentation before detection to determine the decision threshold, this method does not rely on a pre-set threshold, offering greater flexibility and practicality. Furthermore, this method exhibits excellent noise immunity and is not susceptible to interference. Its low computational complexity ensures real-time and high efficiency in practical applications.

[0043] Based on the above reasons, the present invention can be widely promoted in fields such as physical layer security. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0045] Figure 1 is a flow chart of the method of the present invention;

[0046] Figure 2 (a) is the energy spectrum coefficient entropy trajectory diagram of signal 1, Figure 2 (b) is the posterior probability density curve of the entropy trajectory of the energy spectrum coefficient of signal 1;

[0047] Figure 3 (a) is the energy spectrum coefficient entropy trajectory diagram of signal 2, Figure 3 (b) is the posterior probability density curve of the energy spectrum coefficient entropy trajectory of signal 2;

[0048] Figure 4 (a) is the energy spectrum coefficient entropy trajectory diagram of signal 3, Figure 4 (b) is the posterior probability density curve of the energy spectrum coefficient entropy trajectory of signal 3;

[0049] Figure 5 Figure 2 is the average transient onset detection error of different methods under different signal-to-noise ratios. DETAILED DESCRIPTION

[0050] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0051] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0052] like Figure 1 As shown, the present invention provides a method for detecting transient sections of radio frequency signals based on energy spectrum coefficient entropy, the steps comprising:

[0053] S1. Acquire radio frequency signals and pre-process the radio frequency signals.

[0054] Furthermore, a Tektronix TDS7404 high-sampling-rate oscilloscope and an 8-bit analog-to-digital converter were used to capture the smartphone's Bluetooth signal, generating a radio frequency (RF) signal. This RF signal was filtered using a digital bandpass filter that only allowed components in the ISM2400 band to remove non-target signals. The filtered RF signal was normalized to ensure that the extracted features reflected the overall characteristics of the device.

[0055] S2. Use a sliding window to segment the preprocessed RF signal and calculate the energy spectrum coefficient corresponding to the RF signal in each window to form an energy spectrum coefficient trajectory.

[0056] Furthermore, a sliding window with a window length of W is used to segment the preprocessed RF signal s(n). When the starting point of the window moves to the position with index n, a real sequence truncated signal is obtained.

[0057] Perform Hilbert transform on each segment of the real number sequence truncated signal to achieve the truncated signal of the real number sequence Transform the truncated signal into a complex sequence

[0058] Since it is more convenient to observe complex signals and understand signal structures in the frequency domain, the complex signal is converted to the frequency domain, and the discrete Fourier transform is performed on the complex sequence truncated signal in each window. The discrete Fourier transform can convert the short-time complex sequence in each sliding window into the frequency domain space, and then obtain the energy spectrum coefficient vector corresponding to each window from the frequency domain:

[0059]

[0060] in, is the energy spectrum coefficient vector corresponding to each window, W is the window length, that is, the number of samples of the truncated signal, is the complex sequence truncated signal, k is the frequency index, n is the time index within the window, and j is the imaginary unit.

[0061] The energy spectrum coefficient vector corresponding to each window Take the average value and convert the energy spectrum coefficient vector corresponding to each window into The energy spectrum coefficient scalar corresponding to each window is converted into the energy spectrum coefficient scalar corresponding to each window, and the energy spectrum coefficient scalar corresponding to each window is arranged in order to obtain the energy spectrum coefficient trajectory C(n) corresponding to the radio frequency signal of each window.

[0062] S3. Calculate the entropy of each element in the energy spectrum coefficient trajectory in sequence, thereby obtaining the energy spectrum coefficient entropy trajectory.

[0063] Furthermore, to assess the degree of disorder in the energy spectrum coefficient traces of the signal segments obtained by sliding window segmentation, entropy calculation was introduced. The calculation of information entropy requires modeling the probability of the variables, so we chose to reuse the sliding window method to construct the statistical sample.

[0064] The energy spectrum coefficient trajectory is segmented using a sliding window method to construct a statistical sample of the energy spectrum coefficient trajectory, and the proportion probability of each element in the energy spectrum coefficient trajectory is calculated:

[0065]

[0066] Where W is the window length, p i is the proportion probability of each element in the energy spectrum coefficient trajectory, C(i) is the energy spectrum coefficient value at the i-th time point, C(k) is the energy spectrum coefficient value at the k-th time point, i is the current time index, and k is the summation index variable.

