A GNSS step-type spoofing jamming detection method and related product

Abstract

This invention discloses a method and related products for detecting GNSS step-type deception interference, belonging to the field of satellite navigation signal processing technology. The GNSS step-type deception interference detection method provided by this invention constructs a signal power sample set through sliding window segmentation processing and builds a feature sequence to characterize the statistical properties of the signal power samples. It utilizes a long short-term memory network model for time-series modeling, effectively capturing the temporal dynamic characteristics of the signal power samples. By calculating and fusing the distribution difference measure between the current detection window and the historical reference window, a comprehensive distribution drift index is constructed, which can sensitively and accurately reflect the drift of the signal power sample distribution. Based on the comprehensive distribution drift index, a dynamic threshold is constructed for decision-making, enabling the detection standard to adapt to environmental changes, thereby improving the reliability and accuracy of GNSS step-type deception interference detection under dynamic environments and varying environmental noise characteristics.

Description

Technical Field

[0001] This invention relates to the field of satellite navigation signal processing technology, specifically to a GNSS step-type deception interference detection method and related products. Background Technology

[0002] Global Navigation Satellite System (GNSS) is an indispensable spatiotemporal reference infrastructure in modern information society, widely used in many key fields such as power systems, communication networks, transportation, and national defense, providing high-precision positioning, navigation, and timing services for various terminal devices. However, GNSS signals inherently possess low transmission power and weak signal strength upon reaching the ground receiver during long-distance propagation in space. This makes GNSS receivers vulnerable to various malicious interferences, among which spoofing interference is one of the most serious forms. Spoofing interference generates fake navigation signals with a structure similar to real satellite signals and controllable power, inducing receivers to lock onto false signals and output incorrect positioning, velocity, and timing results, thus posing a serious threat to the operational security of critical infrastructure relying on GNSS services.

[0003] Currently, there is considerable research on GNSS spoofing interference detection techniques, among which signal quality monitoring methods are widely used. These methods typically extract a characteristic index of the received signal power and set corresponding empirical thresholds for spoofing interference judgment, achieving a certain degree of detection under static environments or ideal noise conditions. However, in complex real-world applications, when the receiver is in dynamic motion or environmental noise characteristics change, high false detection or false negative rates are prone to occur, leading to insufficient detection reliability. In particular, step-type spoofing interference is a common and representative form of interference in actual GNSS spoofing attack scenarios, characterized by a significant step-like jump in signal power within a very short time. For this type of spoofing interference with sudden power changes, relying solely on signal quality monitoring methods often fails to accurately distinguish the power fluctuations of the spoofing interference signal from those of the normal signal, making stable and reliable detection impossible.

[0004] Therefore, improving the reliability of GNSS step-type deception interference detection has become a technical challenge that urgently needs to be overcome by those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a method and related products for detecting GNSS step-type spoofing interference, so as to overcome the problem of insufficient reliability of existing signal quality monitoring-based methods in detecting GNSS step-type spoofing interference.

[0006] The present invention solves the above-mentioned technical problems through the following technical solution: This invention provides a method for detecting GNSS step-type deception interference, comprising the following steps: Several signal power samples output by the GNSS receiver are collected, and the signal power samples are segmented using a sliding window method to construct a signal power sample set with Z time windows. Based on the signal power sample set for each time window, a feature sequence is constructed to characterize the statistical characteristics of the signal power sample. The feature sequence is then time-series modeled based on a long short-term memory network model to obtain a time-series characterization result sequence for Z time windows. Based on the current time window, according to the first preset time window length N and the second preset time window length M, extract the first time-series representation result sequence of the current window to be detected and the second time-series representation result sequence of the historical reference window from the time-series representation result sequence; let the current time window be the k-th time window, the current window to be detected include the k-N+1, k-N+2, ..., k-th time windows, and the historical reference window includes the kN-M+1, kN-M+2, ..., kN-th time windows; where Z≥k≥N+M, k∈[ZNM, Z]; Based on the first time series characterization result sequence and the second time series characterization result sequence, the distribution difference measure of the current time window is calculated. The distribution difference measure is then standardized and fused to obtain the comprehensive distribution drift index of the current time window. Based on the comprehensive distribution drift index, a dynamic threshold is constructed; if the comprehensive distribution drift index of each time window within the current detection window is greater than the dynamic threshold of the corresponding time window, it is determined that there is step-type deception interference in the current detection window.

[0007] A further improvement of this invention is that, let the signal power sample set of the k-th time window be {P}. k,1 , P k,2 ,..., P k,L}, P k,1 , P k,2 , ..., P k,L For the signal power samples at the 1st, 2nd, ..., Lth sampling points of the kth time window, a feature sequence is constructed to characterize the statistical properties of the signal power samples. This includes the following steps: The mean characteristic of the k-th time window is calculated as follows:

[0008] in, The total number of sampling points in the k-th time window; Let be the mean feature of the k-th time window; i is the index of the sampling point in the k-th time window; This is the signal power sample at the i-th sampling point in the k-th time window; The summation symbol; The standard deviation feature of the k-th time window is calculated as follows:

[0009] in, The standard deviation feature of the k-th time window; The range feature of the k-th time window is calculated as follows:

[0010] in, This represents the range feature of the k-th time window; This represents the maximum signal power sample in the k-th time window; The minimum signal power sample in the k-th time window; The signal power variation trend characteristics in the k-th time window are calculated as follows: Using L signal power samples within the k-th time window as the dependent variable, the least squares method is used to fit the data to obtain the fitted relationship. The slope in the fitted relationship is used as the signal power change trend feature of the k-th time window. The fitting relationship is:

[0011] in, The signal power variation trend characteristics in the k-th time window; The intercept of the fitted relation; The signal power change rate characteristic of the k-th time window is calculated as follows:

[0012] in, The signal power change rate characteristic of the k-th time window; This is the signal power sample at the (i-1)th sampling point in the k-th time window; The acceleration characteristic of the signal power change in the k-th time window is calculated as follows:

[0013] in, The acceleration characteristic of the signal power change in the k-th time window; This is the signal power sample at the (i-3)th sampling point in the k-th time window; Construct the feature sequence for the k-th time window to characterize the statistical properties of the signal power samples, specifically as follows:

[0014] in, The k-th time window is used to characterize the statistical properties of the signal power samples.

