A Radar Angle Measurement Phase Ambiguity Resolution Method and Device Based on Steering Vector Correlation

Abstract

This invention belongs to the field of radar measurement technology, specifically relating to a method and apparatus for de-phase ambiguity resolution in radar angle measurement based on steering vector correlation. The specific process is as follows: An adaptive neighborhood radius and a minimum number of points are set; the initial values ​​of angle and integer ambiguity are estimated using the density-based clustering algorithm DBSCAN; based on the estimated initial values ​​of angle and integer ambiguity, a finite memory algorithm is used to perform recursive outlier detection and correction; during the recursion process, the allowable error probability is dynamically adjusted based on the standard deviation of the angle measurement noise at the current moment and used as an adaptive reset threshold; when the recursive error probability within the window exceeds the adaptive reset threshold, the recursion is terminated and the SVC secondary correction and reset process is entered; after secondary correction of the current angle value using steering vector correlation SVC, a forced reset is performed.

Description

Technical Field

[0001] This invention belongs to the field of radar measurement technology, specifically relating to a radar angle measurement and phase ambiguity resolution method and apparatus based on steering vector correlation. Background Technology

[0002] Radar phase interferometry receives target echoes using spatially distributed antenna elements. It calculates the target's incident angle by utilizing the phase difference caused by the path difference between the elements. This technology offers advantages such as high accuracy and real-time performance, and is widely used in aerospace telemetry and control, space rendezvous and docking, and electronic reconnaissance. Its measurement principle is as follows: when electromagnetic waves arrive at each element under far-field conditions, a phase difference is formed due to the different propagation paths. This phase difference has a definite functional relationship with the element spacing and the incident angle. By obtaining the phase difference measurement value between each element through phase detection, the target's direction of arrival can be deduced.

[0003] The accuracy of an interferometer angle measurement system is directly related to the spacing between antenna elements. Increasing the baseline length can improve the sensitivity of the phase difference to the angle, thereby obtaining higher angle measurement accuracy. However, when the element spacing exceeds half the wavelength of the incident signal, the phase detector can only output the main value range. The phase measurement value within the range cannot determine the corresponding value. The integer ambiguity number. If this integer ambiguity number cannot be correctly solved, the calculated angle will deviate significantly from the true direction; this is the integer ambiguity problem. Therefore, how to balance the unambiguous measurement range with the requirement for high-precision angle measurement is the core engineering challenge facing interferometric angle measurement technology.

[0004] To address the aforementioned issues, multi-baseline phase ratio deambiguity methods have been developed in engineering. Among these, the long-short baseline method provides initial ambiguity values ​​for the long baseline by configuring a physical short baseline that satisfies the half-wavelength unambiguity condition. However, due to space constraints in spacecraft platform antenna layout, the deployment of physical short baselines is often difficult to implement. The virtual baseline method constructs an equivalent short baseline using the phase difference between two real baselines, circumventing the engineering implementation difficulties of physical short baselines. Therefore, it has been applied in scenarios such as rendezvous and docking microwave radar. In his journal article "Research on a New Deambiguity Method," Si Weijian of Harbin Engineering University proposed a phase interferometer direction finding method based on virtual baselines for sequential deambiguity resolution, addressing the problem that the shortest baseline cannot satisfy the half-wavelength constraint in broadband direction finding scenarios for passive radar seekers. This method solves the integer ambiguity number of each real baseline step-by-step, starting from the virtual baseline, achieving high-precision angle measurement under a finite element layout.

[0005] In real-world operating environments, the phase difference measurement output by the phase detector contains significant random errors due to low radar signal power and receiver thermal noise. During the step-by-step recursive process, this phase error amplifies with each baseline length. When the error exceeds the noise tolerance of the rounding operation, the integer ambiguity decision at a certain level may be incorrect. Since the erroneous ambiguity at this level will be used as a known quantity in the subsequent recursive calculation of longer baselines, the error will propagate and accumulate along the solution chain, ultimately manifesting as isolated jumps or continuous patches of anomalous observations in the angle measurement sequence. Summary of the Invention

[0006] To address the aforementioned issues, this invention proposes a radar angle measurement phase ambiguity resolution method and apparatus based on Steering Vector Correlation (SVC). Building upon time-domain finite memory recursive detection, it introduces spatial correlation discrimination of steering vector to perform secondary verification of the correctness of the ambiguity resolution results, effectively improving the error correction reliability of successive recursive ambiguity resolution under strong noise conditions.

