Antenna beam calibration method based on feature extraction

Abstract

The present invention belongs to the technical field of antenna precision correction, and specifically relates to an antenna beam calibration method based on feature extraction. The specific process is: the array antenna rotates on a turntable according to a set rule, measures the real beam pointing of the beam under some designed pointing angles, and performs preprocessing as follows: averages the beam pointing on the azimuth plane, and separates the error data according to the clockwise and counterclockwise measurement directions of the turntable; utilizes a feature extraction algorithm to respectively extract the preprocessed beam pointing error characteristics under different rotation directions of the turntable during measurement; fits the error characteristics to enhance the dimension, and reconstructs all beam pointing errors within the required angle range based on the fitted characteristics, thereby generating the real beam pointing; based on the real beam pointing, constructs a lookup table and writes it into the array antenna wave control to complete the rapid calibration of the array antenna beam pointing.

Description

Technical Field

[0001] The present invention belongs to the technical field of antenna precision correction, and in particular relates to an antenna beam calibration method based on feature extraction. Background Art

[0002] In recent years, the advantages of array antennas have become increasingly prominent, and they have been widely used in military and civilian fields. The actual beam pointing of array antennas generally deviates from the designed beam pointing, which may have serious consequences during radar applications. Traditional beam pointing correction methods require measuring all actual beam pointing within the required angle range and at the specified beam pointing calibration granularity, thereby establishing a pointing lookup table that corresponds one-to-one between the designed beam pointing and the actual beam pointing to perform beam pointing correction. This method has high accuracy but is costly and time-consuming. Interpolation is also often used in engineering to reduce the number of measurements, but the reduction is limited.

[0003] Feature extraction is the process of extracting useful information from raw data to reduce its complexity while retaining important features, enabling data analysis methods or machine learning models to process the data more effectively. Widely used feature extraction algorithms include principal component analysis (PCA), independent component analysis (ICA), and wavelet transform (WT).

[0004] PCA focuses on the variance of the data. It transforms the original data into another dimension and selects a limited number of principal components with large contributions as the basis for characterizing the original data to achieve the purpose of feature extraction and dimensionality reduction.

[0005] ICA focuses on the independence of data. It reveals the hidden structure in the data by finding transformations that make the decomposed data statistically independent of each other, and extracts statistically independent independent components from the original signal.

[0006] WT decomposes the data into a set of wavelet functions, thereby analyzing the data at different scales and locations. Its advantage is that WT can provide multi-scale and multi-resolution analysis of the signal, so that features at different scales can be effectively extracted.

[0007] Current phased array antenna calibration techniques mostly focus on amplitude and phase calibration. However, due to factors such as mutual coupling and quantization errors, even after these calibrations, the actual beam pointing can still deviate from the designed beam pointing. This deviation is unacceptable in applications requiring precise angle measurement.

[0008] Despite the high demand for beam pointing calibration, research on this topic is very limited. Existing beam pointing calibration methods require measuring the beam pointing at small angle intervals within the required angle range, which is very time-consuming. Interpolation techniques, such as cubic spline interpolation, are often used in engineering to reduce the number of angles required to be measured. However, even with the application of interpolation techniques, the number of angles required to be measured is still large while ensuring the required accuracy. Furthermore, the prediction accuracy of interpolation techniques does not increase monotonically with the increase in measured data, which greatly limits the extent to which the number of measurements can be reduced. Furthermore, the distribution of beam pointing errors exhibits certain characteristics that can be used in beam pointing calibration, but these have not been studied in depth in the existing literature. Summary of the Invention

[0009] The purpose of the present invention is to provide a fast calibration method for array antenna beam pointing based on feature extraction, so as to overcome the shortcomings of the existing technology and greatly improve the calibration speed and system real-time performance while ensuring the accuracy of beam pointing correction.

[0010] The technical solutions of the present invention are as follows:

[0011] An antenna beam calibration method based on feature extraction, the specific process is as follows:

[0012] The array antenna rotates on a turntable according to a set pattern, measuring the actual beam pointing at some designed pointing angles. Preprocessing is then performed by averaging the beam pointing in the azimuth plane and separating the error data based on the clockwise and counterclockwise measurement directions of the turntable.

