Waveform-filter joint design method and system considering correlation constraint

By constructing an onboard echo signal model in an onboard radar system and iteratively optimizing the waveform and filter, combining the correlation constraints between peak level and integral level, the problem of weak and slow target detection and calculation time-consuming in onboard radar is solved. The problem of this method realizes more efficient signal-to-interference-noise ratio optimization and correlation constraint application.

CN118915000APending Publication Date: 2024-11-08XIDIAN UNIV

Patent Information

Application Number
CN202411225146.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-03
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The prior art is difficult to effectively detect weak and slow targets in airborne radar systems, and the calculation of the MATLAB CVX toolbox takes a long time, which fails to effectively solve the correlation migration dimension problem and signal-to-noise ratio optimization problem.

Method used

Using a waveform-filter joint design method that considers correlation constraints, by constructing an onboard echo signal model, maximizing the signal-to-interference noise ratio, using a cyclic algorithm to iteratively optimize the waveform and filter, combining peak level and integral level Relevance constraints, dimensionality reduction constraint models to improve efficiency.

Benefits of technology

In the airborne radar system, it effectively solves the problem of weak and slow target detection caused by high side lobes after echo mismatch filtering, expands the application range of correlation constraints, and improves the efficiency of waveform-filter joint design, and reduces the The amount of data calculation during the iteration process is calculated.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118915000A_ABST
    Figure CN118915000A_ABST
Patent Text Reader

Abstract

The invention discloses a waveform-filter joint design method and system considering correlation constraints, relates to the field of radars, and is used for solving the problem of difficulty in detecting a weak, small and low-speed target caused by high sidelobe after echo mismatched filtering. The method comprises the following steps: constructing an airborne echo signal model based on configured system parameters; taking maximization of the signal to interference plus noise ratio of the airborne echo signal model as a target, constructing a problem model according to correlation constraint and constant modulus constraint, and adopting a loop algorithm to iteratively optimize the waveform and the filter respectively until the signal to interference plus noise ratio of the airborne echo signal model is converged; and finally outputting the optimized waveform and filter. According to the method, the problem of difficulty in weak and small slow target detection caused by high sidelobe after echo mismatch filtering is solved, and meanwhile, the application range of correlation constraint is expanded.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of radar, and in particular to a waveform-filter joint design method and system considering correlation constraints. Background Art

[0002] There are many studies on the correlation of radar emission waveforms. The existing technologies can be divided into three categories:

[0003] The first category is based on matched filtering and studies the low autocorrelation sidelobes of waveforms or low autocorrelation sidelobes and low cross-correlation sidelobes of waveform sets. Since the correlation performance can be expressed as integral correlation and peak correlation, these studies are also divided into these two optimization forms.

[0004] The second type is to use the degree of freedom of non-matched filters to further reduce the autocorrelation sidelobes (and cross-correlation sidelobes) when taking the filter into consideration. At the same time, since the mismatch between the filter and the waveform will affect the main lobe, causing a certain degree of main lobe reduction, this type of optimization problem is carried out under the premise of main lobe loss.

[0005] The third category is the process of joint optimization of waveform and filter, which not only takes into account the improvement of signal-to-noise ratio, but also realizes the constraint of correlation, forming low-distance sidelobes. It is the most complex scenario among the three categories.

[0006] The third type of research has better optimization effects, so the third type of research has higher practical value. For example, Sadjad Imani et al. introduced in the article "Colocated MIMO Radar SINR MaximizationUnder ISL and PSL Constraints" that under the MIMO system, the joint design of waveform filters was considered to optimize the signal-to-noise ratio under the correlation constraint, where the correlation considers the integral level and the peak level respectively. At the algorithm level, this method uses the matlab cvx toolbox to solve the quadratic constrained quadratic programming problem. The optimized output waveform and filter group can achieve an improvement in the signal-to-noise ratio when the correlation meets the requirements.

[0007] However, the above literature considers the correlation work under the MIMO system, and does not conduct further research on airborne radar. In addition, it is found that there are dimensional difficulties in the correlation migration under the airborne system, which makes it difficult to detect weak and slow targets. Secondly, the work is solved using the MATLAB CVX toolbox, which is time-consuming. Summary of the invention

[0008] The object of the present invention is to provide a waveform-filter joint design method taking into account correlation constraints to address all or part of the above-mentioned problems, so as to solve the problem of difficulty in detecting weak and slow targets caused by high side lobes after echo mismatch filtering.

