A GNSS spoofing jamming detection method based on multi-step inertial navigation prediction
By adding an independent inertial navigation INS unit to the tightly coupled GNSS/INS model, uncontaminated pseudorange estimates are generated and a deception detection factor is constructed, solving the problems of poor GNSS deception interference detection performance and high false alarm rate, and achieving faster and more accurate deception interference detection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIHANG UNIV
- Filing Date
- 2026-03-11
- Publication Date
- 2026-06-05
Smart Images
Figure CN122151116A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of navigation and positioning technology, and in particular to a GNSS spoofing interference detection method based on multi-step inertial navigation prediction. Background Technology
[0002] In Global Navigation Satellite Systems (GNSS), spoofing and jamming attacks pose a serious security threat. Due to the openness and relatively low power of GNSS signals, the system is vulnerable to spoofing attacks, leading to a significant decrease in positioning accuracy and, in severe cases, even affecting the normal operation of the system. With the widespread application of high-precision navigation systems such as autonomous driving and air navigation, effectively detecting and preventing spoofing and jamming has become a key research focus in this field.
[0003] Currently, research on deception and interference detection technologies can be mainly divided into two categories: 1. Receiver Autonomous Detection Method: This method utilizes the GNSS receiver's own signal processing to detect spoofing interference, primarily relying on signal quality and characteristic parameters for judgment. These include signal power, carrier-to-noise ratio, pseudorange change rate, and the quality of correlation peaks. Existing research indicates that signal power detection methods can leverage the typically high power of spoofing signals to determine if the signal has been spoofed. However, this method suffers from poor reliability in complex, time-varying environments and is susceptible to multipath effects and environmental interference.
[0004] 2. Externally Auxiliary Detection Methods: These methods improve the detection accuracy of deception interference by using auxiliary information such as multiple antennas, inertial navigation, ultra-wideband sensors, and speed logs. Combining an Inertial Navigation System (INS) with GNSS can provide stronger anti-interference capabilities. INS-assisted deception detection methods are widely used and fall into two main categories: the Direct Comparison Method (DCM) and innovation-based deception detection methods. The Direct Comparison Method determines the presence of a deception attack by comparing the differences in acceleration, velocity, and other observations between GNSS and INS. Although this method can provide effective detection results, the false alarm rate is high and detection performance degrades when GNSS observation quality is poor or INS accuracy is insufficient. On the other hand, innovation-based detection methods determine the presence of deception interference by analyzing the innovation quantity in the Kalman filter, which has lower computational complexity. However, in some application scenarios (such as induced deception interference signals), this method may face long alarm delays and poor detection performance.
[0005] While both of these methods can detect deception interference to some extent, existing technologies still have some limitations. First, receiver-autonomous detection methods are often only effective under specific signal characteristics and are easily affected by complex environmental factors, resulting in insufficient reliability and adaptability. Second, while detection methods based on external auxiliary information can improve accuracy, they have high requirements for equipment and environment, and often require additional sensor equipment or complex algorithms, increasing system complexity and cost.
[0006] In addition, traditional innovation-based detection methods have the following drawbacks: (1) Spoofing attacks can deceive the prior estimates of the Extended Kalman Filter (EKF), and the INS will also be affected, thus contaminating the value of the innovation. Traditional detection methods directly use the contaminated innovation for detection, resulting in lower spoofing detection performance. The Time To Alarm (TTA) will also increase. (2) Innovations always contain biases such as ionospheric errors, tropospheric errors, multipath or receiver hardware delays, and therefore cannot follow a Gaussian distribution. This bias significantly increases the false alarm rate. However, traditional innovation-based detection methods have not taken effective measures to deal with the impact of innovation bias. Summary of the Invention
[0007] This invention proposes a GNSS spoofing interference detection method based on multi-step inertial navigation prediction, which solves the problems of poor detection performance, long alarm time and high false alarm rate of traditional innovation-based tightly coupled GNSS / INS integrated navigation system spoofing detection methods.
[0008] To achieve the above objectives, this invention provides a GNSS spoofing interference detection method based on multi-step inertial navigation prediction, comprising: A tightly coupled GNSS / INS model is constructed, and time and observation updates are performed using an extended Kalman filter (EKF). A new independent inertial navigation INS unit is added to the tightly coupled GNSS / INS model to construct a deception detection method model based on innovation and generate uncontaminated pseudorange estimates within N steps. Replace the contaminated elements in the EKF observation matrix and state transition matrix to obtain the reconstructed observation matrix and state transition matrix; Based on the reconstructed observation matrix and state transition matrix, the improved innovation value is calculated, wherein the improved innovation value is the difference between the pseudorange measurement value of the GNSS receiver and the pseudorange estimate value; Utilizing the current time before The average value of the innovation in each epoch predicts the innovation deviation, and an unbiased improved innovation value is obtained. A deception detection factor is constructed based on the unbiased improved innovation value, and a detection statistic is generated. The detection statistic is compared with a preset threshold to determine deception interference.