[0007] The technical solution for implementing the present invention is as follows:

[0008] In a first aspect, the present invention provides a radar angle measurement and phase ambiguity resolution method based on steering vector correlation, the specific process of which is as follows: Initial angle and integer ambiguity number estimation: setting an adaptive neighborhood radius and minimum points The initial values ​​of angle and integer ambiguity number are estimated using the density-based clustering algorithm DBSCAN; where the adaptive neighborhood radius... The calculation process is as follows: based on the current carrier-to-noise ratio estimated by the receiver. Carrier tracking loop bandwidth and coherent integration time Estimate the standard deviation of phase measurement noise The standard deviation of phase measurement noise Mapped to the standard deviation of angle measurement noise The standard deviation of noise is measured based on the angle. Set adaptive neighborhood radius ; Outlier detection and correction: Based on the estimated angle and the initial value of integer ambiguity, the finite memory algorithm is used to enter the recursive outlier detection and correction process. During the recursion process, the allowable error probability is dynamically adjusted according to the standard deviation of the angle measurement noise at the current moment and used as the adaptive reset threshold. When the recursive error probability within the window exceeds the adaptive reset threshold, the recursion is terminated and the process enters the secondary correction and reset process of the guide vector related SVC. SVC Secondary Correction and Reset: After using SVC secondary correction to correct the current angle value, a forced reset is performed. The reset process is as follows: clear the status flag array, discard all intermediate variables accumulated in the current recursive process, and jump back to the initial angle and integer fuzzy number estimation.

[0009] Optionally, the present invention adapts the neighborhood radius. The calculation process is as follows: First, based on the current carrier-to-noise ratio estimated by the receiver. Carrier tracking loop bandwidth and coherent integration time Estimate the standard deviation of phase measurement noise ;

[0010] Secondly, the standard deviation of the phase measurement noise Mapped to the standard deviation of angle measurement noise :

[0011] in, For the signal wavelength, The longest baseline length, This is a rough estimate of the angle at the current moment; Finally, the standard deviation of the noise was measured based on the angle. Set adaptive neighborhood radius for:

[0012] in, is the clustering confidence coefficient.

[0013] Optionally, the clustering confidence coefficient of the present invention The value range is [2,3], minimum number of points .

[0014] Optionally, after the clustering is completed, the cluster containing the most points is selected as the normal value cluster. Let the set of points contained in the normal value cluster be... The initial angle is taken as the mean within the cluster; the initial integer fuzzy number Using point sets The majority voting method was determined.

[0015] Optionally, in the recursive process, the present invention uses the successively recursively defuzzified output value at the current moment. Recent in the window The average of the corrected angle values The comparison is made, and the deviation is defined as... ; Define a fixed-length array of status flags to record the most recent The result of each detection in the recursion; the flag is defined as:

[0016] in, This indicates that a threshold value is set; after each iteration, [the threshold value will be set]. Storage length is In the sliding buffer; If the buffer is not full, count the number of errors that have occurred with the stored data. When the buffer is full, count the most recent... Number of errors in the next iteration ;

[0017] And further calculate the recursive error probability within the window. ;

[0018] like At this point, it is determined that the true angle can no longer be effectively tracked, so the recursion is terminated and the process of secondary correction and reset of the guide vector related SVC is initiated. This indicates an adaptive reset threshold.

[0019] Optionally, the present invention adaptively resets the threshold. for:

[0020] in, This represents the upper limit of the error probability under high carrier-to-noise ratio. This represents the upper limit of the error probability under low carrier-to-noise ratio. The standard deviation of the noise is measured for the angle at the current moment. The standard deviation of the angular measurement noise of the system at the nominal carrier-to-noise ratio.

[0021] Optionally, during the SVC secondary correction of the present invention, a cosine matching correlation criterion is selected to calculate the candidate angles at the current time. Angle correction from the previous moment Correlation coefficient and , For the most recent window The average of the corrected angle values The angle corresponding to the one with the larger correlation coefficient is selected as the final angle.

[0022] Optionally, the correlation coefficient expression of the present invention is as follows:

[0023] in, For the current moment, the first Phase difference measurements corresponding to the solid baselines are performed. Point complex vector average, Represents angle value The corresponding phase difference, .

[0024] Optionally, the present invention also resets the status flag array and inverts the data validity flag to indicate that the current data is the result after secondary correction.

[0025] In a second aspect, the present invention provides a radar angle measurement and phase ambiguity resolution device based on steering vector correlation, comprising: Initial angle and integer ambiguity estimation module: setting adaptive neighborhood radius and minimum points The initial values ​​of angle and integer ambiguity number are estimated using the density-based clustering algorithm DBSCAN; where the adaptive neighborhood radius... The calculation process is as follows: based on the current carrier-to-noise ratio estimated by the receiver. Carrier tracking loop bandwidth and coherent integration time Estimate the standard deviation of phase measurement noise The standard deviation of phase measurement noise Mapped to the standard deviation of angle measurement noise The standard deviation of noise is measured based on the angle. Set adaptive neighborhood radius ; Outlier detection and correction module: Based on the estimated angle and the initial value of integer ambiguity, the module uses a finite memory algorithm to enter the recursive outlier detection and correction process. During the recursion process, the module dynamically adjusts the allowable error probability based on the standard deviation of the angle measurement noise at the current moment and uses it as an adaptive reset threshold. When the recursive error probability within the window exceeds the adaptive reset threshold, the recursion is terminated and the module enters the secondary correction and reset process of the guide vector related SVC. SVC Secondary Correction and Reset Module: After using SVC secondary correction to correct the current angle value, a forced reset is performed. The reset process is as follows: clear the status flag array, discard all intermediate variables accumulated in the current recursive process, and jump back to the initial angle and integer fuzzy number estimation.