[0067] According to the proportion probability of each element, the entropy of each element in the energy spectrum coefficient trajectory is calculated in turn, thereby constructing the energy spectrum coefficient entropy trajectory:

[0068] H(n)=-P n log2P n

[0069] Among them, H(n) is the entropy trajectory of the energy spectrum coefficient, P n is the probability value corresponding to time point n.

[0070] S4. Calculate the posterior probability density corresponding to the energy spectrum coefficient entropy trajectory.

[0071] Furthermore, Bayesian estimation is used to detect the step changes on both sides of the entropy trajectory to obtain the posterior probability density curve:

[0072]

[0073] Among them, H(n) is the energy spectrum coefficient entropy trajectory, N is the sequence length, m is the index variable, and p({m}|H) is the posterior probability density.

[0074] The conditional probability p({m}|H) represents the probability that the start or end time of the transient part is at index. It should be noted that when N is very large, it may lead to power exponential Therefore, the power exponent needs to be set to a reasonable constant value to ensure a valid calculation result.

[0075] S5. Perform peak detection on the posterior probability density curve to obtain the start and end times of the transient segment of the RF signal.

[0076] Furthermore, the posterior probability density curve presents two peak points, and the index numbers corresponding to the peak points of the curve are the start and end times of the transient segment respectively; the part before the start point is the noise part, and the part after the end point is the steady-state part of the RF signal.

[0077] The above steps S1-S5 are performed sequentially in the present invention.

[0078] The present invention also includes a radio frequency signal transient section detection device based on energy spectrum coefficient entropy, which is used to implement the above radio frequency signal transient section detection method based on energy spectrum coefficient entropy. The device has the following modules:

[0079] Signal acquisition module: used to acquire radio frequency signals.

[0080] Preprocessing module: used to preprocess the acquired RF signal.

[0081] The first calculation module is used to calculate the entropy of the energy spectrum coefficient of the truncated signal after the sliding window segmentation and construct the entropy trajectory.

[0082] The second calculation module is used to calculate the posterior probability density of the energy spectrum coefficient entropy trajectory.

[0083] Peak detection module: used to detect the peak of the posterior probability density curve and obtain the start and end times of the transient section of the RF signal.

[0084] The present invention verifies the effectiveness of the proposed method for detecting transient sections of radio frequency signals based on energy spectrum coefficient entropy through experiments. Figure 2-4 The analysis results of three different signals are shown respectively, where part (a) of each figure is the entropy trajectory of the energy spectrum coefficient, and part (b) is the corresponding posterior probability density curve. Figure 5 Figure 2 is the average transient onset detection error of different methods under different signal-to-noise ratios.

[0085] From the results, we can see that Figure 2 (b) Figure 3 (b) and Figure 4 The probability density curve shown in (b) shows two consecutive rising and falling trends, forming two maximum points. These two maximum points correspond to the transient starting point and end point detected by the present invention, and are in good agreement with the actual starting point and end point, proving that the present invention is accurate in detecting the start and end time of the transient segment. Since the existing transient detection algorithm mainly identifies the starting time of the signal, the detection error average value experiment only detects the transient starting point to maintain consistency. Figure 5 It can be seen that as the signal-to-noise ratio gradually decreases, the detection errors of all the algorithms involved in the test increase significantly. Although VFDTD shows the best anti-noise performance, its detection accuracy is highly dependent on the preset threshold. In contrast, the present invention does not require a preset threshold and shows more superior comprehensive performance.

[0086] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting transient sections of radio frequency signals based on energy spectrum coefficient entropy, characterized in that the steps include: Acquire a radio frequency signal, and pre-process the radio frequency signal; Use a sliding window to segment the preprocessed RF signal, calculate the energy spectrum coefficient corresponding to the RF signal in each window, and thus form an energy spectrum coefficient trajectory; Calculating the entropy of each element in the energy spectrum coefficient trajectory in sequence, thereby obtaining the energy spectrum coefficient entropy trajectory; Calculating the posterior probability density corresponding to the energy spectrum coefficient entropy trajectory; Peak detection is performed on the posterior probability density curve to obtain the start time and end time of the transient section of the radio frequency signal at one time.

