An Adaptive Suppression Method for Meteorological Target Sidelobes Based on Optimal Component Extraction

Abstract

This invention provides an adaptive sidelobe suppression method for meteorological targets based on optimal component extraction. The method includes: constructing a waveform transmission matrix; constructing a cost function based on the waveform transmission matrix; calculating an optimal extraction matrix based on the cost function; calculating a fluctuation qualification rate based on the optimal extraction matrix; and achieving adaptive sidelobe suppression of meteorological targets based on the optimal extraction matrix and the fluctuation qualification rate. This invention can model the expected energy output function of each range library to construct a cost function and obtain the optimal extraction matrix, making the optimized signal energy close to the main lobe energy of the actual target in the range library, thus achieving suppression of sidelobes of meteorological targets with large reflectivity gradients. When the reflectivity gradient is large, this method can significantly improve the basic data estimation error in weak target areas; when the reflectivity gradient is small, this method can improve the basic data estimation accuracy and effectively improve data quality.

Description

Technical Field

[0001] This invention relates to the field of meteorological radar technology, and in particular to an adaptive suppression method for meteorological target sidelobes based on optimal component extraction. Background Technology

[0002] With the application of solid-state transmitters in weather radar, pulse compression technology has been widely used in various radar systems. LFM waveform signals are the most widely used pulse compression waveform signals. For LFM waveform signals, the commonly used sidelobe suppression method is windowing. However, weather targets have a large dynamic range (-10 to 75 dBZ), requiring high radar sensitivity. Although windowing can suppress sidelobes to some extent, it causes energy loss, affecting the detection of weak targets. To reduce the signal-to-noise ratio loss after windowing LFM waveform signals, NLFM waveform signals are also increasingly used. NLFM waveform signals can achieve lower sidelobes by transferring the spectral weighting function to the frequency modulation function and directly performing matched filtering. However, the sidelobe performance of NLFM waveforms is insufficient to meet the accuracy requirements for quantitative detection of distributed scattering targets, necessitating further sidelobe suppression.

[0003] To achieve ultra-low range sidelobes, a mismatch filter can be used to pulse compress the echoes at the cost of a certain signal-to-noise ratio. However, current performance is only achievable when the observed target is a point target or a uniformly distributed target. When the reflectivity gradient reaches a certain range, the sidelobe energy of a strong target becomes comparable to the main lobe energy of a weak target. In this case, the weak signal will be overwhelmed by the sidelobes of the strong reflectivity signal, leading to an overestimation of dBZ, incorrect velocity spectrum estimation, and even the generation of false targets. Summary of the Invention

[0004] Therefore, it is necessary to provide an adaptive suppression method for meteorological target sidelobes based on optimal component extraction to address the aforementioned technical problems.

[0005] An adaptive suppression method for meteorological target sidelobes based on optimal component extraction includes the following steps:

[0006] Constructing the waveform transmission matrix:

[0007] S MF,k =[S MF,k (1),S MF,k (2),…,S MF,k (p),…,S MF,k (P)]

[0008] S SP,k =[S SP,k (1),S SP,k (2),…,S SP,k (p),…,S SP,k (P)]

[0009] Among them, S MF,k S represents the pulse compression waveform transfer function. SP,k Let p represent the p-th range library, P represent the total number of range libraries in the radar echo range direction, and k represent the normalized echo amplitude of the k-th target acting on that range library.

[0010] Construct a cost function based on the waveform transmission matrix;

[0011] Calculate the optimal extraction matrix based on the cost function;

[0012] Calculate the fluctuation pass rate based on the optimal extraction matrix;

[0013] Adaptive suppression of meteorological target sidelobes is achieved based on the optimal extraction matrix and the fluctuation qualification rate.

[0014] In one embodiment, constructing the cost function based on the waveform transmission matrix includes:

[0015] Construct a K×P dimensional weighted matrix based on the waveform transmission matrix:

[0016]

[0017] Among them, w opt Let w represent the weighted matrix. opt,p =[w opt,p (1)...w opt,p (K)] T Let represent the optimization coefficient of the p-th range library, where p represents the p-th range library, P represents the total number of range libraries in the radar echo range direction, k represents the normalized echo amplitude of the k-th target acting on the range library, and K represents the normalized echo amplitude of the target acting on the range library.

[0018] The signals for each distance library are weighted according to the following formula:

[0019] P est =A(w opt ⊙S MF )

[0020] Among them, P est Let A represent the estimated echo signal, and w represent the target characteristic matrix. opt Let S represent the weighted matrix. MF Represents the pulse compression waveform transfer function;

[0021] Using the simple pulse waveform transfer function S SP The ideal echo signal is calculated using the following formula:

[0022]

[0023] Among them, P opt Let S represent the ideal echo signal, A represent the target characteristic matrix, and S represent the target characteristic matrix. SP A represents the transfer function of a simple pulse waveform. k This represents the pointer element for the k-th target characteristic, where k represents the normalized echo amplitude of the k-th target acting on the range library, and K represents the normalized echo amplitude of the target acting on the range library. A P This represents the P-th target characteristic pointer element, where P represents the total number of range data in the radar echo range direction;

[0024] The echo function and the weighted matrix model satisfy the following formula relationship:

[0025]

[0026] in, This means that p(w) can be made opt The smallest weighted matrix, p(w) opt P represents the signal function to be optimized. est P represents the estimated echo signal. opt This represents the ideal echo signal.