[0013] The feature extraction algorithm is used to extract the pre-processed beam pointing error characteristics at different rotation directions of the turntable during measurement.

[0014] Fitting the error features to improve dimensionality, reconstructing all beam pointing errors within the required angle range based on the fitted features, and then generating the true beam pointing;

[0015] Based on the actual beam pointing, a lookup table is constructed and written into the array antenna beam control, so as to complete the rapid calibration of the array antenna beam pointing.

[0016] Furthermore, the array antenna of the present invention measures the actual pointing direction of the beam at some designed pointing angles according to a set rule on a turntable, including:

[0017] (1) Set the ranges of azimuth and elevation planes to be calibrated respectively. and [-θ M ,θ M ], set the angular intervals for azimuth and elevation planes to be calibrated as and Δθ;

[0018] (2) From the initial pointing angle of the turntable Initially, the angular velocities of the turntable in the azimuth and elevation planes are set to and θ V＝ 0° / s and rotate to complete a line scan;

[0019] (3) Then the pitch pointing angle is stepped once by Δθ, and the angular velocity of the turntable in the azimuth and pitch planes is reset to and θ V＝ 0° / s and rotate to complete a line scan;

[0020] (4) Then step the pitch pointing angle by Δθ once, and set the angular velocity of the turntable in the azimuth and pitch planes to and θ V＝ 0° / s and rotate to complete a line scan;

[0021] (5) Repeat the above steps (3)-(4) until the pitch pointing angle is greater than θ M , end the scan.

[0022] Furthermore, the feature extraction algorithm of the present invention extracts beam pointing error features, including:

[0023] (1) The measured beam pointing error data are averaged in the azimuth plane, and the data are divided into clockwise measurement error data and counterclockwise measurement error data according to the different rotation directions of the turntable during measurement;

[0024] (2) The feature extraction algorithm is used to extract the data features from the two sets of data. The azimuth error and pitch angle error in each set of data are processed separately, for a total of four sets of error data.

[0025] Furthermore, the present invention fits the error features to improve the dimension, and the specific process is: the features of each group of error data are fitted and dimensionally improved by a smoothing spline fitting algorithm. When fitting the features, the smoothing parameters are traversed and selected to obtain the best fitting result of the features, specifically: the smoothing parameters are traversed and selected to perform smoothing spline fitting on the features; reconstruction is performed based on different fitting results, and three times the standard deviation of part of the data is calculated; the three times the standard deviation of the part of the data is: three times the standard deviation of the difference between the reconstructed pointing error and the measured pointing error at the measurement angle; the fitting result corresponding to the minimum three times the standard deviation of the part of the data is selected as the best fitting result.

[0026] Furthermore, all beam pointing errors within the required angle range are reconstructed based on the fitted features to generate a true beam pointing, including: performing reconstruction according to the best fitting result of the features; reconstructing clockwise error data and counterclockwise error data respectively; rearranging the clockwise measurement error and the error predicted based on the clockwise measurement error, and the error predicted based on the counterclockwise measurement error in the order of the pitch angles to obtain all beam pointing errors; and obtaining the predicted all beam pointings based on the designed beam pointing.

[0027] Furthermore, in one implementation, PCA is used for feature extraction, which has the following characteristics:

[0028] PCA is used to process the beam pointing errors in different rotation directions of the turntable separately to obtain the principal components of each group of data. Each group of error data contains azimuth error and pitch angle error, which are processed separately. Each principal component is fitted and upgraded by smoothing spline fitting, and the entire beam pointing error is predicted based on the reconstructed principal components after fitting.

[0029] Furthermore, the principal components are fitted and dimensionally upgraded by smoothing spline fitting, and reconstruction is performed based on the fitted principal components, including:

[0030] First, the principal component corresponding to the maximum eigenvalue is selected, and the smoothing parameters are traversed to perform smoothing spline fitting on the principal component to obtain the fitting results under various smoothing parameters;

[0031] Perform PCA reconstruction based on the fitted principal components and calculate three times the standard deviation of the difference between the reconstructed pointing error and the measured pointing error at the measured angle;

[0032] The fitting result corresponding to the minimum three times standard deviation is selected as the best fitting result of the principal component, and the three times standard deviation is recorded for subsequent operations;

[0033] Then select the principal components in turn and repeat the above operation until the difference between three adjacent standard deviations is less than 0.01°, and output the beam pointing prediction result at this time.

