GNSS spoofing interference detection method and system based on LSTM

Abstract

The present invention provides a GNSS deception interference detection method and system based on LSTM, which relates to the technical field of global satellite navigation system. The target signal is obtained through the receiver in the GNSS tracking loop model; the output information of the tracking loop is obtained; four basic SQM detection quantities of Ratio, Delta, ELP, and PCS are selected as the detection quantity set; a feature parameter set is generated according to the output information and the detection quantity set; the feature parameter set is preprocessed by using the sliding window technology to generate standardized data; the standardized data is input into the preset LSTM model to obtain the detection result corresponding to the target signal. The technical effect of effectively identifying navigation deception is achieved.

Description

Technical Field

[0001] The present application relates to the technical field of global satellite navigation systems, and in particular to a GNSS spoofing interference detection method and system based on LSTM. Background Art

[0002] The Global Navigation Satellite System (GNSS) is widely used in various fields. However, due to the weak strength of navigation signals reaching the ground and the openness of the system structure, GNSS is extremely susceptible to interference. According to the different interference production mechanisms, GNSS interference mainly includes suppression and deception interference. Suppression interference causes the receiver to lose lock by transmitting high-power interference signals covering the GNSS working frequency band. At present, its detection and suppression technologies are relatively mature. Deception interference generates false satellite signals, which can cause positioning and timing errors without the knowledge of the receiver, and has strong concealment. Deception interference is divided into forwarding and generative deception: forwarding deception forwards the received real satellite signal to the target receiver after power amplification. Due to the inevitable time delay, it is detected by simple methods such as time verification and multi-correlation peak detection; using the openness of civil GNSS signals, generative deception simulates and generates navigation signals with false positions and false times, so that the receiver tracking loop is locked to the deception signal. It has stronger concealment and destructiveness, and seriously threatens the security of navigation space-time information.

[0003] Therefore, how to effectively prevent GNSS spoofing has become a technical problem that needs to be solved urgently. Summary of the invention

[0004] In order to effectively prevent GNSS spoofing, the present application provides a GNSS spoofing interference detection method and system based on LSTM.

[0005] In the first aspect, the present application provides a GNSS spoofing interference detection method based on LSTM, which adopts the following technical solution:

[0006] A GNSS spoofing interference detection method based on LSTM, comprising:

[0007] Acquire the target signal through the receiver in the GNSS tracking loop model;

[0008] Obtaining output information of a tracking loop in the GNSS tracking loop model;

[0009] Four basic SQM test quantities, Ratio, Delta, ELP, and PCS, were selected as the test quantity set;

[0010] Generate a characteristic parameter set according to the output information and the detection quantity set;

[0011] Preprocessing the characteristic parameter set by using sliding window technology to generate standardized data;

[0012] The standardized data is input into a preset LSTM model to obtain the detection result corresponding to the target signal.

[0013] Optionally, the receiver includes a radio frequency front-end processing module, a baseband digital signal processing module and a positioning and navigation solution module.

[0014] Optionally, the output information of the tracking loop includes: carrier phase, code phase, carrier-to-noise ratio and carrier Doppler frequency shift.

[0015] Optionally, the step of selecting four basic SQM detection quantities of Ratio, Delta, ELP, and PCS as the detection quantity set includes:

[0016] Select the Ratio basic SQM test quantity;

[0017]

[0018] Select the Delta basic SQM test volume;

[0019]

[0020] Select the ELP-based SQM test volume;

[0021]

[0022] Select PCS-based SQM test quantity

[0023]

[0024] in , and The outputs of the early code, late code and prompt code of the in-phase correlator; and are the outputs of the early code and late code of the orthogonal correlator; , and are the relevant output amplitudes of early code, late code and immediate code respectively;

[0025] The four basic SQM detection quantities, namely Ratio, Delta, ELP and PCS, are used as the detection quantity set.

[0026] Optionally, the step of preprocessing the feature parameter set using a sliding window technology to generate standardized data includes:

[0027] Performing standardization preprocessing on the characteristic parameter set;

[0028]

[0029] in, is a set of characteristic parameters; is the feature mean; is the characteristic standard deviation; is the standardized value;

[0030] The windowing technique is used to split the long time series into subsequences of fixed length and slide them with a given step size. Assuming the time length is time series, It is The subsequences generated by sliding windows

[0031]

[0032] is the window length, is the step length, and the subsequence generated by the sliding window is

[0033]

[0034] Set the window length to 100 and the step size to 10ms.