[0026] Beneficial effects: First, the method of the present invention can accurately identify and correct outliers in the phase deblurring process. Compared with existing outlier removal methods such as finite memory method, wavelet transform method and least squares B spline method, it significantly improves the deblurring accuracy under low signal-to-noise ratio and does not discard valid measurement data.

[0027] Second, compared with clustering methods using fixed initialization parameters, this invention adaptively determines the neighborhood radius of DBSCAN clustering based on the receiver's current carrier-to-noise ratio, carrier tracking loop bandwidth, and coherent integration time, allowing the initialization process to adjust according to changes in the actual noise environment. Under low carrier-to-noise ratio conditions, appropriately increasing the neighborhood radius can prevent normal measurements from being misjudged as outliers due to noise diffusion; under high carrier-to-noise ratio conditions, appropriately decreasing the neighborhood radius can improve the ability to distinguish initial outliers.

[0028] Third, compared with the method of triggering secondary error correction using a fixed anomaly ratio threshold, this invention sets an adaptive reset threshold based on the current receiver state, enabling steering vector correlation (SVC) verification to be triggered as needed. When phase noise is strong and the risk of deambiguity errors is high, the trigger threshold is lowered, allowing spatial domain verification to intervene earlier. When measurements are relatively stable, the trigger threshold is raised, reducing unnecessary correlation calculations and erroneous corrections. Therefore, this invention balances deambiguity reliability, computational complexity, and real-time engineering performance. Attached Figure Description

[0029] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 This is a schematic diagram of the principle of one-dimensional multi-baseline non-uniform linear array angle measurement.

[0031] Figure 2 This is a flowchart of the SVC error correction algorithm.

[0032] Figure 3 This is a comparison chart showing the effects of SVC and other algorithms on outlier removal under static conditions.

[0033] Figure 4 This is a comparison chart showing the outlier removal effects of SVC and other algorithms under uniform speed conditions.

[0034] Figure 5 This is a comparison chart of the outlier removal effects of SVC and other algorithms under sinusoidal condition 1.

[0035] Figure 6 This is a comparison chart showing the outlier removal effects of SVC and other algorithms under sinusoidal condition 2.

[0036] Figure 7 This is a comparison chart showing the effects of static motion SVC and other algorithms before and after removing outliers at different carrier-to-noise ratios. Detailed Implementation

[0037] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0038] It should be noted that, in the absence of conflict, the following embodiments and features can be combined with each other; and, based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0039] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this disclosure, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0040] This invention proposes a radar angle measurement phase ambiguity resolution method based on steering vector correlation, which is used to identify and correct ambiguity errors in the successive recursion process, thereby assisting in phase ambiguity resolution based on virtual baseline successive recursion.

[0041] like Figure 1 As shown, the present invention is applicable to sparse, non-uniform, multi-baseline array configurations, which consist of... Each actual array element (hereinafter referred to as a real array element) The structure is such that the interval between any two real array elements is the real baseline, and the length of the real baseline between adjacent real array elements is... , This represents the number of real baselines. By performing differential operations on the phase differences of different real baselines, a [formula / formula] can be constructed. Equivalent virtual baselines, with lengths of respectively ,in With antenna array elements The reference element is located at distances of 1 / 2 to 2 / 3. The position can be regarded as a virtual array element To meet the algorithm requirements of successive recursive defuzzification and ensure the system's angle measurement accuracy, the lengths of all baselines must be arranged in ascending order, i.e., satisfying... .

[0042] Consider two lines of length... and The length of the virtual baseline constructed from the real baseline is , ,in and According to the fundamental interferometric angle measurement equation, the phase difference between two real baselines can be expressed as: (1) in, and For length and The unambiguous phase difference corresponding to the real baseline For length and The unambiguous phase difference value corresponding to the real baseline. For length and The length of the virtual baseline constructed from the real baseline, For length and The difference between the integer ambiguity numbers corresponding to the real baseline. and For length and The integer ambiguity number corresponding to the real baseline. and For length and The real baseline corresponds to an ambiguous phase difference. For length and The real baseline corresponds to the fuzzy phase difference value.

