CN113917421A - A Distributed Radar Main Lobe Interference Suppression Method Based on Cascaded LMS Filters - Google Patents

Abstract

The invention discloses a distributed radar main lobe interference suppression method based on a cascaded LMS filter, which can effectively and steadily suppress main radar echo interference through two-stage filter cascade. Firstly, receiving a target and an interference signal by using a main radar, and receiving the interference signal by using a large-aperture coherent array with extremely narrow beams, which is formed by sparse deployment of a plurality of auxiliary radars; then, the main radar and the large-aperture coherent receiving array signals are used as input, and the fast time-slow time cascade LMS filter is adopted to carry out interference cancellation (wiener filtering) processing on the main radar echo. The fast time LMS filter can realize fast time stable tracking and filtering of the interference parameters, the slow time LMS filter can realize slow time stable tracking and filtering of the interference parameters, and the two-stage filter cascade can realize effective and steady suppression of the echo interference of the main radar.

Description

A Distributed Radar Main Lobe Interference Suppression Method Based on Cascaded LMS Filters

technical field

The invention relates to the field of radar main lobe interference suppression processing, in particular to a distributed radar main lobe interference suppression method based on cascaded LMS filters.

Background technique

The electromagnetic environment of modern battlefield is increasingly complex, and electromagnetic interference has become an important bottleneck problem that restricts the efficiency of radar. As a typical interference pattern, the main lobe interference is modulated by the gain of the radar antenna, the interference energy is larger, and the interference effect is more obvious, and the traditional side lobe interference suppression processing method cannot effectively suppress it.

Aiming at the problem of radar main lobe interference suppression, the traditional airspace processing method is to form a zero point of the pattern in the interference direction, but due to the constraints of the single-base radar aperture, the pattern after adaptive beamforming is distorted, and the target energy loss is relatively large. In view of this, Yang X., Yin P, Zeng T. et al. (Yang X., Yin P, Zeng T., et al.Applyingauxiliary array to suppress mainlobe interference for ground-based radar[J].IEEE Antennas and Wireless Propagation Letters, 2013, 12: 433-436.) proposed a distributed radar architecture using an auxiliary receiving array, which can effectively suppress main lobe interference. H.Zhang,J.Luo(H.Zhang,J.Luo,X.Chen,Q.Liu and T.Zeng,Whitening filter for mainlobeinterference suppression in distributed array radar,Proc.CIE Int.Conf.Radar(RADAR), pp.1-5, 2016.) On the basis of distributed radar, a whitening filter is used to achieve robust main lobe interference suppression performance. Jiang Tiezhen, Liao Tongqing A Least MeanSquare (LMS) filter suppresses the main lobe interference. The above studies all show that the distributed radar system has obvious performance advantages and broad application prospects in the main lobe interference suppression. On the one hand, by increasing the array aperture, the distributed radar system can effectively improve the angular resolution of the system and narrow the main lobe width of the synthetic pattern: when the spatial adaptive filtering is used, the main lobe of the pattern originally located in the single-base radar pattern is narrowed. The internal interference signal falls into the side lobes of the synthetic array pattern, which can effectively reduce the target energy loss caused by the interference suppression processing and achieve better interference suppression effect; when the time domain adaptive filtering is used, due to the distributed Due to the high angular resolution characteristics of radar, interference and targets can be distinguished from the airspace, so a more "pure" interference sample can be obtained as a reference signal, and the target energy loss caused by time-domain filtering can also be effectively reduced. But on the other hand, because the interference signal may have slow changes in parameters such as Doppler and power in slow time, neither spatial or temporal adaptive filtering processing can effectively solve this problem, resulting in loss of multi-frame accumulation processing performance.

Therefore, it is of great practical significance and application value to study a robust main lobe interference suppression method on the basis of distributed radar system.

SUMMARY OF THE INVENTION

In view of this, the present invention provides a distributed radar main lobe interference suppression method based on cascaded LMS filters, which can effectively and robustly suppress main radar echo interference by cascading two-stage filters.

In order to achieve the above object, the technical scheme of the present invention is: a method for suppressing the main lobe interference of a distributed radar based on cascaded LMS filters, the distributed radar system is composed of a main radar and a plurality of auxiliary radars; the plurality of auxiliary radars The sparsely dispersed arrangement is around the main radar to form a sparse coherent receiving array, the aperture of the sparse coherent receiving array is larger than the aperture of the main radar, and the sparsely dispersed arrangement of the plurality of auxiliary radars means that two adjacent auxiliary radars are spatially independent from each other , the distance between the antenna phase centers on it is ten to one hundred times the half wavelength, and the time-frequency and timing synchronization equipment is connected between the main radar and each auxiliary radar to ensure the coherence of work; the distributed radar main lobe interference suppression method It includes the following steps:

Step 1: The main radar beam is aimed at a certain detection airspace to transmit signals, and the main radar echo signal is obtained; the sparse coherent receiving array forms a very narrow beam, and scans the airspace within the main radar main lobe beam range to obtain sparse The array echo of the coherent receiving array at each beam pointing angle; wherein the beam of the sparse coherent array is that the beam width of the extremely narrow beam is much smaller than the main radar beam width (in the embodiment of the present invention, far smaller than refers to, The extremely narrow beam width is a few tenths of the main radar beam, specifically 0.05 degrees).