[0035] Optionally, before the step of inputting the standardized data into a preset LSTM model to obtain the detection result corresponding to the target signal, the step further includes:

[0036] Constructing a preset LSTM model, wherein the preset LSTM model includes 2 LSTM layers, 1 Dropout layer and 1 fully connected layer;

[0037] The Dropout layer is located between the LSTM layer and the fully connected layer, and the output result of the preset LSTM model is a classification category.

[0038] Optionally, after the step of inputting the standardized data into a preset LSTM model to obtain the detection result corresponding to the target signal, the method further includes:

[0039] The evaluation indicators are calculated through the confusion matrix: accuracy, recall, precision, and F1 value. Among them, accuracy is the ratio of correctly predicted samples to the total number of samples, recall is the ratio of samples that actually belong to a certain category to those that are successfully predicted to belong to that category, precision is the ratio of samples that are predicted to belong to a certain category to those that actually belong to that category, and F1 value is the harmonic mean of precision and recall, which is used to balance precision and recall. The accuracy, recall, precision and F1 value of the categories are respectively;

[0040]

[0041]

[0042]

[0043]

[0044] in, For the actual category And the prediction is class The number of samples; For categories that are not actually But the prediction is class The number of samples; For categories that are not actually And the prediction is non-class The number of samples; For the actual category But the prediction is non-class The number of samples.

[0045] In a second aspect, the present application provides a GNSS spoofing interference detection system based on LSTM, and the GNSS spoofing interference detection system based on LSTM includes:

[0046] A target signal acquisition module, used for acquiring a target signal through a receiver in a GNSS tracking loop model;

[0047] An output information module, used to obtain output information of the tracking loop in the GNSS tracking loop model;

[0048] The detection quantity set module is used to select four basic SQM detection quantities, namely Ratio, Delta, ELP and PCS, as the detection quantity set;

[0049] A characteristic parameter set module, used to generate a characteristic parameter set according to the output information and the detection quantity set;

[0050] A standardization module, used for preprocessing the characteristic parameter set by using a sliding window technique to generate standardized data;

[0051] The detection result module is used to input the standardized data into a preset LSTM model to obtain the detection result corresponding to the target signal.

[0052] In a third aspect, the present application provides a computer device, comprising: a memory and a processor, wherein the processor executes the method described above when running computer instructions stored in the memory.

[0053] In a fourth aspect, the present application provides a computer-readable storage medium, comprising instructions, which, when executed on a computer, enable the computer to execute the method described above.

[0054] In summary, the present application includes the following beneficial technical effects:

[0055] The present application obtains the target signal through the receiver in the GNSS tracking loop model; obtains the output information of the tracking loop; selects four basic SQM detection quantities, namely Ratio, Delta, ELP and PCS, as the detection quantity set; generates a feature parameter set according to the output information and the detection quantity set; uses the sliding window technology to pre-process the feature parameter set to generate standardized data; and inputs the standardized data into the preset LSTM model to obtain the detection result corresponding to the target signal. The technical effect of effectively identifying navigation deception is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a schematic diagram of the computer device structure of the hardware operating environment involved in the embodiment of the present application;

[0057] Figure 2 It is a flowchart of the first embodiment of the LSTM-based GNSS spoofing interference detection method of the present application;

[0058] Figure 3 It is a tracking loop diagram in the first embodiment of the GNSS spoofing interference detection method based on LSTM of the present application;

[0059] Figure 4 This is a schematic diagram of data sliding window processing of the first embodiment of the GNSS spoofing interference detection method based on LSTM in the present application;

[0060] Figure 5 This is a preset LSTM classification model diagram of the first embodiment of the GNSS spoofing interference detection method based on LSTM of the present application;

[0061] Figure 6 This is a confusion matrix diagram of the LSTM model test set of the first embodiment of the GNSS spoofing interference detection method based on LSTM in this application;

[0062] Figure 7 It is a structural block diagram of the first embodiment of the LSTM-based GNSS spoofing interference detection system of the present application. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below through the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0064] Reference Figure 1 , Figure 1 A schematic diagram of the computer device structure of the hardware operating environment involved in the embodiment of the present application.

[0065] like Figure 1 As shown, the computer device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (Wireless-Fidelity, Wi-Fi) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM), or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk storage. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0066] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the computer device, and may include more or less components than shown in the figure, or combine some components, or arrange the components differently.

