CN112881982B - Method for restraining dense forwarding interference by frequency agile radar - Google Patents

Abstract

The invention belongs to the field of radar anti-jamming. Aiming at the problem that radar frequency-agility can reduce the number of interference signals entering as much as possible, but a small amount of interference signals entering the radar system will also affect the detection of real targets, a method of frequency-agile radar suppressing intensive forwarding interference is proposed. Three key problems of interference suppression, including range sidelobe suppression, interference rejection, and coherent accumulation, are analyzed for frequency-agile radar, and solutions are given in combination with nonlinear processing, outlier detection, and slow-time matched filtering. The invention can significantly improve the target detection probability in a strong interference environment. Even if the jamming party is fully aware of the radar agility frequency, the critical interference-to-signal ratio for effective detection of real targets after interference suppression can still be improved by about 20dB compared to before suppression.

Description

A method for suppressing dense forwarding jamming in frequency-agile radar

technical field

The invention belongs to the field of radar anti-jamming, and is suitable for solving the problem of intensive forwarding interference suppressed by frequency-agile radar under the condition of high interference-to-signal ratio.

Background technique

Modern warfare is a confrontation between the systems of the two warring parties. With the development and application of advanced weapon platforms, the role of equipment performance in the system is becoming more and more obvious. Radar, as a sensor for sensing targets, is used in shore-based, ship-based and other weapon systems widely. Electromagnetic interference to radar is usually divided into suppression interference and deception interference according to the interference effect. Due to the rapidly changing battlefield situation and the complex deployment of troops, suppression interference, which is less dependent on radar information, is easier to achieve release conditions than deception interference in theory. Dense forwarding jamming is a new type of suppressing jamming based on digital radio frequency storage devices. The jamming signal is generated by intercepting full-pulse radar signals. It does not require radar repetition frequency prior information. Continuous copying and forwarding can effectively suppress the radar system. Compared with noise-like jamming, its power The utilization rate is much higher. Modern radar mostly adopts the fully coherent system. Frequency agility is a common technique used in coherent radar. Detection, MTD) compatibility problem, radar frequency agility can reduce the number of interference signals entering as much as possible, but a small amount of strong interference signals entering the radar system will also affect the real target detection.

At present, the anti-intensive forwarding jamming algorithm of frequency-agile radar is mainly based on the signal layer. The basic idea is to limit the interference by using the characteristic that the peak slow time after the target pulse pressure is linear, and improve the detection of real targets by designing a frequency-agile coherent processing method. probability. Literature [Quan Yinghui, Chen Xiada, Ruan Feng, etc. A frequency-agile joint Hough transform anti-dense false target interference algorithm [J]. Journal of Electronics and Information Technology, 2019, 41(11): 2639-2645], literature [ Dong Shuxian, Quanyinghui, Chen Xiada, et al. Interference suppression algorithm based on frequency-agile combined mathematical morphology [J]. Value voting and corrosion expansion calculation, the above algorithm is easy to set the real echo data to "0" under the condition of low signal-to-noise ratio (SNR), and due to the lack of side lobe suppression link, high interference signal ratio (Jamming to signal ratio) , JSR) it is easy to set the interference side lobe to "1", which affects the scope of application of the algorithm. At the same time, the two-dimensional high-resolution sparse reconstruction theory is introduced to solve the problem of frequency-agile and radar MTD compatibility. This method requires a "distance-velocity" fine search for the slow time signal of the target range unit, which requires a large amount of calculation.

Contents of the invention

The purpose of the present invention is to propose a method for suppressing intensive forwarding interference by frequency-agile radar to solve the problem of suppressing intensive forwarding interference under the condition of high interference-to-signal ratio by frequency-agile radar. The method mainly includes the following steps: step (1) compressing a CPI echo fast time pulse, and performing range sidelobe suppression; step (2) using the interquartile range criterion to sequentially perform interference outliers on the echoes of each range unit Point detection, according to the detection result, the echo IQ data is set to 0 at the same time; step (3) adopts the method of matched filtering, and performs coherent processing on the echo signal after the interference is eliminated.

