Method for identifying chaff interference based on improved support vector machine

Abstract

The invention discloses a chaff interference identification method based on an improved support vector machine. The chaff interference identification method comprises the following steps: acquiring a target echo signal and a chaff interference echo signal; performing down-conversion on the target echo signal to obtain a first down-conversion signal; performing down-conversion on the chaff interference echo signal to obtain a second down-conversion signal; respectively performing pulse compression processing on the first down-conversion signal and the second down-conversion signal to obtain a first pulse compression signal and a second pulse compression signal; performing dimension reduction processing on the first pulse compression signal and the second pulse compression signal by using a uniform manifold approximation and projection method to obtain a first dimension reduction signal and a second dimension reduction signal; and inputting the first dimension-reduced signal and the second dimension-reduced signal as input signals into a support vector machine identification method for identification. According to the identification method, the problem of low identification efficiency of a traditional support vector machine method is solved, and the overall identification performance of the chaff interference background radar target is improved.

Description

Technical Field

[0001] The invention relates to the field of radar signal processing, and in particular to a method for identifying chaff interference based on an improved support vector machine. Background Art

[0002] When the radar detects a target, it radiates the electromagnetic waves generated by the transmitting antenna to a specific direction in the air, and then detects the target by receiving the echo. The radar target echo information mainly includes: the target's distance, speed, angle and other motion information, radar cross-sectional area, high-resolution one-dimensional image, two-dimensional time-frequency spectrum, polarization characteristics and other target scattering information.

[0003] As an important passive electronic countermeasure, chaff can suppress or deceive radars of different frequencies, directions, and systems, causing radar target detection failure. Chaff is usually made of very thin metal wires or metal-plated glass wires, nylon wires, etc. When the length of the chaff is close to half the wavelength of the interfered radar, it will produce strong resonant reflection of the radar wave.

[0004] A large number of chaff strips form a cloud-like distribution in space. When radar waves shine on the chaff strip cloud, the chaff strip cloud will reflect the radar waves like a real target, thereby forming numerous false target echoes on the radar screen, interfering with the radar's detection and tracking of real targets.

[0005] Chaff jamming technology has been widely used in radar passive countermeasure systems due to its advantages of low price, large jamming range, significant jamming effect and simple operation. Therefore, it is of great practical significance to carry out research on chaff jamming identification technology. Summary of the invention

[0006] The invention aims to provide a method for identifying chaff interference based on an improved support vector machine, which has the advantage of high recognition rate.

[0007] To achieve the above object, the present invention provides a method for identifying chaff interference based on an improved support vector machine, which comprises:

[0008] S10, acquiring a target echo signal and a chaff interference echo signal;

[0009] S20, down-converting the target echo signal to obtain a first down-converted signal; down-converting the foil interference echo signal to obtain a second down-converted signal;

[0010] S30, performing pulse compression processing on the first down-converted signal and the second down-converted signal respectively to obtain a first pulse compression signal and a second pulse compression signal;

[0011] S40, performing dimensionality reduction processing on the first pulse compression signal and the second pulse compression signal by using a uniform manifold approximation and projection method to obtain a first dimensionality reduction signal and a second dimensionality reduction signal;

[0012] S50: Input the first dimension reduction signal and the second dimension reduction signal as input signals into a support vector machine recognition method for recognition.

[0013] Optionally, the first down-converted signal is expressed as:

[0014] S rTarget (t) = A(t)exp[-j2πf c τ+jπk(t-τ) 2 ], where A(t) represents the target echo amplitude; Indicates the target echo delay; c = 3 × 10 8 m / s is the speed of light, R0 is the initial radial distance of the target; v is the target movement speed.

[0015] Optionally, the second down-converted signal may be expressed as: , where N is the total number of foil strips, represents the Doppler shift of the i-th foil strip, a i (t), and v i represent the echo amplitude, radial velocity and radial distance of the i-th foil strip respectively.

[0016] Optionally, the total echo signal of the radar is the sum of the first down-converted signal and the second down-converted signal, and the total echo signal can be expressed as: S r (t) = S rTarget (t)+S sChaff (t).

[0017] Optionally, after step S50, the method further includes step S60, inputting the first pulse compression signal and the second pulse compression signal as input signals into the K nearest neighbor method, the random forest method and the traditional support vector machine method for identification, and comparing the identification results obtained by the K nearest neighbor method, the random forest method and the traditional support vector machine with the results obtained in step S50.

[0018] Optionally, comparing the recognition obtained by the K nearest neighbor method, the random forest method and the traditional support vector machine with the result obtained in step S50 includes comparing the AUC value under the ROC curve.

[0019] In summary, compared with the prior art, the improved support vector machine chaff interference recognition method provided by the present invention has the following beneficial effects:

[0020] The improved support vector machine based foil interference recognition method of the present invention down-converts the echo signal, performs pulse compression processing on the down-converted signal, performs dimensionality reduction processing on the pulse compressed signal, and then uses the support vector machine recognition method to recognize it. The recognition method of the present invention overcomes the problem of low recognition efficiency of the traditional support vector machine method and improves the overall recognition performance of the foil interference background radar target. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 The present invention is a flow chart of the method for identifying chaff interference based on an improved support vector machine.