[0067] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and a LSTM-based GNSS spoofing interference detection program.

[0068] exist Figure 1 In the computer device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the present application can be set in the computer device, and the computer device calls the LSTM-based GNSS spoofing interference detection program stored in the memory 1005 through the processor 1001, and executes the LSTM-based GNSS spoofing interference detection method provided in the embodiment of the present application.

[0069] The present application embodiment provides a GNSS spoofing interference detection method based on LSTM, referring to Figure 2 , Figure 2This is a flow chart of the first embodiment of the LSTM-based GNSS spoofing interference detection method of the present application.

[0070] In this embodiment, the LSTM-based GNSS spoofing interference detection method includes the following steps:

[0071] Step S10: Acquire the target signal through the receiver in the GNSS tracking loop model.

[0072] It is understandable that GNSS spoofing detection technology includes six categories based on navigation data information, spatial processing, RF front-end, baseband digital signal processing, positioning and navigation calculation results, and machine learning. According to the number of parameters, it can be divided into single-parameter detection method and multi-parameter detection method. The single-parameter detection method includes spatial signal direction of arrival (DOA) information based on array signal processing, automatic gain control (AGC) parameters based on RF front-end, and baseband signal SQM. Compared with the single-parameter detection method, the multi-parameter detection method comprehensively utilizes multiple parameter information, which can effectively make up for the shortcomings of each single parameter and has the ability to accurately identify and distinguish deception categories. The machine learning algorithm can use multiple detection quantities as feature parameters, establish a classification model based on the correlation between data, and detect deception interference. Junzhi Li et al. used the K-Nearest Neighbor (KNN) algorithm to monitor the changes in the number of correlation peaks and the phase difference in small delay deception scenarios, calculated the occurrence frequency of the categories to which the K nearest neighbor points of the test data belonged, predicted the data category, and detected deception interference. The collected GNSS data set verified that this method is superior to traditional multi-peak deception detection in small delay deception scenarios.

[0073] It should be noted that the receiver consists of three signal processing functional modules: RF front-end processing, baseband digital signal processing, and positioning and navigation solution. When there is no deception, the signal received by the receiver contains satellite signals and Gaussian white noise.

[0074]

[0075] in, For satellite signals, It means the mean is 0 and the variance is Gaussian white noise, and

[0076]

[0077] in, , and are signal power, pseudo-random spread spectrum code and navigation information data bit level values ​​respectively; , and are code phase, carrier Doppler shift and carrier phase respectively; is the carrier center frequency of the navigation signal, and j is the imaginary unit in the complex number.

[0078] In order to achieve carrier and spread spectrum code synchronization and obtain information such as carrier phase, code phase, Doppler frequency shift, etc., the tracking loop in baseband digital signal processing is generally designed as follows: Figure 3 The structure shown.

[0079] When there is no deception interference, the signal enters the tracking loop and is divided into in-phase (I) and quadrature-phase (Q) branches through the mixer. The output results of the I branch and the Q branch are

[0080]

[0081]

[0082] in, is the carrier center frequency of the satellite navigation signal;

[0083] After low-pass filtering, the in-phase component and the orthogonal component are obtained.

[0084]

[0085]

[0086] The phase difference between the two signals is obtained by using the loop discriminator. The phase difference between the local carrier and the received signal carrier is

[0087]

[0088] The local carrier NCO is adjusted according to the phase difference to align the local carrier with the input signal carrier phase. Similar to the carrier loop, the I / Q branches of the early code, late code, and prompt code are correlated with the locally generated pseudo-random code. When the receiver is locked, the received signal has exactly the same frequency as the locally generated signal. Assuming that no data bit jump occurs during the coherent integration accumulation time, and the effects of code phase and Doppler frequency shift are independent of each other, the coherent integration output

[0089]

[0090]

[0091] in, , are the outputs of the in-phase and quadrature correlators respectively; , are the noise of the in-phase and quadrature branches respectively; is the code phase difference between the real signal and the locally generated signal; is the cross-correlation function between the real signal and the locally generated signal. Taking the GPS L1 band as an example, the expression is:

[0092]

[0093] In a specific implementation, when there is deception, the signal received by the receiver contains the real signal, the deception signal and Gaussian white noise.