[0027] In one embodiment, calculating the optimal extraction matrix based on the cost function includes:

[0028] The optimal extraction matrix is ​​calculated using the following formula:

[0029] W opt =(I⊙B⊙BI) T ) -1 (I⊙BD)

[0030] Among them, W opt Let I represent the optimal extraction matrix, and let B = S. MF Denotes the pulse compression waveform transfer function, D = S sp This represents the simple pulse waveform transfer function, where T represents the transpose calculation.

[0031] In one embodiment, calculating the volatility compliance rate based on the optimal extraction matrix includes:

[0032] Calculate the optimized echo signal data based on the optimal extraction matrix;

[0033] Calculate the base data based on the optimized echo signal data;

[0034] Calculate the data offset error based on the base data, and obtain the base data in response to the data offset error being less than a preset offset threshold.

[0035] Calculate the standard deviation of the base data based on the base data;

[0036] The data fluctuation pass rate is calculated based on the standard deviation of the base data. In response to the data fluctuation pass rate being greater than the preset fluctuation pass rate, the data fluctuation pass rate is obtained.

[0037] In one embodiment, calculating the optimized echo signal data based on the optimal extraction matrix includes:

[0038] The optimized echo signal data is calculated using the following formula.

[0039] S est =A(W opt ⊙S MF )

[0040] Among them, S est This represents the optimized echo signal data, where A represents the target characteristic matrix, and W represents the target characteristic matrix. opt S represents the optimal extraction matrix. MF represents the pulse compression waveform transfer function, and ⊙ represents the Hadamard product.

[0041] In one embodiment, calculating the base data based on the optimized echo signal data includes:

[0042] The optimized echo signal data is subjected to spectral moment estimation to obtain basic data; wherein, the basic data includes: reflectivity factor, average velocity and spectral width.

[0043] In one embodiment, calculating the data offset error based on the base data includes:

[0044] Calculate the data offset error using the following formula:

[0045]

[0046] Among them, RMSE E This represents the data offset error, p represents the p-th range database, P represents the total number of range databases in the radar echo range direction, and E est (p) represents the estimated baseline data, E gt (p) represents the ideal base data.

[0047] In one embodiment, calculating the standard deviation of the base data based on the base data includes:

[0048] Calculate the standard deviation of the base data using the following formula:

[0049]

[0050]

[0051] Among them, SD pE represents the standard deviation of the base data. p+j-n This represents the (p+jn)th base data. Let P represent the moving average calculated using 2n+1 range libraries, where P represents the total number of range libraries in the radar echo range direction, p represents the p-th range library, n represents the n-th range library, and j represents the j-th range library.

[0052] In one embodiment, calculating the data volatility compliance rate based on the standard deviation of the base data includes:

[0053] The data fluctuation pass rate is calculated using the following formula:

[0054]

[0055] Where AR represents the data fluctuation pass rate, sum() represents cumulative calculation, P represents the total number of range databases in the radar echo range direction, p represents the p-th range database, and SD p This represents the standard deviation of the base data.

[0056] Compared to existing technologies, the advantages and beneficial effects of this invention are as follows: This invention can model the expected energy output function of each range library to construct a cost function and obtain the optimal extraction matrix, making the optimized signal energy close to the main lobe energy of the actual target in the range library, thus suppressing the side lobes of meteorological targets with large reflectivity gradients. When the reflectivity gradient is large, this method can significantly improve the basic data estimation error in weak target regions; when the reflectivity gradient is small, this method can improve the basic data estimation accuracy and effectively improve data quality. Attached Figure Description

[0057] Figure 1 This is a flowchart illustrating an adaptive suppression method for meteorological target sidelobes based on optimal component extraction in one embodiment.

[0058] Figure 2A This is a schematic diagram of the reflectance factor distribution when the reflectance gradient is 15 dBZ / kms in one embodiment.

[0059] Figure 2B This is a schematic diagram of the reflectivity factor distribution when the reflectivity gradient is 25 dBZ / km in one embodiment.

[0060] Figure 2C This is a schematic diagram of the reflectivity factor distribution when the reflectivity gradient is 40 dBZ / km in one embodiment;

[0061] Figure 3A This is the result of estimating the reflectivity factor of a one-dimensional distributed meteorological target echo before and after optimization when the reflectivity gradient is 15 dBZ / kms in one embodiment.

[0062] Figure 3BThis is the result of estimating the reflectance factor of a one-dimensional distributed meteorological target echo before and after optimization when the reflectance gradient is 25 dBZ / km in one embodiment.

[0063] Figure 3C This is the result of estimating the reflectance factor of a one-dimensional distributed meteorological target echo before and after optimization when the reflectance gradient is 25 dBZ / km in one embodiment.

[0064] Figure 4A This is the velocity estimation result of a one-dimensional distributed meteorological target echo before and after optimization when the reflectivity gradient is 15 dBZ / kms in one embodiment.