[0015] A further improvement of this invention lies in that, based on a long short-term memory network model, time-series modeling of the feature sequence is performed to construct a time-series representation result sequence with Z time windows, specifically: The system determines whether the time window corresponding to the feature sequence to be modeled is the first time window. If the result is yes, the feature sequence to be modeled is input into the Long Short-Term Memory (LSTM) network model. After processing by the gating mechanism of the LSM network model, the time representation result is output. If the result is no, the feature sequence to be modeled and the time representation result of the time window preceding the time window corresponding to the feature sequence to be modeled are input into the LSM network model. After processing by the gating mechanism of the LSM network model, the time representation result is output. The time series representation results of all time windows are constructed into a time series representation result sequence; The gating mechanism of the Long Short-Term Memory (LSTM) network model includes the forget gate, the input gate, and the output gate.

[0016] A further improvement of this invention is that, assuming the sequence of time-series characterization results is {h1, h2, ..., h...} Z}, h1, h2, ..., h Z The time series representation results are for the 1st, 2nd, ..., Zth time windows, respectively. Based on the current time window, and according to the first preset time window length N and the second preset time window length M, the first time-series representation result sequence of the current window to be detected and the second time-series representation result sequence of the historical reference window are extracted from the time-series representation result sequence, specifically as follows: Extract the first temporal representation result sequence of the current window to be detected from the temporal representation result sequence, specifically {h k-N+1 ,h k-N+2 ,...,h k}, h k-N+1 ,h k-N+2 ,...,h k These are the time series representation results for the (k-N+1), (k-N+2), ..., (k) time windows, respectively; Extract the second time series representation result sequence of the historical reference window from the time series representation result sequence, specifically {h k-N-M+1 ,h k-N-M+2 ,...,h k-N}, hk-N-M+1 ,h k-N-M+2 ,...,h k-N The time series representation results for the kN-M+1, kN-M+2, ..., kN time windows are respectively; The distributional differences within the current time window include mean drift, Kolmogorov–Smirnov distance drift, Jensen–Shannon divergence drift, and Wasserstein distance drift. Based on the first and second time-series characterization result sequences, the distribution difference measure of the current time window is calculated, specifically including the following steps: Based on the first time-series representation result sequence, the feature distribution of the current window to be detected is constructed as follows:

[0017] in, The feature distribution of the current window to be detected; is the benchmark value for comparing the time series characterization results; Π() is an indicator function, which takes the value 1 when the condition in ( ) is true, and 0 otherwise; This represents the time series representation result for the i-th time window; Based on the second time-series representation result sequence, the feature distribution of the historical reference window is constructed as follows:

[0018] in, The characteristic distribution of the historical reference window; The mean shift metric for the k-th time window is as follows:

[0019] in, This is a measure of the mean shift in the k-th time window; The Kolmogorov–Smirnov distance drift metric for the k-th time window is as follows:

[0020] in, The Kolmogorov–Smirnov distance drift metric for the k-th time window; This indicates traversing all Take later The maximum value; The Jensen–Shannon divergence drift metric for the k-th time window is obtained through the following steps: The range of values ​​for the time-series characterization results in the first and second time-series characterization result sequences is evenly divided into several intervals. By statistically analyzing the proportion of time-series characterization results in each interval, the discrete probability distribution of the current detection window and the discrete probability distribution of the historical reference window are obtained. The Jensen-Shannon divergence drift metric for the k-th time window is then calculated, specifically as follows:

[0021] in, The Jensen–Shannon divergence drift metric for the k-th time window; For range index; The total number of intervals; For falling into the historical reference window Discrete probabilities of each interval; For the current window to be detected, the first one to fall into the detection window Discrete probabilities of each interval; It is a logarithmic function; for and The average value; The Wasserstein distance drift metric for the k-th time window is obtained through the following steps: Arrange the time-series representation results in the first and second time-series representation result sequences in ascending order. Take a number of quantiles uniformly within the interval [0,1] to calculate the Wasserstein distance drift metric for the k-th time window. Specifically:

[0022] in, The Wasserstein distance drift metric for the k-th time window. The total number of quantiles; For quantile index; The quantile point values ​​are as follows:

[0023] The first value is obtained through the following steps: The N time-series characterization results of the current detection window are arranged in ascending order, and the first value is located at... The temporal representation result at the location is the first value; The second value is obtained through the following steps: The M time-series characterization results of the historical reference window are arranged in ascending order, located at... The temporal representation result at the location is the second value.

[0024] A further improvement of this invention lies in standardizing and fusing the distribution difference measure to obtain a comprehensive distribution drift index for the current time window, specifically including the following steps: The mean drift metric is standardized as follows:

[0025] in, This is a measure of mean shift after standardization. This is a median function; This is the function for calculating the absolute median deviation. This is the correction coefficient for the median absolute deviation calculation function; To prevent extremely small positive numbers with a denominator of zero; This is the set of mean shift measures for the N+M+1 to k-1 time windows; The Kolmogorov–Smirnov distance drift metric is standardized as follows:

[0026] in, The standardized Kolmogorov–Smirnov distance drift metric; For the set of Kolmogorov–Smirnov distance drift metrics for the N+M+1 to k-1 time windows; The Jensen–Shannon divergence drift metric is standardized as follows:

[0027] in, The Jensen–Shannon divergence drift metric after standardization; For the set of Jensen–Shannon divergence drift metrics for the N+M+1 to k-1 time windows; The Wasserstein distance drift metric is standardized as follows:

[0028] in, The standardized Wasserstein distance drift metric; For the set of Wasserstein distance drift metrics for the N+M+1 to k-1 time windows; The standardized mean drift metric, Kolmogorov–Smirnov distance drift metric, Jensen–Shannon divergence drift metric, and Wasserstein distance drift metric are fused to obtain the comprehensive distribution drift index for the k-th time window, specifically:

[0029] in, This is the comprehensive distribution drift index for the k-th time window; These are the weighting coefficients for the standardized mean shift metric. The weighting coefficients for the standardized Kolmogorov–Smirnov distance drift metric; Weighting coefficients for the standardized Jensen–Shannon divergence drift metric; These are the weighting coefficients for the standardized Wasserstein distance drift metric.