Step 2: Perform interference identification on the echoes of the sparse large-aperture coherent receiving array to determine whether there is interference and the type of interference; if there are multiple interferences, give the angle of each interference and use the sparse large-aperture at the angle. The coherent reception array echo is used as the corresponding interference reference signal.

Step 3. Under each interference angle, perform the following steps:

Taking the fast time echo of the main radar as the main channel signal and the fast time echo of the sparse coherent receiving array as the reference signal, the fast time LMS filter is used to perform the fast time LMS filtering process on the fast time echo of the main radar, and the filtering intermediate processing result is obtained. , repeating the fast-time LMS filtering process multiple times to obtain multiple pulse fast-time LMS filtering processing results.

Then, the slow-time echo of the main radar is used as the main channel signal, and the slow-time echo of the intermediate processing result of the fast-time LMS filtering is used as the reference signal, and the slow-time LMS filter is used to perform the slow-time LMS filtering process on the slow-time echo of the main radar. , repeating the slow-time LMS filtering process multiple times to obtain the slow-time LMS filtering processing results of multiple sampling points, and then obtaining the final cascaded filtering processing results.

The cascaded filtering processing results obtained under all interference angles constitute the final interference cancellation processing result.

Step 4: Perform pulse compression processing on each sub-pulse in the interference cancellation processing result to obtain multi-frame one-dimensional range images.

Step 5: Perform coherent processing on the multi-frame one-dimensional range images to obtain the multi-frame joint accumulation processing result.

Further, the fast-time LMS filter is an N-order digital filter with a step size μ.

Further, the fast-time LMS filtering process specifically includes the following steps:

Step 1: Set the current processing pulse number to m, and the initial value of m is 1.

Step 2: Record the intermediate result of the fast-time LMS filtering weight vector operation at the n-th sampling point of the m-th pulse as

Among them, Q is the number of sampling points included in the pulse, and the fast time sampling instant is t _n , n∈[1,...,Q].

And 0≤t ₁ <t ₂ <...<t _Q ≤T, where T is the pulse repetition time PRT.

快时间LMS滤波器的自适应滤波输出为：When n=1, that is, time t ₁ ,

The adaptive filtering output of the fast-time LMS filter is:

e(m, _tn )=d(m, _tn ), d(m,tn) is the main channel signal of the fast-time LMS filter at the _nth sampling point (time _tn ) of the mth pulse;

Step 3: Determine the reference signal of the fast-time LMS filter at the n-th sampling point of the m-th pulse (time t _n ) as:

x(m,t _n )=[x(m,t _n ) x(m,t _n-1 ) ... x(m,t _n-N+1 )] ^T ;

where x(m,t _n ) x(m,t _n-1 ) … x(m,t _n-N+1 ) are the fast time echoes of the sparse coherent receiving array, respectively, t _n t _n-1 … t _n-N+1 is the fast time sampling point;

Then the intermediate result of the fast-time LMS filtering weight vector operation at the n+1th sampling point of the mth pulse is:

其中e^*(m,t_n)为e(m,t_n)的共轭；

where e ^* (m,t _n ) is the conjugate of e(m,t _n );

Then at the n+1th sampling point of the mth pulse (time t _n+1 ), the corresponding output signal after the fast-time LMS filter weighted summation, that is, the intermediate processing result of the fast-time LMS filter is y(m, t _n+1 ):

The adaptive filtering output of the n+1th sampling point (time tn ₊₁ ) of the mth pulse is e(m,tn ₊₁ )=d(m,tn ₊₁ )-y(m, _{tn +1} )

Step 4: Let n increment by 1, if n=Q, go to Step 5, otherwise return to Step 3;

重新计算第m个脉冲第n个采样点(t_n时刻)对应的所述快时间LMS滤波器加权求和后的结果为Step 5: Make the final fast-time LMS filter weight vector

The result of recalculating the weighted summation of the fast-time LMS filter corresponding to the n-th sampling point (time t _n ) of the m-th pulse is:

Step 6: If m=M, terminate the processing; otherwise, make m increment by 1 and return to Step 2.

Further, the slow-time LMS filter is a K-order digital filter with a step size of n.

Further, the slow-time LMS filtering process specifically includes the following steps:

SS1: Set the current processing fast time sampling point number to n, and the initial value of n is 1.

