CN-116467670-B - MP and Root-MUSIC algorithm-based radar band fusion expansion method

CN116467670BCN 116467670 BCN116467670 BCN 116467670BCN-116467670-B

Abstract

The invention belongs to the technical field of radar target identification, and particularly relates to a radar frequency band fusion expansion method based on MP and Root-MUSIC algorithm. The method comprises the steps of 1, setting two adjacent radars, wherein the observation angles are the same, and the working frequencies are different, so that a full-pole GTD model of low-frequency and high-frequency electromagnetic echoes is built for a static target consisting of scattering centers, 2, utilizing echo data and conjugate data thereof to estimate incoherent parameters, namely a linear phase and a fixed phase, and 3, utilizing an improved algorithm to process the coherent data to estimate poles and amplitudes of a broadband full-polarization model, and then using original data of a low-frequency sub-band and a high-frequency sub-band to calculate to obtain a full-frequency fused signal. The multiband fusion method provided by the invention constructs a new covariance matrix by utilizing the original data and the conjugate form thereof, so that the modified algorithm has better parameter estimation performance under the same signal-to-noise ratio.

Inventors

XU JIAHUA
ZHANG XIAOKUAN
ZHENG SHUYU
Ma Qiankuo
WANG YANG
ZONG BINFENG
ZHANG JINGWEI

Assignees

中国人民解放军空军工程大学

Dates

Publication Date: 20260508
Application Date: 20230323

Claims (4)

1. The incoherent parameter estimation and multiband fusion method based on MP and Root-MUSIC algorithm is characterized by comprising the following steps: step 1, setting two adjacent radars, wherein the observation angles are the same and the working frequencies are different, and establishing a full-pole GTD model of low-frequency and high-frequency electromagnetic echoes for a stationary target consisting of a scattering center; Step 2, estimating incoherent parameters, namely a linear phase and a fixed phase, by utilizing echo data and conjugate data thereof on the basis of the step 1, wherein the step 2 specifically comprises the following steps: step 2.1, firstly, assume that the low-frequency subband and the high-frequency subband signal echo are respectively: (7) (8); Is provided with Then And Is expressed as: (9) (10); Wherein: representing the variance of the noise and, Representation of Is used for the matrix of units of (a), For the matrix eigenvalue of R Y1Y1 , Matrix eigenvalues for R Y2Y2 ; Step 2.2, constructing an echo data matrix of the formula (11-12): (11) (12) Wherein: is a matrix beam parameter that is a function of the beam parameters, Is that The expression of which is as follows: (13) Combining the cross covariance and the covariance of different subbands to construct a matrix as follows: (14) (15); Step 2.3 in the construction And Then, SVD processing is carried out on the complex phase-locked loop, the complex phase-locked loop and the complex phase-locked loop are repeated in the same steps as the traditional MP algorithm and the traditional Root-MUSIC algorithm, so that linear phase and fixed phase can be obtained, and sub-band confusion of different frequency bands is completed, wherein the step 2.3 specifically comprises: Step 2.3.1 pair matrix And And (3) performing eigenvalue decomposition: (16) (17); Based on the MDL method, the signal is decomposed into a signal subspace a 11 of the low frequency subbands and a noise subspace a 22 of the high frequency subbands: (18) (19) Wherein: And Representing the signal subspace and the noise subspace of the low frequency subbands respectively, And A signal subspace and a noise subspace representing high frequency subbands, respectively; Step 2.3.2 calculation according to A 11 and A 22 of step 2.3.1 , : (20) (21); Step 2.3.3 to carry out Or (b) Representative of Or (b) Is arranged in the first column of the (c) column, , The re-expression is as follows: (22) (23); step 2.3.4 calculation , Selecting the root closest to the unit circle as the pole of the full polarization GTD model, and finally calculating the amplitude by using the least squares method And ; Step 2.3.5, calculating linear phase and fixed phase, and calculating echo data of two sub-bands; and 3, based on the linear phase and the fixed phase obtained in the step 2, processing coherent data by utilizing an improved algorithm, estimating poles and amplitudes of a broadband full-polarization model, and then calculating to obtain a full-band fusion signal by using the original data of the low-frequency sub-band and the high-frequency sub-band.
2. The method for estimating and performing multi-band fusion based on the incoherent parameters of the MP and Root-MUSIC algorithm according to claim 1, wherein the step 1 comprises: Step 1.1, two adjacent radars are arranged to work at the same observation angle and different in working frequency, so that for a stationary target consisting of a scattering center, electromagnetic echoes of two sub-bands can be expressed as: (1) (2) Wherein S 1 is a low-frequency wavelet echo, S 2 is a high-frequency wavelet echo, And , And Representing the initial operating frequencies of the two sub-bands of low and high frequencies respectively, And Respectively the m-th sum Frequency points, M 1 and M 2 represent total frequency points of the two sub-bands of low frequency and high frequency, respectively, r i represents the position of the scattering center, alpha i represents the scattering type, A i represents the scattering intensity, Representing a different typical scattering structure, Is the speed of light, beta represents a fixed phase, eta represents a linear phase, beta and eta represent an incoherent relationship between the low frequency sub-band and the high frequency sub-band, And And respectively representing Gaussian white noise of a low-frequency sub-band and a high-frequency sub-band, wherein I is the number of scattering centers, j is an imaginary unit, Step by step for frequency; step 1.2, when the working bandwidth is smaller than 10% of the center frequency, the following approximate values are realized: (3) (4); step 1.3, substituting the formulas (3) - (4) obtained in the step 2 into the formulas (1) - (2), and obtaining the full-pole GTD model of the low-frequency and high-frequency electromagnetic echoes, wherein the full-pole GTD model comprises the following steps: (5) (6)。
3. the method for estimating and performing multi-band fusion based on the incoherent parameters of the MP and Root-MUSIC algorithm according to claim 2, wherein the step 2.3.5 is specifically as follows: The amplitude coefficients estimated by the MP algorithm are: (24); the phase angle is: (25); Based on expressions (24) and (25), it can be derived that: (26); The fixed phase is calculated as: (27) Wherein: , ; Calculating incoherent subband phase in high frequency region : (28)。
4. The method for estimating and performing multi-band fusion based on the incoherent parameters of the MP and Root-MUSIC algorithm as set forth in claim 3, wherein the step 3 specifically includes: step 3.1, calculating a high-frequency subband based on the linear phase and the fixed phase which are estimated in the step 2, wherein the subband is as follows: (29); And 3.2, processing the coherent data in the step 2 by using an improved algorithm to estimate poles and amplitudes of a broadband full polarization model, wherein the expression is as follows: (30) Wherein: And Representing the estimated amplitude and pole of the signal, Representing the number of frequencies of the wideband; Thus, the sub-bands Can be expressed as: (31); Step 3.3, to reduce the estimation error, the fusion signal of the full frequency band is calculated by using the original data of the low or high frequency sub-band, and the expression is: (32)。