[0043] When the virtual baseline length satisfies the unambiguous condition At that time, the difference in the number of fuzzy numbers throughout the whole week At this point, equation (1) degenerates into (2) That is, the unambiguous phase difference corresponding to the virtual baseline It can be directly derived from the real baseline with fuzzy phase difference mode Obtain. Obtain the unambiguous phase difference corresponding to the virtual baseline. Following the recursive approach of the long and short baseline method, we assume the baseline length of the next level of real baseline is... The ratio of the actual baseline to the virtual baseline length is: By combining the phase difference formulas for the real baseline and the virtual baseline, we can obtain the first... Integer blur value of the real baseline

[0044] (3) in, The virtual baseline corresponds to an ambiguous phase difference. The real baseline corresponds to an ambiguous phase difference. This is the rounding function.

[0045] Then, the integer blur value, unblurred phase difference, and corresponding incident angle of the longer baseline can be obtained by recursion step by step.

[0046] (4) in, , and They are respectively Figure 1 In the antenna array shown, the baseline length, integer ambiguity value, and ambiguous phase difference corresponding to the longest real baseline are given.

[0047] In engineering practice, phase difference measurements are inevitably affected by phase noise. Considering the influence of phase noise, the actual output of the phase detector will have an ambiguous phase. , It can be represented as (5) in, , The corresponding phase noise error is due to the virtual baseline integer ambiguity value. Then the integer ambiguity value of the next level real baseline Represented as (6) As can be seen from equation (6), as long as the magnitude of the amplified equivalent phase error does not exceed This will prevent de-ambiguity errors. To correctly resolve the next level of baseline ambiguity values... The noise term must not affect the result of the rounding operation, i.e. (7) Consider the most extreme case, assuming , Take the maximum value respectively and minimum value , Since is a positive real number, the phase noise has the most severe impact on defuzzification. Substituting into equation (7), we can obtain the constraint condition for correct defuzzification on phase noise error as follows: (8) Based on the above analysis, it can be seen that as the baseline length ratio increases, the phase noise error that the system can tolerate decreases sharply. This means that in a strong noise environment, the probability of a real baseline exhibiting a de-ambiguity error is significantly higher than that of a virtual baseline. Once a certain level of de-ambiguity fails, the ambiguity number output by that level will deviate from the true value and be substituted as a known quantity into the recursive solution of subsequent longer baselines, causing the error to be propagated and amplified step by step, ultimately manifesting as continuous or isolated outliers in the angle measurement results. In order to reduce the proportion of outliers in the angle de-ambiguity results and improve the accuracy of the incident angle measurement, it is necessary to perform outlier detection and correction on the angle value output by equation (4). Therefore, this application provides a radar angle measurement de-phase ambiguity method based on steering vector correlation, such as Figure 2 As shown, the specific implementation steps are as follows: Step 1: Initial Angle and Integer Blur Number Estimation: This process is based on the density-based clustering algorithm DBSCAN (Density-Based Spatial Clustering of Applications with Noise).

[0048] Assume the initial window contains Angle measurement values The DBSCAN algorithm introduces two key parameters: adaptive neighborhood radius. and minimum points .

[0049] In noisy environments, the statistical characteristics of angle measurements fluctuate drastically with changes in the carrier-to-noise ratio (CNR) of the received signal. Using a fixed neighborhood radius may result in normal values ​​failing to cluster at low CNRs, while noise may be misclassified as valid points at high CNRs. Therefore, this invention proposes a carrier-to-noise ratio adaptive clustering method based on one-dimensional DBSCAN, where the neighborhood radius is adjusted according to the standard deviation of the angle measurement noise. Adaptive settings.

[0050] First, based on the current carrier-to-noise ratio estimated by the receiver. Carrier tracking loop bandwidth and coherent integration time Estimate the standard deviation of phase measurement noise .

[0051] (9) Secondly, the standard deviation of phase measurement noise is calculated using the interferometer angle measurement formula. Mapped to the standard deviation of angle measurement noise : (10) in, For the signal wavelength, The longest baseline length, This is a rough estimate of the angle at the current moment.

[0052] Finally, based on this, the standard deviation of the noise was measured at the angle. Set adaptive neighborhood radius for: (11) In the formula, The clustering confidence coefficient is typically set to 2-3. This adaptive mechanism ensures that the neighborhood radius automatically expands at low load-to-noise ratios to prevent normal values ​​from being misclassified as noise; and automatically shrinks at high load-to-noise ratios to prevent outliers from being absorbed into the main cluster, thereby significantly improving the robustness of the initial value estimation.

[0053] In this embodiment, the minimum number of points is usually taken as... To capture the smallest cluster.