[0034] Furthermore, the predicting of all beam pointings of the present invention includes:

[0035] PCA reconstruction requires principal components, corresponding eigenvectors, and mean vectors;

[0036] Reconstruct clockwise error data and counterclockwise error data respectively;

[0037] Rearrange the clockwise measurement error and the error based on its prediction, the counterclockwise measurement error and the error based on its prediction in the order of their pitch angles (by θ M to -θ M ) and then the total beam pointing error is obtained;

[0038] According to the designed beam pointing, the predicted total beam pointing is obtained.

[0039] Furthermore, in one implementation, ICA is used for feature extraction, which has the following characteristics:

[0040] ICA is used to process the beam pointing errors of the turntable in different rotation directions, obtaining the independent components and corresponding mixing matrices of each data set. Each set of error data contains azimuth error and elevation error, which are processed separately. Each set of independent components is fitted and dimensionally upgraded using smoothing spline fitting, and reconstruction is performed based on the fitted independent components to predict the total beam pointing error.

[0041] The step of fitting the independent components and increasing the dimension by smoothing spline fitting includes:

[0042] The independent components of each group of data are fitted and dimension-raised by smoothing spline fitting;

[0043] The dimension after fitting the upgraded dimension depends on the actual needs;

[0044] When fitting independent components, the smoothing parameters are traversed to obtain the best fitting results for each group of independent components.

[0045] Furthermore, the present invention traverses and selects smoothing parameters to obtain the best fitting result for each group of independent components, including:

[0046] Traverse and select smoothing parameters, and perform smoothing spline fitting on the group of independent components;

[0047] Perform ICA reconstruction with the mixing matrix of the data set and calculate the three times standard deviation of some data;

[0048] The fitting result corresponding to the minimum three times standard deviation is selected as the best fitting result;

[0049] The best fitting result is taken as the fitting result of this group of independent components.

[0050] Furthermore, the predicting of all beam pointings of the present invention includes:

[0051] Perform ICA reconstruction based on the fitted independent components;

[0052] ICA reconstruction requires a mixing matrix, an independent component matrix, and a mean vector;

[0053] Reconstruct clockwise error data and counterclockwise error data respectively;

[0054] Rearrange the clockwise measurement error and the error based on its prediction, the counterclockwise measurement error and the error based on its prediction in the order of their pitch angles (by θ M to -θM ) and then the total beam pointing error is obtained;

[0055] According to the designed beam pointing, the predicted total beam pointing is obtained.

[0056] In one implementation, WT is used for feature extraction, which has the following characteristics:

[0057] A wavelet basis is selected and the WT is used to process the beam pointing errors in different rotation directions of the turntable to obtain the wavelet coefficients of each data set. Each set of error data contains azimuth error and elevation error, which are processed separately. Each set of wavelet coefficients is fitted and dimensionally upgraded using smoothing spline fitting, and reconstruction is performed based on the fitted wavelet coefficients to predict the total beam pointing error.

[0058] The step of fitting and dimension-raising each group of wavelet coefficients by smoothing spline fitting includes:

[0059] The wavelet coefficients of each group of data are fitted and dimension-raised by smoothing spline fitting;

[0060] The dimension after fitting the upgraded dimension depends on the actual needs;

[0061] When fitting wavelet coefficients, smoothing parameters are traversed and selected to obtain the best fitting result for each set of wavelet coefficients.

[0062] Furthermore, the traversal selection of smoothing parameters in the present invention to obtain the best fitting result of the wavelet coefficients includes:

[0063] Traverse and select smoothing parameters, and perform smoothing spline fitting on the group of wavelet coefficients;

[0064] WT reconstruction was performed separately, and three times the standard deviation of some data was calculated;

[0065] The fitting result corresponding to the minimum three times standard deviation is selected as the best fitting result;

[0066] The best fitting result is used as the fitting result of the set of wavelet coefficients.