The above steps are specifically:

Step (1): fast-time matched filtering and range sidelobe suppression for one coherent process interval (CPI) echo to reduce the influence of interference sidelobes on the real echo. Among them, in order to suppress the distance side lobe, an improved side lobe suppression method without widening the main lobe is proposed, and the specific steps are as follows:

(1) Take the echo x _r (t,t ₀ ) of the first repetition period and perform matched filtering without windowing to obtain y _r (t,t ₀ );

(2) Select the window function, and perform windowed matching filtering on the echo x _r (t,t ₀ ), to obtain

取包络，归一化得到z_r(t,t₀)、

(3) For y _r (t,t ₀ ),

Take the envelope and normalize to get z _r (t,t ₀ ),

数值大小，将快时间划分成X、Y两部分，当t∈X时，

(4) According to z _r (t,t ₀ ),

The numerical value divides the fast time into two parts X and Y. When t∈X,

沿快时间拼接，完成第1个重复周期回波旁瓣抑制；(5) Put y _r (t,t ₀ ) in X and y in Y

Stitching along the fast time to complete echo sidelobe suppression in the first repetition period;

(6) According to steps 1-5, complete the side lobe suppression of all echoes within one CPI.

t in step (1) to step (6) is fast time, t _m =mT _r is slow time, m=0,1,2...M-1, M is the number of coherent accumulation, T _r is pulse repetition cycle. In addition, step (2) involves window function selection. Steps (1)-(5) show that the suppressed sidelobe level mainly depends on the window function. In order to minimize the influence of interference sidelobes on the real echo, From the perspective of interference suppression, the window function needs to have low sidelobe and other ripple characteristics. Fixed windows such as Nuttall windows, Flattop windows, Blackman-Harris windows, and parametric windows such as Chebyshev windows are available.

Step (2): Use the modified interquartile range criterion to detect interference outliers in the echoes of each distance unit in turn, and set the interference IQ data to 0 at the same time according to the detection results. The specific steps are:

(1) Take the echo on one distance unit after sidelobe suppression as the sample set X _N ={x ₁ ,x ₂ ,…,x _n ,…,x _N }, and its interquartile range (Inter-Quartile Range ,IQR) is:

IQR _{＝ Q3} -Q1

Among them, Q ₁ and Q ₃ are the 25% and 75% quantiles of the sample set X _N respectively.

(2) At the same time, calculate the normalized skewness of the sample set (Medcouple, MC):

为样本集的50％分位数，即中位数，x_i、x_j为样本值。Among them, med[ ] means to find the median,

is the 50% quantile of the sample set, namely the median, and x _i and x _j are the sample values.

(3) According to IQR and MC, determine the interference outlier detection interval as:

Among them, L is the lower bound of the interval, and U is the upper bound.

(4) When the sample x _n in X _N > U, it is judged that x _n is an interference outlier point. It should be noted that when detecting an outlier point, it is necessary to obtain the envelope of the echo after sidelobe suppression, and remove the interference outlier. In case of cluster points, the echo I and Q data should be set to 0 at the same time according to the position of the interference outlier point, so as to facilitate subsequent coherent processing.

Step (3): Adopting the method of slow-time matched filtering, performing coherent processing on the echo signal after interference elimination, improving the target signal-to-noise ratio and improving the detection efficiency. The specific principle is:

(1) The real echo (RF signal) when the radar frequency is agile is:

Among them, T _p is the pulse width of the radar transmission signal, k=B/T _p is the frequency modulation slope of the transmission signal, B is the bandwidth; σ is the target reflection coefficient, t _d =2R(t _m )/c is the time delay, R(t _m )=R _t -v _t t _m is the radial distance function between the target and the radar, R _t is the initial distance, v _t is the radial velocity, c is the speed of light; f _z (t _m )=f ₁ +a(m) Δf is the radar agile carrier frequency, f ₁ is the initial carrier frequency, a(m) is the frequency code, and Δf is the agile step size.

(2) Distinguish between fixed frequency local oscillator and frequency automatic tracking local oscillator. For fixed frequency local oscillator, local oscillator frequency f _L = f ₁ , for frequency automatic tracking local oscillator f _L = f _z (t _m ), return Wave down-conversion to get:

Among them, s _r1 (t, t _m ) and s _r2 (t, t _m ) are the down-converted echo signals of the fixed-frequency local oscillator and the frequency automatic tracking local oscillator, respectively. It can be seen that the fixed-frequency local oscillator down-converted echo signal has a certain initial frequency, while the frequency automatically tracks the initial frequency of the local oscillator down-converted echo signal to zero (usually referred to as baseband).

(3) According to the radar transmission signal, design two kinds of matching signals, namely:

Wherein, s ₁ (t) is a fixed-frequency local oscillator matching signal, and s ₂ (t) is a frequency automatic tracking local oscillator matching signal.

(4) Using two kinds of matching signals to fix the frequency of the local oscillator and automatically track the frequency of the local oscillator to perform pulse compression on the down-converted echo signal to obtain:

Among them, y _s1 (t, t _m ) and y _s2 (t, t _m ) are the echo signals after the pulse pressure of the fixed-frequency local oscillator and the frequency automatic tracking local oscillator, respectively.