[0094]

[0095] in, It is a deceptive signal. According to the principle of deceptive interference, and The structure is similar, expressed as

[0096]

[0097] in, , and are signal power, pseudo-random spread spectrum code and navigation information data bit level values ​​respectively; , and are code phase, carrier Doppler shift and carrier phase respectively; is the carrier center frequency of the navigation signal, and j is the imaginary unit in the complex number. According to the GNSS tracking loop model, the injection of spoofing signals will cause changes in the tracking loop output information, such as carrier phase, code phase, carrier-to-noise ratio, carrier Doppler shift, etc. The in-phase component and orthogonal component of the carrier loop output are

[0098]

[0099] Taking intermediate GNSS navigation spoofing as an example, the spoofing process is divided into four stages: Stage 1: the spoofing signal enters the tracking loop and gradually aligns with the real signal; Stage 2: the spoofing signal and the real signal are completely synchronized in carrier frequency and code phase; Stage 3: the spoofing signal is pulled with a power higher than the real signal; Stage 4: the receiver tracking loop is locked on the spoofing signal, and the target receiver obtains incorrect time or position information through positioning solution.

[0100] The I / Q branch correlator output is

[0101]

[0102] The cross-correlation function after the correlation operation is

[0103]

[0104] in, is the cross-correlation function between the spoofing signal and the local generated signal. When the code phase difference between the spoofing signal and the real signal is reduced to within 1 chip, the spoofing signal makes The result is distortion.

[0105] Step S20: Obtain output information of the tracking loop in the GNSS tracking loop model. It can be understood that the output information of the tracking loop includes: carrier phase, code phase, carrier-to-noise ratio and carrier Doppler frequency shift.

[0106] Step S30: Select four basic SQM detection quantities, namely Ratio, Delta, ELP and PCS, as the detection quantity set.

[0107] In a specific implementation, the steps of selecting four basic SQM detection quantities, namely Ratio, Delta, ELP and PCS, as the detection quantity set include:

[0108] Select the Ratio basic SQM test quantity;

[0109]

[0110] Select the Delta basic SQM test volume;

[0111]

[0112] Select the ELP-based SQM test volume;

[0113]

[0114] Select PCS-based SQM test quantity

[0115]

[0116] in , and The outputs of the early code, late code and prompt code of the in-phase correlator; and are the outputs of the early code and late code of the orthogonal correlator; , and are the relevant output amplitudes of early code, late code and immediate code respectively;

[0117] The four basic SQM detection quantities, namely Ratio, Delta, ELP and PCS, are used as the detection quantity set.

[0118] Step S40: Generate a characteristic parameter set according to the output information and the detection quantity set.

[0119] In practice, the problem of navigation spoofing detection is essentially a classification problem, in which the selection and configuration of feature parameters have a significant impact on the performance of spoofing detection. The injection of spoofing signals will change the carrier phase, code phase, carrier-to-noise ratio, and carrier Doppler shift output by the tracking loop, so it is necessary to select the direct information output by these tracking loops as the feature parameters for classification.

[0120] At the same time, as a commonly used detection quantity in detecting GNSS multipath and deception, the SQM parameter directly reflects the change of the deception signal on the tracking loop correlation peak. By processing the early code, instant code, and late code outputs of the tracking loop correlator, various types of SQM detection quantities are obtained. Without loss of generality, this embodiment selects four basic SQM detection quantities such as Ratio, Delta, ELP, and PCS as feature parameters for subsequent machine learning. These four types of detection quantities are respectively expressed as

[0121]

[0122]

[0123]

[0124]

[0125] in , and The outputs of the early code, late code and prompt code of the in-phase correlator; and are the outputs of the early code and late code of the orthogonal correlator; , and They are the relevant output amplitudes of early code, late code and immediate code respectively.

[0126] In summary, in order to make full use of baseband signal processing information, this embodiment uses direct information output by the tracking loop, such as carrier phase, code phase, carrier-to-noise ratio and carrier Doppler shift, and SQM detection quantity as characteristic parameters for deception detection.

[0127] Step S50: preprocessing the feature parameter set using a sliding window technique to generate standardized data.