[0067] Furthermore, the present invention predicts all beam pointings, including:

[0068] WT reconstruction is performed based on the fitted wavelet coefficients;

[0069] WT reconstruction requires wavelet coefficients, wavelet basis, and decomposition levels;

[0070] Reconstruct clockwise error data and counterclockwise error data respectively;

[0071] Rearrange the clockwise measurement error and the error based on its prediction, the counterclockwise measurement error and the error based on its prediction in the order of their pitch angles (by θM to -θ M ) and then the total beam pointing error is obtained;

[0072] According to the designed beam pointing, the predicted total beam pointing is obtained.

[0073] Beneficial effects:

[0074] First, the array antenna beam pointing rapid calibration method provided by the present invention can predict all beam pointings within the required angle range and within the specified beam pointing calibration granularity while only measuring the actual beam pointing of the beam at a small part of the designed pointing angles; and if the three times the standard deviation of the deviation between the predicted error and the actual error is less than 0.25° as the indicator requirement, this method can meet the standard.

[0075] Second, the predicted beam pointing and the corresponding designed beam pointing are imported into a lookup table and written into the array antenna to achieve correction of the array antenna pointing accuracy.

[0076] Third, the method provided by the present invention effectively improves the beam pointing accuracy and real-time performance of the array antenna system, and has low cost and low system complexity. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0078] Figure 1 FIG2 is a flow chart of the antenna beam calibration method based on feature extraction provided by the present invention;

[0079] Figure 2 FIG2 shows a trajectory diagram of beam pointing in space using a conventional calibration method provided by an embodiment of the present invention, taking a partial angle range of the pitch plane as an example;

[0080] Figure 3 FIG2 shows a trajectory diagram of beam pointing in space using a fast calibration method provided by an embodiment of the present invention, taking a partial angle range of the pitch plane as an example;

[0081] Figure 4 FIG2 is a schematic diagram of an experimental verification process of a PCA-based beam pointing fast calibration method provided by an embodiment of the present invention;

[0082] Figure 5 FIG2 is a schematic diagram of a process of fitting principal components and reconstructing in a PCA-based beam pointing fast calibration method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0083] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0084] It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments may be combined with each other; and, based on the embodiments in this disclosure, all other embodiments obtained by persons of ordinary skill in the art without creative work are within the scope of protection of this disclosure.

[0085] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on this disclosure, it should be understood by those skilled in the art that an 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 described herein can be used to implement an apparatus and / or practice a method. In addition, other structures and / or functionalities other than one or more of the aspects described herein can be used to implement this apparatus and / or practice this method.

[0086] like Figure 1 As shown, the antenna beam calibration method based on feature extraction in the embodiment of the present application has the following specific process:

[0087] S1: The array antenna rotates on a turntable according to a set pattern, and the actual beam pointing direction is measured at some designed pointing angles.

[0088] S2: Using feature extraction algorithms to extract beam pointing error features at different rotation directions of the turntable during measurement;

[0089] S3: Fitting the error feature to improve the dimension;

[0090] S4: Reconstruct all beam pointing errors within the required angle range to generate the true beam pointing;

[0091] S5: Based on the real beam pointing, a lookup table is constructed and written into the array antenna for calibration.

[0092] After obtaining the actual beam pointing of the measurement beam at some of the designed pointing angles in step S1, the beam pointing error is preprocessed. The preprocessing includes: averaging the beam pointing error on the azimuth plane; and separating the error data according to the clockwise and counterclockwise measurement directions of the turntable.

[0093] The trajectory diagram of measuring the total beam pointing error using a partial angle range of the pitch plane as an example is as follows: Figure 3As shown, the intervals in the figure are only for illustration and are not the only implementation form.

[0094] 1) Beam pointing measurements in the azimuth plane are time-sampled, with the number of points depending on the turntable speed. The measured azimuth angle intervals are much smaller than the required angle intervals. The averaging of the beam pointing error in the azimuth plane during preprocessing is performed to eliminate random jitter generated during measurement and reduce the error to the required angle intervals.