[0009] Preferably, generating the uncontaminated pseudorange estimate within the N steps includes: Based on the user equipment location, velocity, latitude, and attitude output by the newly added INS unit, calculate the... m Distance vector from each satellite to the user equipment; By combining the prior estimates of clock bias and clock drift from EKF, a new pseudorange estimate is generated, namely the uncontaminated pseudorange estimate.
[0010] Preferably, the newly added INS unit specifically comprises: ; In the formula, This indicates the user equipment speed calculated by the newly added INS unit. This represents the user equipment dimension calculated from the newly added INS unit. This indicates the location of the user equipment calculated by the newly added INS unit. This represents the user equipment attitude calculated by the newly added INS unit. This represents the inertial navigation unit processing function. The output ratio of the inertial navigation unit, Output angular velocity to the inertial navigation unit. This is the propagation interval.
[0011] Preferably, the new pseudorange estimate is: ; In the formula, For the first The distance vector from each satellite to the newly added INS unit is used to calculate the location of the user equipment. and Let represent the EKF prior estimates of clock bias and clock drift, respectively. This is the new pseudorange estimate.
[0012] Preferably, replacing the contaminated elements in the EKF observation matrix and state transition matrix includes: Replace the EKF prior estimated location with the location output by the newly added INS unit; Based on the position output of the newly added INS unit, update the coordinate transformation matrix from the inertial coordinate system to the geocentric-ground-fixed coordinate system to obtain the updated coordinate transformation matrix; The EKF observation matrix and state transition matrix are reconstructed based on the updated coordinate transformation matrix.
[0013] Preferably, the improved innovation value is calculated as follows: ; In the formula, For the first m The pseudorange from the satellite to the user equipment For the first m The deceptive component of satellite pseudorange, The first number calculated by the proposed method m The pseudorange from the satellite to the user equipment The actual location of the user equipment to the number m The true distance between the satellites This refers to the distance error between the predicted position and the actual position of the newly added INS element, caused by the cumulative error of the newly added INS element. k For the current epoch, K To deceive the attack moment, For the first m News about satellite-to-user equipment improvements.
[0014] Preferably, using the time before the current moment The average new interest rate forecast bias for each epoch includes: Extract the current time before The improved innovation value of each epoch is calculated, and the average of the improved innovation values is used as the innovation bias estimate. The unbiased improved innovation value is obtained by subtracting the innovation deviation estimate from the current improved innovation value.
[0015] Preferably, the unbiased improvement innovation value is: ; In the formula, For the first epoch-improved innovation vector For the first The new information vector of the epoch improvement.
[0016] Preferably, the deception detection factor is: ; In the formula, The first number calculated by the proposed method The deception detection factor of the era, For the first Transpose of the epoch-improved innovation vector It is the inverse matrix of the improved information vector covariance matrix.
[0017] Preferably, the newly added independent inertial navigation (INS) unit is corrected every N steps by the output of the GNSS receiver. The correction process is as follows: ; In the formula, For the first The estimated position of the newly added INS unit at the next epoch. For the first The estimated rate of new INS units added at each epoch. For the first The pose estimated by the newly added INS unit at each epoch. For the first The clock drift estimated by the newly added INS unit under the epoch. For the first Clock bias estimated by the newly added INS unit at the epoch. For the first The position estimated by EKF at each epoch. For the first The velocity estimated by EKF at each epoch. For the first Attitude estimated by EKF at epoch. For the first EKF-estimated clock drift at different epochs For the first Clock bias estimated by EKF at epoch.
[0018] Compared with the prior art, the present invention has the following advantages and technical effects: Compared to traditional methods, the method proposed in this invention can detect low-intensity and medium-intensity induced deception interference more quickly and accurately, effectively solves the problem of increased false alarm rate caused by innovation bias, and has better robustness. Attached Figure Description
[0019] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a trajectory diagram of the deceived GNSS receiver, the real user equipment, the tightly coupled system, and the improved method in a deception scenario according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the improved innovation change without innovation deviation in the deception scenario of the simulation scenario of this invention embodiment; Figure 3 This is a schematic diagram showing the changes of the improved deception detection factor without information deviation under three deception intensities in the simulation scenario of this invention, where (a) represents low intensity, (b) represents medium intensity, and (c) represents high intensity. Figure 4 This is a schematic diagram illustrating the change in alarm time at a fixed speed as the deception angular rate increases, according to an embodiment of the present invention. Figure 5 This is a schematic diagram illustrating the change in alarm time with increasing speed at a fixed deception angular rate, according to an embodiment of the present invention. Figure 6 The following is a diagram of a test scenario according to an embodiment of the present invention, wherein (a) is a schematic diagram of the hardware experimental platform, and (b) is a schematic diagram of the real trajectory and the deception trajectory; Figure 7 This is a schematic diagram of the improved innovation change without innovation deviation in a deception scenario in a real-world test scenario according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the changes in the improved deception detection factor without information deviation under three deception scenarios in the actual test scenario of this invention, where (a) is low intensity, (b) is medium intensity, and (c) is high intensity; Figure 9 This is a schematic diagram of the information based on multi-step inertial navigation prediction in an embodiment of the present invention. Detailed Implementation
[0020] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.