[0065] Figure 4B This is the velocity estimation result of a one-dimensional distributed meteorological target echo before and after optimization when the reflectivity gradient is 25 dBZ / km in one embodiment.

[0066] Figure 4C This is the velocity estimation result of a one-dimensional distributed meteorological target echo before and after optimization when the reflectivity gradient is 40 dBZ / km in one embodiment.

[0067] Figure 5A This is the result of spectral width estimation of a one-dimensional distributed meteorological target echo before and after optimization when the reflectivity gradient is 15 dBZ / kms in one embodiment.

[0068] Figure 5B This is the result of spectral width estimation of a one-dimensional distributed meteorological target echo before and after optimization when the reflectivity gradient is 25 dBZ / km in one embodiment.

[0069] Figure 5C This is the spectral width estimation result of a one-dimensional distributed meteorological target echo before and after optimization when the reflectivity gradient is 40 dBZ / km in one embodiment;

[0070] Figure 6A In one embodiment, the RMSE of the baseline data before and after optimization of the one-dimensional distributed meteorological target echo is given as a function of reflectivity gradient when the reflectivity gradient is 15 dBZ / kms.

[0071] Figure 6B In one embodiment, the RMSE of the baseline data before and after optimization of the one-dimensional distributed meteorological target echo is given as a function of reflectivity gradient when the reflectivity gradient is 25 dBZ / km.

[0072] Figure 6C In one embodiment, the RMSE of the baseline data before and after optimization of the one-dimensional distributed meteorological target echo is given as a function of reflectivity gradient when the reflectivity gradient is 40 dBZ / km.

[0073] Figure 7AIn one embodiment, the RMSE of the baseline data before and after optimization of the one-dimensional distributed meteorological target echo is given as a function of reflectivity gradient when the reflectivity gradient is 15 dBZ / kms.

[0074] Figure 7B In one embodiment, the RMSE of the baseline data before and after optimization of the one-dimensional distributed meteorological target echo is given as a function of reflectivity gradient when the reflectivity gradient is 25 dBZ / km.

[0075] Figure 7C The RMSE of the baseline data before and after optimization of the one-dimensional distributed meteorological target echo is shown as a function of the reflectivity gradient when the reflectivity gradient is 40 dBZ / km in one embodiment. Detailed Implementation

[0076] Before describing the specific embodiments of the present invention, the overall concept of the present invention will be explained as follows:

[0077] This invention is primarily developed for meteorological analysis processes, currently using LFM and NLFM waveform signals. For LFM waveform signals, the commonly used sidelobe suppression method is windowing. However, meteorological targets have a large dynamic range (-10 to 75 dBZ), requiring high radar sensitivity. While windowing can suppress sidelobes to some extent, it causes energy loss, affecting the detection of weak targets. For NLFM waveform signals, transferring the spectral weighting function to the frequency modulation function and directly performing matched filtering can yield lower sidelobes. However, the sidelobe performance of NLFM waveforms is insufficient to meet the accuracy requirements for quantitative detection of distributed scattering targets, necessitating further sidelobe suppression. Suppressing sidelobes comes at the cost of a certain signal-to-noise ratio, using a mismatch filter to pulse-compress the echo. However, current performance is only achievable when the observed target is a point target or uniformly distributed. When the reflectivity gradient reaches a certain range, the sidelobe energy of strong targets becomes comparable to the main lobe energy of weak targets. The weak signal is then overwhelmed by the sidelobes of the strong reflectivity signal, leading to an overestimation of dBZ, incorrect velocity spectrum estimation, and even the generation of false targets.

[0078] Therefore, this invention proposes an adaptive sidelobe suppression method for meteorological targets based on optimal component extraction. The method optimizes the signal of each range library after pulse compression by modeling the expected energy output function of each range library and constructing a cost function using a simple pulse transfer function to obtain the optimal extraction matrix. This matrix is ​​then used to reweight the pulse-compressed signal. The optimized echo is then used to calculate relevant data quality indicators, which are set as the iteration stopping condition. Through continuous iteration, the most contributing main lobe signal in each range library can be extracted from the mixed signal, making the optimized signal energy close to the main lobe energy of the actual target in the range library, thus achieving adaptive sidelobe suppression for meteorological targets with large gradients.

[0079] After introducing the overall concept of the present invention, in order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below through specific embodiments and in conjunction with the accompanying drawings.

[0080] It should be noted that, unless otherwise defined, the technical or scientific terms used in one or more embodiments of this specification should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in one or more embodiments of this specification do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word covers the element or object listed following the word and its equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0081] For ease of understanding, the terms used in the embodiments of this application are explained below:

[0082] LFM signal: Linear frequency modulation signal, refers to a signal whose instantaneous frequency changes linearly with time.

[0083] NLFM signal: Non-linear frequency modulation signal, which refers to a signal whose frequency and time are not non-linear.

[0084] RMSE: Residual Standard Deviation, reflects the degree of data discretization, and is mainly used as a standard for introducing and removing variables in distributed regression analysis.

[0085] AR: Volatility Acceptance Rate, describes the degree of volatility in the data.