[0095] The characteristics of the measurement data under different scanning directions of the turntable are different, so the error data of different scanning directions are separated in the preprocessing;

[0096] 2) At the same time, the average processing in the above preprocessing process is a moving average processing.

[0097] like Figure 2 As shown in Figure 1, the trajectory of measuring all beam pointing errors within a certain angle range of the pitch plane is used as an example. The specific operations are as follows:

[0098] Set the ranges that need to be calibrated for the azimuth and elevation planes to be and [-θ M ,θ M ], the angle intervals that need to be calibrated are and Δθ;

[0099] The array antenna is installed on a two-dimensional turntable, and the initial pointing angle of the turntable is Initially, the angular velocities of the turntable in the azimuth and elevation planes are set to and θ V＝ 0° / s and rotate to complete a line scan;

[0100] Then the pitch pointing angle is stepped once by Δθ, and the angular velocity of the turntable in the azimuth and pitch planes is reset to and θ V＝ 0° / s and rotate to complete a line scan;

[0101] Then step the pitch pointing angle by Δθ once, and set the angular velocity of the turntable in the azimuth and pitch planes to and θ V＝ 0° / s and rotate to complete a line scan;

[0102] Repeat the above steps until the pitch pointing angle is greater than θ M , end the scan.

[0103] Another embodiment of the present application utilizes a feature extraction algorithm to extract the beam pointing error features at different rotation directions of the turntable during measurement, specifically:

[0104] First, the measured beam pointing error data is averaged in the azimuth plane, and the data is divided into clockwise measurement error data and counterclockwise measurement error data according to the different rotation directions of the turntable during measurement;

[0105] Secondly, feature extraction algorithms are used on the two sets of data to extract data features. The azimuth error and pitch angle error in each set of data are processed separately, for a total of four sets of error data.

[0106] The feature extraction algorithms used in this step include, but are not limited to, principal component analysis (PCA), independent component analysis (ICA), wavelet transform (WT), and the like.

[0107] In another embodiment of the present application, fitting the error feature to improve the dimension includes:

[0108] The smoothing spline fitting algorithm is used to fit and upgrade the features of each group of error data. The dimension after fitting and upgrading depends on the actual needs. When fitting the features, the smoothing parameters are traversed and selected to obtain the best fitting result of the features.

[0109] In this embodiment, smoothing parameters are traversed and selected to obtain the best fitting result of the feature, including: traversing and selecting smoothing parameters to perform smoothing spline fitting on the feature; reconstructing based on different fitting results respectively, and calculating three times the standard deviation of part of the data; the three times the standard deviation of the part of the data is: three times the standard deviation of the difference between the reconstructed pointing error and the measured pointing error in the measurement angle; and selecting the fitting result corresponding to the minimum three times the standard deviation of the part of the data as the best fitting result.

[0110] In another embodiment of the present application, all beam pointing errors within a required angle range are reconstructed based on the fitted feature, thereby generating a beam pointing, including: reconstructing according to the best fitting result of the feature; reconstructing clockwise error data and counterclockwise error data respectively; rearranging the clockwise measurement error and the error predicted based on it, and the counterclockwise measurement error and the error predicted based on it in the order of their pitch angles (by θ M to -θ M ) and obtain the total beam pointing error; according to the designed beam pointing, obtain the predicted total beam pointing.

[0111] Example 1:

[0112] This embodiment implements rapid beam pointing calibration based on PCA.

[0113] It should be noted that, in this exemplary embodiment, all beam pointings need to be measured in advance in order to subsequently verify the effectiveness of the method. The verification flow diagram of this exemplary embodiment based on MATLAB R2023a is shown in FIG. Figure 4 Shown, including:

[0114] SS1: Preprocess all beam pointing errors within the required angle range and the specified beam pointing calibration granularity obtained by turntable measurement to obtain the beam pointing errors at some design angles.

[0115] SS2: Perform PCA processing on the beam pointing error to extract error features and obtain the principal components of the error data;

[0116] SS3: Use smoothing spline fitting to fit each principal component and increase the dimensionality. Reconstruct based on the fitted principal components and predict all beam pointing errors within the required angle range.

[0117] SS4: Compare the predicted error with the true error to verify the effectiveness of the method.

