A radar jamming effect evaluation method under non-cooperative condition with adaptive parameter adjustment

Abstract

The application discloses a radar jamming effect evaluation method under non-cooperative condition with adaptive parameter adjustment, which comprises the following steps: S1, detecting and collecting original radar signal parameters; S2, establishing a time sequence parameter index system before jamming; S3, releasing jamming, and detecting and collecting radar signal parameters after jamming; S4, establishing a time sequence parameter index system after jamming; S5, obtaining signal feature vectors before and after jamming by using a common principal component analysis method; and S6, calculating the change of the signal feature vectors before and after jamming by using an adaptive weighted dynamic time warping method, so as to realize jamming effect evaluation. The application establishes a time sequence parameter index system by using radar signal parameters before and after jamming, reduces the dimension of the time sequence parameter index system by using a common principal component analysis method, and finally calculates the change of the signal feature vectors before and after jamming by using an adaptive weighted dynamic time warping method, so as to realize jamming effect evaluation. The application has the advantages of adaptive parameter adjustment, rapidness and strong generalization ability.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of radar countermeasures, and particularly relates to a radar jamming effect evaluation method under non-cooperative conditions with adaptive parameter adjustment. BACKGROUND

[0002] The radar jamming effect evaluation specifically refers to global analysis and evaluation of the jamming received by the radar system, which can provide direct guidance for jamming mode selection and strategy optimization, and is helpful for guiding operational decision and tactical deployment. In electronic warfare, how to quickly and accurately comprehensively analyze and evaluate the effect of the jamming implemented by the own side has become a prerequisite for accurate jamming.

[0003] The traditional radar jamming effect evaluation method mainly relies on cooperation, that is, the jamming effect is evaluated according to the labels and sample quantities given by the target radar. The current main cooperative evaluation methods include template matching, decision tree (DT), support vector regression (SVR), convolutional neural network (CNN) and autoencoder method, etc. However, in modern warfare, the target radar and the jammer are under non-cooperative conditions, and show high antagonism. Under non-cooperative conditions, the jammer cannot determine the degree of jamming and the change of the target radar working mode, and it is unrealistic to obtain the labels and sample quantities given by the target radar, we can only evaluate the jamming effect through the parameter change of the target radar. For example, when the target radar implements anti-jamming measures, the parameters such as bandwidth, pulse width, transmission power, frequency agility range and rate will change; when the target radar switches the working mode, the parameters such as pulse repetition frequency and beam dwell time will change. It has become an important task to evaluate the jamming effect of the target radar only from the received radar signal parameters of the jammer.

[0004] The document "T. Tian, Z. Feng, Y. Li, et al. Performance evaluation of deception against synthetic aperture radar based on multifeature fusion. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing. 14 (2020): 103-115." adopts a method based on a multi-feature fusion convolutional neural network to evaluate the deception jamming performance of synthetic aperture radar. The method extracts explicit indicators and implicit indicators for typical explicit deception jamming evaluation based on a convolutional neural network, and uses multi-feature fusion to evaluate the jamming effect. The method has high effectiveness on the moving target and static target capture recognition database. However, the deep learning method can only be used under cooperative conditions. When the target radar label is unknown, the target radar data presents dynamic pseudo-random variation characteristics for the jammer, and the performance will decrease significantly, and the model can only be retrained.

[0005] The document "G. Xu, Y. Zhang, W. Huo, et al. A Cognitive Jamming Decision-making Method for Multi-functional Radar Based on Threat Assessment. 2023 IEEE Radar Conference (RadarConf23). San Antonio, USA, 2023." proposes a jamming effect evaluation method based on threat assessment. It assumes that the jamming system can detect the three-dimensional spatial motion state of the target radar, and uses the change of the target radar motion information and operational parameters to propose a threat assessment model for airborne radar to evaluate the jamming effect. The method can achieve good results in the specific environment modeled by them. However, the method used has a large number of hyperparameters. The setting of hyperparameters is mainly based on expert knowledge, and depends on the experience and subjective judgment of experts, which may cause overfitting. Once the war environment changes, the performance will be greatly reduced. SUMMARY

[0006] The purpose of the present application is to overcome the shortcomings of the prior art and provide a radar jamming effect evaluation method under non-cooperative conditions with adaptive parameter adjustment. The present application can quickly evaluate the jamming effect under non-cooperative conditions while realizing the characteristics of adaptive parameter adjustment.