[0086] In one embodiment, such as Figure 1 As shown, an adaptive suppression method for meteorological target sidelobes based on optimal component extraction is provided, including the following steps:

[0087] Step S101, construct the waveform transmission matrix:

[0088] S MF,k =[S MF,k (1),S MF,k (2),…,S MF,k (p),…,S MF,k (P)]

[0089] S SP,k =[SSP,k (1),S SP,k (2),…,S SP,k (p),…,S SP,k (P)]

[0090] Among them, S MF,k S represents the pulse compression waveform transfer function. SP,k Let represent a simple pulse waveform transfer function, p represent the p-th range library, P represent the total number of range libraries in the radar echo range direction, and k represent the normalized echo amplitude of the k-th target acting on that range library.

[0091] Specifically, considering a range-oriented signal, a modulated transmit waveform with an amplitude of 1 (such as LFM and NLFM waveforms) is defined, and the normalized signal echo matrix after pulse compression of the echo of the k-th target is the pulse compression waveform transfer function S. MF,k The normalized signal echo matrix obtained after detecting the k-th target using a rectangular pulse waveform is defined as the simple pulse waveform transfer function S. SP,k Therefore, a waveform transmission matrix is ​​established.

[0092] Step S102: Construct a cost function based on the waveform transmission matrix.

[0093] Based on this, the cost function constructed according to the waveform transfer matrix includes:

[0094] Construct a K×P dimensional weighted matrix based on the waveform transmission matrix:

[0095]

[0096] Among them, w opt Let w represent the weighted matrix. opt,p =[w opt,p (1)...w opt,p (K)] T Let represent the optimization coefficient of the p-th range library, where p represents the p-th range library, P represents the total number of range libraries in the radar echo range direction, k represents the normalized echo amplitude of the k-th target acting on the range library, and K represents the normalized echo amplitude of the target acting on the range library.

[0097] The signals for each distance library are weighted according to the following formula:

[0098] P est =A(w opt ⊙S MF )

[0099] Among them, P est Let A represent the estimated echo signal, and w represent the target characteristic matrix. opt Let S represent the weighted matrix. MFRepresents the pulse compression waveform transfer function;

[0100] Since the echoes of simple pulse waveforms from different distances do not superimpose, theoretically, the signal energy calculated using the echoes obtained from simple pulse waveforms is optimal. Therefore, the transfer function S of the simple pulse waveform can be utilized. SP Model the desired echo function.

[0101] Using the simple pulse waveform transfer function S SP The ideal echo signal is calculated using the following formula:

[0102]

[0103] Among them, P opt Let S represent the ideal echo signal, A represent the target characteristic matrix, and S represent the target characteristic matrix. SP A represents the transfer function of a simple pulse waveform. k This represents the pointer element for the k-th target characteristic, where k represents the normalized echo amplitude of the k-th target acting on the range library, and K represents the normalized echo amplitude of the target acting on the range library. A P This represents the P-th target characteristic pointer element, where P represents the total number of range data in the radar echo range direction;

[0104] The echo function and the weighting matrix satisfy the following formula relationship:

[0105]

[0106] in, This means that p(w) can be made opt The smallest weighted matrix, p(w) opt P represents the signal function to be optimized. est P represents the estimated echo signal. opt This represents the ideal echo signal.

[0107] Step S103: Calculate the optimal extraction matrix based on the cost function.

[0108] Based on this, calculating the optimal extraction matrix according to the cost function includes:

[0109] The optimal extraction matrix is ​​calculated using the following formula:

[0110] W opt =(I⊙B⊙BI) T ) -1 (I⊙BD)

[0111] Among them, W opt Let I represent the optimal extraction matrix, and let B = S. MF Denotes the pulse compression waveform transfer function, D = Ssp This represents the simple pulse waveform transfer function, where T represents the transpose calculation.

[0112] Specifically, the optimal extraction matrix is ​​calculated based on the relationship between the echo function and the weighting matrix.

[0113] Expanding the relationship between the echo function and the weighting matrix, we can obtain:

[0114]

[0115] in, This means that p(w) can be made opt The smallest weighted matrix, p(w) opt Let S denote the signal function to be optimized, A denote the target characteristic matrix, which is a matrix that does not change with time, and S... MF S represents the pulse compression waveform transfer function. SP This represents a simple pulse waveform transfer function.

[0116] Since A is a time-invariant matrix, and for distributed meteorological targets, we can assume P = K, thus introducing an identity matrix I, transforming the above expression into the following expression:

[0117]

[0118] Let S MF =B,S SP =D, then:

[0119]

[0120] Let S = w opt I⊙B, g = 2(w opt I⊙B)D=2SD

[0121] Where : denotes the matrix inner product.

[0122] Because B and w opt It is irrelevant; based on the property of differentiating the Hadamard product, we can conclude that:

[0123] dS=(dw opt )I⊙B

[0124] Based on the properties of matrix inner product and matrix differential, we can conclude that:

[0125]

[0126] Similarly, it can be concluded from this that:

[0127]

[0128] Because D and wopt It is irrelevant, therefore we can conclude that:

[0129]

[0130] Because D and w opt Irrelevant The reciprocal of is 0. Therefore:

[0131]

[0132] make The optimal extraction matrix can be obtained:

[0133] W opt =(I⊙B⊙BI) T ) -1 (I⊙BD)

[0134] Among them, W opt Let I represent the optimal extraction matrix, and let B = S. MF Denotes the pulse compression waveform transfer function, D = S sp This represents the simple pulse waveform transfer function, where T represents the transpose calculation.