[0128] It is understandable that if Figure 4 The data sliding window processing schematic diagram shown in the figure, the steps of preprocessing the feature parameter set using the sliding window technology to generate standardized data include:

[0129] Performing standardized preprocessing on the feature parameter set;

[0130]

[0131] in, is a set of characteristic parameters; is the feature mean; is the characteristic standard deviation; is the standardized value;

[0132] The windowing technique is used to split the long time series into subsequences of fixed length and slide them with a given step size. Assuming the time length is time series, It is The subsequences generated by sliding windows

[0133]

[0134] is the window length, is the step length, and the subsequence generated by the sliding window is

[0135]

[0136] Set the window length to 100 and the step size to 10ms.

[0137] In the specific implementation, the subsequences generated by the sliding window are the input of LSTM. Each input sequence contains the feature values ​​of the current time point and the previous 99 time points, and the label of the last time point of the subsequence is used as the label of the entire input sequence. The sliding window processing data can effectively capture and memorize the information in the entire time window and is applied in combination with LSTM.

[0138] Step S60: input the standardized data into a preset LSTM model to obtain a detection result corresponding to the target signal.

[0139] It should be noted that before the step of inputting the standardized data into the preset LSTM model to obtain the detection result corresponding to the target signal, the following steps are also included:

[0140] Build a preset LSTM model, which includes 2 LSTM layers, 1 Dropout layer, and 1 fully connected layer;

[0141] The Dropout layer is located between the LSTM layer and the fully connected layer, and the output result of the preset LSTM model is the classification category.

[0142] In the specific implementation, the preset LSTM classification model in this embodiment is as follows Figure 5 shown.

[0143] It is understandable that LSTM is a special recurrent neural network (RNN) that aims to solve the problem of gradient explosion and vanishing when traditional RNN processes long sequence data. LSTM adopts a nonlinear structure and has excellent long sequence data processing capabilities. By introducing three gating units (input gate, forget gate, and output gate), LSTM is allowed to selectively retain or discard information and control the flow of gradients when transmitting information. The impact of deception on features is time-related, that is, the detection quantity in the deception detection problem is manifested as time-related long sequence data, so LSTM is selected as the model for GNSS deception interference detection.

[0144] In a specific implementation, after the step of inputting the standardized data into the preset LSTM model to obtain the detection result corresponding to the target signal, it also includes: calculating the evaluation indicators through the confusion matrix: accuracy, recall, precision, F1 value, wherein the accuracy is the ratio of the number of correctly predicted samples to the total number of samples, the recall is the ratio of samples that actually belong to a certain category to those that are successfully predicted to belong to the category, the precision is the ratio of samples that are predicted to belong to a certain category to those that actually belong to the category, and the F1 value is the harmonic mean of the precision and recall, which is used to balance the precision and recall, and the accuracy, recall, precision and F1 value of the category are respectively;

[0145]

[0146]

[0147]

[0148]

[0149] in, For the actual category And the prediction is class The number of samples; For categories that are not actually But the prediction is class The number of samples; For categories that are not actually And the prediction is non-class The number of samples; For the actual category But the prediction is non-class The number of samples.

[0150] In this embodiment, the experimental language is python 3.8, the deep learning framework is pytorch 1.10.2, the GPU is NVIDIA RTX A6000, the CPU is Intel Core i9-13900k, and the operating system is Windows 10.

[0151] The TEXBAT dataset released by the Radio Navigation Laboratory of the University of Texas at Austin is used. This dataset is a classic experimental data for studying GNSS spoofing. It contains eight sets of spoofing scenarios and two sets of real sampling data of clean scenarios. The spoofing scenario settings of the TEXBAT dataset are shown in Table 1:

[0152]

[0153] Due to the multipath interference in dynamic scenes, the SQM detection method is only applicable to static scenes. To ensure the effectiveness of the comparative experiment, the ds2, ds3, ds4, and ds7 data sets are selected. Each data set corresponds to a different deception scene. Each data set is processed by a software receiver to obtain the characteristic parameters of the tracking process. The processing period of each data set is selected as 0-450s, and the characteristic parameter output interval is 10ms. Each data set can obtain 45,000 sets of characteristic parameters. According to different experimental purposes, ds2, ds3, ds4, ds7 and clean data sets are used as separate data sets or constitute a total data set. Without destroying the time order, the data sets are divided into layers according to the data labels, with the training set accounting for 80% and the test set accounting for 20%.