[0007] The object of the application is achieved by the following technical solution: a radar jamming effect evaluation method under non-cooperative conditions with adaptive parameter adjustment, comprising the following steps:

[0008] S1, detecting original radar signal parameters, the radar signal parameters including time of arrival, pulse width, carrier frequency, bandwidth, pulse amplitude;

[0009] S2, establishing a pre-jamming time sequence parameter index system for the detection parameters in step S1;

[0010] S3, releasing jamming and detecting post-jamming radar signal parameters;

[0011] S4, establishing a post-jamming time sequence parameter index system for the detection parameters in step S3;

[0012] S5, using a common principal component analysis method to obtain signal feature vectors before and after jamming;

[0013] S6, using an adaptive weighted dynamic time warping method to calculate the change of the signal feature vectors before and after jamming, thereby realizing jamming effect evaluation.

[0014] The radar jamming effect evaluation method under non-cooperative conditions with adaptive parameter adjustment of the application uses radar signal parameters detected before and after jamming to establish a time sequence parameter index system based on the time sequence parameter index system. A common principal component analysis method is used to reduce the dimension of the time sequence parameter index system, and signal feature vectors before and after jamming are obtained. An adaptive weighted dynamic time warping method is used to calculate the change of the signal feature vectors before and after jamming, thereby realizing jamming effect evaluation. The application realizes the characteristics of adaptive parameter adjustment while quickly evaluating the jamming effect under non-cooperative conditions, and has the advantages of adaptive parameter adjustment, rapidness and strong generalization ability. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 A radar jamming effect evaluation scheme flowchart is provided for the embodiments of the application.

[0016] Figure 2 A jamming success scenario schematic diagram is provided for the embodiments of the application.

[0017] Figure 3 A jamming failure scenario schematic diagram is provided for the embodiments of the application.

[0018] Figure 4 A scenario jamming effect simulation schematic diagram is provided for the embodiments of the application. DETAILED DESCRIPTION

[0019] The technical solutions of the application are further described below with reference to the drawings.

[0020] As Figure 1As shown, the adaptive parameter adjusting radar jamming effect evaluation method under non-cooperative condition of the present application comprises the following steps:

[0021] S1, detecting original radar signal parameters, the confrontation between the target radar and the own jammer can be regarded as a game process, and the beam dwell time of the radar can be regarded as a round of game, which is called "jamming round". In each jamming round, the radar tries to transmit several pulse signals to detect the target, while the jammer tries to interfere with the radar signal to hinder its function. The original radar signal is the time domain signal of the target radar without being interfered. The radar signal parameters specifically include time of arrival, pulse width, carrier frequency, bandwidth and pulse amplitude.

[0022] S2, establishing a jamming before timing parameter index system for the detection parameters of step S1; the timing parameter index system is denoted as JEIS, the time of arrival is denoted as TOA, the pulse width is denoted as PW, the carrier frequency is denoted as CF, the pulse amplitude is denoted as PA, and the bandwidth is denoted as BW.

[0023] A jamming round contains several radar pulses, and the total number of pulses in different jamming rounds is different. Therefore, JEIS is modeled as a two-dimensional array, where each row represents the number of pulses in a jamming round, and each column represents the category of the parameter, which is in the form of:

[0024]

[0025] Wherein represents the time of arrival of all pulses in a jamming round, represents the width of all pulses in a jamming round, represents the carrier frequency of all pulses in a jamming round, represents the bandwidth of all pulses in a jamming round, represents the pulse amplitude of all pulses in a jamming round, and n is the number of pulses.