[0054] For any point Adaptive neighborhood radius The area is defined as " Neighborhood, neighborhood points for (12) like Then it is called As the core point. Starting from the core point, clusters are formed by expanding through density reachability relationships: if the point Located at a certain core point Within the neighborhood, If the point belongs to the cluster where the core point is located, It is itself a core point, therefore... of The remaining points in the neighborhood are judged and the cluster is expanded; for non-core points, if they are located in the neighborhood of any core point, they are absorbed as boundary points, otherwise they are marked as noise.

[0055] The specific execution steps are as follows: a) Traverse all unvisited points. If the current point is unvisited, calculate its... The number of points in the neighborhood.

[0056] b) If the number of neighboring points If the current point is the core point, a new cluster is created based on it, and all points in the neighborhood are added to the seed set.

[0057] c) Recursively process the points in the seed set: For each seed point, if it has not been visited, mark it as visited and calculate its... Neighborhood; if the point is also a core point, add the points in its neighborhood that are not yet included in the current cluster to the seed set; add the seed point to the current cluster.

[0058] d) If the number of neighboring points If the current point is temporarily marked as noise (it may be absorbed by other core points later), then the current point is temporarily marked as noise.

[0059] e) Repeat the above process until all points have been visited.

[0060] After clustering, the cluster containing the most points is selected as the normal value cluster. Let the set of points contained in the normal value cluster be . The initial angle estimate is then taken as the cluster mean: (13) Initial integer fuzzy number Using point sets The majority voting method was determined.

[0061] Step 2: Outlier Detection and Correction Based on Finite Memory Obtain the initial value through step one. and initial integer fuzzy number Then, the recursive error correction stage is entered using a finite memory algorithm. The core of this stage is to use a sliding window to store historically corrected angle data, detect outliers by the deviation between the window mean and the current measurement value, and correct for anomalies by using the integer ambiguity information from the previous time step.

[0062] Let the current time be The length of the sliding window is The window stores the most recent An angle value after correction Calculate the nearest [item] within the window. The average of the corrected angle values : (14) The angle value obtained by successively defuzzifying the current moment The deviation is defined by comparing it with the window mean. (15) like (threshold) Pick ,in If the value is a multiple of half the wavelength corresponding to the longest baseline in a non-uniform linear array, then the current measurement is considered correct and output directly. , The integer number of the current time period. This is the measured value of the integer fuzzy number at the current moment; if If the current measurement value is determined to be an outlier, then the integer ambiguity number corrected at the previous time step is used. and its neighboring values Corrections are made. Based on the interferometer angle measurement formula, three candidate angles are constructed. (16) in, This represents the phase difference corresponding to the longest baseline in the non-uniform linear array at the current moment.

[0063] Calculate the three candidate angles and the correction angle from the previous moment. absolute deviation (17) Select the candidate angle with the smallest deviation as the correction value at the current moment. (18) in, The increment of the fuzzy number for the corresponding candidate angle If the deviation of all three candidate angles is greater than the threshold... Then the correction result from the previous moment remains unchanged. .

[0064] In the process of finite memory recursion, if the duration of consecutive outliers is too long, the error correction algorithm may deviate from the true value due to accumulated errors, leading to continuous errors in subsequent estimates. To address this, an error correction and reset mechanism needs to be introduced. When the proportion of errors within the window exceeds the adaptive reset threshold, the current recursion is automatically terminated and the initial value presetting process is restarted, thereby preventing the long-term accumulation of errors.

[0065] Define a fixed-length array of status flags to record the most recent The result of each detection in the recursion. Define the flag bit. (19) After each iteration, Storage length is The sliding buffer is used. If the buffer is not full, the number of errors in the stored data is counted. When the buffer is full, count the most recent... Number of errors in the next iteration : (20) If the buffer is not full, calculate the probability of error in the already stored data. When the buffer is full, calculate the nearest... Probability of error in the next iteration ; (twenty one) Regarding the selection of the error probability decision threshold, traditional methods typically set a fixed error proportion threshold (e.g., 50%). However, this threshold cannot adapt to the measurement uncertainties caused by dynamic changes in the carrier-to-noise ratio: when the phase noise is high, a fixed threshold may lead to premature reset and loss of effective tracking information; when the phase noise is low, a fixed threshold may lead to delayed reset and error accumulation. Therefore, this invention proposes an adaptive reset threshold that dynamically adjusts the allowable upper limit of the error probability based on the standard deviation of the angle measurement noise at the current moment. Its expression is: (twenty two) in, , , The standard deviation of the noise is measured for the angle at the current moment. The standard deviation of the angular measurement noise of the system at the nominal carrier-to-noise ratio.

[0066] This threshold design follows the following physical intuition: when the standard deviation of noise is measured at the current angle... When the value is small, the measurement is highly reliable, and the upper limit of the allowable error probability should be close to... (That is, tolerating at most half of the recursive results as errors occur); as noise increases, the uncertainty of the measurement increases, and the algorithm should trigger a reset earlier to avoid error propagation. Therefore, the upper limit of the allowable error probability adaptively decreases to... This adaptive mechanism ensures that the reset conditions match the signal quality in real time, improving the robustness of the algorithm over a wide dynamic carrier-to-noise ratio range.