[0118] Figure 5 The schematic diagram of the process of fitting the principal components and reconstructing the data in this embodiment is shown, including:

[0119] First, the principal component with the largest eigenvalue is selected, and smoothing parameters are selected to perform smoothing spline fitting on the principal component to obtain the fitting results under various smoothing parameters;

[0120] Secondly, PCA reconstruction is performed based on the fitted principal components, and three times the standard deviation of the difference between the reconstructed pointing error and the measured pointing error at the measured angle is calculated;

[0121] Again, select the fitting result corresponding to the minimum three times standard deviation as the best fitting result of the principal component, and record the three times standard deviation for subsequent operations;

[0122] Finally, the principal components are selected in sequence and the above operation is repeated until the difference between three adjacent standard deviations is less than 0.01°, and the beam pointing prediction result at this time is output.

[0123] The measured beam pointing is subtracted from the designed beam pointing to obtain the beam pointing error data at all designed angles.

[0124] It is understandable that the specific steps of angle measurement are well known to those skilled in the art and will not be described in detail in this specification.

[0125] Furthermore, the preprocessing steps in step SS1 are: the beam pointing error is sampled at 1 / 2 equal intervals on the azimuth plane and averaged every 5 points; the error data is separated according to the clockwise measurement and counterclockwise measurement directions of the turntable; the obtained clockwise measurement data and counterclockwise measurement data are sampled at 1 / 10 equal intervals on the pitch plane respectively.

[0126] The trajectory of the beam pointing error at some angles obtained after preprocessing is similar to the beam pointing measurement trajectory of the traditional method, except that the angle interval is changed to alternating between Δθ and 19Δθ. The schematic diagram of the trajectory of measuring the full beam pointing error using a partial angle range of the pitch plane as an example is shown in the figure below. Figure 3 shown.

[0127] 1) Each embodiment is to verify the effectiveness of the fast calibration method. It is necessary to measure the actual beam pointing at all angles in advance so as to compare it with the predicted beam pointing to evaluate the prediction effect.

[0128] However, when the fast calibration method is actually used, the only available data is the true beam pointing at some angles measured according to established rules. Therefore, it is necessary to sample the true beam pointing at all angles in the pitch plane to simulate the actual usage scenario, that is, the purpose of equally spaced sampling in the pitch plane during the preprocessing process. In this embodiment, 1 / 10 equally spaced sampling is used.

[0129] Beam pointing measurements in the azimuth plane are time-sampled, with the number of points depending on the turntable speed. The measured azimuth angle interval is much smaller than the required angle interval. During preprocessing, the beam pointing error is sampled at equal intervals of 1 / 2 on the azimuth plane and averaged every five points (here, the measured azimuth angle interval is 0.01°, while the required angle interval is 0.1°). This eliminates random jitter generated during measurement and reduces the required angle interval.

[0130] The characteristics of the measurement data under different scanning directions of the turntable are different, so the error data of different scanning directions are separated in the preprocessing;

[0131] If all original data are directly sampled 1 / k in the pitch plane, the sampled data may be all measurement data in a certain scanning direction. Therefore, the measurement data in different scanning directions are separated before sampling.

[0132] The error data after preprocessing can be regarded as Figure 5 The measurement results are obtained by measuring the measurement trace shown.

[0133] It should be noted that in step SS2, the clockwise measurement error data and the counterclockwise measurement error data need to be processed separately by PCA. Each set of data is further divided into azimuth error data and pitch error data, which are processed separately.

[0134] PCA transforms the original data into another dimension and selects a limited number of principal components with large contributions as the basis for characterizing the original data to achieve the purpose of feature extraction and dimensionality reduction. The steps of PCA processing are as follows:

[0135] Decentralize the error data X to obtain X′;

[0136] Construct the covariance matrix V of X′;

[0137] Perform eigendecomposition on the covariance matrix V to obtain the eigenvalue α and the corresponding eigenvector W;

[0138] The extracted main component F=W T X′.

[0139] Furthermore, the obtained principal components are fitted and dimension-upgraded in turn, and all beam pointing error data are reconstructed. The specific steps are the same as Figure 3 Same as shown.