[0022] Figure 2 This is the target echo image after pulse compression processing.

[0023] Figure 3 This is the echo image at the initial stage of foil strip release and diffusion.

[0024] Figure 4 This is the echo image during the stable period of foil diffusion.

[0025] Figure 5 Schematic diagram of the distribution of the original data set.

[0026] Figure 6 Schematic diagram of the distribution of the data set after dimensionality reduction.

[0027] Figure 7 The figure is a schematic diagram of ROC curve comparison between the method of the present invention and other methods. DETAILED DESCRIPTION

[0028] The following will be combined with the attached embodiment of the present invention Figure 1 ~Attached Figure 7 , the technical solutions, structural features, objectives achieved and effects in the embodiments of the present invention are described in detail.

[0029] It should be noted that the drawings are in a very simplified form and use non-precise proportions. They are only used to conveniently and clearly assist in explaining the embodiments of the present invention, and are not used to limit the conditions for the implementation of the present invention. Therefore, they have no substantive technical significance. Any structural modification, change in proportional relationship or adjustment of size, without affecting the effects and purposes that can be achieved by the present invention, should still fall within the scope of the technical content disclosed by the present invention.

[0030] It should be noted that, in the present invention, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only the elements explicitly listed, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0031] like Figure 1 As shown, the present invention provides a method for identifying chaff interference based on an improved support vector machine, comprising the following steps:

[0032] S10, obtaining the target echo signal and the foil interference echo signal; assuming that the radar transmits a linear frequency modulation signal, the radar transmission signal can be expressed as: s(t) = u(t) exp(j2πf c t+jπkt 2 ), where u(t) represents the complex envelope of the transmitted signal, k represents the frequency modulation slope, and f c Indicates the radar operating center frequency.

[0033] S20, down-converting the target echo signal to obtain a first down-converted signal; down-converting the chaff interference echo signal to obtain a second down-converted signal.

[0034] In this embodiment, the first down-converted signal is expressed as: rTarget (t) = A(t)exp[-j2πf c τ+jπk(t-τ) 2 ], where A(t) represents the target echo amplitude; Indicates the target echo delay; c = 3 × 10 8 m / s is the speed of light, R0 is the initial radial distance of the target; v is the target movement speed.

[0035] The second down-converted signal can be expressed as:

[0036] , where N is the total number of foil strips, represents the Doppler shift of the i-th foil strip, a i (t), v i and r i represent the echo amplitude, radial velocity and radial distance of the i-th foil strip respectively.

[0037] The total echo signal of the radar is the sum of the first down-conversion signal and the second down-conversion signal. The total echo signal can be expressed as: S r(t) = S rTarget (t)+S sChaff (t).

[0038] S30, respectively perform pulse compression processing on the first down-converted signal and the second down-converted signal to obtain a first pulse compression signal and a second pulse compression signal. In this embodiment, the parameters used for simulation are shown in Table 1. The target echo after pulse compression processing is as follows: Figure 2 As shown, the interference effect of chaff cloud is as follows Figure 3 and Figure 4 As shown, Figure 3 This is the echo image of the early stage of foil strip release and diffusion. Figure 4 This is the echo image of the foil strip diffusion stability period. Figure 2 , Figure 3 and Figure 4 It can be seen that before the radar is interfered, the echo can clearly see the target after pulse compression. In the early stage of chaff release, the target signal is submerged by the interference signal, resulting in target detection failure. Figure 3 It can be seen that the foil cloud echo has multiple strong scattering centers and the one-dimensional range image is sparse. Figure 4 This is the mixed echo signal of the chaff cloud in the stable stage. At this time, the chaff cloud is in a stable stage, the diffusion radius of the chaff cloud in the air has basically reached the maximum, the density is small, and the target is still submerged by the chaff interference. Figure 2 , Figure 3 and Figure 4 As shown, the one-dimensional range image of the foil interference is denser than the target echo in terms of strong scattering points.

[0039] S40, using uniform manifold approximation and projection method to perform dimensionality reduction processing on the first pulse compression signal and the second pulse compression signal to obtain a first dimensionality reduction signal and a second dimensionality reduction signal, and constructing an interference recognition data set with the first dimensionality reduction signal and the second dimensionality reduction signal after dimensionality reduction.

[0040] S50: Input the first dimension reduction signal and the second dimension reduction signal as input signals into a support vector machine recognition method for recognition. Figure 5 This is a schematic diagram of the distribution of the original data set. Figure 6 This is a schematic diagram of the distribution of the data set after dimensionality reduction. Figure 5 and Figure 6 As shown in the comparison, the echo data distribution of the first dimensionality reduction signal and the second dimensionality reduction signal after dimensionality reduction is more concentrated, and the number of discrete points is greatly reduced.