SS2: Record the intermediate result of the slow-time LMS filtering weight vector operation at the mth pulse at the nth sampling point as

where M is the total number of pulses, the fast-time sampling point is t _n , n∈[1,...,Q], Q is the number of sampling points included in the pulse; and 0≤t ₁ <t ₂ <...<t _Q ≤T, T is the pulse repetition time PRT.

t₁时刻第m个脉冲对应的慢时间LMS滤波器的自适应滤波输出为：When m=1, that is, time t ₁ ,

The adaptive filtering output of the slow-time LMS filter corresponding to the mth pulse at time _t1 is:

f(m, _tn )=d(m, _tn ), d(m, _tn ) is the main channel signal of the slow-time LMS filter corresponding to the mth pulse at the _nth sampling point (time tn).

SS3: Use the slow-time echo of the intermediate processing result of the fast-time LMS filter as the reference signal of the slow-time LMS filter, specifically:

y(m,t _n )=[y(m,t _n ) y(m-1,t _n ) ... y(m-K+1,t _n )] ^T ;

y(m,t _n ) y(m-1,t _n ) … y(m-K+1,t _n ) are the corresponding values of K pulses before m at time t _n respectively.

Then at the m+1th pulse of the nth sampling point, the intermediate result of the slow-time LMS filtering weight vector operation is:

其中f^*(m,t_n)为f(m,t_n)的共轭。

where f ^* (m, _tn ) is the conjugate of f(m, _tn ).

其中y(m+1,t_n)为第n个采样点(t_n时刻)第m+1个脉冲对应的参考信号；Then at the m+1th pulse at the nth sampling point, the result of the weighted summation of the slow-time LMS filter is:

where y(m+1,t _n ) is the reference signal corresponding to the m+1th pulse at the _nth sampling point (time tn);

The output result of the slow-time LMS filter adaptive filter corresponding to the m+1th pulse at the _nth sampling point (time tn) is: f(m+1, _tn )=d(m+1, _tn ) -z(m+1, _tn ); d(m+1, _tn ) is the main channel signal of the slow-time LMS filter corresponding to the m+1th pulse at the _nth sampling point (time tn).

SS4: Let m increment by 1, if m=M, enter SS5, otherwise repeat SS3.

重新计算t_n时刻第m个脉冲对应的所述慢时间LMS滤波器加权求和后的结果为：SS5: Let the final slow-time LMS filter weight vector

The result of recalculating the weighted summation of the slow-time LMS filter corresponding to the mth pulse at time _tn is:

The adaptive filtering output of the slow-time LMS filter corresponding to the m-th pulse at time t _n is recalculated as follows:

f(m,t _n )=d(m,t _n )-z(m,t _n )m=1,2,...,M

SS6: If n=Q, terminate the process, otherwise set n=n+1, and return to SS2.

Beneficial effects:

The invention provides a robust distributed radar main lobe interference suppression method. The method first uses the main radar to receive the target and jamming signals, and uses the large-aperture coherent array with extremely narrow beams formed by the sparse deployment of multiple auxiliary radars to receive the jamming signals; The main radar echo is processed by interference cancellation (Wiener filter) by using fast-time-slow-time cascaded LMS filters. Among them, the fast-time LMS filter can realize fast-time stable tracking and filtering of interference parameters, the slow-time LMS filter can realize slow-time stable tracking and filtering of interference parameters, and the cascaded two-stage filter can realize the main radar feedback. Effective robust suppression of wave interference.

Description of drawings

Figure 1 is a schematic diagram of a typical distributed radar system detection;

Figure 2 is a schematic diagram of the relationship between the main radar and the baseline of the sparse large-aperture coherent receiving array;

FIG. 3 is a schematic structural diagram of a fast-time-slow-time cascaded LMS filter;

4 is a schematic diagram of the structure of a fast-time LMS filter;

FIG. 5 is a schematic structural diagram of a slow-time LMS filter;

FIG. 6 is a schematic diagram of a one-dimensional range image before and after interference suppression.

Detailed ways

The present invention will be described in detail below with reference to the accompanying drawings and embodiments.

A typical distributed radar system detection scenario is shown in Figure 1. The distributed radar system is composed of a main radar and a plurality of auxiliary radars sparsely arranged around the main radar, wherein a plurality of auxiliary radars are sparsely arranged around the main radar to form a sparse coherent receiving array, and the sparse coherent receiving The array aperture is larger than the main radar aperture. The sparse and dispersed arrangement of the multiple auxiliary radars means that two adjacent auxiliary radars are spatially independent from each other, the distance between the antennas on them is ten to one hundred times the half wavelength, and the distance between the main radar and each auxiliary radar is Time-frequency and timing synchronization equipment is connected to ensure the coherence of work, and multiple scattered auxiliary radars can form a sparse large-aperture coherent receiving array with extremely narrow beams.