[0030] A further improvement of this invention lies in the dynamic threshold corresponding to the time window, specifically:

[0031] in, is the dynamic threshold for the k-th time window; S is the sensitivity coefficient; This is the set of comprehensive distribution drift indicators for the N+M+1 to k-1 time windows.

[0032] The present invention also provides a GNSS step-type deception interference detection system, comprising: The first module is used to collect several signal power samples output by the GNSS receiver. The signal power samples are segmented using a sliding window method to construct a signal power sample set with Z time windows. The second module is used to construct a feature sequence to characterize the statistical characteristics of signal power samples based on the signal power sample set of each time window, and to perform time-series modeling on the feature sequence based on the long short-term memory network model to construct a time-series characterization result sequence of Z time windows. The third module is used to extract the first temporal representation result sequence of the current window to be detected and the second temporal representation result sequence of the historical reference window from the temporal representation result sequence based on the current time window and according to the first preset time window length N and the second preset time window length M respectively; let the current time window be the k-th time window, the current window to be detected include the k-N+1, k-N+2, ..., k-th time windows, and the historical reference windows include the kN-M+1, kN-M+2, ..., kN-th time windows; where Z≥k≥N+M, k∈[ZNM, Z]; The fourth module is used to calculate the distribution difference measure of the current time window based on the first time series representation result sequence and the second time series representation result sequence, and to standardize and fuse the distribution difference measure to obtain the comprehensive distribution drift index of the current time window. The fifth module is used to construct a dynamic threshold based on the comprehensive distribution drift index. If the comprehensive distribution drift index of each time window in the current detection window is greater than the dynamic threshold of the corresponding time window, it is determined that there is step-type deception interference in the current detection window.

[0033] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the GNSS step-type deception interference detection method described above.

[0034] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the GNSS step-type deception interference detection method described above.

[0035] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the GNSS step-type deception interference detection method described above.

[0036] Compared with the prior art, the positive and progressive effects of the present invention are as follows: This invention provides a GNSS step-type deception interference detection method. It constructs a signal power sample set through sliding window segmentation and builds a feature sequence to characterize the statistical properties of the signal power samples. A long short-term memory network model is used for time-series modeling, effectively capturing the temporal dynamic characteristics of the signal power samples. By calculating and fusing the distribution difference measure between the current detection window and the historical reference window, a comprehensive distribution drift index is constructed, which can sensitively and accurately reflect the drift of the signal power sample distribution. Finally, a dynamic threshold is constructed based on the comprehensive distribution drift index for decision-making, enabling the detection standard to adapt to environmental changes, thereby improving the reliability and accuracy of GNSS step-type deception interference detection under dynamic environmental conditions and varying environmental noise characteristics. Attached Figure Description

[0037] The accompanying drawings are provided to further understand the invention and constitute a part of this invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0038] Figure 1 This is a flowchart illustrating a GNSS step-type deception interference detection method according to the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] In the description of this invention, it should be understood that the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0041] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0042] It should be understood that although terms such as first, second, third, etc., may be used in the embodiments of the present invention to describe the preset range, these preset ranges should not be limited to these terms. These terms are only used to distinguish the preset ranges from one another. For example, without departing from the scope of the embodiments of the present invention, the first preset range may also be referred to as the second preset range, and similarly, the second preset range may also be referred to as the first preset range.

[0043] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0044] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. This is an explanation of the present invention and not a limitation thereof.

[0045] Existing signal quality monitoring methods typically rely on setting empirical thresholds based on a single characteristic index of signal power for judgment. This method is effective in static or ideal noise environments, but it is difficult to accurately distinguish between power spikes caused by deceptive interference and normal signal fluctuations when the receiver is in dynamic motion or the ambient noise changes. This results in insufficient detection reliability and high false detection and false negative rates.

[0046] For the above issues, please refer to Figure 1 This invention provides a GNSS step-type deception interference detection method, comprising the following steps: Several signal power samples output by the GNSS receiver are collected, and the signal power samples are segmented using a sliding window method to construct a signal power sample set with Z time windows. Based on the signal power sample set for each time window, a feature sequence is constructed to characterize the statistical characteristics of the signal power sample. The feature sequence is then time-series modeled based on a long short-term memory network model to obtain a time-series characterization result sequence for Z time windows. Based on the current time window, according to the first preset time window length N and the second preset time window length M, extract the first time-series representation result sequence of the current window to be detected and the second time-series representation result sequence of the historical reference window from the time-series representation result sequence; let the current time window be the k-th time window, the current window to be detected include the k-N+1, k-N+2, ..., k-th time windows, and the historical reference window includes the kN-M+1, kN-M+2, ..., kN-th time windows; where Z≥k≥N+M, k∈[ZNM, Z]; Based on the first time series characterization result sequence and the second time series characterization result sequence, the distribution difference measure of the current time window is calculated. The distribution difference measure is then standardized and fused to obtain the comprehensive distribution drift index of the current time window. Based on the comprehensive distribution drift index, a dynamic threshold is constructed; if the comprehensive distribution drift index of each time window within the current detection window is greater than the dynamic threshold of the corresponding time window, it is determined that there is step-type deception interference in the current detection window.

[0047] This invention constructs a signal power sample set through a sliding window segmentation process and builds a feature sequence to characterize the statistical properties of the signal power samples. It utilizes a long short-term memory network model for time-series modeling, effectively capturing the temporal dynamics of the signal power samples. By calculating and fusing the distribution difference measure between the current detection window and the historical reference window, a comprehensive distribution drift index is constructed, which can sensitively and accurately reflect the drift of the signal power sample distribution. Finally, a dynamic threshold is constructed based on the comprehensive distribution drift index for decision-making, enabling the detection standard to adapt to environmental changes. This improves the reliability and accuracy of GNSS step-type deception interference detection under dynamic environments and varying environmental noise characteristics.