[0067] Step 3: Secondary correction of steering vector related SVC In this embodiment, the error probability is recursively derived within a finite memory sliding window. If the algorithm is deemed unable to effectively track the true angle, a second SVC correction needs to be triggered, and a forced reset is required. The reset operation includes: clearing the state flag array, discarding all intermediate variables accumulated during the current recursion process, jumping back to the initial value preset stage, re-collecting data, and estimating the new initial angle using the one-dimensional DBSCAN algorithm. With the initial integer fuzzy number The SVC secondary correction process is as follows: For a certain angle of incidence Its phase steering vector can be constructed as .in For the first The phase difference corresponding to the solid baselines, , The number of real baselines. Because the complex exponential function... With a periodicity, this steering vector is naturally insensitive to phase jumps, thus avoiding matching errors caused by phase entanglement.

[0068] To reduce the impact of noise from a single measurement, the current time of the first... Phase difference measurements corresponding to the solid baselines are performed. Point complex vector average (twenty three) in, Indicates the first At the [time]th moment The phase difference corresponding to the solid baseline.

[0069] This constructs the receiving guidance vector. The correlation coefficient is defined as the matching metric. In this step, the cosine correlation criterion is selected, and its expression is: (twenty four) in, Represents the correlation coefficient. express The corresponding phase difference.

[0070] Calculate the candidate angles at the current time. Angle correction from the previous moment Correlation coefficient and The one with the larger correlation coefficient is selected as the final output. (25) Simultaneously, the status flag array is reset, and the data validity flag is inverted to indicate that the current data is the result after secondary correction. This mechanism utilizes the spatial structure information of the antenna array to introduce spatial domain constraints on top of time-domain detection. It can provide additional discrimination criteria in critical situations where recursion with limited memory is difficult to determine, significantly improving the error correction capability for persistent outliers.

[0071] This invention proposes a steering vector correlation error correction algorithm to identify and correct erroneous phase ambiguity estimates, effectively improving the deambiguity accuracy of radar phase interferometers under low signal-to-noise ratio conditions. Specifically: The method of this invention estimates the current phase noise level based on the carrier-to-noise ratio, carrier tracking loop bandwidth, and coherent integration time, and adaptively determines the one-dimensional DBSCAN neighborhood radius accordingly, thereby improving the reliability of the initial angle estimation under low carrier-to-noise ratio or strong phase noise conditions.

[0072] The method of this invention uses the proportion of anomalies within a finite memory window as a recursive reliability index, and dynamically adjusts the SVC verification trigger adaptive reset threshold according to the carrier-to-noise ratio or short-time phase residual, so that secondary correction can be triggered in advance under low carrier-to-noise ratio conditions, and unnecessary secondary correction can be reduced under high carrier-to-noise ratio conditions.

[0073] This invention comprises a closed-loop robust defuzzification process consisting of adaptive initialization clustering, finite memory outlier detection, risk probability triggering, and guide vector correlation verification. This process can suppress the propagation of phase fuzziness errors in multiple stages, including initial estimation, recursive detection, anomaly triggering, and spatial verification, thus preventing the continuous accumulation of single-level fuzziness number decision errors during the successive recursion of long baselines.

[0074] To verify the effectiveness of this invention, the following simulation experiments were conducted: In this example, the simulation parameters of the interferometer system are as follows: The phase interferometer uses a one-dimensional four-baseline non-uniform linear array, and the antenna array layout has the following baseline lengths. Set to respectively ,in Indicates the first The positional offset of each array element relative to the reference point, and the actual spacing between each array element can be determined by the difference in position between adjacent array elements. Confirmed, among which Specifically, array elements With the array element and array elements With the array element The distance between them is Half-wavelength, array element With the array element The distance between them is Half-wavelength, array element With the array element The distance between them is Half a wavelength. By receiving carrier phase information through the aforementioned four channels, and with each antenna channel containing independent zero-mean Gaussian white noise, angle calculation can be achieved using a successive recursive deambiguation method based on a virtual baseline. The virtual baseline is constructed as follows: (26) The virtual baseline constructed in this way has a length exactly half the carrier wavelength, enabling unambiguous angle measurement with one-fold accuracy. Based on this, using the virtual baseline successive recursive deambiguation algorithm described above, an unambiguous angle estimate with 23 times the accuracy can be obtained. This estimate will be used as input data for the SVC error correction algorithm. The simulation signal parameters are set as shown in Table 1.