Considering the existence of grating lobes in the pattern of the sparse large-aperture coherent receiving array, generally, array configuration optimization or other methods can be used to reduce the influence of the grating lobes. Accordingly, the specific steps of a distributed radar main lobe interference suppression method based on cascaded LMS filters are as follows:

Step 1: The main radar beam is aimed at a certain detection airspace to transmit signals, and the main radar echo signal is obtained; the sparse coherent receiving array forms a very narrow beam, and scans the airspace within the main radar main lobe beam range to obtain sparse The array echo of the coherent receiving array at each beam pointing angle; wherein the beam width of the sparse coherent array is that the beam width of the extremely narrow beam is much smaller than the beam width of the main radar. The extremely narrow beam is a few tenths of the main radar beam, specifically 0.05 degrees.

Distributed radar echo acquisition.

Let the main radar echo be

为对应接收到的干扰信号功率，

为主雷达第m个脉冲第l个干扰信号的多普勒及相移相位项，w_M(m,t)为主雷达第m个脉冲中的接收机本底高斯白噪声且满足E[|w_M(m,t)|²]＝σ²，σ²为噪声功率。Among them, m is the slow time pulse number, t is the fast time axis, θ ₀ is the current beam pointing angle of the main radar, s(t) is the main radar transmit signal, A _k (m) is the k-th pulse of the m-th pulse of the main radar The echo amplitude of the target, f ₀ is the operating frequency of the system, R _k (m) is the distance between the k-th target of the m-th pulse and the main radar, and ψ _k (m) is the echo of the k-th target of the m-th pulse of the main radar. Wave Doppler and phase-shift phase terms, I _l (m,t) is the complex envelope of the l-th interfering signal of the m-th pulse of the main radar,

In order to correspond to the received interference signal power,

Doppler and phase-shift phase terms of the l-th jamming signal of the m-th pulse of the main radar, w _M (m,t) is the receiver background white Gaussian noise in the m-th pulse of the main radar and satisfies E[| w _M (m,t)| ² ]=σ ² , where σ ² is the noise power.

Assuming that the sparse large-aperture coherent receiving array contains N auxiliary radars, the relationship with the array baseline of the main radar is shown in Figure 2, then the steering vector of the target or interference signal entering the coherent receiving array at an angle θ is:

表示角度θ条件下第i个辅助雷达与主雷达的传播路程差，

分别为角度θ条件下信源到主雷达和第i个辅助雷达的传播路程，[·]^T为矩阵转置运算。则稀疏大孔径相参接收阵列回波可表示为where λ is the operating wavelength of the system,

represents the propagation distance difference between the ith auxiliary radar and the main radar under the condition of angle θ,

are the propagation distances from the source to the main radar and the i-th auxiliary radar under the condition of angle θ, respectively, [·] ^T is the matrix transposition operation. Then the sparse large aperture coherent receiving array echo can be expressed as

where ⊙ is the Hadamard product,

)条件下，稀疏大孔径相参接收阵列各辅助雷达第m个脉冲的目标信号回波包络，

为各辅助接收雷达天线增益，R_k(m)的定义与E_M(m,t,θ₀)相同，

为稀疏大孔径相参接收阵列第m个脉冲第k个目标的回波多普勒及相移相位项。

为第l个干扰(角度

)条件下，稀疏大孔径相参接收阵列各辅助雷达第m个脉冲的干扰信号回波包络，

稀疏大孔径相参接收阵列第m个脉冲第l个干扰信号的多普勒及相移相位项。n(m,t)＝[…,n_i(m,t),…]^T，i∈[1,…,N]，为稀疏大孔径相参接收阵列第m个脉冲各辅助接收雷达的本底噪声信号，n_i(m,t)相互独立且功率相同。is the kth target (angle

), the target signal echo envelope of the mth pulse of each auxiliary radar of the sparse large-aperture coherent receiving array,

is the antenna gain of each auxiliary receiving radar, the definition of R _k (m) is the same as that of E _M (m,t,θ ₀ ),

is the echo Doppler and phase-shift phase term of the k-th target of the m-th pulse of the sparse large-aperture coherent receiving array.

is the l-th interference (angle

), the echo envelope of the jamming signal of the mth pulse of each auxiliary radar of the sparse large-aperture coherent receiving array,

The Doppler and phase-shift phase terms of the l-th interfering signal of the m-th pulse of the sparse large-aperture coherent receiving array. n(m,t)=[…,n _i (m,t),…] ^T , i∈[1,…,N], is the original value of each auxiliary receiving radar of the mth pulse of the sparse large-aperture coherent receiving array Noise floor signals, n _i (m,t) are independent of each other and have the same power.

Ignoring the propagation path difference and time delay of the target and interference to each auxiliary radar, the echo of the sparse large aperture coherent receiving array can be simplified as

为

简化后的复包络。

为

简化后的复包络。in,

for

Simplified complex envelope.

for

Simplified complex envelope.