[0009] The technical solution adopted by the present invention is as follows:

[0010] A waveform-filter joint design method considering correlation constraints, comprising:

[0011] S1. Building an airborne echo signal model based on the configured system parameters;

[0012] S2. With the goal of maximizing the signal-to-interference-to-noise ratio of the airborne echo signal model, a problem model is constructed according to configured constraints, wherein the problem model includes a filter optimization problem model and a waveform optimization problem model; the constraints include a correlation constraint and a constant modulus constraint;

[0013] S3. Based on the problem model, a cyclic algorithm is used to iteratively optimize the waveform and the filter respectively until the signal to interference noise ratio of the airborne echo signal model converges;

[0014] S4. Output optimized waveform and filter.

[0015] Furthermore, the airborne echo signal model is constructed as follows:

[0016] y=y t +y c +y j +n,

[0017] Among them, y represents the airborne echo signal model, y t Represents the received echo vector, y c represents the clutter vector, y j represents the interference vector, and n represents the noise vector.

[0018] Furthermore, the correlation constraint includes a correlation constraint expressed in terms of an integrated sidelobe level and a correlation constraint expressed in terms of a peak sidelobe level.

[0019] Furthermore, in the problem model, the waveform optimization problem model is constructed as follows:

[0020]

[0021] in,

[0022] x=vec(S * ), S represents the waveform in matrix form, vec() represents vectorization processing, and x represents the waveform in vector form;

[0023] W is L×MN R dimensional matrix, N R represents the number of receiving antennas, L represents the number of snapshots, and M represents the total number of pulse received echoes; w = vec(W), where w represents the filter weight;

[0024]

[0025] t represents the echo number, p represents the clutter number, θ represents the azimuth, A t is the echo amplitude, Indicates dimension N T ×N T The identity matrix, N T Indicates the number of transmitting antennas;

[0026]

[0027] f t 、f p represents the normalized Doppler frequency of the echo and clutter, a(θ t ), a(θ p ) represents the launch steering vector, b(θ t ), b(θ p ) represents the receive steering vector, represents the Kronecker product operation;

[0028] [-Q,Q] represents the range of the ranging unit, N c Indicates the number of azimuth divisions of a single ranging unit; A means to seek expectation. (q,p) is the backscatter amplitude; L×L is the shift matrix J of the qth ranging unit q Dimensions;

[0029] {x(t)] = x, t∈[0,L-1], E represents the emission energy,

[0030] w k represents the filter weight matrix of the k-th pulse received echo, I L is the identity matrix of dimension L×L, ∈ represents the upper limit of correlation;

[0031]

[0032] ISL kIt represents the integrated sidelobe level of the k-th pulse received echo, m and n represent the serial numbers of the transmitting antennas, τ is the time delay, and 0≤ω≤1 represents the sidelobe average rate.

[0033] Further, in the iterative optimization of the waveform and the filter respectively using the problem model, the iterative optimization of the waveform includes:

[0034] The main function of the waveform optimization problem model is converted into a second-order function using the Dinklebach transformation method;

[0035] The transformed main function is converted into a sub-problem form using the generalized Lagrangian form;

[0036] The sub-problems are iteratively solved under the constraints of the waveform optimization problem model to achieve iterative optimization of the waveform.

[0037] Furthermore, for each round of iteration, the waveform of the previous round of iteration is kept unchanged, the filter of the current round is optimized, and then the filter of the current round is kept unchanged and the waveform of the current round is optimized. The waveform of the first round of iteration adopts the initialization value.

[0038] Furthermore, the method of optimizing the filter in each iteration includes:

[0039] Based on the waveform S of the previous iteration, the problem model is converted to a filter optimization problem model;

[0040] The closed-form analytical expression of the filter optimization problem model is used to obtain the filter weights of the current round of optimization.

[0041] Furthermore, the iterative optimization of the waveform also includes:

[0042] Perform dimensionality reduction on correlation constraints.