[0021] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.
[0022] This invention proposes a GNSS spoofing interference detection method based on multi-step inertial navigation prediction, comprising: A tightly coupled GNSS / INS model is constructed, and time and observation updates are performed using an extended Kalman filter (EKF). A new independent inertial navigation INS unit is added to the tightly coupled GNSS / INS model to construct a deception detection method model based on innovation and generate uncontaminated pseudorange estimates within N steps. Replace the contaminated elements in the EKF observation matrix and state transition matrix to obtain the reconstructed observation matrix and state transition matrix; Based on the reconstructed observation matrix and state transition matrix, the improved innovation value is calculated, wherein the improved innovation value is the difference between the pseudorange measurement value of the GNSS receiver and the pseudorange estimate value; Utilizing the current time before The average value of the innovation in each epoch predicts the innovation deviation, and an unbiased improved innovation value is obtained. A deception detection factor is constructed based on the unbiased improved innovation value, and a detection statistic is generated. The detection statistic is compared with a preset threshold to determine deception interference.
[0023] This embodiment adds an additional INS calculation unit to the original detection framework. Instead of being corrected every N epochs by EKF, it is corrected at every epoch. Then, the output of the new INS unit is used to replace the deceptive prior estimate of the pseudorange, thereby constructing an improved innovation. Furthermore, a sliding window is used to estimate the average value of the innovation in real time, thereby reducing innovation bias.
[0024] Further, generating the uncontaminated pseudorange estimate within the N steps includes: Based on the user equipment location, velocity, latitude, and attitude output by the newly added INS unit, calculate the... m Distance vector from each satellite to the user equipment; By combining the prior estimates of clock bias and clock drift from EKF, a new pseudorange estimate is generated, namely the uncontaminated pseudorange estimate.
[0025] Specifically, the contaminated EKF prior estimates and the contaminated elements in the EKF matrix are replaced with uncontaminated INS output values within N steps: The newly added INS unit can be represented as: (1); in, This indicates the user equipment speed calculated by the newly added INS unit. This represents the user equipment dimension calculated from the newly added INS unit. This indicates the location of the user equipment calculated by the newly added INS unit. This represents the user equipment attitude calculated by the newly added INS unit. This represents the inertial navigation unit processing function. For the inertial navigation output ratio, For the inertial navigation system output angular velocity, This is the propagation interval.
[0026] A new pseudorange estimate It can be represented as: (2); in, For the first The distance vector from each satellite to the newly added INS unit is used to calculate the location of the user equipment. and These represent the prior estimates of clock error and clock drift by EKF in this embodiment.
[0027] Furthermore, replace the contaminated elements in the EKF observation matrix and state transition matrix, including: Replace the EKF prior estimated location with the location output by the newly added INS unit; Based on the position output of the newly added INS unit, update the coordinate transformation matrix from the inertial coordinate system to the geocentric-ground-fixed coordinate system to obtain the updated coordinate transformation matrix; Reconstruct the EKF observation matrix and state transition matrix based on the updated coordinate transformation matrix.
[0028] Specifically, the solution obtained using the newly added INS unit Replacing the contaminated EKF prior estimate with the correct location yields a result that is uncontaminated within N steps. and : (3); (4); in, express Liyuan Shidi The location of the satellite.
[0029] Similarly, using Replace the contaminated coordinate transformation matrix from the inertial coordinate system to the geocentric coordinate system. The distances from each satellite to its EKF prior estimated position are used to obtain the coordinate transformation matrix. : (5); in, Represents the Earth's angular velocity of rotation. It represents the speed of light.
[0030] Replacing the contaminated elements in the observation matrix and state transition matrix of the EKF yields a new observation matrix. and the new state transition matrix (6); (7); in , , ; Therefore, the mathematical expression for the improved innovation is: (8); In the formula, For the first m The pseudorange from the satellite to the user equipment For the first mThe deceptive component of satellite pseudorange, The first number calculated by the proposed method m The pseudorange from the satellite to the user equipment The actual location of the user equipment to the number m The true distance between the satellites This refers to the distance error between the predicted position and the actual position of the newly added INS element, caused by the cumulative error of the newly added INS element. k For the current epoch, K To deceive the attack moment, For the first m News about satellite-to-user equipment improvements.
[0031] Furthermore, utilizing the current moment before The average new interest rate forecast bias for each epoch includes: Extract the current time before The improved innovation value of each epoch is calculated, and the average of the improved innovation values is used as the innovation bias estimate. The unbiased improved innovation value is obtained by subtracting the innovation deviation estimate from the current improved innovation value.