The echo of the sparse large-aperture coherent receiving array after spatial matching filtering (space vector weighted summation) can be expressed as

为稀疏大孔径相参接收阵列经过空域匹配滤波后第m个脉冲第k个目标的回波幅度，

为稀疏大孔径相参接收阵列经过空域匹配滤波后第m个脉冲第k个目标的回波多普勒及相移相位项，

为稀疏大孔径相参接收阵列经过空域匹配滤波后第m个脉冲接收到的第l个干扰信号的干扰信号功率，

为稀疏大孔径相参接收阵列经过空域匹配滤波后第m个脉冲第l个干扰信号的多普勒及相移相位项，w_A(m,t)为稀疏大孔径相参接收阵列第m个脉冲中的接收机本底高斯白噪声且满足E[|w_A(m,t)|²]＝σ²，σ²为噪声功率。Among them, θ _A is the beam pointing angle of the sparse large-aperture coherent receiving array, [ ] ^H is the matrix conjugate transpose operation, m, t, s(t), R _k (m), I _l (m,t) The meaning is the same as the main radar echo E _M (m,t,θ ₀ ),

is the echo amplitude of the k-th target of the m-th pulse after the sparse large-aperture coherent receiving array is subjected to spatial matching filtering,

is the echo Doppler and phase-shift phase term of the k-th target of the m-th pulse after the sparse large-aperture coherent receiving array is subjected to spatial matching filtering,

is the interfering signal power of the l-th interfering signal received by the m-th pulse after the sparse large-aperture coherent receiving array is spatially matched and filtered,

is the Doppler and phase-shift phase terms of the l-th interfering signal of the m-th pulse after the sparse large-aperture coherent receiving array is spatially matched, and w _A (m,t) is the m-th interfering signal of the sparse large-aperture coherent receiving array The receiver background white Gaussian noise in the pulse satisfies E[|w _A (m,t)| ² ]=σ ² , where σ ² is the noise power.

By changing θ _A and traversing the scanning space, the sparse large-aperture coherent receiving array echoes at different angles θ _A can be obtained.

Step 2: Perform interference identification on the echoes of the sparse large-aperture coherent receiving array to determine whether there is interference and the type of interference; if there are multiple interferences, give the angle of each interference and use the sparse large-aperture at the angle. The coherent reception array echo is used as the corresponding interference reference signal.

均不为0。稀疏大孔径相参接收阵列以主雷达波束指向角θ₀为中心，主雷达3dB波束宽度θ_w为范围，改变θ_A遍历扫描主雷达主瓣空域。对不同θ_A角度的相参接收阵列回波E_A(m,t,θ_A)进行干扰辨识，鉴别是否存在干扰以及干扰类型。Assuming that the pointing angle of the main radar beam is θ ₀ , E _M (m,t, θ ₀ ) contains both target and interference echoes, that is, A _k (m) and

are not 0. The sparse large-aperture coherent receiving array takes the main radar beam pointing angle θ ₀ as the center, the main radar 3dB beam width θ _w as the range, and changes θ _A to traverse and scan the main radar main lobe airspace. Interference identification is carried out on the coherent receiving array echoes EA (m,t, θ _{A )} at different angles of θ _A to identify whether there is interference and the type of interference.

选定角度

的稀疏大孔径相参接收阵列回波

作为干扰参考信号，则此时

可近似为Assuming that under the extremely narrow beam condition of the sparse large-aperture coherent receiving array, there are P interferences within the main lobe beam range of the main radar that can be distinguished from the airspace by the coherent receiving array, and the corresponding angles are

selected angle

The sparse large-aperture coherent receiving array echoes

As the interference reference signal, then at this time

can be approximated as

的干扰信号，目标回波幅度A_k(m)及其余干扰的回波功率

均近似为0，即通过稀疏大孔径相参接收阵列空间匹配接收可获得某一干扰更为纯净的参考信号。Among them, I _l (m, t) is the corresponding angle

The interference signal of , the target echo amplitude A _k (m) and the echo power of the rest of the interference

Both are approximately 0, that is, a reference signal with a purer interference can be obtained through the spatial matching of the sparse large-aperture coherent receiving array.

Step 3. Under each interference angle, perform the following steps:

Taking the fast time echo of the main radar as the main channel signal and the fast time echo of the sparse coherent receiving array as the reference signal, the fast time LMS filter is used to perform the fast time LMS filtering process on the fast time echo of the main radar, and the filtering intermediate processing result is obtained. , repeating the fast-time LMS filtering process multiple times to obtain multi-frame fast-time LMS filtering processing results.

Then, the slow-time echo of the main radar is used as the main channel signal, and the slow-time echo of the intermediate processing result of the fast-time LMS filtering is used as the reference signal, and the slow-time LMS filter is used to perform the slow-time LMS filtering process on the slow-time echo of the main radar. , repeating the slow-time LMS filtering process multiple times to obtain the slow-time LMS filtering processing results of multiple sampling points, and then obtaining the final cascaded filtering processing results.