[0043] The present invention also provides a waveform-filter joint design system considering correlation constraints, which comprises:

[0044] The parameter initialization module is configured to: receive system parameters configured for the radar system composed of a waveform and a filter;

[0045] The airborne echo signal model building module is configured to: build an airborne echo signal model based on configured system parameters;

[0046] The problem model building module is configured to: build a problem model according to configured constraints with the goal of maximizing the signal-to-interference-to-noise ratio of the airborne echo signal model, wherein the problem model includes a filter optimization problem model and a waveform optimization problem model; the constraints include a correlation constraint and a constant modulus constraint;

[0047] An optimization module, based on the problem model, adopts a loop algorithm to iteratively optimize the waveform and the filter respectively until the signal to interference noise ratio of the airborne echo signal model converges;

[0048] Output module, outputs optimized waveform and filter.

[0049] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0050] 1. In the process of joint optimization of waveform and filter in the design of airborne radar system, the present invention fully considers the correlation constraint and constant modulus constraint, and adopts two forms of peak level and integral level to express the correlation constraint, solves the adaptation problem caused by it, solves the problem of difficulty in detecting weak and slow targets caused by the high sidelobe after echo mismatch filtering, and expands the application scope of the correlation constraint.

[0051] 2. In the process of optimizing waveforms and filters, the present invention simplifies the problem model and reduces the dimension of the constraint model, thereby reducing the amount of data calculation required in the iterative process and improving the efficiency of the waveform-filter joint design. In addition, when considering the correlation constraint, the airborne radar STAP filter contains three dimensions of space-time-Doppler, while the waveform only contains two dimensions of space-time. The filter is divided according to Doppler to become a sub-filter group containing two dimensions of space-time, completing the matching with the waveform dimension, so that the cyclic algorithm can be smoothly used in the present invention for iteration. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The present invention will now be described by way of example with reference to the accompanying drawings, in which:

[0053] Figure 1 It is a flow chart of the waveform-filter joint design method considering correlation constraints.

[0054] Figure 2 This is the clutter signal distribution diagram of the airborne MIMO-STAP radar.

[0055] Figure 3 It is a graph showing the change of SINR performance with the number of iterations for different ∈ and ω parameters.

[0056] Figure 4 It is a graph showing the normalized correlation characteristics of the waveform.

[0057] Figure 5 2 is a performance test diagram of a waveform, where (a) represents the autocorrelation performance and (b) represents the cross-correlation performance. DETAILED DESCRIPTION

[0058] All features disclosed in this specification, or steps in all methods or processes disclosed, except mutually exclusive features and / or steps, can be combined in any manner.

[0059] Any feature disclosed in this specification (including any additional claims and abstract), unless otherwise stated, may be replaced by other equivalent or alternative features having similar purposes. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.

[0060] Example 1

[0061] This embodiment introduces a waveform-filter joint design method considering correlation constraints. The design concept of the method includes:

[0062] S1. Building an airborne echo signal model based on the configured system parameters;

[0063] S2. With the goal of maximizing the signal-to-interference-to-noise ratio of the airborne echo signal model, a problem model is constructed according to configured constraints, wherein the problem model includes a filter optimization problem model and a waveform optimization problem model; the constraints include a correlation constraint and a constant modulus constraint;

[0064] S3. Based on the problem model, a cyclic algorithm is used to iteratively optimize the waveform and the filter respectively until the signal to interference noise ratio of the airborne echo signal model converges;

[0065] S4. Output optimized waveform and filter.

[0066] In some specific embodiments, Figure 1 As shown, the waveform-filter joint design method includes the following steps:

[0067] Step 1: Establish an airborne echo signal model.

[0068] The signal received by the receiver includes received echo, clutter, interference and noise. Therefore, the airborne echo signal model is expressed as:

[0069] y=y t +y c +y j +n,

[0070] y represents the airborne echo signal model, y t Represents the received echo vector, y c represents the clutter vector, y j represents the interference vector, and n represents the noise vector.