[0048] During operation, a GNSS receiver continuously outputs signal power observations, which are collected at a preset sampling frequency as signal power samples. To perform local statistical analysis on the continuous signal stream, a sliding window approach is used to segment the long-term signal power samples. The sliding window is truncated according to a preset time length or sample size, with each segment corresponding to a time window. By segmenting the signal power samples, the continuous signal power changes are discretized into statistical characteristic analysis problems within Z time windows, thereby capturing the fluctuation characteristics of the signal power samples at local time scales. It should be understood that the length of the sliding window can be adjusted according to the rate of signal power change in the actual application scenario. For example, in high-speed dynamic scenarios, the window length can be shortened to improve time resolution, while in static or slow-moving scenarios, the window length can be increased to improve statistical stability.

[0049] In a specific embodiment of the present invention, several signal power samples output by a GNSS receiver are collected, and the signal power samples are segmented using a sliding window method to construct a signal power sample set with Z time windows, specifically: Z×L signal power samples are collected from the GNSS receiver output. These Z×L samples are then segmented using a sliding window approach, with a total of Z sliding windows. This results in Z time-window signal power sample sets, each containing L sampling points. Each sampling point yields one signal power sample, and each signal power sample set contains L signal power samples collected from the L sampling points. The signal power sample set for the Zth time window is P. Z ={P Z,（Z-1）L+1 , P Z,（Z-1）L+2 , ..., P Z,（Z-1）L+L}, where P Z,（Z-1）L+1 , P Z,（Z-1）L+2 , ..., P Z,（Z-1）L+L These are the signal power samples at the 1st, 2nd, ..., Lth sampling points of the Zth time window.

[0050] Long Short-Term Memory (LSTM) network models possess memory capabilities, enabling them to capture the temporal dependencies of feature sequences. Through temporal modeling, they can more accurately reflect the true state of signal power samples in dynamic environments, avoiding misjudgments caused by instantaneous noise or normal dynamic fluctuations.

[0051] This method defines two key time windows: one is the current detection window, with a length of N. The first time-series representation result sequence contains the time-series representation results of N time windows, representing the state of the signal power samples in the current detection window; the other is the historical reference window, with a length of M, which immediately precedes the current detection window and represents the state benchmark of the signal power samples not affected by step-type spoofing interference. To detect whether step-type spoofing interference exists in the current detection window, it is necessary to compare the current detection window with the historical reference window. By setting the lengths of N and M, the detection sensitivity and benchmark stability can be flexibly adjusted. For example, a smaller N value can capture abrupt changes more quickly, while a larger M value can provide a more robust state benchmark.

[0052] By calculating the distribution difference measure between the first time-series characterization result sequence and the second time-series characterization result sequence, the degree of deviation of the state of the signal power sample in the current detection window from the state benchmark of the signal power sample that has not been affected by step-type spoofing can be quantified.

[0053] Considering that GNSS receivers may operate in dynamically changing environments in practical applications, the normal fluctuation range of signal power samples will also change accordingly, making it difficult for fixed thresholds to adapt to such variations. Therefore, this invention constructs a dynamic threshold based on a comprehensive distribution drift index. This dynamic threshold can adaptively adjust according to the statistical characteristics of historical signal power samples, thereby maintaining high detection sensitivity while ensuring a low false detection rate. In terms of the judgment logic, the method of this invention requires that the comprehensive distribution drift index of each time window within the current detection window is greater than the dynamic threshold. This judgment condition effectively eliminates instantaneous pulse interference or accidental abnormal fluctuations, ensuring that only continuous step changes are judged as step-type deception interference, thus significantly improving the reliability of the detection results.

[0054] Specifically, let the signal power sample set for the k-th time window be {P}. k,1 , P k,2 , ..., P k,L}, P k,1 ,P k,2 , ..., P k,LFor the signal power samples at the 1st, 2nd, ..., Lth sampling points of the kth time window, a feature sequence is constructed to characterize the statistical properties of the signal power samples. This includes the following steps: The mean characteristic of the k-th time window is calculated as follows:

[0055] in, The total number of sampling points in the k-th time window; Let be the mean feature of the k-th time window; i is the index of the sampling point in the k-th time window; This is the signal power sample at the i-th sampling point in the k-th time window; The summation symbol; The standard deviation feature of the k-th time window is calculated as follows:

[0056] in, The standard deviation feature of the k-th time window; The range feature of the k-th time window is calculated as follows:

[0057] in, This represents the range feature of the k-th time window; This represents the maximum signal power sample in the k-th time window; The minimum signal power sample in the k-th time window; The signal power variation trend characteristics in the k-th time window are calculated as follows: Using L signal power samples within the k-th time window as the dependent variable, the least squares method is used to fit the data to obtain the fitted relationship. The slope in the fitted relationship is used as the signal power change trend feature of the k-th time window. The fitting relationship is:

[0058] in, The signal power variation trend characteristics in the k-th time window; The intercept of the fitted relation; The signal power change rate characteristic of the k-th time window is calculated as follows:

[0059] in, The signal power change rate characteristic of the k-th time window; This is the signal power sample at the (i-1)th sampling point in the k-th time window; The acceleration characteristic of the signal power change in the k-th time window is calculated as follows:

[0060] in, The acceleration characteristic of the signal power change in the k-th time window; This is the signal power sample at the (i-3)th sampling point in the k-th time window; Construct the feature sequence for the k-th time window to characterize the statistical properties of the signal power samples, specifically as follows:

[0061] in, The k-th time window is used to characterize the statistical properties of the signal power samples.

[0062] The mean feature reflects the central tendency of the signal power sample within the current time window and is a fundamental indicator for judging the overall rise of the signal power sample. The standard deviation feature measures the dispersion of the signal power sample around the mean feature. At the moment of step-type deception interference, the fluctuation characteristics of the signal power sample often change, and the standard deviation feature can effectively capture this dynamic change. The range feature can reflect the sudden peak of the signal power sample within the current time window, making up for the slow response of the standard deviation feature to local extreme changes. The signal power change trend feature extracts the overall trend of the signal power sample change through linear fitting and can distinguish between normal linear drift and abnormal step-type sudden changes. The signal power change rate feature reflects the first derivative characteristics of the signal power sample change and can intuitively reflect the speed of change. The signal power change acceleration feature reflects the second derivative characteristics of the signal power sample change. For step-type deception interference, its typical feature is that the signal power sample undergoes a drastic change in a very short time. This change will show a significant abnormal peak in the acceleration dimension. Compared with the mean feature or the signal power change rate feature, the signal power change acceleration feature has a stronger forward-looking capture ability and can improve the timeliness of detection.