[0032] Specifically, in practical applications, ionospheric errors, tropospheric errors, and multipath propagation are very common, leading to significant innovation bias. Innovation bias varies over time; if the noise is relatively small, but the innovation bias is relatively large compared to the noise, the deception detection factor may mistakenly perceive a deception attack, thus increasing the false alarm rate. This embodiment improves the innovation bias before the current time step. The mean at time 1 is used as an estimate of the improved information bias at the current time.
[0033] The process of reducing the improved information bias can be expressed as: (9); In the formula, For the first epoch-improved innovation vector For the first The new information vector of the epoch improvement.
[0034] Furthermore, the deception detection factor is: (10); In the formula, The first number calculated by the proposed method The deception detection factor of the epoch. For the first Transpose of the epoch-improved innovation vector It is the inverse matrix of the improved information vector covariance matrix.
[0035] Furthermore, if the newly added INS unit is not corrected by GNSS, it will not be affected by induced deception attacks. However, the error of the newly added INS unit will accumulate continuously. After a period of time, the accumulated error of the newly added INS unit will be large, and the positioning and velocity measurement results of the newly added INS unit will no longer be reliable. The positioning results of the GNSS receiver must be used to correct the newly added INS unit, thereby ensuring that the accumulated error of the newly added INS unit remains at a low level. The newly added INS unit is corrected by the output results of the GNSS receiver every N steps. The correction process can be represented as follows: (11); In the formula, For the first The estimated position of the newly added INS unit at the next epoch. For the first The estimated rate of new INS units added at each epoch. For the first The pose estimated by the newly added INS unit at each epoch. For the first The clock drift estimated by the newly added INS unit under the epoch. For the first Clock bias estimated by the newly added INS unit at the epoch. For the first The position estimated by EKF at each epoch. For the first The velocity estimated by EKF at each epoch. For the first Attitude estimated by EKF at epoch. For the first EKF-estimated clock drift at different epochs For the first Clock bias estimated by EKF at epoch.
[0036] This embodiment adds an INS unit to the traditional tightly coupled framework, ensuring that the improved innovation remains uncontaminated within N steps. This makes the deception detection factor calculated based on the improved innovation more sensitive to low- and medium-intensity induced deception interference, and utilizes the previous time step... The average value of the innovation of each epoch is used to mitigate innovation bias, thereby solving the problems of traditional innovation-based detection methods.
[0037] To more clearly illustrate the technical solution of the present invention, specific embodiments are provided below for description: Step 1: Establish a tightly coupled GNSS / INS model: Tightly coupled systems use the error equations of GNSS and INS as their state equations. The error state vectors of GNSS and INS are used as the state vectors, i.e.: (12); In the formula, express k The system state vector at any given time. For attitude error, For speed error, For positional error, To achieve zero bias in the accelerometer, For zero bias of the gyroscope, For clock difference, This is due to clock drift error.
[0038] The discrete state equations of the system are: (13); In the formula, express k The system state vector at any given time. for k The state transition matrix at time -1 and for k The noise vector of the discrete system at time -1, in this embodiment the noise covariance matrix of the system is... , for k The system state vector at time -1.
[0039] The system's discrete observation equation is: (14); In the formula, It is the observation matrix, which is determined by the characteristics of the tightly coupled system; In order to be in k White noise source at any moment and The standard deviation is the observation noise covariance matrix ; This is the system's discrete observation equation.
[0040] The Kalman filtering process of a tightly coupled GNSS / INS system can be represented by time updates and observation updates as follows: The time has been updated to: (15); (16); Observations updated as follows: (17); (18); (19); Among them, symbols Indicates the estimated value, symbol This represents the value after the time update. This indicates the updated value after measurement. Let be the error covariance matrix. Here is the Kalman gain matrix. for Prior state estimation at time 10:00 for State estimation at time 10:00 for State estimation at time 10:00 This is the time-updated error covariance matrix. for The updated error covariance matrix is observed at each time step. for The system noise variance matrix at time 1. for The transpose of the state transition matrix at each time step. for The transpose of the observation matrix at any given time. for The updated error covariance matrix is observed at each time step. for The inverse of the error covariance matrix after time-time update.
[0041] k -1 hour arrives k State transition matrix at time step and observation matrix It can be defined as: (20); (twenty one); in, , , as well as ; In the formula, It is represented as the unit vector from the Mth satellite to the user's receiver antenna. Let be the skew-symmetric matrix of the Earth's velocity. To estimate the attitude prior, For comparison, The equatorial radius is in the WGS84 coordinate system. To estimate latitude a priori, It is a three-dimensional identity matrix. For the propagation interval, For position Acceleration due to gravity, For prior estimates of user equipment, Transpose of the prior estimate for user equipment. for Distance to the center of the Earth Let be the vector from the Mth satellite to the user's receiver antenna. Let M be the distance from the Mth satellite to the user's receiver antenna.