[0154] The LSTM model parameter settings are shown in Table 2:

[0155]

[0156] The PCS detection amount of different scenarios is used as the input feature, and the data set is set and divided. The number of samples in each data set is 45,000, of which the number of samples in the test set is 36,000 and the number of samples in the test set is 9,000. The LSTM model is trained and tested respectively. The LSTM model test accuracy and threshold detection accuracy are calculated according to the test set confusion matrix and binary judgment results of each data set. The LSTM model test accuracy and threshold detection accuracy under different data sets are shown in Table 3:

[0157]

[0158] Compared with threshold detection, the LSTM model has a higher accuracy rate in deception detection. However, whether using traditional threshold detection or LSTM classification detection, single SQM detection has problems such as insufficient universality and poor immediacy in deception scenarios. To address this problem, the direct information output by the tracking loop, such as carrier phase, code phase, carrier-to-noise ratio, and carrier Doppler shift, is used together with the SQM detection quantity as the input feature parameters of LSTM. According to the model parameter setting, the LSTM output is whether there is deception, that is, 1 / 0. The original data set samples are 180,000 groups, which are divided into layers according to the labels. The test set accounts for 20%, that is, the test data samples are 36,000 groups. The confusion matrix of the LSTM model test set is as follows: Figure 6 shown.

[0159] Depend on Figure 6 It can be seen that compared with the detection volume of a single SQM, the LSTM classification model is used to detect deception by integrating multiple parameters, with fewer misclassified samples and an accuracy rate close to 100%.

[0160] This embodiment obtains the target signal through the receiver in the GNSS tracking loop model; obtains the output information of the tracking loop; selects four basic SQM detection quantities, namely Ratio, Delta, ELP and PCS, as the detection quantity set; generates a feature parameter set according to the output information and the detection quantity set; uses the sliding window technology to pre-process the feature parameter set to generate standardized data; and inputs the standardized data into the preset LSTM model to obtain the detection result corresponding to the target signal. The technical effect of effectively identifying navigation deception is achieved.

[0161] In addition, an embodiment of the present application also proposes a computer-readable storage medium, on which a program for GNSS spoofing interference detection based on LSTM is stored. When the program for GNSS spoofing interference detection based on LSTM is executed by a processor, the steps of the method for GNSS spoofing interference detection based on LSTM as described above are implemented.

[0162] Reference Figure 7 , Figure 7 This is a structural block diagram of the first embodiment of the LSTM-based GNSS spoofing interference detection system of the present application.

[0163] like Figure 7 As shown, the LSTM-based GNSS spoofing interference detection system proposed in the embodiment of the present application includes:

[0164] A target signal acquisition module 10 is used to acquire a target signal through a receiver in a GNSS tracking loop model;

[0165] An output information module 20, used to obtain output information of a tracking loop in a GNSS tracking loop model;

[0166] The detection quantity set module 30 is used to select four basic SQM detection quantities, namely Ratio, Delta, ELP and PCS, as the detection quantity set;

[0167] A characteristic parameter set module 40, used to generate a characteristic parameter set according to the output information and the detection quantity set;

[0168] A standardization module 50, for preprocessing the feature parameter set using a sliding window technique to generate standardized data;

[0169] The detection result module 60 is used to input the standardized data into a preset LSTM model to obtain the detection result corresponding to the target signal.

[0170] It should be understood that the above is only an example and does not constitute any limitation on the technical solution of the present application. In specific applications, technicians in this field can make settings as needed, and the present application does not impose any limitation on this.

[0171] This embodiment obtains the target signal through the receiver in the GNSS tracking loop model; obtains the output information of the tracking loop; selects four basic SQM detection quantities, namely Ratio, Delta, ELP and PCS, as the detection quantity set; generates a feature parameter set according to the output information and the detection quantity set; uses the sliding window technology to pre-process the feature parameter set to generate standardized data; and inputs the standardized data into the preset LSTM model to obtain the detection result corresponding to the target signal. The technical effect of effectively identifying navigation deception is achieved.

[0172] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of the present application. In practical applications, technicians in this field can select part or all of it according to actual needs to achieve the purpose of the present embodiment, and no limitation is made here.

[0173] In addition, for technical details not described in detail in this embodiment, please refer to the LSTM-based GNSS spoofing interference detection method provided in any embodiment of the present application, which will not be repeated here.

[0174] In addition, it should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0175] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0176] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory (ROM) / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods of each embodiment of the present application.