The cascaded filtering processing results obtained under all interference angles constitute the final interference cancellation processing result.

In this embodiment, the fast-time-slow-time cascaded LMS filtering process is specifically:

将x(m,t)作为参考信号，对d(m,t)进行快时间-慢时间级联LMS滤波(维纳滤波)处理，实现对d(m,t)中干扰信号的自适应对消。设每个PRT回波包含的采样点数均为Q，采样时刻为t_n(n∈[1,…,Q]且0≤t₁＜t₂＜…＜t_Q≤T，T为PRT)，则实际的数字回波为d(m,t_n)和x(m,t_n)的定义域为{(m,t_n)|m＝1,2,…,M；n＝1,2,…,Q}，M为多帧相参处理脉冲个数。Let d(m,t)=EM( _m ,t,θ ₀ ),

Taking x(m,t) as the reference signal, fast-time-slow-time cascaded LMS filtering (Wiener filtering) processing is performed on d(m,t) to realize the adaptive adjustment of the interference signal in d(m,t). remove. Assuming that the number of sampling points contained in each PRT echo is Q, and the sampling time is t _n (n∈[1,…,Q] and 0≤t ₁ <t ₂ <…<t _Q ≤T, T is PRT), Then the actual digital echoes are d(m, t _n ) and x(m, t _n ) whose definition domain is {(m, t _n )|m=1,2,...,M; n=1,2, ...,Q}, M is the number of multi-frame coherent processing pulses.

The fast-time-slow-time cascaded LMS filter structure is shown in Figure 3. The fast-time LMS filter is an N-order digital filter with a step size of μ, and its processing process is shown in Figure 4. y(m,t _n ) is the output signal corresponding to x(m,t _n ) after the fast-time LMS filter weighted summation, which satisfies

Among them, w _m =[w _m0 ,w _m1 ,...,w _m(N-1) ] ^T is the weight vector of the mth pulse fast time LMS filter, w ₀ ,w ₁ ,...,w _N-1 is the corresponding The N-th order weight vector coefficients. x(m,t _n )=[x(m,t _n ) x(m,t _n-1 ) ... x(m,t _n-N+1 )] ^T .

e(m,t _n ) is the adaptive filter output, which satisfies

e(m, _tn )=d(m, _tn )-y(m, _tn )m=1,2,...,M; n=1,2,...,Q

The processing flow of fast-time LMS filtering is as follows:

In this embodiment, the adopted slow-time LMS filter is a K-order digital filter with a step size of n, and its processing process is shown in FIG. 5 . Let z(m, t _n ) be the output signal corresponding to y(m, t _n ) after the weighted summation of the slow-time LMS filter, which satisfies

Among them, v _n =[v _n0 ,v _n1 ,...,v _n(K-1) ] ^T is the weight vector of the slow-time LMS filter at the nth sampling point, v _n0 ,v _n1 ,...,v _{n(K- 1)} is the corresponding K-order weight vector coefficient. y(m,t _n )=[y(m,t _n ) y(m-1,t _n ) ... y(m-K+1,t _n )] ^T .

f(m,t _n ) is the adaptive filter output, which satisfies

f(m,t _n )=d(m,t _n )-z(m,t _n )

The process flow of slow-time LMS filtering is as follows:

z(m,t) is the result of approximating the interference components in d(m,t) by x(m,t) in the slow time independent variable m domain and the fast time independent variable t domain respectively, so it can be realized that d(m , t) Adaptive cancellation of interference components.

下的稀疏大孔径相参接收阵列回波

作为新的干扰参考信号。Return to "Step 2", select the angle

sparse large-aperture coherent receiving array echoes under

as a new interference reference signal.

将x(m,t)作为参考信号，对d(m,t)进行快时间-慢时间级联LMS滤波(维纳滤波)处理，得到角度

干扰对消后的新的滤波结果f(m,t)。Return to "Step 3", and d(m,t)=f(m,t),

Using x(m,t) as a reference signal, perform fast-time-slow-time cascaded LMS filtering (Wiener filtering) on d(m,t) to obtain the angle

The new filtering result f(m,t) after interference cancellation.

Repeat "step 2" and "step 3" multiple times, traverse P interference angles within the main lobe beam range of the main radar, obtain the interference reference signal of the corresponding angle, and use the last filtering result f(m, t) as the new The main radar echo is input, and the fast-time-slow-time cascade LMS filtering process is performed until the P interferences have been filtered and cancelled, and the final main lobe interference cancellation echo f(m, t) is obtained.

Step 4, echo pulse compression.

Using the main radar transmit signal s(t) as the reference signal, perform pulse compression on each sub-pulse in the f(m,t) result filtered by the fast-time-slow-time cascade LMS in "Step 3" to obtain the interference The multi-frame coherent one-dimensional distance image P _c (m, t) after the cancellation is completed.