[0042] Step 2: Construct a deception detection method model based on new information: The definition of new information is as follows: (twenty two); It follows a Gaussian distribution with zero mean and known covariance, that is: (twenty three); (twenty four); In the formula, It is the covariance matrix. The mean of the new information. Let be the covariance of the new information.
[0043] In a classic tightly coupled system, the physical meaning of the innovation in EKF is the difference between the pseudorange calculated by the GNSS receiver and the predicted pseudorange, which can be expressed as: (25); In the formula, M This means that the number of visible stars is also equal to the dimension of the news. Represented as the first m The pseudorange from the satellite to the user receiver, It is the pseudo-range of the prior estimate calculated in EKF. For the first Era 1 A new piece of information.
[0044] Deception detection factor Defined as the sum of squares of the normalized innovation: (26); In a scenario without deception, It follows a chi-square distribution with M degrees of freedom: (27); Preset false alarm rate The threshold for deception detection can be calculated as follows: (28); in, It is the chi-square inverse cumulative distribution function, which can be defined as: (29); if Compare If the value is too large, it will trigger a deception alarm.
[0045] In the formula, To deceive the detection threshold, The default false alarm probability is set.
[0046] Step 3: Replace the contaminated EKF prior estimates and the contaminated elements in the EKF matrix with uncontaminated INS output values from within N steps: A schematic diagram of the improved information is shown below. Figure 9 As shown, after the induced deception occurs, the predicted pseudorange is not deceived within N steps, and the improved information is not contaminated. Since the newly added INS unit is not deceived within N steps, it can be guaranteed that the predicted pseudorange is not deceived within N steps. Therefore, the information that is not contaminated within N steps reflects the true deception strength better than the contaminated information.
[0047] The newly added INS unit can be represented as: (30); in, This indicates the user equipment speed calculated by the newly added INS unit. This represents the user equipment dimension calculated from the newly added INS unit. This indicates the location of the user equipment calculated by the newly added INS unit. This represents the user equipment attitude calculated by the newly added INS unit. This represents the inertial navigation unit processing function. For the inertial navigation output ratio, For the inertial navigation system output angular velocity, This is the propagation interval.
[0048] A new pseudorange estimate It can be represented as: (31); in, For the first The distance vector from each satellite to the newly added INS unit is used to calculate the location of the user equipment. and These represent the prior estimates of clock error and clock drift by EKF in this embodiment, respectively. The first number calculated by the proposed method The pseudorange from a satellite to user equipment.
[0049] The solution calculated using the newly added INS unit Replacing the contaminated EKF prior estimate with the correct location yields a sequence of locations that are uncontaminated within N steps. and : (32); (33); in, Represented as Liyuan Shidi The location of the satellite, Here is the coordinate transformation matrix from the ECEF to the ECI coordinate system using the proposed method. The location of the user equipment is calculated using the proposed method. The first number calculated by the proposed method Vector from satellite to user equipment The first number calculated by the proposed method The distance between a satellite and the user's equipment.
[0050] Similarly, using Replace the contaminated coordinate transformation matrix from the inertial coordinate system to the geocentric coordinate system. The distances from each satellite to its EKF prior estimated position are used to obtain the coordinate transformation matrix. : (34); in, Represents the Earth's angular velocity of rotation. It represents the speed of light.
[0051] Replacing the contaminated elements in the observation matrix and state transition matrix of the EKF yields a new observation matrix. and the new state transition matrix : (35); (36); in , , After replacing the deceived elements in the matrix, equations (16)-(18) can be expressed as: (37); (38); (39); Therefore, the mathematical expression for the improved innovation is: (40); In the formula, For the first m The pseudorange from the satellite to the user equipment For the first m The deceptive component of satellite pseudorange, The first number calculated by the proposed method m The pseudorange from the satellite to the user equipment The actual location of the user equipment to the number m The true distance between the satellites This refers to the distance error between the predicted position and the actual position of the newly added INS element, caused by the cumulative error of the newly added INS element. k For the current epoch, K To deceive the attack moment, For the first m News about satellite-to-user equipment improvements.
[0052] Step 4: Utilize the time before the current moment Using the average of the new interest over epochs to predict the new interest bias: In practical applications, ionospheric errors, tropospheric errors, and multipath propagation are very common, leading to significant innovation bias. Innovation bias varies over time; if the noise is relatively small, but the innovation bias is relatively large compared to the noise, the deception detection factor may mistakenly perceive a deception attack, thus increasing the false alarm rate. This invention improves the innovation bias before the current time step. The mean at time 1 is used as an estimate of the improved information bias at the current time.
[0053] The process of reducing the improved information bias can be expressed as: (41); In the formula, For the first epoch-improved innovation vector For the first The new information vector of the epoch improvement.
[0054] Step 5: Construct new deception detection factors: (42); In the formula, The first number calculated by the proposed method The deception detection factor of the epoch. For the first Transpose of the epoch-improved innovation vector It is the inverse matrix of the improved information vector covariance matrix.