[0075] To comprehensively evaluate the error correction performance of the SVC algorithm under different motion states, three typical working conditions were designed for simulation verification. The static working condition was used to examine the steady-state error correction capability of the algorithm in the absence of dynamic disturbances; the uniform motion working condition covered the complete angle measurement range to verify the algorithm's suppression effect on continuous outliers in stable relative motion; the sinusoidal motion working condition further examined the algorithm's adaptability to changes in angle rate of change and acceleration. Among them, sinusoidal motion 1 focuses on the error correction performance under different angular velocity conditions, while sinusoidal motion 2 focuses on the algorithm's response characteristics under different angular acceleration conditions. The specific parameter settings for the three working conditions are shown in Table 2.

[0076] Given the large number of sampling points and the complex and diverse forms of outliers in the test cases of the operational design, three methods with continuous outlier processing capabilities were selected for comparison to evaluate the defuzzification and error correction performance of the SVC error correction algorithm: wavelet transform coefficient detection, least squares B-spline approximation, and finite memory defuzzification and error correction algorithm. To ensure fairness in the comparison, all test conditions were kept consistent except for the defuzzification and error correction stage, and the aforementioned simulation parameter settings were uniformly adopted. To visually demonstrate the outlier removal effect of the SVC algorithm, the original carrier phase data was downsampled by 15 points. The corresponding results are as follows: Figure 3-6 As shown, the proportion of outliers after processing by each algorithm is listed in Table 3.

[0077] Simulation results show that the SVC error correction algorithm exhibits excellent outlier detection and correction capabilities under four typical operating conditions, with a significantly lower outlier ratio compared to the other three comparative algorithms. Under static conditions, the algorithm effectively identifies and corrects isolated outliers caused by phase noise, maintaining the stability of angle estimation. Under uniform motion conditions, the algorithm demonstrates significant suppression of continuously occurring outliers, effectively preventing angle estimation deviations caused by error accumulation. Under sinusoidal motion conditions, regardless of changes in the rate of angle change or acceleration, the algorithm maintains high detection and correction accuracy, demonstrating its robustness to dynamic motion modes. The SVC error correction algorithm can adapt to various operating conditions ranging from static to complex dynamic environments, effectively identifying and accurately correcting both isolated and continuous outliers, providing a reliable guarantee for subsequent high-precision angle data processing.

[0078] To further examine the performance of each algorithm under different carrier-to-noise ratio conditions, the simulation parameters for static motion in Table 2 were used to conduct 1000 independent Monte Carlo experiments for various carrier-to-noise ratio scenarios. The probability of outliers before and after processing by each algorithm was statistically analyzed, and the results are shown in Table 3.

[0079] contrast Figure 7 The processing results show that under strong phase noise interference (i.e. The original measurement sequence contains significant outliers. Finite memory, wavelet transform, and least squares B-spline methods can all suppress abnormal fluctuations and reduce the proportion of outliers to some extent, but their outlier identification accuracy and filtering thoroughness are significantly inferior to the SVC algorithm. Under weak phase noise conditions (i.e.... When the outlier contamination of the original observation sequence approaches zero, all the algorithms mentioned above exhibit good feature fidelity and do not erroneously remove valid measurement data. This shows that the SVC algorithm not only has excellent outlier removal capabilities in environments with severe phase noise, but also ensures zero misjudgments within ideal observation intervals, demonstrating strong environmental adaptability.

[0080] Table 1 Signal Processing Parameter Settings

[0081] Table 2 Simulation Condition Design Table

[0082] Table 3. Statistics on the proportion of wild values ​​after wild removal for each algorithm.

[0083] In summary, the above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A radar angle measurement phase ambiguity resolution method based on steering vector correlation, characterized in that, The specific process is as follows: Initial angle and integer ambiguity number estimation: setting an adaptive neighborhood radius and minimum points We use the density-based clustering algorithm DBSCAN to estimate the initial values ​​of angle and integer ambiguity number. Adaptive neighborhood radius The calculation process is as follows: based on the current carrier-to-noise ratio estimated by the receiver. Carrier tracking loop bandwidth and coherent integration time Estimate the standard deviation of phase measurement noise The standard deviation of phase measurement noise Mapped to the standard deviation of angle measurement noise ; The standard deviation of noise is measured based on the angle. Set adaptive neighborhood radius ; Outlier detection and correction: Based on the estimated angle and the initial value of the integer fuzzy number, the finite memory algorithm is used to perform recursive outlier detection and correction; During the recursion process, the allowable error probability is dynamically adjusted based on the standard deviation of the angle measurement noise at the current moment and used as the adaptive reset threshold. When the recursion error probability within the window exceeds the adaptive reset threshold, the recursion is terminated and the process of secondary correction and reset of the guide vector related SVC is entered. SVC Secondary Correction and Reset: After using SVC secondary correction to correct the current angle value, a forced reset is performed. The reset process is as follows: clear the status flag array, discard all intermediate variables accumulated in the current recursive process, and jump back to the initial angle and integer fuzzy number estimation.