Step 5: Joint accumulation of multiple frames.

Multi-frame joint accumulation processing is performed on the coherent one-dimensional distance image P _c (m, t) to obtain the multi-frame joint processing distance image P _cI (m, t). P _cI (m,t) can be expressed as

It should be noted that the interferences assumed in the above "Step 2" and "Step 3" are all located within the main lobe beam range of the main radar. If the interference angle _θJ is located in the side lobe of the main radar, it will not affect the processing flow in the above "Step 2" and "Step 3". It is only necessary to expand the interference identification and reference signal selection range in "Step 2" to the entire airspace. That is, the above processing method can deal with multiple main lobe and side lobe interference at the same time.

The present invention provides the following examples to illustrate the inventive method:

The implementation example verification conditions given by the present invention are shown in Table 1. Here, the sparse large-aperture coherent receiving array only contains one auxiliary radar, and its receiving pattern does not have extremely narrow beam conditions, that is, the fast-time-slow-time cascaded LMS filtering processing in the present invention will have certain targets. energy loss; in addition, the interference type used in the implementation example is noise suppression, but the related method of the present invention is also applicable to deceptive interference. That is, the relevant conditions of this embodiment do not affect the specific implementation and technical essence of the present invention.

Table 1 Implementation Example Verification Conditions

Figure 6 shows the one-dimensional range image before and after interference suppression. It can be seen from the figure that the desired signal component in the original echo signal is annihilated by the interference signal. Comparing the filtering results of time-domain LMS and cascaded LMS, it can be seen that the target peak is more obvious after cascaded LMS, which has better interference suppression effect. Table 2 presents the SINR improvement statistics of different filtering processing methods. It can be seen that the SINR after fast-time-slow-time cascaded LMS filtering is improved by more than 7dB compared to the SINR after time-domain LMS filtering.

Table 2 Signal-to-interference and noise ratio (SINR) improvement of different filtering methods

It can be seen from the above results that after using the method proposed by the present invention for interference suppression, the SINR improvement and other indicators are significantly improved, which shows the superiority of the method of the present invention.

To sum up, the above are only preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims (5)

1. a distributed radar main lobe interference suppression method based on cascaded LMS filters, is characterized in that, described distributed radar system is made up of main radar and multiple auxiliary radars; A sparse coherent receiving array is formed around the main radar, and the aperture of the sparse coherent receiving array is larger than that of the main radar. The sparse and dispersed arrangement of the plurality of auxiliary radars means that two adjacent auxiliary radars are spatially independent from each other, and the antennas on them are independent of each other. The spacing between the phase centers is ten to one hundred times the half wavelength, and the time-frequency and timing synchronization equipment is connected between the main radar and each auxiliary radar to ensure the coherence of work; the distributed radar main lobe interference suppression method includes the following steps: Step 1: The main radar beam is aimed at a certain detection airspace to transmit signals, and the main radar echo signal is obtained; the sparse coherent receiving array forms a very narrow beam, and scans the airspace within the main radar main lobe beam range to obtain sparse The array echo of the coherent receiving array at each beam pointing angle; wherein the beam of the sparse coherent array is an extremely narrow beam, and the beam width of the extremely narrow beam is much smaller than the main radar beam width; Step 2: Perform interference identification on the echoes of the sparse large-aperture coherent receiving array to determine whether there is interference and the type of interference; if there are multiple interferences, give the angle of each interference and use the sparse large-aperture at the angle. The coherent receiving array echo is used as the corresponding interference reference signal; Step 3. Under each interference angle, perform the following steps: Taking the fast time echo of the main radar as the main channel signal and the fast time echo of the sparse coherent receiving array as the reference signal, the fast time LMS filter is used to perform the fast time LMS filtering process on the fast time echo of the main radar, and the filtering intermediate processing result is obtained. , repeating the fast-time LMS filtering process multiple times to obtain multiple pulse fast-time LMS filtering processing results; Then, the slow-time echo of the main radar is used as the main channel signal, and the slow-time echo of the intermediate processing result of the fast-time LMS filtering is used as the reference signal, and the slow-time LMS filter is used to perform the slow-time LMS filtering process on the slow-time echo of the main radar. , loop the slow-time LMS filtering process multiple times to obtain the slow-time LMS filtering processing results of multiple sampling points, and then obtain the final cascaded filtering processing results; The cascaded filtering processing results obtained under all interference angles constitute the final interference cancellation processing result; Step 4: Perform pulse compression processing on each sub-pulse in the interference cancellation processing result to obtain multi-frame one-dimensional range images; Step 5: Perform coherent processing on the multi-frame one-dimensional range images to obtain the multi-frame joint accumulation processing result. 2. The method of claim 1, wherein the fast-time LMS filter is an N-order digital filter with a step size of μ. 3. The method of claim 2, wherein the fast-time LMS filtering process specifically comprises the steps: Step1: Set the current processing pulse number to m, and the initial value of m is 1; Step 2: Record the intermediate result of the fast-time LMS filtering weight vector operation at the n-th sampling point of the m-th pulse as

n=1,2,...,Q; where Q is the number of sampling points included in the pulse, and the fast-time sampling time is t _n , n∈[1,…,Q]; And 0≤t ₁ <t ₂ <...<t _Q ≤T, T is the pulse repetition time PRT; When n=1, that is, time t ₁ ,