[0026] In addition to the explicitly listed parameters, JEIS also implicitly reflects the number of pulses, beam dwell time, frequency agility range, pulse repetition interval (PRI), received power and other parameters related to jamming. Assuming that the number of pulses in the jamming round before releasing the jamming is n1, then the signal before jamming is

[0027] S3, releasing the jamming, detecting the radar signal parameters after jamming, the radar signal parameters are the same as S1, and also include the time of arrival, pulse width, carrier frequency, bandwidth and pulse amplitude;

[0028] S4, a system of time sequence parameter indexes after interference is established for the intercepted parameters in step S3, and the expression method is the same as that in step S2. Assuming that the number of pulses in the interference round after the interference is released is n2, the time sequence parameter indexes of the signal after the interference are

[0029] S5, the dimension of the time sequence parameter index system established in steps S2 and S4 is reduced by using a common principal component analysis method to obtain the signal feature vectors before and after the interference. The specific implementation method is as follows: the JEIS before and after the interference is projected onto a common subspace, and the Euclidean distance between the projections of the JEIS before and after the interference on the common subspace is the farthest. The feature vector of the common subspace is the common principal component. This method reduces the computational complexity and ensures the isomorphism between the same dimensions of before and after the interference, that is, the physical meaning of the same dimensions of is the same.

[0030] The two JEIS before and after the interference are denoted as JEIS1 and JEIS2, and the signal feature vectors before and after the interference are denoted as and The specific calculation process includes the following steps:

[0031] S51, each column of JEIS1 and each column of JEIS2 is normalized to the range of (-1, 1) by using the Z-score method:

[0032]

[0033] i1 and i2 are the i-th elements in and respectively, i'1 and i'2 are the corresponding normalized elements, n1 is the number of pulses in the interference round before the interference is released, and n2 is the number of pulses in the interference round after the interference is released. That is, each column parameter is subtracted by the mean value and divided by the variance to complete the normalization, so that the processed parameter index system approximately conforms to the standard normal distribution.

[0034] S52, the covariance matrices of the time sequence parameter index system JEIS1 before the interference and the time sequence parameter index system JEIS2 after the interference are calculated:

[0035]

[0036] wherein represents a centering operation, that is, each column vector of JEIS is subtracted by the mean value of the column vector.

[0037] S53, the common covariance matrix is calculated:

[0038] Σ = (Σ1+Σ2) / 2

[0039] S54, calculate the common principal component:

[0040] S=SVD(∑)(:,1)

[0041] where SVD(∑) represents the singular value decomposition matrix of ∑, and (:,1) represents taking the first column vector of the matrix as the common principal component;

[0042] S55, calculate the signal feature vectors before and after interference:

[0043]

[0044] where · represents matrix multiplication. The obtained dimension reduction.

[0045] S6, use the adaptive weighted dynamic time warping method to calculate the change of the signal feature vectors before and after interference, so as to realize the interference effect evaluation.

[0046] The specific implementation method of step S6 is: convert the interference effect evaluation problem into the similarity measurement problem of the feature vectors before and after interference; let define the interference distance matrix D The calculation method of the element D(i,j) in the matrix is: the distance from a certain component of after interference to a certain component of before interference; the formula is expressed as:

[0047] D(i,j)=y j -x i

[0048] where

[0049] The mathematical modeling of the dynamic time warping algorithm is: solve the dynamic time warping distance using the interference distance matrix D. In the present interference effect evaluation problem, the dynamic time warping distance is the interference effect evaluation value r. First, define the warping path Route in the interference distance matrix D, which represents the matching relationship between each component of and . The warping path is a set containing K component pairs, which is defined as:

[0050] Route={(i1,j1),(i2,j2),…,(i k ,j k ),…,(i K ,j K )}

[0051] where (i k ,j kLet ) represent the row and column numbers corresponding to a certain element in the interference distance matrix D; the distorted path Route needs to satisfy the following three conditions:

[0052] Boundary conditions: i1 = j1 = 1, i K =n1,j K =n2;

[0053] Monotonic condition: i k ≤i k+1 ,j k ≤j k+1 ;

[0054] Continuity condition: 0≤i k+1 -i k ≤1,0≤j k+1 -j k ≤1;

[0055] Secondly, it is necessary to find a twisted path (Route) whose sum of distances represents the twisted path. The minimum value for this twisted path (Route) represents the optimal twisted path. The sum of the distances represented by the optimal twisted path corresponds to the dynamic time twist distance, which is the interference effect evaluation value r, expressed as:

[0056]

[0057] The dynamic time warp distance described above is solved using dynamic programming.