[0063] The six features described above constitute a complementary feature sequence. The mean and standard deviation features describe the static statistical properties of the distribution, the range feature enhances the sensitivity to local anomalies, and the signal power change trend, signal power change rate, and signal power change acceleration features reflect the changing patterns of the signal power samples from a dynamic evolution perspective. This construction of a multi-dimensional feature sequence effectively solves the problem of insufficient reliability in single-feature detection.

[0064] Specifically, based on the Long Short-Term Memory (LSTM) network model, time-series modeling of the feature sequence is performed to construct a time-series representation sequence with Z time windows, as follows: The system determines whether the time window corresponding to the feature sequence to be modeled is the first time window. If the result is yes, the feature sequence to be modeled is input into the Long Short-Term Memory (LSTM) network model. After processing by the gating mechanism of the LSM network model, the time representation result is output. If the result is no, the feature sequence to be modeled and the time representation result of the time window preceding the time window corresponding to the feature sequence to be modeled are input into the LSM network model. After processing by the gating mechanism of the LSM network model, the time representation result is output. The time series representation results of all time windows are constructed into a time series representation result sequence; The gating mechanism of the Long Short-Term Memory (LSTM) network model includes the forget gate, the input gate, and the output gate.

[0065] The algorithm determines whether the time window corresponding to the feature sequence to be modeled is the first time window. If the result is yes, it means there is no historical time window information available for reference. In this case, the feature sequence to be modeled is directly input into the Long Short-Term Memory (LSTM) network model. After processing by the LTM network model's gating mechanism, the time representation result is output. If the result is no, it means that historical time window information has been accumulated before. In this case, the feature sequence to be modeled and the time representation result of the previous time window corresponding to the time window to be modeled are input together into the LTM network model. After processing by the LTM network model's gating mechanism, the time representation result is output. It should be understood that the time representation result of the previous time window carries the memory of the signal statistical characteristics of the historical time window. Inputting it together with the feature sequence to be modeled allows the LTM network model to effectively suppress the interference of instantaneous noise.

[0066] Specifically, let the sequence of time series representation results be {h1, h2, ..., h...} Z}, h1, h2, ..., h Z The time series representation results are for the 1st, 2nd, ..., Zth time windows, respectively. Based on the current time window, and according to the first preset time window length N and the second preset time window length M, the first time-series representation result sequence of the current window to be detected and the second time-series representation result sequence of the historical reference window are extracted from the time-series representation result sequence, specifically as follows: Extract the first temporal representation result sequence of the current window to be detected from the temporal representation result sequence, specifically {h k-N+1 ,h k-N+2 ,...,h k}, h k-N+1 ,h k-N+2 ,...,h kThese are the time series representation results for the (k-N+1), (k-N+2), ..., (k) time windows, respectively; Extract the second time series representation result sequence of the historical reference window from the time series representation result sequence, specifically {h k-N-M+1 ,h k-N-M+2 ,...,h k-N}, h k-N-M+1 ,h k-N-M+2 ,...,h k-N The time series representation results for the kN-M+1, kN-M+2, ..., kN time windows are respectively; The distributional differences within the current time window include mean drift, Kolmogorov–Smirnov distance drift, Jensen–Shannon divergence drift, and Wasserstein distance drift. Based on the first and second time-series characterization result sequences, the distribution difference measure of the current time window is calculated, specifically including the following steps: Based on the first time-series representation result sequence, the feature distribution of the current window to be detected is constructed as follows:

[0067] in, The feature distribution of the current window to be detected; is the benchmark value for comparing the time series characterization results; Π() is an indicator function, which takes the value 1 when the condition in ( ) is true, and 0 otherwise; This represents the time series representation result for the i-th time window; Based on the second time-series representation result sequence, the feature distribution of the historical reference window is constructed as follows:

[0068] in, The characteristic distribution of the historical reference window; The mean shift metric for the k-th time window is as follows:

[0069] in, This is a measure of the mean shift in the k-th time window; The Kolmogorov–Smirnov distance drift metric for the k-th time window is as follows:

[0070] in, The Kolmogorov–Smirnov distance drift metric for the k-th time window; This indicates traversing all Take later The maximum value; The Jensen–Shannon divergence drift metric for the k-th time window is obtained through the following steps: The range of values ​​for the time-series characterization results in the first and second time-series characterization result sequences is evenly divided into several intervals. By statistically analyzing the proportion of time-series characterization results in each interval, the discrete probability distribution of the current detection window and the discrete probability distribution of the historical reference window are obtained. The Jensen-Shannon divergence drift metric for the k-th time window is then calculated, specifically as follows:

[0071] in, The Jensen–Shannon divergence drift metric for the k-th time window; For range index; The total number of intervals; For falling into the historical reference window Discrete probabilities of each interval; For the current window to be detected, the first one to fall into the detection window Discrete probabilities of each interval; It is a logarithmic function; for and The average value; The Wasserstein distance drift metric for the k-th time window is obtained through the following steps: Arrange the time-series representation results in the first and second time-series representation result sequences in ascending order. Take a number of quantiles uniformly within the interval [0,1] to calculate the Wasserstein distance drift metric for the k-th time window. Specifically:

[0072] in, The Wasserstein distance drift metric for the k-th time window. The total number of quantiles; For quantile index; The quantile point values ​​are as follows:

[0073] The first value is obtained through the following steps: The N time-series characterization results of the current detection window are arranged in ascending order, and the first value is located at... The temporal representation result at the location is the first value; The second value is obtained through the following steps: The M time-series characterization results of the historical reference window are arranged in ascending order, located at... The temporal representation result at the location is the second value.

[0074] The method of this invention constructs a multi-dimensional quantity system including mean drift metric, Kolmogorov–Smirnov distance drift metric, Jensen–Shannon divergence drift metric, and Wasserstein distance drift metric. These four drift metrics comprehensively reflect the distribution differences between the current detection window and the historical reference window from the perspectives of location center, cumulative distribution pattern, probability similarity, and geometric distance, thereby improving the robustness of step-type deception interference detection.