[0055] Step Six: Use the positioning results from the GNSS receiver to correct the newly added INS units: If the newly added INS unit is not corrected by GNSS, it will not be affected by induced deception attacks. However, the error of the newly added INS unit will continue to accumulate. After a period of time, the accumulated error of the newly added INS unit will be large, and the positioning and velocity measurement results of the newly added INS unit will no longer be reliable. The positioning results of the GNSS receiver must be used to correct the newly added INS unit, so as to ensure that the accumulated error of the newly added INS unit remains at a low level.
[0056] The newly added INS unit is corrected every N steps by the output of the GNSS receiver. The correction process can be represented as follows: (43); In the formula, For the first The estimated location of the newly added INS unit at the next epoch. For the first The estimated rate of new INS units added at each epoch. For the first The pose estimated by the newly added INS unit at each epoch. For the first The clock drift estimated by the newly added INS unit under the epoch. For the first Clock bias estimated by the newly added INS unit at the epoch. For the first The position estimated by EKF at each epoch. For the first The velocity estimated by EKF at each epoch. For the first Attitude estimated by EKF at epoch. For the first EKF-estimated clock drift at different epochs For the first Clock bias estimated by EKF at epoch.
[0057] Compared to traditional innovation-based spoofing detection methods (where the INS unit is corrected by EKF at every epoch), this embodiment corrects the INS unit only every N epochs. This ensures that the predicted pseudorange is not spoofed within N steps, thus preventing the improved innovation from being contaminated and better reflecting the true spoofing strength. This improves detection performance and shortens alarm time. Furthermore, it utilizes the previous time step... The average value of the new information over each epoch is used to mitigate the new information bias, thereby reducing the false alarm rate.
[0058] Figure 1The diagram shows the position calculated by the GNSS receiver before and after a deception attack, the actual user equipment, the tightly coupled calculation, and the trajectory calculated by the method proposed in this embodiment. Before the induced deception occurs, the user equipment moves at a constant speed in a straight line eastward. The deception attack occurs at 60 seconds. After the induced deception attack, the position calculated by the GNSS receiver turns at a constant angular rate for 4 seconds, then continues to move at a constant speed in a straight line. The GNSS receiver is deceived into another direction, causing the position calculated by the tightly coupled calculation to be gradually deviated and gradually approach the deception trajectory. Compared to the traditional position calculated based on tightly coupled GNSS / INS calculation, the method proposed in this invention still travels in the correct direction for a short period after being subjected to an induced deception attack. During this short period, the improved information can reflect the true deception intensity and can detect the induced deception attack more quickly and sensitively. The technical solution section mentions that since the error of the newly added INS unit will continue to accumulate, it needs to be corrected to ensure that its accumulated error remains at a low level. The time indicated by the blue circle in the diagram is the time when the EKF corrects the newly added inertial navigation unit.
[0059] Figure 2 The graph shows the changes in the improved innovations of six satellites before and after a low-intensity induced deception attack. As can be seen from the graph, before the induced deception, the mean innovation value was 0. After the induced deception, the improved innovations fluctuated significantly. Furthermore, the graph also shows that the fluctuations in the improved innovations of different satellites were inconsistent. The red dots in the graph represent the peak innovation value after the deception, and the blue dots represent the maximum innovation value in the no-deception scenario. The ratio of the red dot value to the blue dot value represents the magnitude of the deception attack, which can be expressed as: (44); Taking PRN5 as an example, the traditional method The value is 2.9614, and the method proposed in this invention... The result is 6.3117, indicating that the method proposed in this invention is more sensitive to low-intensity induced deception interference and the deception detection performance is improved.
[0060] Figure 3 (a)- Figure 3 (c) The changes in the deception detection factor of the traditional method and the method proposed in this invention were compared under three induced deception intensities. As can be seen from the figure, the peak value of the deception detection factor gradually increases with the increase of deception intensity. Compared with the traditional method, which detects high-intensity deception interference, the method proposed in this invention successfully detects deception interference of all three intensities. Simulation results show that, compared with the traditional method, the method proposed in this invention can effectively detect low-intensity and medium-intensity induced deception interference.
[0061] Figure 4The changes in alarm time were compared under different deception angular rates at the same speed. The angular rate ranged from 1° / s to 40° / s in 1° / s increments. When the pseudorange measurement noise was 0.05 m, the minimum detectable angular rate of the conventional method was 11° / s, and the alarm time was 4.286 s. The method proposed in this invention reduces the minimum detectable angular rate to 6° / s. When the pseudorange measurement noise was 0.15 m, the minimum detectable angular rate was 32° / s, and the alarm time was 4.286 s. The method proposed in this invention reduces the minimum detectable angular rate to 13° / s. When the pseudorange noise increased from 0.05 m to 0.15 m, the minimum detectable deception angular rate of the conventional method increased by 21° / s, while the method proposed in this invention only increased by 7° / s, indicating that the method proposed in this invention has better robustness. Furthermore, when the angular rate is 15° / s and the pseudorange measurement noise is 0.05m, the alarm time of the conventional method is 3.083s, while the alarm time of the method proposed in this invention is 2.08s, a reduction of 1.003 seconds. The results show that the method proposed in this invention is not only more sensitive to weaker spoofing attacks, but also detects spoofing interference more quickly.