[0058] Specifically, the steps for solving the above dynamic time warp distance using dynamic programming are as follows:

[0059] S61. Calculate the pulse number factor l:

[0060] l = min(n1,n2) / max(n1,n2)

[0061] Where n1 represents The length of n2 represents Length;

[0062] S62. Calculate the interference range matrix D;

[0063] S63, Calculate the dynamic programming matrix

[0064]

[0065] Where i = 1, 2, ..., n1, j = 1, 2, ..., n2. The adaptive weight w belongs to each pulse. i =e 0.05lkwhere k represents the number of times a certain pulse is matched during the dynamic programming. When a certain pulse is reused, the value of k increases, which increases the distance of the path to improve the abnormal matching problem. Due to the agility of radar parameters (residence and switching, sliding, jitter, interleaving, etc.), using DTW can cause a certain component of one to match a plurality of components on the other , resulting in inaccurate similarity measurements, referred to as abnormal matching. An adaptive weighting method is used to improve the dynamic time warping method to measure the similarity of before and after the interference. The method includes automatically weighting each component of , and dynamically adjusting the weight of each component to optimize dynamic programming. As the matching frequency of the components of increases, the weight gradually decreases, resulting in an increased matching distance with other points and a reduced likelihood of re-matching.

[0066] S64, get the interference effect evaluation value r:

[0067] r = DP(n1, n2).

[0068] The interference effect evaluation performance of a successful interference case scenario and a failed interference case scenario is experimentally tested. It is assumed that the target radar has four working modes: velocity search (VS), range while scanning (RWS); scan while tracking (TWS) and single target tracking (STT). At the same time, the target radar can counteract the jamming by parameter agility in the same mode. The working mode parameters of the target radar are shown in Table 1, and there are five parameters: pulse repetition interval (PRI), pulse width (PW), carrier frequency (CF), bandwidth (BW) and pulse amplitude (PA). Each parameter can be parameter agile, and the radar parameter agility method includes: residence switching A, B, which means that any one of the parameter values A, B changes to the other after appearing continuously for several times within the pulse number n, and so on; interleaving [A, B, C], which means that the parameter value cycles with A, B, and C as the list; sliding A:C, which means that the parameter value uniformly increases from A to C, a total of n; jitter (A, B), which means that the parameter value is randomly selected from the range A to B.

[0069] It is worth mentioning that the jammer does not know this parameter template. On the contrary, it releases jamming according to the received radar parameters and uses the proposed method to evaluate the interference effect, which conforms to the assumption of non-cooperative conditions. Before the appearance of the jammer, the target radar is in VS mode, which means a low alert state. From a certain interference round, the radar detects an anomaly and starts to measure the range of the jammer, switching to RWS mode. As Figure 2 and Figure 3As shown, the present application simulates two scenarios of jammer interference success and failure in a non-cooperative environment: in the first scenario, the jammer successfully interferes with the target radar, causing it to return to VS mode; in the second scenario, the jammer fails to interfere, causing the target radar to switch to STT mode.

[0070] After generating the timing parameters for the two scenarios according to Table 1, a complex electromagnetic environment is simulated using Gaussian white noise to obtain the parameters of each jamming round at the receiver of the jammer. For each scenario, there are six jamming round transitions, so six jamming effect evaluation values r for each scenario can be calculated, as shown in Equation (1). Figure 4 As can be seen from the figure, the jamming effect evaluation using the method is consistent with the experimental scenario setting. If the radar operating mode switches to a higher threat level, r will be a relatively large negative number. Conversely, if the radar operating mode switches to a lower threat level, r will be a relatively large positive number. In addition, when the target radar does not switch the operating mode, the method can also analyze the anti-jamming measures taken by the radar, thereby providing an exact jamming evaluation value.