[0071] Specifically, assume that for the t-th pulse receiving echo, the azimuth angle θ tThere is a target at , and after being received by the receiving antenna for analog-to-digital conversion and down-sampling, the kth (k = 1, ..., M, M represents the total number of pulse reception echoes) pulse reception echo is:

[0072]

[0073] Among them A t is the echo amplitude of the t-th pulse received echo, t∈[0,L-1], f t is the normalized Doppler frequency of the received echo of the tth pulse, a(θ t ) and b(θ t ) represents the transmission and reception steering vectors of the echo, and the waveform required is N T represents the number of transmitting antennas, and L represents the number of snapshots. Therefore, the received echo consisting of M pulses is:

[0074]

[0075] in, and the Doppler steering vector Indicates that the matrix dimension is N R ×N R The identity matrix, I M Similarly, the matrix dimension is M×M, N R Indicates the number of receiving antennas, Represents the Kronecker product operation.

[0076] Without loss of generality, clutter is distributed in both range and azimuth.

[0077] Figure 1 The clutter model of airborne MIMO-STAP radar is described, and the clutter patch partitioning strategy is explained. Figure 1 As shown in the figure, the normal direction of the semicircle represents different range rings. The target is located on the 0th range ring. There are Q range rings in front of the target and Q range rings behind the target, so the range of the range unit is [-Q, Q]. The tangent direction of the semicircle represents the azimuth, and there are N range rings in the azimuth. c patches, that is, the number of azimuth divisions of a single ranging unit is N c The target is located on the pth azimuth patch. Therefore, the relevant target is located at (0, p). The clutter model is the superposition of many independent clutter patches evenly distributed in different range-azimuth directions. The echo consists of 2Q+1 range bins, each of which includes N c Azimuth patches. The clutter is expressed as:

[0078]

[0079] Among them, q=-Q,-Q+1,...,0,...,Q-1,Q,p=1,2,...,N c represents the increment of the ranging unit, is the clutter of the ranging-azimuth unit.

[0080] The clutter of the qth ranging loop and the pth azimuth is θ still represents the azimuth, θ p represents the azimuth of the pth clutter, f p represents the normalized Doppler frequency of the clutter, A (q,p) is the backscattering amplitude, a(θ p ) represents the transmission steering vector of the clutter, b(θ p ) represents the receiving steering vector of the clutter, J q is a shift matrix with:

[0081]

[0082] L×L is the shift matrix J of the qth ranging unit q Dimension.

[0083] In addition, the airborne echo signal model also considers the blocking interference y j The interference is expressed as The interference Indicates that in θ i ,i=1,2,...,I J The blocking interference at J is the interference number. It is assumed that the interference is spatially correlated between the receiving antenna elements, but uncorrelated in fast time and pulse to pulse. Therefore, the interference covariance can be expressed as:

[0084]

[0085] in is the power of the ith jammer.

[0086] Step 2: Establish a problem model.

[0087] Step 2a), in this embodiment, a multi-dimensional filter-waveform joint design is used, with the signal-to-noise ratio as the main function, and the signal-to-noise ratio is in the form of:

[0088]

[0089] This function group is used as the main function of the problem model, where w represents the filter (weight), represents the covariance of the clutter-related waveform, represents the covariance of the interference-related waveform, represents the noise covariance.

[0090] Step 2b), using a cyclic algorithm to iteratively solve the waveform S and the filter w respectively, for each round of iteration, first keep the waveform S of the previous round of iteration unchanged, optimize the filter w of the current round, then keep the filter w of the current round unchanged, optimize the waveform S of the current round, and the waveform S of the first round of iteration uses the initialization value. Among them, based on the waveform S of the previous round of iteration, the problem model is converted to the filter optimization problem model, and the closed-form solution of the filter optimization problem model is solved to obtain the filter optimized in the current round. When the waveform S is fixed, the waveform optimization problem model for solving the filter w is expressed as:

[0091]

[0092] in It can be seen that the above formula is a generalized Rayleigh quotient, so the closed-form solution is:

[0093]

[0094] In order to facilitate the SINR optimization function and the following correlation waveform constraint unified parameters, in the following embodiments, x is used to represent the vector form of the waveform, and: x = s * =vec(S * ), vec() represents vectorized processing, and the waveforms appearing below refer to the vectorized waveforms represented by x.

[0095] Step 3: Propose record relevance constraints.

[0096] The dependency constraint can be expressed as:

[0097]

[0098] Among them, x m is the code transmitted by the mth transmitting antenna, the number of code elements is L, l is the code element number, x n is the code transmitted by the nth transmitting antenna. When m=n, the above formula represents the autocorrelation function; otherwise, it represents the cross-correlation function, and τ is the time delay.