[0062] Figure 5 The changes in alarm time under different speeds with the same deception angular rate were compared. Speeds ranged from 1 m / s to 16 m / s in increments of 0.4 m / s. When the pseudorange measurement noise was 0.05 m, the minimum detectable speed of the conventional method was 4.4 m / s, and the alarm time was 3.885 s. The proposed method reduces the minimum detectable speed to 2 m / s. When the pseudorange measurement noise was 0.15 m, the minimum detectable speed was 12.4 m / s, and the alarm time was 4.688 s. The proposed method reduces the minimum detectable speed to 5.4 m / s. When the pseudorange noise increased from 0.05 m to 0.15 m, the minimum detectable speed of the conventional method increased by 8 m / s, while the proposed method only increased by 3.4 m / s, indicating that the proposed method has better robustness. Furthermore, when the velocity is 6 m / s and the pseudorange measurement noise is 0.05 m, the alarm time of the traditional method is 3.083 s, while the alarm time of the method proposed in this invention is 2.08 s, a reduction of 1.003 seconds. The results show that the method proposed in this invention is not only more sensitive to weaker spoofing attacks, but also detects spoofing interference faster. Meanwhile, Figure 6 and Figure 7 The conclusions indicate that the greater the velocity or deception angular rate, the faster the deception interference can be detected.
[0063] In the practical analysis, this embodiment designed two different scenarios: a no-spoofing scenario and a spoofing scenario. The no-spoofing scenario was used to evaluate the false alarm rate in real-world scenarios, while the spoofing scenario was used to evaluate the spoofing detection performance in real-world scenarios. Figure 6In (a), both the user equipment and the antenna are stationary, which is a non-spoofing scenario. Figure 6 (b) For the deception scenario, before the induced deception occurs, the user equipment and the omnidirectional antenna are fixed at point C. In order to simulate the induced deception attack, after the induced deception attack occurs, the user equipment is kept stationary, and the antenna is moved eastward at a fixed speed of 0.5m / s. The distances of 1m, 2m and 5m represent different intensities of induced deception: low intensity, medium intensity and high intensity, respectively.
[0064] Figure 7 This figure shows the changes in the improved innovation observed by six satellites before and after a low-intensity induced spoofing attack. Consistent with traditional innovations, the improved innovation is also susceptible to ionospheric delay, tropospheric delay, and receiver hardware delay, resulting in innovation bias. When calculating the improved spoofing detection factor, this bias must be eliminated first. The improved innovation shown in the figure represents the innovation after the bias has been eliminated. The spoofing attack occurred at 60 s; before the spoofing, the mean of the improved innovation was approximately 0. The definitions of the red and blue dots are as follows: Figure 4 Consistent, taking PRN 5 as an example, the improved information =7.1697, compared to traditional new information The improvement increased by 520.1%, indicating that the improved information is more sensitive to deceptive interference.
[0065] Figure 8 (a)- Figure 8 (c) The changes in the deception detection factor of the traditional method and the proposed method were compared under three induced deception intensities in the actual test scenario. As the induced deception intensity increases, the deception detection factor also gradually increases. Consistent with the simulation results, compared to the traditional method which only detects high-intensity induced deception interference, the proposed method successfully detects all three intensities of induced deception interference. Simulation results show that, compared to the traditional method, the proposed method can effectively detect lower-intensity induced deception interference.
[0066] Peak-to-peak ratio of new information Similarly, the peak-to-peak ratio is defined as the deception detection factor. Table 1 summarizes the different N Value And alarm time. The table shows the optimal values for low-intensity and medium-intensity deception. N The value is approximately 20, which is optimal under high intensity. N The value is 25. When N is 20, the values for low and medium strength are... The largest are 45.3958 and 80.7014 respectively. Furthermore, as can be seen from Table 1, when... NAt a value of 20, the alarm time is the shortest under the three induced deception scenarios, at 2.602s and 1.992s respectively. The alarm time remains the same even under high-intensity induced deception attacks because 1.992s is already the minimum time limit for the improved method to detect deception. Increasing the GNSS sampling frequency will further improve the deception detection performance. Therefore, N A value of 20 represents the optimal value for achieving the best performance in this experiment. In practical applications, it is important to select the appropriate value based on the specific GNSS and INS configuration. N value.