[0135] Step S104: Calculate the fluctuation pass rate based on the optimal extraction matrix.

[0136] Based on this, the fluctuation pass rate is calculated according to the optimal extraction matrix, including:

[0137] Calculate the optimized echo signal data based on the optimal extraction matrix;

[0138] Calculate the base data based on the optimized echo signal data;

[0139] Calculate the data offset error based on the base data, and obtain the base data in response to the data offset error being less than a preset offset threshold.

[0140] Calculate the standard deviation of the base data based on the base data;

[0141] The data fluctuation pass rate is calculated based on the standard deviation of the base data. In response to the data fluctuation pass rate being greater than the preset fluctuation pass rate, the data fluctuation pass rate is obtained.

[0142] Specifically, firstly, the optimized echo signal data is calculated, and then the base data E is calculated based on the optimized echo signal data. The base data E includes: reflectivity factor Z. est Average velocity V est and spectral width W est The data offset error is calculated based on the base data, and a preset offset threshold ε1 is set. When the offset error is less than this preset offset threshold, the data offset error satisfies the following conditions: The iteration should stop when required, and the base data E should be obtained. Specifically, the reflectivity factor distribution diagram when the reflectivity gradient is 15 dBZ / kms is shown in the figure below. Figure 2A The diagram shows the reflectivity factor distribution when the reflectivity gradient is 25 dBZ / km. Figure 2B The diagram shows the reflectivity factor distribution when the reflectivity gradient is 40 dBZ / km. Figure 2C As shown.

[0143] Calculate the standard deviation of the baseline data. Define the iteration requirements for the data volatility compliance rate (AR). When the true value is unavailable, the moving average can be used as the true value to calculate the AR. Since this data reflects the degree of data volatility, it can be called data volatility, and this indicator can be used to set the iteration conditions.

[0144] The data fluctuation compliance rate is calculated based on the standard deviation of the base data. In response to the data fluctuation compliance rate being greater than the preset fluctuation compliance rate ε2, i.e. AR>ε2, 0<ε2<100%, the data fluctuation compliance rate is obtained.

[0145] Based on this, the optimized echo signal data calculated according to the optimal extraction matrix includes:

[0146] The optimized echo signal data is calculated using the following formula:

[0147] S est =A(W opt ⊙S MF )

[0148] Among them, S est This represents the optimized echo signal data, where A represents the target characteristic matrix, and W represents the target characteristic matrix. opt S represents the optimal extraction matrix. MF represents the pulse compression waveform transfer function, and ⊙ represents the Hadamard product.

[0149] Specifically, for initializing the target characteristic matrix, assume the actual received echo signal data is as follows:

[0150] S r =AS MF

[0151] Among them, S r This represents the actual received echo signal data, where A represents the target characteristic matrix, and S represents the target characteristic matrix. MF This represents the pulse compression waveform transfer function.

[0152] Then we have:

[0153]

[0154] The optimized echo signal data is as follows:

[0155] S est =A(W opt ⊙S MF )

[0156] Among them, S est This represents the optimized echo signal data, where A represents the target characteristic matrix, and W represents the target characteristic matrix. opt S represents the optimal extraction matrix. MF represents the pulse compression waveform transfer function, and ⊙ represents the Hadamard product.

[0157] Based on this, the base data calculated according to the optimized echo signal data includes:

[0158] The optimized echo signal data is subjected to spectral moment estimation to obtain basic data; wherein, the basic data includes: reflectivity factor, average velocity and spectral width.

[0159] Based on this, the data offset error calculated according to the base data includes:

[0160] Calculate the data offset error using the following formula:

[0161]

[0162] Among them, RMSE E This represents the data offset error, p represents the p-th range database, P represents the total number of range databases in the radar echo range direction, and E est (p) represents the estimated baseline data, E gt (p) represents the ideal base data.

[0163] Based on this, the standard deviation of the base data is calculated using the base data, including:

[0164] Calculate the standard deviation of the base data using the following formula:

[0165]

[0166]

[0167] Among them, SD p E represents the standard deviation of the base data. p+j-n This represents the (p+jn)th base data. Let P represent the moving average calculated using 2n+1 range libraries, where P represents the total number of range libraries in the radar echo range direction, p represents the p-th range library, n represents the n-th range library, and j represents the j-th range library.

[0168] Based on this, the data fluctuation compliance rate is calculated according to the standard deviation of the base data, including:

[0169] The data fluctuation pass rate is calculated using the following formula:

[0170]

[0171] Where AR represents the data fluctuation pass rate, sum() represents cumulative calculation, P represents the total number of range databases in the radar echo range direction, p represents the p-th range database, and SD p This represents the standard deviation of the base data.

[0172] Step S105: Adaptive suppression of meteorological target sidelobes is achieved based on the optimal extraction matrix and the fluctuation qualification rate.

[0173] Specifically, by using the optimal extraction matrix to make the optimized signal energy close to the main lobe energy of the actual target in the range database, the side lobes of meteorological targets with large gradient reflectivity can be suppressed.