[0099] In this embodiment, the correlation constraint includes a correlation constraint expressed in terms of an integrated sidelobe level (ISL) and a correlation constraint expressed in terms of a peak sidelobe level (PSL).

[0100] Step 3a), express the airborne correlation constraint in peak level form:

[0101]

[0102] Among them, w k represents the filter weight matrix of the k-th pulse receiving echo, let |r k,m,n (τ)| 2 Less than a mask value ∈ 2 , where ∈ represents the upper limit of the correlation, is called the selection matrix, C n Similarly, The signal is selected to come from the mth transmitting antenna. The airborne STAP space-time-Doppler three-dimensional is divided according to the Doppler dimension, so that the dimension of the sub-filter is the same as the waveform dimension, which is convenient for the subsequent iteration of the waveform coding-sub-filter weight. Thus, the problem of different dimensions of the STAP filter and the waveform filter is solved.

[0103] Step 3b) uses the integrated sidelobe level to express the airborne correlation constraint:

[0104]

[0105] ISL k It represents the integrated sidelobe level of the k-th pulse received echo. For simplicity, it is expressed as: ISL k ≤ζ,k=1,2,..,M, the value of ζ can be expressed as:

[0106]

[0107] Where 0≤ω≤1 represents the sidelobe average rate, which is the ratio coefficient between PSL and average ISL.

[0108] Step 3c), in order to make the radar amplifier work at the maximum efficiency, consider the constant modulus constraint:

[0109]

[0110] Among them, {x(t)}=x, that is, x(t) represents the elements in the vector x, E represents the emission energy,

[0111] Step 3d), build a waveform optimization problem model:

[0112]

[0113] When the filter w is solved, the waveform optimization problem model can be expressed as:

[0114]

[0115] Where W represents an L×MN R Dimensional matrix, N Rrepresents the number of receiving antennas, and the relationship between W and w is w = vec(W),

[0116]

[0117] in, Indicates dimension N T ×N T The identity matrix, N T represents the number of transmitting antennas, Expressing hope, f p represents the normalized Doppler frequency of the clutter, a(θ p ) represents the clutter transmission steering vector, b(θ p ) represents the receiving steering vector of the clutter, Represents the Kronecker product operation.

[0118] Step 4, decouple w and x.

[0119] As can be seen from step 2a), the problem model contains two mutually coupled variables w and x. We first use CA (Cyclic algorithm) to decouple and solve the waveform x and filter weight w separately. When the waveform is fixed, we solve the filter weight w, and then solve the waveform x when the filter weight w is fixed.

[0120] Step 4a), solving the filter w.

[0121] The method of solving the filter w when the waveform x is fixed has been introduced in step 2b) above, so it will not be repeated here.

[0122] Step 4b), solving the waveform x.

[0123] The waveform optimization problem model obtained by step 3d) is:

[0124]

[0125] It can be seen that the waveform optimization problem model is a quadratic constrained quadratic programming (QCQP) multi-constraint problem decomposition problem. In order to facilitate the solution, the constraints are reduced in dimension here:

[0126] Step 4c), reduce the dimension of the correlation constraint.

[0127] The dependency constraint is expressed as:

[0128]

[0129] The two constraints in the above formula are the peak level and the integral level in the optimization function. It can be seen that it is about wk and a quadratic function of x. In this embodiment, G is used to represent the constraint element Its values ​​are:

[0130]

[0131] in because and For offset.

[0132] Introduce Theorem 1:

[0133] Let B be a Hermitian matrix D is another Hermitian matrix and for For any point x in H Bx is replaced at x by

[0134] in Represents a real number.

[0135] Based on this theorem, here, B represents H k,m,n,τ , and the value of D is ν max (B)I, where ν max (B) represents the maximum eigenvalue of B, and I represents the dimension N T L×N T The identity matrix of L. Then, after applying this theorem, a linear function can be used to convert Simplified to:

[0136]

[0137] Implement dimensionality reduction and simplify constraints. Since the elements appear in ISL and PSL constraints, the above simplification process is used to simplify ISL and PSL respectively.