[0071] Table 1 Target Radar Parameter Table

[0072]

[0073] Those skilled in the art will realize that the embodiments described herein are for the purpose of helping the reader understand the principles of the present application and should be understood as not limiting the scope of protection of the present application to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations according to the technical inspiration disclosed by the present application without departing from the spirit of the present application, and these modifications and combinations are still within the scope of protection of the present application.

Claims

1. A radar jamming effect evaluation method under non-cooperative conditions with adaptive parameter adjustment, characterized in that, The method comprises the following steps: S1, intercepting original radar signal parameters, the radar signal parameters including time of arrival, pulse width, carrier frequency, bandwidth, pulse amplitude; S2, establishing a pre-interference timing parameter index system for the intercepted parameters in step S1; S3, releasing interference, intercepting radar signal parameters after interference; S4, establishing a post-interference timing parameter index system for the intercepted parameters in step S3; S5, using the public principal component analysis method, the signal feature vector before and after the interference is obtained; the specific implementation method is: the two signals before and after the interference are respectively recorded as and , and the signal feature vectors before and after the interference are respectively recorded as and ;​ The specific calculation process comprises the following steps: S51, to each column of and each column of was normalized to the range (-1, 1) using the Z-score method: ； ； and are respectively and the i-th element in and are the corresponding normalized elements; is the number of pulses of the interfering round before the release of the interference, is the number of pulses of the interfering round after the release of the interference; S52, compute and covariance matrix of ； ； wherein represents a decentralized operation, i.e. each column vector of is subtracted by the mean of that column vector; S53, calculating a common covariance matrix: ； S54, calculating common principal components: ； wherein denotes the singular value decomposition matrix of denotes taking the first column vector of the matrix as the common principal component; S55, calculating signal feature vectors before and after interference: ； ； wherein denotes matrix multiplication; S6, using the dynamic time warping method of adaptive weighting, the change of signal feature vector before and after the interference is calculated, so as to realize the interference effect evaluation; The specific implementation method is: the interference effect evaluation problem is converted into the similarity measurement problem of feature vectors before and after the interference; Set , ; define the interference distance matrix , the calculation method of the element in the matrix is: ； wherein ; The mathematical modeling of the dynamic time warping algorithm is: using the interference distance matrix to solve the dynamic time warping distance; first, define the warping path in the interference distance matrix , which represents the matching relationship between each component of and ; the warping path is a set containing component pairs, defined as: ； wherein is the interference distance matrix the row number and column number corresponding to a certain element in the interference distance matrix; the twisted path need to meet the following three conditions: Boundary conditions: ; Monotonicity condition: ; Continuous condition: ; Second, a warping path needs to be found that minimizes the sum of the distances represented by the warping path The warping path that minimizes the sum of the distances is the optimal warping path; the sum of the distances represented by the optimal warping path is the dynamic time warping distance, which is the interference effect assessment value ； The dynamic programming method is used to solve the dynamic time warping distance; the specific steps are as follows: S61, calculate the pulse number factor : ； wherein represents the length of represents the length of S62, compute interference distance matrix ; S63, compute dynamic programming matrix : ； wherein , ; adaptive weight belonging to each pulse wherein denotes the number of times a certain pulse is matched during the dynamic programming; S64, obtain an interference effect evaluation value : ； The timing parameter indicator system in steps S2 and S4 is denoted by ; is modeled as a two-dimensional array, where each row represents the number of pulses within an interference round and each column represents a category of parameters.

2. The radar jamming effect evaluation method under non-cooperative condition with adaptive parameter adjustment according to claim 1, characterized in that, The timing parameter indicator system in the steps S2 and S4 is denoted as , the time of arrival is denoted as TOA, the pulse width is denoted as PW, the carrier frequency is denoted as CF, the pulse amplitude is denoted as PA, and the bandwidth is denoted as BW. In the form of: ； wherein denotes the arrival time of all pulses within an interfering round, denotes the width of all pulses within an interfering round, denotes the carrier frequency of all pulses within an interfering round, denotes the bandwidth of all pulses within an interfering round, denotes the pulse amplitude of all pulses within an interfering round, n being the number of pulses.

Images