[0174] The adaptive sidelobe suppression method for meteorological targets based on optimal component extraction of this invention constructs a cost function by modeling the expected energy output function of each range database, and obtains the optimal extraction matrix. This makes the optimized signal energy close to the main lobe energy of the actual target in the range database, thus achieving the suppression of sidelobes of meteorological targets with large reflectivity gradients. When the reflectivity gradient is large, this method can significantly improve the basic data estimation error in weak target areas; when the reflectivity gradient is small, this method can improve the basic data estimation accuracy and effectively improve data quality. The method in this invention processes the pulse-compressed echo signal and has no requirements on the radar transmitted waveform. It only needs to model the transmission function of the pulse-compressed signal based on the transmitted waveform. In actual radar systems, this can be achieved by adding a step after pulse compression. Therefore, this invention has a good effect on sidelobe suppression in meteorological radar scenarios with large reflectivity gradients.

[0175] The technical solutions of this invention will be clearly and completely described below with reference to the embodiments thereof. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0176] Example 1

[0177] In this example, the effectiveness of the invention is verified by simulating one-dimensional range-oriented distributed target echoes with different reflectivity gradient distributions. The one-dimensional range-oriented transmitted waveform is constructed based on a nonlinear frequency modulation (NLFM) waveform and subjected to mismatch filtering to obtain a pulse-compressed signal, which serves as the input to the invention. The simulation parameters are shown in the table below, where the signal-to-noise ratio of the input signal is 60 dB.

[0178] parameter numerical values parameter numerical values Frequency (GHz) 13.6 Pulse width (µs) 64 Bandwidth (MHz) 4 Sampling frequency (MHz) 16 Number of pulses accumulated 64 PRF(Hz) 1860 / 1395 Distance to the library 350 Reflectivity gradient (dBZ / km) 15 / 25 / 40 Target spectrum (m / s) [1,5] random integers Target velocity (m / s) Integers in the range [-7, 7]

[0179] Table 1 Simulation Parameters

[0180] Step S301: Construct the waveform transmission matrix:

[0181]

[0182]

[0183] In this matrix, each row represents a range sampling sequence, and K = P = 350 represents the number of radial targets or range data detected by the radar.

[0184] Step S302: Construct a K×P dimensional weighted matrix based on the waveform transmission matrix:

[0185]

[0186] Among them, w opt Let w represent the weighted matrix. opt,p =[w opt,p (1)...w opt,p (350)] T Let represent the optimization coefficient of the p-th range library, where p represents the p-th range library, P represents the total number of range libraries in the radar echo range direction, and k represents the normalized echo amplitude of the k-th target acting on that range library.

[0187] Step 303: Weight the signal of each distance library according to the following formula:

[0188] P est =A(w opt ⊙S MF )

[0189] Among them, P est Let A represent the estimated echo signal, and w represent the target characteristic matrix. opt Let S represent the weighted matrix. MF This represents the pulse compression waveform transfer function.

[0190] Step 304: Utilize the simple pulse waveform transfer function S SP The ideal echo signal is calculated using the following formula:

[0191]

[0192] Among them, P opt Let S represent the ideal echo signal, A represent the target characteristic matrix, and S represent the target characteristic matrix. SP A represents the transfer function of a simple pulse waveform. kThis represents the pointer element for the k-th target characteristic, where k represents the normalized echo amplitude of the k-th target acting on the range library, and K represents the normalized echo amplitude of the target acting on the range library. A P This represents the pointer element of the Pth target characteristic.

[0193] Step 304: The echo function and the weighting matrix satisfy the following formula relationship:

[0194]

[0195] in, This means that p(w) can be made opt The smallest weighted matrix, p(w) opt P represents the signal function to be optimized. est P represents the estimated echo signal. opt This represents the ideal echo signal.

[0196] Step 305: Calculate the optimal extraction matrix according to the following formula:

[0197] W opt =(I⊙B⊙BI) T ) -1 (I⊙BD)

[0198] Among them, W opt Let I represent the optimal extraction matrix, and let B = S. MF Denotes the pulse compression waveform transfer function, D = S sp This represents the simple pulse waveform transfer function, where T represents the transpose calculation.

[0199] Step 306: Calculate the optimized echo signal data using the following formula.

[0200] S est =A(W opt ⊙S MF )

[0201] Among them, S est This represents the optimized echo signal data, where A represents the target characteristic matrix, and W represents the target characteristic matrix. opt S represents the optimal extraction matrix. MF represents the pulse compression waveform transfer function, and ⊙ represents the Hadamard product.

[0202] Step 307: Perform spectral moment estimation on the optimized echo signal data to obtain basic data; wherein, the basic data includes: reflectivity factor, average velocity and spectral width.

[0203] Step 308: Calculate the data offset error using the following formula:

[0204]

[0205] Among them, RMSE E This represents the data offset error, p represents the p-th range database, P represents the total number of range databases in the radar echo range direction, and E est (p) represents the estimated baseline data, E gt (p) represents the ideal base data.