[0075] Specifically, the distribution difference measure is standardized and fused to obtain the comprehensive distribution drift index for the current time window, which includes the following steps: The mean drift metric is standardized as follows:

[0076] in, This is a measure of mean shift after standardization. This is a median function; This is the function for calculating the absolute median deviation. This is the correction coefficient for the median absolute deviation calculation function; To prevent extremely small positive numbers with a denominator of zero; This is the set of mean shift measures for the N+M+1 to k-1 time windows; The Kolmogorov–Smirnov distance drift metric is standardized as follows:

[0077] in, The standardized Kolmogorov–Smirnov distance drift metric; For the set of Kolmogorov–Smirnov distance drift metrics for the N+M+1 to k-1 time windows; The Jensen–Shannon divergence drift metric is standardized as follows:

[0078] in, The Jensen–Shannon divergence drift metric after standardization; For the set of Jensen–Shannon divergence drift metrics for the N+M+1 to k-1 time windows; The Wasserstein distance drift metric is standardized as follows:

[0079] in, The standardized Wasserstein distance drift metric; For the set of Wasserstein distance drift metrics for the N+M+1 to k-1 time windows; The standardized mean drift metric, Kolmogorov–Smirnov distance drift metric, Jensen–Shannon divergence drift metric, and Wasserstein distance drift metric are fused to obtain the comprehensive distribution drift index for the k-th time window, specifically:

[0080] in, This is the comprehensive distribution drift index for the k-th time window; These are the weighting coefficients for the standardized mean shift metric. The weighting coefficients for the standardized Kolmogorov–Smirnov distance drift metric; Weighting coefficients for the standardized Jensen–Shannon divergence drift metric; These are the weighting coefficients for the standardized Wasserstein distance drift metric.

[0081] Through the standardization process described above, all four drift metrics are converted into dimensionless standardized scores, eliminating the influence of differences in numerical magnitude and laying the foundation for subsequent fusion processing. The comprehensive distribution drift index obtained through fusion processing can more comprehensively and robustly reflect the true drift of signal power samples, avoiding misjudgments or omissions that may occur with a single index. The fusion process is equivalent to constructing a multi-dimensional joint decision mechanism, with each drift metric assigned a corresponding weight based on its sensitivity to distribution differences. For example, if the step change in signal power is significant in the application scenario, the weight of the mean drift metric can be appropriately increased; if the change in signal distribution pattern is more critical in the application scenario, the weight of the Kolmogorov–Smirnov distance drift metric or the Jensen–Shannon divergence drift metric can be increased.

[0082] Specifically, the dynamic threshold corresponding to the time window is as follows:

[0083] in, is the dynamic threshold for the k-th time window; S is the sensitivity coefficient; This is the set of comprehensive distribution drift indicators for the N+M+1 to k-1 time windows.

[0084] In traditional signal quality monitoring methods, empirical thresholds are set and then fixed. However, in practical applications, GNSS receivers may be in different motion states or environments, and fixed empirical thresholds are insufficient to meet the detection requirements of different environments: if the empirical threshold is set too high, it is easy to miss detections in low-noise environments; if the empirical threshold is set too low, it is easy to make false detections in high-noise or dynamic environments. This method introduces a dynamic threshold and uses the statistical characteristics (median and MAD) of the comprehensive distribution drift index set of historical windows to construct the decision criterion for the current time window. That is, the dynamic threshold can adaptively adjust with changes in the environment. This adaptive mechanism fundamentally solves the problem of high false detection rate of fixed thresholds in dynamic environments.

[0085] Based on the same inventive concept, this invention also provides a GNSS step-type deception interference detection system, comprising: The first module is used to collect several signal power samples output by the GNSS receiver. The signal power samples are segmented using a sliding window method to construct a signal power sample set with Z time windows. The second module is used to construct a feature sequence to characterize the statistical characteristics of signal power samples based on the signal power sample set of each time window, and to perform time-series modeling on the feature sequence based on the long short-term memory network model to construct a time-series characterization result sequence of Z time windows. The third module is used to extract the first temporal representation result sequence of the current window to be detected and the second temporal representation result sequence of the historical reference window from the temporal representation result sequence based on the current time window and according to the first preset time window length N and the second preset time window length M respectively; let the current time window be the k-th time window, the current window to be detected include the k-N+1, k-N+2, ..., k-th time windows, and the historical reference windows include the kN-M+1, kN-M+2, ..., kN-th time windows; where Z≥k≥N+M, k∈[ZNM, Z]; The fourth module is used to calculate the distribution difference measure of the current time window based on the first time series representation result sequence and the second time series representation result sequence, and to standardize and fuse the distribution difference measure to obtain the comprehensive distribution drift index of the current time window. The fifth module is used to construct a dynamic threshold based on the comprehensive distribution drift index. If the comprehensive distribution drift index of each time window in the current detection window is greater than the dynamic threshold of the corresponding time window, it is determined that there is step-type deception interference in the current detection window.

[0086] The GNSS step-type deception interference detection system provided by this invention can effectively improve the reliability and accuracy of GNSS step-type deception interference detection. Specifically, by collecting signal power samples output by the GNSS receiver, segmenting them using a sliding window method, a signal power sample set for each time window is constructed. Based on the signal power sample set, a feature sequence representing statistical characteristics is constructed. A long short-term memory network model is used to perform time-series modeling on the feature sequence to obtain a time-series representation result sequence. The time-series representation result sequences of the current detection window and the historical reference window are extracted from the time-series representation result sequence. The distribution difference metric of the current time window is calculated. The distribution difference metric is standardized and fused to obtain a comprehensive distribution drift index. A dynamic threshold is constructed based on the comprehensive distribution drift index. If the comprehensive distribution drift index of each time window within the current detection window is greater than the corresponding dynamic threshold, then step-type deception interference is determined to exist.

[0087] Based on the same inventive concept, this application provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of a GNSS step-type deception interference detection method. The memory may include main memory, such as high-speed random access memory, or it may also include non-volatile memory, such as at least one disk storage device. The processor, network interface, and memory are interconnected via an internal bus, which may be an industry-standard architecture bus, a peripheral component interconnection standard bus, an extended industry-standard architecture bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. The memory stores the program; specifically, the program may include program code, which includes computer operation instructions. The memory may include main memory and non-volatile memory, and provides instructions and data to the processor.