(5) Let t=t _d , the target peak slow time signals are obtained as follows:

Among them, y _s1 (t _d , t _m ) and y _s2 (t _d , t _m ) are the fixed frequency local oscillator and the frequency automatic tracking local oscillator target peak slow time signal respectively. For the fixed-frequency local vibration agile frequency conversion radar, the target peak slow-time signal is a single-frequency signal, which can be directly integrated in theory, but since the echo signal after down-conversion has an initial frequency, the echo digital sampling does not satisfy the sampling theorem , the sampling frequency is required to satisfy f _s ≥ 2{max[|f _z (t _m )-f ₁ |]+B}, and for the frequency automatic tracking local oscillator agility radar, the sampling frequency f _s ≥ 2B is sufficient.

(6) The frequency automatically tracks the slow time signal of the local oscillator radar echo:

与目标径向速度、捷变载频有关，相位求导得到其瞬时频率为

当雷达频率不捷变即f_z(t_m)＝f₁时，

为单频信号(频率为目标多普勒频率)，回波慢时间做快速傅里叶变换(Fast FourierTransform,FFT)可实现相参积累；当雷达步进频即f_z(t_m)＝f₁+t_mΔf时，

为LFM信号，利用分数阶傅里叶变换(FractionalFourier Transform，FRFT)时频旋转特性可使回波能量慢时间上有效聚焦；当雷达频率随机捷变，

内部形式复杂，利用FFT、FRFT等工具无法实现相参积累。从目标检测的角度，对于已知的信号形式，匹配滤波是最优的检测方法，而

可理解部分信息已知的信号(雷达捷变载频已知，目标速度未知)，设定速度搜索范围[v_min,v_max]，构造参考信号

v_x∈[v_min,v_max]，利用

对

进行匹配滤波，当v_x＝v_t时，匹配输出呈现冲击特性，且由于

不存在延时，输出峰值位于慢时间t₀时刻。Among them, the first part φ(t _m , R _t ) is related to the radial distance of the target and the agile carrier frequency, which can be eliminated by compensating the echo phase of each range unit. The second part

It is related to the target radial velocity and the agile carrier frequency, and its instantaneous frequency obtained by phase derivation is

When the radar frequency is not agile, that is, f _z (t _m ) = f ₁ ,

It is a single-frequency signal (the frequency is the target Doppler frequency), and fast Fourier transform (FFT) can be performed on the echo slow time to realize coherent accumulation; when the radar step frequency is f _z (t _m )=f ₁ +t _m Δf,

For LFM signals, using the time-frequency rotation characteristics of Fractional Fourier Transform (FRFT) can effectively focus the echo energy in slow time; when the radar frequency changes randomly,

The internal form is complex, and coherent accumulation cannot be realized by using tools such as FFT and FRFT. From the point of view of target detection, for known signal forms, matched filtering is the optimal detection method, while

Can understand signals with known partial information (radar agility carrier frequency is known, target speed is unknown), set the speed search range [v _min ,v _max ], and construct reference signals

v _x ∈[v _min ,v _max ], using

right

Perform matched filtering, when v _x = v _t , the matching output presents an impact characteristic, and due to

There is no delay, and the output peak is at the slow time t ₀ .

Beneficial effect description of the present invention:

(1) Before the interference is eliminated, the echo is suppressed by the non-range widening sidelobe, and the influence of the interference sidelobe on the real echo is effectively reduced under the premise of maintaining the high range resolution requirement of the radar.

(2) The actual working environment of the radar is complex, and the echo distribution is usually difficult to accurately model. The present invention uses the modified interquartile range criterion to detect interference outliers, and the detection interval is adjusted accordingly with the degree of sample skew.

(3) A frequency-agile coherent processing method based on slow-time matched filtering is proposed. Compared with the two-dimensional high-resolution sparse reconstruction theory, the proposed method only needs to search for the speed, and the computational complexity is lower.

Description of drawings

Accompanying drawing 1 is method step flowchart of the present invention;

Accompanying drawing 2 is the echo pulse compression result after side lobe suppression;

Accompanying drawing 3 is the partial enlargement diagram of side lobe suppression effect;

Accompanying drawing 4 is echo slow time normalized skewness;

Accompanying drawing 5 is an outlier detection interval upper bound;

Accompanying drawing 6 is the echo pulse compression result after interference elimination;

Accompanying drawing 7 is target range unit echo slow time matched filtering result;

Accompanying drawing 8 is the coherent processing result of 1 CPI echo;

Accompanying drawing 9 is the change curve of true target detection rate with interference-to-signal ratio;

Accompanying drawing 10 is the change curve of true target detection rate with the number of interference;

Accompanying drawing 11 is the change curve of the real target detection rate with the interference ratio.