[0177] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A GNSS spoofing interference detection method based on LSTM, characterized in that: include: Acquire the target signal through the receiver in the GNSS tracking loop model; Obtaining output information of a tracking loop in the GNSS tracking loop model; Four basic SQM test quantities, Ratio, Delta, ELP, and PCS, were selected as the test quantity set; Generate a characteristic parameter set according to the output information and the detection quantity set; Preprocessing the characteristic parameter set by using sliding window technology to generate standardized data; The standardized data is input into a preset LSTM model to obtain the detection result corresponding to the target signal.

2. The LSTM-based GNSS spoofing interference detection method according to claim 1, characterized in that: The receiver includes a radio frequency front-end processing module, a baseband digital signal processing module and a positioning and navigation solution module.

3. The LSTM-based GNSS spoofing interference detection method according to claim 1, characterized in that: The output information of the tracking loop includes: carrier phase, code phase, carrier-to-noise ratio and carrier Doppler frequency shift.

4. The LSTM-based GNSS spoofing interference detection method according to claim 1, characterized in that: The step of selecting four basic SQM detection quantities of Ratio, Delta, ELP, and PCS as the detection quantity set includes: Select the Ratio basic SQM test quantity; Select the Delta basic SQM test volume; Select the ELP-based SQM test volume; Select PCS-based SQM test quantity in , and The outputs of the early code, late code and prompt code of the in-phase correlator; and are the outputs of the early code and late code of the orthogonal correlator; , and are the relevant output amplitudes of early code, late code and immediate code respectively; The four basic SQM detection quantities, namely Ratio, Delta, ELP and PCS, are used as the detection quantity set.

5. The LSTM-based GNSS spoofing interference detection method according to claim 1, characterized in that: The step of preprocessing the characteristic parameter set by using the sliding window technology to generate standardized data includes: Performing standardization preprocessing on the characteristic parameter set; in, is a set of characteristic parameters; is the feature mean; is the characteristic standard deviation; is the standardized value; The windowing technique is used to split the long time series into subsequences of fixed length and slide them with a given step size. Assuming the time length is time series, It is The subsequences generated by sliding windows is the window length, is the step length, and the subsequence generated by the sliding window is Set the window length to 100 and the step size to 10ms.

6. The LSTM-based GNSS spoofing interference detection method according to claim 1, characterized in that: Before the step of inputting the standardized data into a preset LSTM model to obtain the detection result corresponding to the target signal, the method further includes: Constructing a preset LSTM model, wherein the preset LSTM model includes 2 LSTM layers, 1 Dropout layer and 1 fully connected layer; The Dropout layer is located between the LSTM layer and the fully connected layer, and the output result of the preset LSTM model is a classification category.

7. The LSTM-based GNSS spoofing interference detection method according to claim 1, characterized in that: After the step of inputting the standardized data into a preset LSTM model to obtain the detection result corresponding to the target signal, the method further includes: The evaluation indicators are calculated through the confusion matrix: accuracy, recall, precision, and F1 value. Among them, accuracy is the ratio of correctly predicted samples to the total number of samples, recall is the ratio of samples that actually belong to a certain category to those that are successfully predicted to belong to that category, precision is the ratio of samples that are predicted to belong to a certain category to those that actually belong to that category, and F1 value is the harmonic mean of precision and recall, which is used to balance precision and recall. The accuracy, recall, precision and F1 value of the categories are respectively; in, For the actual category And the prediction is class The number of samples; For categories that are not actually But the prediction is class The number of samples; For categories that are not actually And the prediction is non-class The number of samples; For the actual category But the prediction is non-class The number of samples.

8. A GNSS spoofing interference detection system based on LSTM, characterized in that: The LSTM-based GNSS spoofing interference detection system includes: A target signal acquisition module, used for acquiring a target signal through a receiver in a GNSS tracking loop model; An output information module, used to obtain output information of the tracking loop in the GNSS tracking loop model; The detection quantity set module is used to select four basic SQM detection quantities, namely Ratio, Delta, ELP and PCS, as the detection quantity set; A characteristic parameter set module, used to generate a characteristic parameter set according to the output information and the detection quantity set; A standardization module, used for preprocessing the characteristic parameter set by using a sliding window technique to generate standardized data; The detection result module is used to input the standardized data into a preset LSTM model to obtain the detection result corresponding to the target signal.

9. A computer device, characterized in that: The device comprises: a memory and a processor, and when the processor runs the computer instructions stored in the memory, the processor executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The method comprises instructions, which, when executed on a computer, cause the computer to execute the method according to any one of claims 1 to 7.

Images