[0138] Step 4d), transforming the main function of the waveform optimization problem model.

[0139] In this embodiment, iterative optimization of the waveform includes:

[0140] The main function of the waveform optimization problem model is converted into a second-order function using the Dinklebach transformation method:

[0141]

[0142] Among them, f (i) represents the iterative result of the signal-to-noise ratio in the i-th round (current round), Ω=Σ(W)-f (i-1) Σ cn(W); f (i-1) Represents the signal-to-interference-noise ratio obtained in the previous iteration.

[0143] The transformed main function is converted into a subproblem form using the generalized Lagrangian form:

[0144] The mathematical form of the generalized Lagrangian form is:

[0145]

[0146] The generalized multiplier method result is:

[0147]

[0148] This embodiment introduces a slack variable z, assuming z = x, and then transforms the main function x of the signal to noise ratio optimization problem into H Ωx is converted into Lagrangian form:

[0149]

[0150] in, represents -Ω+μI, which transforms the maximum value problem into a minimum value problem; μ is the diagonal loading value, ensuring is a (semi) positive definite Hermitian matrix; ρ>0 represents the multiplier vector and penalty coefficient respectively.

[0151] In this way, the solution to the waveform optimization problem model is converted into the solution of multiple sub-problems, which reduces the difficulty of solution. Iterative optimization of the waveform can be achieved by iteratively solving the sub-problems under the constraints of the waveform optimization problem model. The sub-problem solving process includes:

[0152] Step 4e), sub-problem solving iteration criteria.

[0153] 1) Given x (i-1) ,λ (i-1) ,ρ (i-1) , solve:

[0154]

[0155] in, Express z (i) Solve; i represents the i-th inner layer iteration value.

[0156] 2) Given z (i) ,λ (i-1) ,ρ (i-1) , solve:

[0157]

[0158] In the formula, Express x (i) Solve, where the constraints have been reduced to linear constraints,

[0159]

[0160] The QP problem is easy to solve.

[0161] 3) Given x (i) ,z (i) ,λ (i-1) , solve:

[0162]

[0163] Among them, u >1,0<υ d <1,υ u , d All are preset values; Δd (i) =||z (i) -x (i) || 2 Represents the original residual; initial penalty function and Respectively represent Maximum and minimum eigenvalues.

[0164] 4) Given x (i) ,z (i) ,ρ (i) , solve:

[0165]

[0166] in, Λ is the upper limit of the multiplier value, The initial multiplier vector is

[0167] Step 5, iteratively optimize the waveform and filter.

[0168] Solve the sub-problems according to the method in step 4e), iteratively solve the waveform, and then continue to solve the filter according to step 4a), and repeat the cycle until the improvement in the signal-to-noise ratio is less than the set value and the cycle ends.

[0169] Step 6: Output the optimized waveform and filter.

[0170] Example 2

[0171] This embodiment verifies the method of embodiment 1. This embodiment considers a side-looking airborne co-located multiple-input multiple-output radar, which has N T= 4 transmit antennas and N R = 4 receiving antennas, which are linearly evenly distributed with element spacing d T =2λ and d R =λ / 2. The number of pulses in one CPI is M = 16. The carrier frequency and corresponding bandwidth of the transmission waveform are f 0 = 1 GHz and B = 1 MHz. The code length of each pulse is L = 32. There are two pulses from the direction of arrival (DOA) θ 1 =60 and θ 2 = -45, the jammer to noise ratio (JNR) is 35dB, and the variance of Gaussion white noise is The clutter-to-noise ratio (CNR) is 30 dB, assuming Q = 1. q=-Q,...,Q,p=1,.,N c , where N c = 361 is the number of clutter patches. The target of interest moves at a constant speed at azimuth 0, with a normalized frequency of 0.3 and a total transmitted energy of E = 1.

[0172] In all cases, Orthogonal Frequency Division Multiplexing Linear Frequency Modulation (OFDM-LFM) waveform coding is chosen as the initial waveform sequence considering its good properties in terms of pulse compression and fuzziness characteristics.