[0206] Step S309: The preset offset threshold is 0.5. In response to the data offset error being less than 0.5, the base data is obtained, i.e.:

[0207]

[0208] Step S310: Calculate the standard deviation of the base data according to the following formula:

[0209]

[0210]

[0211] Among them, SD p E represents the standard deviation of the base data. p+j-3 This represents the (p+j-3)th base data. Let j represent the moving average calculated using 7 distance libraries, where j represents the j-th distance library.

[0212] The standard deviation SD of the p-th distance from the basic data in the library p The standard deviation can be calculated using data from the three distance databases before and after the given distance database, and the three distance databases before and after the given distance database do not participate in the calculation of the standard deviation.

[0213] Step 311: Calculate the data fluctuation pass rate according to the following formula:

[0214]

[0215] Where AR represents the data fluctuation pass rate, sum() represents cumulative calculation, P represents the total number of range databases in the radar echo range direction, p represents the p-th range database, and SD p This represents the standard deviation of the base data.

[0216] Step 312: The preset fluctuation pass rate is 80%. In response to the data fluctuation pass rate being greater than 80%, the data fluctuation pass rate is obtained, that is:

[0217]

[0218] Step 313: Implement adaptive suppression of meteorological target sidelobes based on the optimal extraction matrix and the fluctuation qualification rate.

[0219] Figure 3AThe above are the estimation results of the reflectivity factor of a one-dimensional distributed meteorological target echo before and after optimization when the reflectivity gradient is 15 dBZ / kms. Figure 3B The above are the estimation results of the reflectivity factor of a one-dimensional distributed meteorological target echo before and after optimization when the reflectivity gradient is 25 dBZ / km. Figure 3C The present invention provides the estimation results of reflectivity factor for one-dimensional distributed meteorological target echoes before and after optimization when the reflectivity gradient is 25 dBZ / km. Figure 4A 4B shows the velocity estimation results of the one-dimensional distributed meteorological target echo before and after optimization when the reflectivity gradient is 15 dBZ / kms. 4C shows the velocity estimation results of the one-dimensional distributed meteorological target echo before and after optimization when the reflectivity gradient is 25 dBZ / kms. Figure 5A The present invention provides the spectral width estimation results for the echo of a one-dimensional distributed meteorological target before and after optimization when the reflectivity gradient is 15 dBZ / kms. Figure 5B The present invention provides the spectral width estimation results for the echo of a one-dimensional distributed meteorological target before and after optimization when the reflectivity gradient is 25 dBZ / km. Figure 5C The figures show the spectral width estimation results for one-dimensional distributed meteorological target echoes before and after optimization when the reflectivity gradient is 40 dBZ / km. It can be seen that the baseline data calculated using this invention is closer to the true value and can improve the estimation performance for weak target baseline data. Figure 6A This invention presents the change in RMSE of the baseline data before and after optimization of one-dimensional distributed meteorological target echoes with reflectivity gradient when the reflectivity gradient is 15 dBZ / kms. Figure 6B This invention presents the change in RMSE of the baseline data before and after optimization of one-dimensional distributed meteorological target echoes with reflectivity gradient when the reflectivity gradient is 25 dBZ / km. Figure 6C The RMSE of the baseline data before and after optimization of the one-dimensional distributed meteorological target echo is given by the reflectivity gradient when the reflectivity gradient is 40 dBZ / km.

[0220] Figure 7A This invention presents the change in RMSE of the baseline data before and after optimization of one-dimensional distributed meteorological target echoes with reflectivity gradient when the reflectivity gradient is 15 dBZ / kms. Figure 7B This invention presents the change in RMSE of the baseline data before and after optimization of one-dimensional distributed meteorological target echoes with reflectivity gradient when the reflectivity gradient is 25 dBZ / km. Figure 7C This invention presents the change in RMSE (Reflectance Scale) of the baseline data before and after optimization of one-dimensional distributed meteorological target echoes with reflectance gradient, when the reflectance gradient is 40 dBZ / km. RMSE reflects the degree to which the data deviates from the true value, while data volatility reflects the degree of data fluctuation.

[0221] When the reflectivity factor gradient is large (≥30 dBZ / km), the RMSE of the unoptimized reflectivity factor and velocity is large, and the deviation from the true value is serious. At this time, the data is invalid, and the calculated AR is not a reference. However, the method based on optimal component extraction can greatly reduce the RMSE of the data, especially the RMSE of the reflectivity factor, and turn invalid data into valid data. When the reflectivity factor gradient is small (<30 dBZ / km), the RMSE of the data is small, and the data is valid. At this time, the focus is on data volatility. Compared with the unoptimized data, the qualified rate of data volatility is greatly increased after optimization, which effectively improves the data quality.

[0222] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the invention (including the claims) is limited to these examples; within the framework of the invention, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of the invention as described above, which are not provided in the details for the sake of brevity.

[0223] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0224] Additionally, to simplify the description and discussion, and to avoid obscuring the embodiments of the invention, the well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. Furthermore, the apparatus may be shown in block diagram form to avoid obscuring the embodiments of the invention, and this also takes into account the fact that the details of implementation of these block diagram apparatuses are highly dependent on the platform on which the embodiments of the invention will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuits) have been set forth to describe exemplary embodiments of the invention, it will be apparent to those skilled in the art that the embodiments of the invention may be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0225] Although the invention has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0226] The embodiments of this invention are intended to cover all such substitutions, modifications, and variations falling within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the embodiments of this invention should be included within the protection scope of this invention.