[0140] It should be noted that the formula for PCA reconstruction is:

[0141] X″=W n F n +μ

[0142] Among them, F n is the selected principal component, W n is the corresponding eigenvector, and μ is the mean vector of the samples in the data.

[0143] In this method, the principal components, eigenvectors, and mean vectors after fitting can be selected for PCA reconstruction to predict the entire beam pointing error.

[0144] Based on the above verification process, this embodiment verifies 6 sets of measured data, and the verification results are shown in Table 1.

[0145] If three times the standard deviation of the deviation between the designed beam pointing and the predicted beam pointing is less than 0.25°, the standard is considered to be met.

[0146] It should be noted that, in this embodiment, the deviation between the designed beam pointing and the predicted beam pointing is equivalent to the deviation between the actual beam pointing error and the predicted beam pointing error.

[0147] Table 1

[0148]

[0149] In this embodiment, the verification of the five data sets all met the standards, and the processing time for each data set was about 1 minute, which shows the effectiveness of this method.

[0150] Example 2

[0151] This embodiment uses ICA as the feature extraction algorithm, and the specific implementation process is similar to that of the first embodiment.

[0152] It should be noted that, in this embodiment, when fitting independent components, all independent components need to be fitted and then reconstructed.

[0153] It should be noted that ICA assumes that the original data X is a linear mixture of several independent components S, that is:

[0154] X=AS

[0155] Where X is the original data matrix, S is the independent component matrix, and A is the mixing matrix. The ICA algorithm estimates the independent component matrix S and the mixing matrix A from the original data. During reconstruction, the original data X can be estimated by reconstructing the known independent component matrix S and the mixing matrix A using the formula.

[0156] In this method, the fitted independent components can be selected and reconstructed together with the mixing matrix through ICA to predict the entire beam pointing error.

[0157] The results of verification of 5 sets of measured data in this embodiment are shown in Table 2.

[0158] Table 2

[0159]

[0160] In this embodiment, the verification of 5 sets of data basically meets the standards, and the processing time for each set of data is about 4 minutes, which shows the effectiveness of this method.

[0161] Example 3

[0162] This embodiment uses WT as the feature extraction algorithm, and the specific implementation process is similar to that of the first embodiment.

[0163] It should be noted that, in this embodiment, when fitting the wavelet coefficients, all wavelet coefficients need to be fitted and then reconstructed.

[0164] WT analyzes signals by scaling and shifting wavelet basis functions. Scaling determines the frequency of the wavelet basis functions, while translation determines the location of the wavelet basis functions. The wavelet coefficients represent the intensity of the data at different scales and locations.

[0165] During reconstruction, the original signal can be reconstructed from the wavelet coefficients, wavelet basis functions and the number of decomposition layers.

[0166] In this method, the fitted wavelet coefficients can be selected and used together with the set wavelet basis functions and decomposition levels to perform wavelet reconstruction and predict the entire beam pointing error.

[0167] The results of verifying the five sets of measured data in this embodiment are shown in Table 3.

[0168] Table 3

[0169]

[0170] In this embodiment, the verification results of the five groups of data all meet the requirements, and the processing time for each group of data is about 3 minutes, which shows the effectiveness of this method.

[0171] Furthermore, in specific practice, the predicted beam pointing error can be used to obtain all beam pointings within the required angle range and calibration interval according to the corresponding designed beam pointing.

[0172] A lookup table can be established based on the predicted beam pointing and the corresponding designed beam pointing, and written into the array to complete the pointing calibration.

[0173] The rapid beam pointing calibration method provided by this invention can predict the full beam pointing range within a required angle range and within a specified beam pointing calibration granularity, while only measuring the actual beam pointing at a few actual angles. Compared to existing technologies, this method requires fewer measurements, significantly improving efficiency, and offering high accuracy, low cost, and ease of engineering implementation.

[0174] It should be noted that, in the description of the present invention, terms such as "first" and "second" are used for illustrative purposes only and do not indicate or imply relative importance. In addition, in the description of the present application, unless otherwise explicitly stated, the term "plurality" refers to at least two.