The adaptive filtering output of the fast-time LMS filter is e(m, _tn )=d(m, _tn ), where d(m, _tn ) is the fast-time LMS filter at the nth sampling point of the mth pulse the main channel signal; Step 3: Determine the reference signal of the fast time LMS filter at the nth sampling point of the mth pulse as: x(m,t _n )=[x(m,t _n ) x(m,t _n-1 ) ... x(m,t _n-N+1 )] ^T ; where x(m,t _n ) x(m,t _n-1 ) … x(m,t _n-N+1 ) are the fast time echoes of the sparse coherent receiving array, respectively, t _n t _n-1 … t _n-N+1 is the fast time sampling point; Then the intermediate result of the fast-time LMS filtering weight vector operation at the n+1th sampling point of the mth pulse is:

where e ^* (m,t _n ) is the conjugate of e(m,t _n ); Then at time t _n+1 , the output signal corresponding to the fast-time LMS filter weighted and summed at the n+1-th sampling point of the m-th pulse, that is, the intermediate processing result of the fast-time LMS filter is y(m, t _{n +1} ):

At time t _n+1 , the adaptive filtering output of the n+1th sampling point of the mth pulse is e(m,tn ₊₁ )=d(m,tn ₊₁ )-y(m,tn _{+ 1} ); Step 4: Let n increment by 1, if n=Q, go to Step 5, otherwise return to Step 3; Step 5: Make the final fast-time LMS filter weight vector

Recalculate time t _n , the result after the weighted summation of the fast-time LMS filter corresponding to the n-th sampling point of the m-th pulse is:

n=1,2,...,Q; Step 6: If m=M, terminate the processing; otherwise, make m increment by 1 and return to Step 2. 4. The method of claim 1, wherein the slow-time LMS filter is a K-order digital filter with a step size of n. 5. The method of claim 4, wherein the slow-time LMS filtering process specifically comprises the following steps: SS1: Set the current processing fast time sampling point number as n, and the initial value of n is 1; SS2: Record the intermediate result of the slow-time LMS filtering weight vector operation at the mth pulse at the nth sampling point as

m=1,2,...,M; where M is the total number of pulses, the instant of fast-time sampling point is t _n , n∈[1,...,Q], Q is the number of sampling points included in the pulse; and 0≤t ₁ <t ₂ <...<t _Q ≤T , T is the pulse repetition time PRT; When m=1, that is, time t ₁ ,

The adaptive filtering output of the slow-time LMS filter corresponding to the mth pulse at time _t1 is: f(m,t _n )=d(m,t _n ), d(m,t _n ) is the main channel signal of the slow-time LMS filter corresponding to the mth pulse of the mth pulse at the nth sampling point ; SS3: The slow-time echo of the intermediate processing result of the fast-time LMS filter is used as the reference signal of the slow-time echo and the slow-time LMS filter, specifically: y(m,t _n )=[y(m,t _n ) y(m-1,t _n ) ... y(m-K+1,t _n )] ^T ; where y(m,t _n ) y(m-1,t _n ) … y(m-K+1,t _n ) are the corresponding values of K pulses before m at time t _n respectively; Then at the m+1th pulse of the nth sampling point, the intermediate result of the slow-time LMS filtering weight vector operation is:

where f ^* (m,t _n ) is the conjugate of f(m,t _n ); Then at the m+1th pulse at the nth sampling point, the result of the weighted summation of the slow-time LMS filter is:

where y(m+1,t _n ) is the reference signal corresponding to the m+1th pulse at the _nth sampling point (time tn); The output result of the slow-time LMS filter adaptive filter corresponding to the m+1th pulse of the nth sampling point is: f(m+1, _tn )=d(m+1, _tn )-z(m+ 1,t _n ); d(m+1,t _n ) is the main channel signal of the slow-time LMS filter corresponding to the m+1th pulse of the nth sampling point. SS4: Let m increase by 1, if m=M, enter SS5, otherwise repeat SS3; SS5: Let the final slow-time LMS filter weight vector

The result of recalculating the weighted summation of the slow-time LMS filter corresponding to the mth pulse at time _tn is:

m=1,2,...,M; The adaptive filtering output of the slow-time LMS filter corresponding to the m-th pulse at time t _n is recalculated as follows: f(m,t _n )=d(m,t _n )-z(m,t _n )m=1,2,...,M SS6: If n=Q, terminate the process, otherwise set n=n+1, and return to SS2.

Images