2. The radar angle measurement and phase ambiguity resolution method based on steering vector correlation according to claim 1, characterized in that, Adaptive neighborhood radius The calculation process is as follows: First, based on the current carrier-to-noise ratio estimated by the receiver. Carrier tracking loop bandwidth and coherent integration time Estimate the standard deviation of phase measurement noise ; Secondly, the standard deviation of the phase measurement noise Mapped to the standard deviation of angle measurement noise : in, For the signal wavelength, The longest baseline length, This is a rough estimate of the angle at the current moment; Finally, the standard deviation of the noise was measured based on the angle. Set adaptive neighborhood radius for: in, is the clustering confidence coefficient.

3. The radar angle measurement and phase ambiguity resolution method based on steering vector correlation according to claim 2, characterized in that, Cluster confidence coefficient The value range is [2,3], minimum number of points .

4. The radar angle measurement and phase ambiguity resolution method based on steering vector correlation according to claim 2, characterized in that, After clustering, the cluster containing the most points is selected as the normal value cluster. Let the set of points contained in the normal value cluster be . The initial angle is taken as the mean within the cluster; the initial integer fuzzy number Using point sets The majority voting method was determined.

5. The radar angle measurement and phase ambiguity resolution method based on steering vector correlation according to claim 2, characterized in that, During the recursive process, the successively defuzzified output value at the current time step is used. Recent in the window The average of the corrected angle values The comparison is made, and the deviation is defined as... ; Define a fixed-length array of status flags to record the most recent The result of each detection in the recursion; the flag is defined as: in, This indicates that a threshold value is set; after each iteration, [the threshold value will be set]. Storage length is In the sliding buffer; If the buffer is not full, count the number of errors that have occurred with the stored data. When the buffer is full, count the most recent... Number of errors in the next iteration ; And further calculate the recursive error probability within the window. ; like At this point, it is determined that the true angle can no longer be effectively tracked, so the recursion is terminated and the process of secondary correction and reset of the guide vector related SVC is initiated. This indicates an adaptive reset threshold.

6. The radar angle measurement and phase ambiguity resolution method based on steering vector correlation according to claim 5, characterized in that, Adaptive reset threshold for: in, This represents the upper limit of the error probability under high carrier-to-noise ratio. This represents the upper limit of the error probability under low carrier-to-noise ratio. The standard deviation of the noise is measured for the angle at the current moment. The standard deviation of the angular measurement noise of the system at the nominal carrier-to-noise ratio.

7. The radar angle measurement and phase ambiguity resolution method based on steering vector correlation according to claim 5, characterized in that, During the second correction of SVC, the cosine matching correlation criterion is selected to calculate the candidate angle at the current time. Angle correction from the previous moment Correlation coefficient and , For the most recent window The average of the corrected angle values; The angle corresponding to the one with the larger correlation coefficient is selected as the final angle.

8. The radar angle measurement and phase ambiguity resolution method based on steering vector correlation according to claim 7, characterized in that, The expression for the correlation coefficient is: in, For the current moment Phase difference measurements corresponding to the solid baselines are performed Point complex vector average, Represents angle value The corresponding phase difference, .

9. The radar angle measurement and phase ambiguity resolution method based on steering vector correlation according to claim 6, characterized in that, It also resets the status flag array and inverts the data validity flag to indicate that the current data is the result after secondary correction.

10. A radar angle measurement and phase ambiguity resolution device based on steering vector correlation, characterized in that, include: Initial angle and integer ambiguity estimation module: setting adaptive neighborhood radius and minimum points We use the density-based clustering algorithm DBSCAN to estimate the initial values ​​of angle and integer ambiguity number. Adaptive neighborhood radius The calculation process is as follows: based on the current carrier-to-noise ratio estimated by the receiver. Carrier tracking loop bandwidth and coherent integration time Estimate the standard deviation of phase measurement noise The standard deviation of phase measurement noise Mapped to the standard deviation of angle measurement noise ; The standard deviation of noise is measured based on the angle. Set adaptive neighborhood radius ; Outlier detection and correction module: Based on the estimated angle and the initial value of integer fuzziness, the module uses a finite memory algorithm to perform recursive outlier detection and correction. During the recursion process, the allowable error probability is dynamically adjusted based on the standard deviation of the angle measurement noise at the current moment and used as the adaptive reset threshold. When the recursion error probability within the window exceeds the adaptive reset threshold, the recursion is terminated and the process of secondary correction and reset of the guide vector related SVC is entered. SVC Secondary Correction and Reset Module: After using SVC secondary correction to correct the current angle value, a forced reset is performed. The reset process is as follows: clear the status flag array, discard all intermediate variables accumulated in the current recursive process, and jump back to the initial angle and integer fuzzy number estimation.

Images