2. The method for detecting transient sections of radio frequency signals based on energy spectrum coefficient entropy according to claim 1, wherein: Preprocessing the radio frequency signal includes: Performing filtering on the radio frequency signal using a bandpass filter; The filtered RF signal is normalized.

3. The method for detecting transient sections of radio frequency signals based on energy spectrum coefficient entropy according to claim 1, wherein: The method of using a sliding window to segment the preprocessed RF signal and calculating the energy spectrum coefficient corresponding to the RF signal in each window to form an energy spectrum coefficient trajectory includes: The preprocessed RF signal s(n) is segmented using a sliding window with a window length of W. When the starting point of the window moves to the position with index n, a real sequence truncated signal is obtained. Perform Hilbert transform on each segment of the real number sequence truncated signal to achieve the goal of transforming the real number sequence truncated signal Transform the truncated signal into a complex sequence Perform discrete Fourier transform on the complex sequence truncated signal in each window and calculate the energy spectrum coefficient vector corresponding to each window: in, is the energy spectrum coefficient vector corresponding to each window, W is the window length, is the complex sequence truncated signal, k is the frequency index, n is the time index within the window, and j is the imaginary unit; The energy spectrum coefficient vector corresponding to each window Take the average value and convert the energy spectrum coefficient vector corresponding to each window into The energy spectrum coefficient scalar corresponding to each window is converted into the energy spectrum coefficient scalar corresponding to each window, and the energy spectrum coefficient scalar corresponding to each window is arranged in order to obtain the energy spectrum coefficient trajectory C(n) corresponding to the radio frequency signal of each window.

4. The method for detecting transient sections of radio frequency signals based on energy spectrum coefficient entropy according to claim 1, wherein: The step of sequentially calculating the entropy of each element in the energy spectrum coefficient trajectory to obtain the energy spectrum coefficient entropy trajectory includes: The energy spectrum coefficient trajectory is segmented using a sliding window method to construct a statistical sample of the energy spectrum coefficient trajectory, and the proportion probability of each element in the energy spectrum coefficient trajectory is calculated: Where W is the window length, p i is the proportion probability of each element in the energy spectrum coefficient trajectory, C(i) is the energy spectrum coefficient value at the i-th time point, C(k) is the energy spectrum coefficient value at the k-th time point, i is the current time index, and k is the summation index variable; According to the proportion probability of each element, the entropy of each element in the energy spectrum coefficient trajectory is calculated in turn, thereby constructing the energy spectrum coefficient entropy trajectory: H(n)=-P n log2P n Among them, H(n) is the entropy trajectory of the energy spectrum coefficient, P n is the probability value corresponding to time point n.

5. The method for detecting transient sections of radio frequency signals based on energy spectrum coefficient entropy according to claim 1, wherein: The calculating the posterior probability density corresponding to the energy spectrum coefficient entropy trajectory includes: Bayesian estimation is used to detect the step changes on both sides of the entropy trajectory, and the posterior probability density curve is obtained: Among them, H(n) is the energy spectrum coefficient entropy trajectory, N is the sequence length, m is the index variable, and p({m}|H) is the posterior probability density.

6. The method for detecting transient sections of radio frequency signals based on energy spectrum coefficient entropy according to claim 1, wherein: The peak detection of the posterior probability density curve to obtain the start time and the end time of the transient section of the radio frequency signal at one time includes: The posterior probability density curve presents two peak points, and the index numbers corresponding to the peak points of the curve are the starting time and the ending time of the transient segment respectively.

7. A radio frequency signal transient segment detection device based on energy spectrum coefficient entropy, used to implement any radio frequency signal transient segment detection method based on energy spectrum coefficient entropy according to claims 1-6, characterized in that: The device includes: Signal acquisition module: used to acquire radio frequency signals; Preprocessing module: used to preprocess the acquired RF signal; The first calculation module is used to calculate the entropy of the energy spectrum coefficient of the truncated signal after the sliding window segmentation and construct the entropy trajectory; The second calculation module is used to calculate the posterior probability density of the energy spectrum coefficient entropy trajectory; Peak detection module: used to detect the peak of the posterior probability density curve and obtain the starting and ending times of the transient section of the RF signal.

Images