[0173] The mathematical expression of the OFDM-LFM signal transmitted through a single channel is:

[0174]

[0175] Among them, f s =2MHz is the sampling frequency, t p =L / f s is the pulse width, μ=B / t p is the modulation frequency, n=1,...,N T and Δf=K / t p , is a set of positive natural numbers representing the carrier frequency interval between adjacent transmission signals. Therefore, the initial code is

[0176] The termination value of the outer iteration is set to ι = 10 -3 , the parameter of the inner iteration is υ d =0.99,υ u =1.01, Λ=10 4 , e abs =10 -6 , e rel =10 -7 .

[0177] like Figure 3 As shown, compared with the initial OFDM-LFM waveform (SINR generated by the first iteration), the curve obtains a monotonically increasing SINR performance through the method proposed in Example 1 and gradually approaches the upper boundary, which illustrates the increasing behavior of the output SINR processed by the method of Example 1 and the convergence of the proposed algorithm. Figure 3 It is also shown that with different requirements on side lobes, different degrees of freedom lead to changes in the SINR increase.

[0178] like Figure 4 The correlation performance without CP constraints is shown. Under CM constraints, the autocorrelation and cross-correlation results are almost the same. In other words, the signal sent by the transmit array is as strongly correlated as the phased array system, losing the advantage of the MIMO system.

[0179] Figure 5 (a) describes the autocorrelation performance of the waveform, and its parameter value directly reflects the height of the peak level.

[0180] According to the design goal, we only require the maximum value to be reached when the sequence point τ = 0, and from τ = ±1, these autocorrelations are required to have very low levels. Therefore, the narrowest main lobe width is formed, and the side lobes also reach the required level. The cross-correlation performance of the waveform is as follows Figure 3 As shown in (b), compared with the sequence cross-correlation diagram with only constant modulus constraint, the cross-correlation diagram under correlation constraint has no peak and is very flat. Indeed, the lower the cross-correlation between MIMO radar transmitting units, the closer it is to orthogonal waveforms. In addition, the transmission energy diverges in the spatial domain, resulting in low interception performance.

[0181] Example 3

[0182] This embodiment, based on the design concept of Embodiment 1, proposes a waveform-filter joint design system considering correlation constraints, the system comprising:

[0183] The parameter initialization module is configured to: receive system parameters configured for the radar system composed of a waveform and a filter;

[0184] The airborne echo signal model building module is configured to: build an airborne echo signal model based on configured system parameters;

[0185] The problem model building module is configured to: build a problem model according to configured constraints with the goal of maximizing the signal-to-interference-to-noise ratio of the airborne echo signal model, wherein the problem model includes a filter optimization problem model and a waveform optimization problem model; the constraints include a correlation constraint and a constant modulus constraint;

[0186] An optimization module, based on the problem model, adopts a loop algorithm to iteratively optimize the waveform and the filter respectively until the signal to interference noise ratio of the airborne echo signal model converges;

[0187] Output module, outputs optimized waveform and filter.

[0188] The relevant configurations of each module can refer to the relevant features in Example 1 respectively, and will not be repeated here.

[0189] The present invention is not limited to the above-mentioned specific embodiments, but extends to any new features or any new combination disclosed in this specification, as well as any new method or process steps or any new combination disclosed.

Claims

1. A waveform-filter joint design method considering correlation constraints, characterized in that: include: S1. Building an airborne echo signal model based on the configured system parameters; S2. With the goal of maximizing the signal-to-interference-to-noise ratio of the airborne echo signal model, a problem model is constructed according to configured constraints, wherein the problem model includes a filter optimization problem model and a waveform optimization problem model; the constraints include a correlation constraint and a constant modulus constraint; S3. Based on the problem model, a cyclic algorithm is used to iteratively optimize the waveform and the filter respectively until the signal to interference noise ratio of the airborne echo signal model converges; S4. Output optimized waveform and filter.

2. The waveform-filter joint design method considering correlation constraints as claimed in claim 1, characterized in that: The airborne echo signal model is constructed as: y=y t +y c +y j +n, y represents the airborne echo signal model, y t Represents the received echo vector, y c represents the clutter vector, y j represents the interference vector, and n represents the noise vector.

3. The waveform-filter joint design method considering correlation constraints as claimed in claim 2, characterized in that: The correlation constraint includes a correlation constraint expressed in terms of an integrated sidelobe level and a correlation constraint expressed in terms of a peak sidelobe level.