Description

MP and Root-MUSIC algorithm-based radar band fusion expansion method Technical Field The invention belongs to the technical field of radar target identification, and particularly relates to a radar frequency band fusion expansion method based on MP and Root-MUSIC algorithm. Background High resolution plays a critical role in non-cooperative radar target identification, space target surveillance, ballistic missile defense, and other applications. Radar resolution may be improved by designing and manufacturing ultra wideband (UVB) operational radars. This approach is then difficult to implement due to technical and cost constraints. To solve such a problem, a conventional method decomposes a wideband signal into a plurality of narrowband signals in a frequency domain, and then directly performs narrowband processing on each subband. Since the subband signals are measured from operating radars of different frequencies, they are mutually incoherent. Cuomo et al in the lincoln laboratory analyzed multi-band radars to obtain incoherent results between different radars that ultimately introduce linear and fixed phase shifts on the signal. Therefore, these parameters should be estimated and mutually compensated before subband fusion in order to improve accuracy. Cuomo et al propose to estimate incoherent parameters using Root-MUSIC algorithm and least squares method. However, such algorithms have the disadvantage of being too complex to operate. A linear phase shift method has also been proposed, which uses the correlation of the one-dimensional image of the target to obtain a fixed phase shift with a cost function constraint. Compared with the former, the algorithm has higher estimation accuracy, but the algorithm estimation performance is unstable because the algorithm is limited by the signal sampling number. Non-coherent estimation methods have also been proposed by scholars based on Root-MUSIC algorithm and ESPRIT (ESTIMATED SIGNAL PARAMETERS VIA Rotational Invariance Technique) algorithm. However, the method has the problem of poor effect of estimating parameters under the condition of low signal-to-noise ratio. In addition, the full phase fast fourier transform (apFFT) spectral analysis method is also used to estimate incoherent parameters, which avoids the use of a priori knowledge of the scattering centers, but suffers from the limitation of receiving frequency samples. Increasing radar resolution in the last 10 years has been the focus and difficulty of research in the radar field, and multi-band fusion algorithms used in this field have been continuously updated. After the parameter compensation is carried out on the incoherent subband signals, the original incoherent signals are changed into coherent signals, and further model fitting is needed. The model fitting directly determines the precision of multi-band fusion, and is an important step of multi-band fusion. Researchers in the lincoln laboratory have proposed an all-pole model that fits electromagnetic scattering data. At small relative bandwidths of the operating radar, the all-pole model approximates a geometric diffraction theory (Geometrical Theory ofDiffraction, GTD) model. The GTD model can provide a relatively high-precision electromagnetic scattering description of radar targets, and is widely used for radar signal modeling. The parameter estimation algorithm of the GTD model is relatively mature, and typical estimation algorithms such as a Sparse Bayesian Learning (SBL) algorithm, an ESPRIT algorithm, a MUSIC algorithm and the like can perform effective parameter estimation on the GTD model. Disclosure of Invention Aiming at the problem that parameters cannot be estimated accurately in the traditional spatial algorithm such as matrix beams (MP) and Root-MUSIC due to the similarity of noise characteristic values and signal characteristic values under the condition of low signal-to-noise ratio, the invention provides an incoherent parameter estimation and multiband fusion method based on the MP and the Root-MUSIC algorithm, and the proposed algorithm constructs a new covariance matrix by utilizing original data and conjugate forms thereof, so that the modified algorithm has better parameter estimation performance under the condition of the same signal-to-noise ratio. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: A radar frequency band fusion expansion method based on MP and Root-MUSIC algorithm comprises the following steps: step 1, setting two adjacent radars, wherein the observation angles are the same and the working frequencies are different, and establishing a full-pole GTD model of low-frequency and high-frequency electromagnetic echoes for a stationary target consisting of a scattering center; Step 2, estimating incoherent parameters, namely a linear phase and a fixed phase, by utilizing echo data and conjugate data thereof on the basis of the step 1; and 3, based on the linear ph