Claims

1. A method for weather target sidelobe adaptive suppression based on optimal component extraction, characterized in that, The method comprises the following steps: constructing a waveform transmission matrix; ， ， wherein, represents a pulse compressed waveform transfer function, represents a simple pulse waveform transfer function, represents a first distance bin, represents a total number of distance bins in a radar return range profile, represents a first target acting on the distance bin; constructing a cost function according to the waveform transmission matrix, and the echo signal and the weighting matrix satisfying the following formula relationship: ， ， wherein, represents an ability to enable a minimum weighted matrix, represents a signal function to be optimized, represents an estimated echo signal, represents an ideal echo signal; calculating an optimal extraction matrix according to the cost function; calculating optimized echo signal data according to the optimal extraction matrix, calculating base data according to the optimized echo signal data, and calculating a fluctuation pass rate according to the base data; implementing sidelobe adaptive suppression of a meteorological target according to the optimal extraction matrix and the fluctuation pass rate.

2. The method of claim 1, wherein the method is based on optimal component extraction for weather target sidelobe adaptive suppression. The step of constructing a cost function according to the waveform transmission matrix comprises the following steps: Construction The weighting matrix for the Viterbi algorithm: ， wherein, represents a weighting matrix, represents the optimization coefficient of the distance bank, represents the distance bank, represents the total number of distance banks of the radar echo range profile, represents the normalized echo amplitude of the target acting on the distance bank, weighting the signal of each distance bin according to the following formula: ； wherein represents an estimated echo signal, represents a target property matrix, represents a weighting matrix, represents a pulse-compressed waveform transfer function; Utilizing simple pulse waveform transfer functions The ideal echo signal is calculated according to the following equation: ， ， ， in, Represents an ideal echo signal. Represents the target characteristic matrix. Represents the transfer function of a simple pulse waveform. Indicates the first One target characteristic pointer element, Indicates the first The target's effect on the normalized echo amplitude of the distance library. This represents the normalized echo amplitude of the target acting at that distance. Indicates the first One target characteristic pointer element, This indicates the total number of range data in the radar echo range direction.

3. The method of claim 2, wherein the method is based on optimal component extraction for weather target sidelobe adaptive suppression. The step of calculating an optimal extraction matrix according to the cost function comprises the following steps: calculating the optimal extraction matrix according to the following formula: ， wherein, denotes the optimal extraction matrix, denotes the identity matrix, denotes the impulse compressed waveform transfer function, denotes the simple impulse waveform transfer function, denotes the transpose computation.

4. The method of claim 1, wherein the method is based on optimal component extraction for weather target sidelobe adaptive suppression. The step of calculating a fluctuation pass rate according to the base data comprises the following steps: calculating a data offset error according to the base data, and obtaining the base data in response to the data offset error being less than a preset offset threshold; calculating a standard deviation of the base data according to the base data; calculating a data fluctuation pass rate according to the standard deviation of the base data, and obtaining the data fluctuation pass rate in response to the data fluctuation pass rate being greater than a preset fluctuation pass rate.

5. The method of claim 4, wherein the method is based on optimal component extraction for weather target sidelobe adaptive suppression. The step of calculating optimized echo signal data according to the optimal extraction matrix comprises the following steps: calculating the optimized echo signal data according to the following formula: , wherein, represents the optimized echo signal data, represents the target characteristic matrix, represents the optimal extraction matrix, represents the pulse-compressed waveform transmission function, represents the Hadamard product.

6. The method of claim 4, wherein the method is based on optimal component extraction for weather target sidelobe adaptive suppression. The step of calculating base data according to the optimized echo signal data comprises the following steps: performing spectral moment estimation on the optimized echo signal data to obtain the base data; wherein the base data comprises reflectivity factors, average speeds and spectral widths.

7. The method of claim 4, wherein the method is based on optimal component extraction for weather target sidelobe adaptive suppression. The step of calculating a data offset error according to the base data comprises the following steps: calculating the data offset error according to the following formula: ， wherein, represents a data offset error, represents a first distance bank, represents a total number of distance banks of a radar echo range profile, represents an estimated base data, represents an ideal base data.

8. The method of claim 4, wherein the method is based on optimal component extraction for weather target sidelobe adaptive suppression. The step of calculating a standard deviation of the base data according to the base data comprises the following steps: calculating the standard deviation of the base data according to the following formula: ， ， wherein, denotes the standard deviation of the base data, denotes the base data, denotes the sliding average calculated using distance bins, denotes the total number of distance bins of the radar echo range profile, denotes the distance bin, denotes the distance bin, denotes the distance bin.

9. The method of claim 4, wherein the method is based on optimal component extraction for weather target sidelobe adaptive suppression. The step of calculating a data fluctuation pass rate according to the standard deviation of the base data comprises the following steps: calculating the data fluctuation pass rate according to the following formula: ， wherein, represents a data fluctuation qualification rate, represents an accumulated calculation, represents a total number of distance bins in a radar echo distance direction, represents the first distance bin, represents a standard deviation of base data.

Images