[0088] Based on the same inventive concept, embodiments of this application provide a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the steps of the GNSS step-type deception interference detection method. Specifically, the computer-readable storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. The volatile memory may include RAM (Random Access Memory) and / or cache memory, etc. The non-volatile memory may include ROM (Read-Only Memory), hard disk, flash memory, optical disk, magnetic disk, etc.

[0089] Based on the same inventive concept, this application provides a computer program product, which includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions, which, when executed by a computer device, cause the computer device to perform the steps of the above-described GNSS step-type deception interference detection method.

[0090] Those skilled in the art will understand that embodiments of the present invention can be provided as methods or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM (Compact Disc Read-Only Memory), optical storage, etc.) containing computer-usable program code.

[0091] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer apparatus or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0092] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer device or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0093] These computer program instructions may also be loaded onto a computer device or other programmable data processing equipment to cause a series of operational steps to be performed on the computer device or other programmable equipment to produce a process implemented by the computer device, thereby providing instructions that execute on the computer device or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0094] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0095] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for detecting GNSS step-type deception interference, characterized in that, Includes the following steps: Several signal power samples output by the GNSS receiver are collected, and the signal power samples are segmented using a sliding window method to construct a signal power sample set with Z time windows. Based on the signal power sample set for each time window, a feature sequence is constructed to characterize the statistical characteristics of the signal power sample. The feature sequence is then time-series modeled based on a long short-term memory network model to obtain a time-series characterization result sequence for Z time windows. Based on the current time window, according to the first preset time window length N and the second preset time window length M, extract the first time-series representation result sequence of the current window to be detected and the second time-series representation result sequence of the historical reference window from the time-series representation result sequence; let the current time window be the k-th time window, the current window to be detected include the k-N+1, k-N+2, ..., k-th time windows, and the historical reference window includes the kN-M+1, kN-M+2, ..., kN-th time windows; where Z≥k≥N+M, k∈[ZNM, Z]; Based on the first time series characterization result sequence and the second time series characterization result sequence, the distribution difference measure of the current time window is calculated. The distribution difference measure is then standardized and fused to obtain the comprehensive distribution drift index of the current time window. Based on the comprehensive distribution drift index, a dynamic threshold is constructed; if the comprehensive distribution drift index of each time window within the current detection window is greater than the dynamic threshold of the corresponding time window, it is determined that there is step-type deception interference in the current detection window.

2. The GNSS step-type deception interference detection method according to claim 1, characterized in that, Let the signal power sample set for the k-th time window be {P}. k,1 , P k,2 , ..., P k,L }, P k,1 , P k,2 , ..., P k,L For the signal power samples at the 1st, 2nd, ..., Lth sampling points of the kth time window, a feature sequence is constructed to characterize the statistical properties of the signal power samples. This includes the following steps: The mean characteristic of the k-th time window is calculated as follows: in, The total number of sampling points in the k-th time window; Let be the mean feature of the k-th time window; i is the index of the sampling point in the k-th time window; This is the signal power sample at the i-th sampling point in the k-th time window; The summation symbol; The standard deviation feature of the k-th time window is calculated as follows: in, The standard deviation feature of the k-th time window; The range feature of the k-th time window is calculated as follows: in, This represents the range feature of the k-th time window; This represents the maximum signal power sample in the k-th time window; The minimum signal power sample in the k-th time window; The signal power variation trend characteristics in the k-th time window are calculated as follows: Using L signal power samples within the k-th time window as the dependent variable, the least squares method is used to fit the data to obtain the fitted relationship. The slope in the fitted relationship is used as the signal power change trend feature of the k-th time window. The fitting relationship is: in, The signal power variation trend characteristics in the k-th time window; The intercept of the fitted relation; The signal power change rate characteristic of the k-th time window is calculated as follows: in, The signal power change rate characteristic of the k-th time window; This is the signal power sample at the (i-1)th sampling point in the k-th time window; The acceleration characteristic of the signal power change in the k-th time window is calculated as follows: in, The acceleration characteristic of the signal power change in the k-th time window; This is the signal power sample at the (i-3)th sampling point in the k-th time window; Construct the feature sequence for the k-th time window to characterize the statistical properties of the signal power samples, specifically as follows: in, The k-th time window is used to characterize the statistical properties of the signal power samples.

3. The GNSS step-type deception interference detection method according to claim 2, characterized in that, Temporal modeling of the feature sequence is performed based on the Long Short-Term Memory (LSTM) network model, resulting in a temporal representation sequence of Z time windows, specifically: The system determines whether the time window corresponding to the feature sequence to be modeled is the first time window. If the result is yes, the feature sequence to be modeled is input into the Long Short-Term Memory (LSTM) network model. After processing by the gating mechanism of the LSM network model, the time representation result is output. If the result is no, the feature sequence to be modeled and the time representation result of the time window preceding the time window corresponding to the feature sequence to be modeled are input into the LSM network model. After processing by the gating mechanism of the LSM network model, the time representation result is output. The time series representation results of all time windows are constructed into a time series representation result sequence; The gating mechanism of the Long Short-Term Memory (LSTM) network model includes the forget gate, the input gate, and the output gate.