[0175] Any process or method described in any flowchart or other manner described in the present invention can be understood as a code module, segment, or portion including one or more executable instructions for performing steps of a specific logical function or process. The scope of the preferred embodiments of the present invention also includes other implementations, which may not be performed in the order shown, as understood by those skilled in the art to which the present invention relates.

[0176] It should be understood that when describing the present invention, terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" refer to at least one embodiment or example having specific features, structures, materials, or characteristics. The use of these terms does not necessarily refer to the same embodiment or example, and the specific features, structures, materials, or characteristics described in the description may be appropriately combined in one or more embodiments or examples.

[0177] The above specific embodiments further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent replacements or improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An antenna beam calibration method based on feature extraction, characterized in that: The specific process is: The array antenna rotates on a turntable according to a set pattern, measuring the actual beam pointing at some designed pointing angles. Preprocessing is then performed by averaging the beam pointing in the azimuth plane and separating the error data based on the clockwise and counterclockwise measurement directions of the turntable. The feature extraction algorithm is used to extract the pre-processed beam pointing error characteristics at different rotation directions of the turntable during measurement. Fitting the error features to improve dimensionality, reconstructing all beam pointing errors within the required angle range based on the fitted features, and then generating the true beam pointing; Based on the actual beam pointing, a lookup table is constructed and written into the array antenna beam control to complete the rapid calibration of the array antenna beam pointing; The feature extraction algorithm extracts beam pointing error features, including: (1) The measured beam pointing error data are averaged in the azimuth plane, and the data are divided into clockwise measurement error data and counterclockwise measurement error data according to the different rotation directions of the turntable during measurement; (2) The feature extraction algorithm is used to extract the data features of the two sets of data. The azimuth error and pitch angle error of each set of data are processed separately, totaling four sets of error data; The error feature is fitted to improve the dimension, and the specific process is: the features of each group of error data are fitted and improved by a smoothing spline fitting algorithm, and when fitting the features, smoothing parameters are traversed and selected to obtain the best fitting result of the features; All beam pointing errors within a required angle range are reconstructed based on the fitted features to generate a true beam pointing, including: performing reconstruction based on a best-fit result of the features; separately reconstructing clockwise error data and counterclockwise error data; rearranging the clockwise measurement error and the error predicted based thereon, and the counterclockwise measurement error and the error predicted thereon in order of their pitch angles to obtain all beam pointing errors; and obtaining a predicted all beam pointing based on the designed beam pointing.

2. The antenna beam calibration method based on feature extraction according to claim 1, characterized in that: The array antenna measures the actual pointing direction of the beam at some designed pointing angles according to a set rule on the turntable, including: (1) Set the ranges of azimuth and elevation planes to be calibrated respectively. and [-θ M ,θ M ], set the angular intervals for azimuth and elevation planes to be calibrated as and Δθ; (2) From the initial pointing angle of the turntable Initially, the angular velocities of the turntable in the azimuth and elevation planes are set to and θ V＝ 0° / s and rotate to complete a line scan; (3) Then the pitch pointing angle is stepped once by Δθ, and the angular velocity of the turntable in the azimuth and pitch planes is reset to and θ V＝ 0° / s and rotate to complete a line scan; (4) Then step the pitch pointing angle by Δθ once, and set the angular velocity of the turntable in the azimuth and pitch planes to and θ V＝ 0° / s and rotate to complete a line scan; (5) Repeat the above steps (3)-(4) until the pitch pointing angle is greater than θ M , end the scan.

3. The antenna beam calibration method based on feature extraction according to claim 1, characterized in that: The specific process of obtaining the best fitting result is: traversing and selecting smoothing parameters, and performing smoothing spline fitting on the features; reconstructing based on different fitting results, and calculating the three times standard deviation of some data; the three times standard deviation of the said partial data is: three times the standard deviation of the difference between the reconstructed pointing error and the measured pointing error in the measurement angle; selecting the fitting result corresponding to the minimum three times standard deviation of the partial data as the best fitting result.

4. The antenna beam calibration method based on feature extraction according to claim 1, characterized in that: The feature extraction algorithm is principal component analysis PCA, independent component analysis or wavelet transform.

Images