4. The waveform-filter joint design method considering correlation constraints as claimed in claim 3, characterized in that: In the problem model, the waveform optimization problem model is constructed as: in, x=vec(S * ), S represents the waveform in matrix form, vec() represents vectorization processing, and x represents the waveform in vector form; W is L×MN R dimensional matrix, N R represents the number of receiving antennas, L represents the number of snapshots, and M represents the total number of pulse received echoes; w = vec(W), where w represents the filter weight; t represents the echo number, p represents the clutter number, θ represents the azimuth, A t is the echo amplitude, Indicates that the dimension is N T ×N T The identity matrix, N T Indicates the number of transmitting antennas; f t 、f p represents the normalized Doppler frequency of the echo and clutter, a(θ t ), a(θ p ) represents the launch steering vector, b(θ t ), b(θ p ) represents the receive steering vector, represents the Kronecker product operation; -Q,Q] represents the range of the ranging unit, N c Indicates the number of azimuth divisions of a single ranging unit; A means to seek expectation. (q,p) is the backscatter amplitude; L×L is the shift matrix J of the qth ranging unit q Dimensions; {x(t)}=x,t∈[0,L-1],E represents the emission energy, w k represents the filter weight matrix of the k-th pulse received echo, I L is the identity matrix of dimension L×L, ∈ represents the upper limit of correlation; ISL k It represents the integrated sidelobe level of the k-th pulse received echo, m and n represent the serial numbers of the transmitting antennas, τ is the time delay, and 0≤ω≤1 represents the sidelobe average rate.

5. The waveform-filter joint design method considering correlation constraints as claimed in claim 1, characterized in that: In iteratively optimizing the waveform and the filter respectively using the problem model, the iterative optimization of the waveform includes: The main function of the waveform optimization problem model is converted into a second-order function using the Dinklebach transformation method; The transformed main function is converted into a sub-problem form using the generalized Lagrangian form; The sub-problems are iteratively solved under the constraints of the waveform optimization problem model to achieve iterative optimization of the waveform.

6. The waveform-filter joint design method considering correlation constraints as claimed in claim 5, characterized in that: For each round of iteration, the waveform of the previous round of iteration is kept unchanged, and the filter of the current round is optimized. Then, the filter of the current round is kept unchanged and the waveform of the current round is optimized. The waveform of the first round of iteration uses the initialization value.

7. The waveform-filter joint design method considering correlation constraints as claimed in claim 6, characterized in that: The method of optimizing the filter in each iteration includes: Based on the waveform S of the previous iteration, the problem model is converted to a filter optimization problem model; The closed-form analytical expression of the filter optimization problem model is used to obtain the filter weights of the current round of optimization.

8. The waveform-filter joint design method considering correlation constraints as claimed in claim 5, characterized in that: Iterative optimization of the waveform also includes: Perform dimensionality reduction on correlation constraints.

9. A waveform-filter joint design system considering correlation constraints, characterized in that: include: The parameter initialization module is configured to: receive system parameters configured for the radar system composed of a waveform and a filter; The airborne echo signal model building module is configured to: build an airborne echo signal model based on configured system parameters; The problem model building module is configured to: build a problem model according to configured constraints with the goal of maximizing the signal-to-interference-to-noise ratio of the airborne echo signal model, wherein the problem model includes a filter optimization problem model and a waveform optimization problem model; the constraints include a correlation constraint and a constant modulus constraint; An optimization module, based on the problem model, adopts a loop algorithm to iteratively optimize the waveform and the filter respectively until the signal to interference noise ratio of the airborne echo signal model converges; Output module, outputs optimized waveform and filter.

Citation Information

Patent Citations

  • Multi-intermittent sampling interference resisting method based on joint design of cognitive waveform and filter

    CN115219997A

  • Low-PAR radar waveform and filter optimization method based on SCNRA-KKT

    CN117875248A

  • Radar waveform and mismatched filter joint transmit-receive design method

    CN118131135A

Cited By

  • FEKO-based synthetic aperture radar (SAR) waveform and range direction filter joint design method

    CN119740372A

  • Joint Design Method of SAR Waveform and Range Filter Based on FEKO

    CN119740372B

  • Airborne radar waveform optimization method, device and equipment based on PCFM

    CN121899809A