4. The GNSS step-type deception interference detection method according to claim 3, characterized in that, ... The time series representation result sequence is {h1, h2, ..., h Z }, h1, h2, ..., h Z The time series representation results are for the 1st, 2nd, ..., Zth time windows, respectively. Based on the current time window, and according to the first preset time window length N and the second preset time window length M, the first time-series representation result sequence of the current window to be detected and the second time-series representation result sequence of the historical reference window are extracted from the time-series representation result sequence, specifically as follows: Extract the first temporal representation result sequence of the current window to be detected from the temporal representation result sequence, specifically {h k-N+1 ,h k-N+2 ,...,h k }, h k-N+1 ,h k-N+2 ,...,h k These are the time series representation results for the (k-N+1), (k-N+2), ..., (k) time windows, respectively; Extract the second time series representation result sequence of the historical reference window from the time series representation result sequence, specifically {h k-N-M+1 ,h k-N-M+2 ,...,h k-N }, h k-N-M+1 ,h k-N-M+2 ,...,h k-N The time series representation results for the kN-M+1, kN-M+2, ..., kN time windows are respectively; The distributional differences within the current time window include mean drift, Kolmogorov–Smirnov distance drift, Jensen–Shannon divergence drift, and Wasserstein distance drift. Based on the first and second time-series characterization result sequences, the distribution difference measure of the current time window is calculated, specifically including the following steps: Based on the first time-series representation result sequence, the feature distribution of the current window to be detected is constructed as follows: in, The feature distribution of the current window to be detected; is the benchmark value for comparing the time series characterization results; Π() is an indicator function, which takes the value 1 when the condition in ( ) is true, and 0 otherwise; This represents the time series representation result for the i-th time window; Based on the second time-series representation result sequence, the feature distribution of the historical reference window is constructed as follows: in, The characteristic distribution of the historical reference window; The mean shift metric for the k-th time window is as follows: in, This is a measure of the mean shift in the k-th time window; The Kolmogorov–Smirnov distance drift metric for the k-th time window is as follows: in, The Kolmogorov–Smirnov distance drift metric for the k-th time window; This indicates traversing all Take later The maximum value; The Jensen–Shannon divergence drift metric for the k-th time window is obtained through the following steps: The range of values ​​for the time-series characterization results in the first and second time-series characterization result sequences is evenly divided into several intervals. By statistically analyzing the proportion of time-series characterization results in each interval, the discrete probability distribution of the current detection window and the discrete probability distribution of the historical reference window are obtained. The Jensen-Shannon divergence drift metric for the k-th time window is then calculated, specifically as follows: in, The Jensen–Shannon divergence drift metric for the k-th time window; For range index; The total number of intervals; For falling into the historical reference window Discrete probabilities of each interval; For the current window to be detected, the first one to fall into the detection window Discrete probabilities of each interval; It is a logarithmic function; for and The average value; The Wasserstein distance drift metric for the k-th time window is obtained through the following steps: Arrange the time-series representation results in the first and second time-series representation result sequences in ascending order. Take a number of quantiles uniformly within the interval [0,1] to calculate the Wasserstein distance drift metric for the k-th time window. Specifically: in, The Wasserstein distance drift metric for the k-th time window. The total number of quantiles; For quantile index; The quantile point values ​​are as follows: The first value is obtained through the following steps: The N time-series characterization results of the current detection window are arranged in ascending order, and the first value is located at... The temporal representation result at the location is the first value; The second value is obtained through the following steps: The M time-series characterization results of the historical reference window are arranged in ascending order, located at... The temporal representation result at the location is the second value.

5. The GNSS step-type deception interference detection method according to claim 4, characterized in that, The distribution difference measure is standardized and fused to obtain the comprehensive distribution drift index for the current time window. The specific steps include: The mean drift metric is standardized as follows: in, This is a measure of mean shift after standardization. This is a median function; This is the function for calculating the absolute median deviation. This is the correction coefficient for the median absolute deviation calculation function; To prevent extremely small positive numbers with a denominator of zero; This is the set of mean shift measures for the N+M+1 to k-1 time windows; The Kolmogorov–Smirnov distance drift metric is standardized as follows: in, The standardized Kolmogorov–Smirnov distance drift metric; For the set of Kolmogorov–Smirnov distance drift metrics for the N+M+1 to k-1 time windows; The Jensen–Shannon divergence drift metric is standardized as follows: in, The Jensen–Shannon divergence drift metric after standardization; For the set of Jensen–Shannon divergence drift metrics for the N+M+1 to k-1 time windows; The Wasserstein distance drift metric is standardized as follows: in, The standardized Wasserstein distance drift metric; For the set of Wasserstein distance drift metrics for the N+M+1 to k-1 time windows; The standardized mean drift metric, Kolmogorov–Smirnov distance drift metric, Jensen–Shannon divergence drift metric, and Wasserstein distance drift metric are fused to obtain the comprehensive distribution drift index for the k-th time window, specifically: in, This is the comprehensive distribution drift index for the k-th time window; These are the weighting coefficients for the standardized mean shift metric. The weighting coefficients for the standardized Kolmogorov–Smirnov distance drift metric; Weighting coefficients for the standardized Jensen–Shannon divergence drift metric; These are the weighting coefficients for the standardized Wasserstein distance drift metric.

6. The GNSS step-type deception interference detection method according to claim 5, characterized in that, The dynamic threshold corresponding to the time window is as follows: in, is the dynamic threshold for the k-th time window; S is the sensitivity coefficient; This is the set of comprehensive distribution drift indicators for the N+M+1 to k-1 time windows.

7. A GNSS step-type deception interference detection system, characterized in that, include: The first module is used to collect several signal power samples output by the GNSS receiver. The signal power samples are segmented using a sliding window method to construct a signal power sample set with Z time windows. The second module is used to construct a feature sequence to characterize the statistical characteristics of signal power samples based on the signal power sample set of each time window, and to perform time-series modeling on the feature sequence based on the long short-term memory network model to construct a time-series characterization result sequence of Z time windows. The third module is used to extract the first temporal representation result sequence of the current window to be detected and the second temporal representation result sequence of the historical reference window from the temporal representation result sequence based on the current time window and according to the first preset time window length N and the second preset time window length M respectively; let the current time window be the k-th time window, the current window to be detected include the k-N+1, k-N+2, ..., k-th time windows, and the historical reference windows include the kN-M+1, kN-M+2, ..., kN-th time windows; where Z≥k≥N+M, k∈[ZNM, Z]; The fourth module is used to calculate the distribution difference measure of the current time window based on the first time series representation result sequence and the second time series representation result sequence, and to standardize and fuse the distribution difference measure to obtain the comprehensive distribution drift index of the current time window. The fifth module is used to construct a dynamic threshold based on the comprehensive distribution drift index. If the comprehensive distribution drift index of each time window in the current detection window is greater than the dynamic threshold of the corresponding time window, it is determined that there is step-type deception interference in the current detection window.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the GNSS step-type deception interference detection method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the GNSS step-type deception interference detection method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of any one of the GNSS step-type deception interference detection methods according to claims 1 to 6.

Images