Specific implementation method

A method for suppressing dense forwarding interference of a frequency-agile radar according to the present invention will be described in detail below in conjunction with the accompanying drawings. With reference to accompanying drawing 1, specific implementation steps are as follows:

(1) Fast time pulse compression of a CPI echo, and suppressing the main lobe without widening the distance and side lobe, and reducing the influence of the interference side lobe on the real echo while maintaining the radar's high range resolution requirements;

(2) Use the modified interquartile range criterion to detect interference outliers in the echoes of each distance unit in turn, and set the echo IQ data to 0 at the same time according to the detection results;

(3) The method of slow-time matched filtering is adopted to carry out coherent processing on the echo signal after interference elimination, so as to improve the target signal-to-noise ratio and improve the detection efficiency.

Implementation conditions: The simulation experiment is carried out under the following parameter conditions:

Table 1 Radar, target and jamming parameters

The radar pulse repetition frequency is 2000Hz, the number of coherent accumulations is 128, the normal operating frequency is 1.5GHz, and there are 64 frequency points. The maximum agility is 10% of the operating frequency, and the agility method is pseudo-random agility between pulses; the transmitted signal is LFM pulse signal, pulse width 100μs, bandwidth 2MHz, sampling frequency 8MHz; the self-defense jammer is a point target, the initial distance is 31km, and the radial velocity is 70m/s; False targets are randomly distributed within the range of 23km to 54km. Assuming that SNR is -13dB, JSR is 40dB, and the number of intensive forwarding interference in one repetition period is 5, there are still 32 repetition period echoes that are interfered in the state of radar frequency agility. For one CPI echo fast time pulse compression, And carry out non-broadening distance side lobe suppression (frequency domain windowing, window function uses Nuttall window), the result is shown in Figure 2. In order to display the details, the interference outliers in the circles in Figure 2 are taken, and the results are shown in Figure 3. The normalized skewness is calculated along the slow time for the echo after sidelobe suppression, and the results are shown in Figure 4. According to the normalized skewness of the echoes of each distance unit, the upper bound of the interference outlier detection interval is obtained, as shown in Figure 5. Set the echo IQ data greater than the upper limit of the interval to 0, and the result is shown in Figure 6. Set the speed search range from 18 to 182m/s, and the step length is 1.3m/s, and perform slow-time matching filtering on the echo of the range unit where the target is located after the interference is eliminated (the phase error caused by the distance has been compensated), and the result is shown in Figure 7. According to the above steps, the echoes of each range unit after interference elimination are matched and filtered in slow time, and the signal at time t ₀ is taken as the search result. The result is shown in Figure 8. Distinguish before interference, after interference, and after interference suppression, and use the real target detection rate as an indicator to analyze SNR, JSR, number of interference (the number of interference signals in one repetition period), interference ratio (the number of interference signals in a repetition period of the radar is Proportion within 1 CPI) has an impact on the effectiveness of the present invention. Figure 9, Figure 10, and Figure 11 are the variation curves of the true target detection rate with the interference-to-signal ratio, the number of interferences, and the interference ratio, respectively. It can be seen from Figure 2 that the real echoes of different repetition periods are located in the same distance unit, and the interference is "messy" distributed in each repetition period, showing obvious outlier characteristics; it can be seen from Figure 3 that no window is added under noise conditions The interference peak side lobe level is about -13dB, and the interference peak side lobe level after windowing is about -40dB, and the main lobe is broadened, and the present invention has both advantages, and the interference side lobe is effectively suppressed under the premise that the main lobe does not widen ; It can be seen from accompanying drawing 4 that most of the distance unit echoes have obvious right-biased characteristics (MC>0); it can be seen from accompanying drawing 5 that the upper bounds of different distance unit outlier detection intervals are different; 6, it can be seen that the interference outliers are effectively eliminated, and the remaining interference sidelobe energy is equivalent to that of the target; it can be seen from Figure 7 that the peak value of the target is at the slow time t ₀ ; it can be seen from Figure 8 that the real target gets Effective accumulation, the target distance obtained by peak search is 31.01km, and the radial velocity is 69.72m/s, which is basically consistent with the simulation use of 31km and 70m/s; it can be seen from Figure 9 that the present invention has excellent suppression performance under different JSR conditions. Compared with the comparative literature, the probability of real target detection is significantly improved after interference suppression, and the tolerance to JSR is enhanced in turn. It can be seen from the curve distribution that "interference rejection" contributes the most to the suppression performance, followed by "frequency-agile coherent processing", and "sidelobe suppression" contributes relatively little. The effective maintenance of and the effective improvement of the real target detection rate under high JSR conditions are indispensable in the invention; it can be seen from the accompanying drawings 10 and 11 that the number of interference and the interference ratio have a certain impact on the present invention. The more the number of interferences and the greater the interference ratio, the corresponding decrease in the detection rate of the real target. From the downward trend of the detection rate of the real target, it can be seen that the tolerance of the present invention to the interference ratio is higher than that of the number of interferences.