[0067] Table 1 The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A GNSS spoofing interference detection method based on multi-step inertial navigation prediction, characterized in that, include: A tightly coupled GNSS / INS model is constructed, and time and observation updates are performed using an extended Kalman filter (EKF). A new independent inertial navigation INS unit is added to the tightly coupled GNSS / INS model to construct a deception detection method model based on innovation and generate uncontaminated pseudorange estimates within N steps. Replace the contaminated elements in the EKF observation matrix and state transition matrix to obtain the reconstructed observation matrix and state transition matrix; Based on the reconstructed observation matrix and state transition matrix, the improved innovation value is calculated, wherein the improved innovation value is the difference between the pseudorange measurement value of the GNSS receiver and the pseudorange estimate value; Utilizing the current time before The average value of the innovation in each epoch predicts the innovation deviation, and an unbiased improved innovation value is obtained. A deception detection factor is constructed based on the unbiased improved innovation value, and a detection statistic is generated. The detection statistic is compared with a preset threshold to determine deception interference.
2. The GNSS spoofing interference detection method based on multi-step inertial navigation prediction according to claim 1, characterized in that, Generating the uncontaminated pseudorange estimate within the N steps includes: Based on the user equipment location, velocity, latitude, and attitude output by the newly added INS unit, calculate the... m Distance vector from each satellite to the user equipment; By combining the prior estimates of clock bias and clock drift from EKF, a new pseudorange estimate is generated, namely the uncontaminated pseudorange estimate.
3. The GNSS spoofing interference detection method based on multi-step inertial navigation prediction according to claim 2, characterized in that, The newly added INS unit is specifically: ； In the formula, This indicates the user equipment speed calculated by the newly added INS unit. This represents the user equipment dimension calculated from the newly added INS unit. This indicates the location of the user equipment calculated by the newly added INS unit. This represents the user equipment attitude calculated by the newly added INS unit. This represents the inertial navigation unit processing function. The output ratio of the inertial navigation unit, Output angular velocity to the inertial navigation unit. This is the propagation interval.
4. The GNSS spoofing interference detection method based on multi-step inertial navigation prediction according to claim 2, characterized in that, The new pseudorange estimate is: ； In the formula, For the first The distance vector from each satellite to the newly added INS unit is used to calculate the location of the user equipment. and Let represent the EKF prior estimates of clock bias and clock drift, respectively. This is the new pseudorange estimate.
5. The GNSS spoofing interference detection method based on multi-step inertial navigation prediction according to claim 1, characterized in that, Replace the contaminated elements in the EKF observation matrix and state transition matrix, including: Replace the EKF prior estimated location with the location output by the newly added INS unit; Based on the position output of the newly added INS unit, update the coordinate transformation matrix from the inertial coordinate system to the geocentric-ground-fixed coordinate system to obtain the updated coordinate transformation matrix; The EKF observation matrix and state transition matrix are reconstructed based on the updated coordinate transformation matrix.
6. The GNSS spoofing interference detection method based on multi-step inertial navigation prediction according to claim 1, characterized in that, The innovation value of the improvement is calculated as follows: ； In the formula, For the first m The pseudorange from the satellite to the user equipment For the first m The deceptive component of satellite pseudorange, The first number calculated by the proposed method m The pseudorange from the satellite to the user equipment The actual location of the user equipment to the number m The true distance between the satellites This refers to the distance error between the predicted position and the actual position of the newly added INS element, caused by the cumulative error of the newly added INS element. k For the current epoch, K To deceive the attack moment, For the first m News about satellite-to-user equipment improvements.
7. The GNSS spoofing interference detection method based on multi-step inertial navigation prediction according to claim 1, characterized in that, Utilizing the current time before The average new interest rate forecast bias for each epoch includes: Extract the current time before The improved innovation value of each epoch is calculated, and the average of the improved innovation values is used as the innovation bias estimate. The unbiased improved innovation value is obtained by subtracting the innovation deviation estimate from the current improved innovation value.
8. The GNSS spoofing interference detection method based on multi-step inertial navigation prediction according to claim 7, characterized in that, The unbiased improvement innovation value is: ； In the formula, For the first epoch-improved innovation vector For the first The new information vector of the epoch improvement.
9. The GNSS spoofing interference detection method based on multi-step inertial navigation prediction according to claim 1, characterized in that, The deception detection factor is: ； In the formula, The first number calculated by the proposed method The deception detection factor of the epoch. For the first Transpose of the epoch-improved innovation vector It is the inverse matrix of the improved information vector covariance matrix.
10. The GNSS spoofing interference detection method based on multi-step inertial navigation prediction according to claim 1, characterized in that, The newly added independent inertial navigation (INS) unit is corrected every N steps by the output of the GNSS receiver. The correction process is as follows: ； In the formula, For the first The estimated location of the newly added INS unit at the next epoch. For the first The estimated rate of new INS units added at each epoch. For the first The pose estimated by the newly added INS unit at each epoch. For the first The clock drift estimated by the newly added INS unit under the epoch. For the first Clock bias estimated by the newly added INS unit at the epoch. For the first The position estimated by EKF at each epoch. For the first The velocity estimated by EKF at each epoch. For the first Attitude estimated by EKF at epoch. For the first EKF-estimated clock drift at different epochs For the first Clock bias estimated by EKF at epoch.