Claims (5)

1. A kind of frequency-agile radar suppresses intensive forwarding interference method, is characterized in that, comprises the following steps: Step (1) compressing the fast time pulse of one CPI echo and suppressing the range sidelobe; Step (2) using the interquartile range criterion to perform interference outlier detection on the echoes of each distance unit in turn, and setting the echo IQ data to 0 at the same time according to the detection results; Step (3) adopting a matched filtering method to perform coherent processing on the echo signal after interference elimination; Among them, the interquartile range criterion in step (2) is specifically: (1) Suppose the sample set X _N ＝{x ₁ ,x ₂ ,…,x _n ,…,x _N }, the interquartile range is: IQR _{＝ Q3} -Q1 Among them, Q ₁ and Q ₃ are respectively the 25% and 75% quantiles of the sample set X _N ; (2) Calculate the normalized skewness of the sample set:

Among them, med[ ] means to find the median,

is the 50% quantile of the sample set, that is, the median, and both x _i and x _j are sample values; (3) According to the interquartile range and normalized skewness of the sample set, the outlier detection interval is determined as:

Among them, L is the lower bound of the interval, and U is the upper bound; (4) When the sample x _n in the sample set > U, judge that x _n is an interference outlier; Wherein, the method of the matched filtering in the step (3) is specifically: (1) Decompose the target echo slow time signal y _s2 (t _m ) after pulse compression into the product of two parts:

in,

f _z (t _m )=f ₁ +a(m)Δf is the radar agile carrier frequency, f ₁ is the initial carrier frequency, a(m) is the frequency code, Δf is the agile step size, t _d =2R(t _m )/c is the target time delay, R(t _m )=R _t -v _t t _m is the radial distance function between the target and the radar, t _m is the slow time, R _t is the initial distance, v _t is the radial velocity, c is the speed of light; (2) Compensate the phase of each range unit echo after pulse compression in turn

Eliminate φ(t _m ,R _t ), where R _x corresponds to different distance units of echo; (3) Set the speed search range [v _min ,v _max ], and construct a reference signal

v _x ∈[v _min ,v _max ], using

The phase-compensated echoes of each range unit are matched and filtered in turn, and the signal at time t ₀ is taken as the search result to complete the coherent processing of one CPI echo of the frequency-agile radar, where t ₀ is a time point of t _m . 2. a kind of frequency-agile radar according to claim 1 suppresses intensive forwarding interference method, it is characterized in that, the method for the distance sidelobe suppression in step (1) is specifically: (1) Take the echo x _r (t,t ₀ ) of the first repetition period and perform matched filtering without windowing to obtain y _r (t,t ₀ ), where t is the fast time, and t _m =mT _r is the slow time Time, m=0,1,2...M-1, M is the number of coherent accumulation, T _r is the pulse repetition period; (2) Select the window function, and perform window matching filtering on the echo x _r (t,t ₀ ) of the first repetition period, and obtain

(3) For y _r (t,t ₀ ),

Take the envelope and normalize to get z _r (t,t ₀ ),

(4) According to z _r (t,t ₀ ),

The numerical value divides the fast time into two parts X and Y. When t∈X,

(5) Put y _r (t,t ₀ ) in X and y in Y

Stitching along the fast time to complete echo sidelobe suppression in the first repetition period; (6) According to (1) to (5), M repetition period echo distance sidelobe suppression is completed. 3. A method for suppressing dense forwarding interference for frequency-agile radar according to claim 2, characterized in that the window function used in step (2) has low sidelobe and other ripple characteristics. 4. A kind of frequency-agile radar suppresses dense forwarding interference method according to claim 3, is characterized in that, window function specifically selects Nuttall window, Flattop window, Blackman-Harris window or Chebyshev window. 5. A method for suppressing dense forwarding interference for frequency-agile radar according to claim 1, characterized in that the velocity search range [v _min , v _max ] is determined according to the maximum unambiguous distance of the radar and the motion characteristics of the target.

Images