[0041] In this embodiment, after step S50, step S60 is also included, the first pulse compression signal and the second pulse compression signal are input as input signals to the K-Nearest Neighbors method (KNN), the Random Forest (RF) method and the traditional Support Vector Machine (SVM) method for recognition, and the recognition obtained by the K-Nearest Neighbors method, the Random Forest method and the traditional Support Vector Machine is compared with the result obtained in step S50. The recognition method proposed in the present invention, i.e., the Modified Support Vector Machine based on Uniform Manifold Approximation and Projection (MSVM-UMAP), is compared with the recognition results of KNN, RF, SVM and other methods. As shown in Tables 2 and 3, the recognition accuracy and running time of each method under different interference-to-signal ratios are compared and analyzed.

[0042] The recognition results obtained by the K nearest neighbor method, the random forest method and the traditional support vector machine are compared with the results obtained in step S50, including comparing the AUC (Area Under the Curve) value under the ROC (Receiver Operating Characteristic Curve) curve. In order to evaluate the overall performance of each method, the ROC curve and AUC value of each method are compared and analyzed. The evaluation results are as follows: Figure 7 As shown in Table 4. It can be seen from Table 2 and Table 3 that the interference recognition accuracy of the MSVM-UMAP method proposed in the present invention and the traditional SVM method are similar and both are better than the other two methods, but the interference recognition efficiency of the MSVM-UMAP method proposed in the present invention is better than the traditional SVM method.

[0043] Depend on Figure 7 As can be seen from Table 4, the ROC curve of the MSVM-UMAP method proposed in the present invention is closest to the upper left corner and has the highest AUC value, that is, the overall recognition performance of the method proposed in the present invention is the best.

[0044] From the comparison results, it can be seen that the identification method of foil interference based on the improved support vector machine of the present invention improves the interference identification efficiency, retains the characteristics of the original data to the greatest extent while greatly reducing the feature dimension, and uses the UMAP (the Modified Support Vector Machine) method to map the original echo data to a low-dimensional space, which reduces the calculation time and helps to improve the overall interference identification performance of the algorithm.

[0045] Table 1 Main simulation parameters

[0046] name Numeric Carrier frequency 17GHz bandwidth 4MHz Pulse Width 20μs Pulse repetition frequency 1kHz Number of foil strips 1 million Wind speed 10m / s Time step 0.005s

[0047] Table 2 Comparison of recognition results of different methods (echo interference signal ratio is 14.3dB)

[0048] method Accuracy Time consumed / s MSVM-UMAP 97.8％ 0.087 SVM 97.9％ 0.389 KNN 97.6％ 0.261 RF 96.4％ 0.869

[0049] Table 3 Comparison of recognition results of different methods (echo interference signal ratio is 17.6dB)

[0050]

[0051]

[0052] Table 4 Comparison of AUC values ​​of various methods (interference-to-signal ratio is 14.3dB)

[0053] method MSVM-UMAP SVM RF KNN AUC value 0.994 0.979 0.973 0.98

[0054] Although the content of the present invention has been described in detail through the above preferred embodiments, it should be appreciated that the above description should not be considered as a limitation of the present invention. After reading the above content, it will be apparent to those skilled in the art that various modifications and substitutions of the present invention will occur. Therefore, the protection scope of the present invention should be limited by the appended claims.

Claims

1. A method for identifying chaff interference based on an improved support vector machine, characterized in that: The identification method comprises: S10, acquiring a target echo signal and a chaff interference echo signal; S20, down-converting the target echo signal to obtain a first down-converted signal; down-converting the foil interference echo signal to obtain a second down-converted signal; S30, performing pulse compression processing on the first down-converted signal and the second down-converted signal respectively to obtain a first pulse compression signal and a second pulse compression signal; S40, performing dimensionality reduction processing on the first pulse compression signal and the second pulse compression signal by using a uniform manifold approximation and projection method to obtain a first dimensionality reduction signal and a second dimensionality reduction signal; S50: Input the first dimension reduction signal and the second dimension reduction signal as input signals into a support vector machine recognition method for recognition.

2. The identification method according to claim 1, characterized in that: The first down-converted signal is expressed as: rTarget (t) = A(t)exp[-j2πf c τ+jπk(t-τ) 2 ], where A(t) represents the target echo amplitude; Indicates the target echo delay; c = 3 × 10 8 m / s is the speed of light, R0 is the initial radial distance of the target; υ is the target movement speed.

3. The identification method according to claim 2, characterized in that: The second down-converted signal can be expressed as: ， Where N is the total number of foil strips, represents the Doppler shift of the i-th foil strip, a i (t), v i and r i represent the echo amplitude, radial velocity and radial distance of the i-th foil strip respectively.

4. The identification method according to claim 3, characterized in that: The total echo signal of the radar is the sum of the first down-converted signal and the second down-converted signal. The total echo signal can be expressed as: S r (t)=S rTarget (t)+S sChaff (t)。 5. The identification method according to claim 1, characterized in that: After step S50, the method also includes step S60, inputting the first pulse compression signal and the second pulse compression signal as input signals into the K nearest neighbor method, the random forest method and the traditional support vector machine method for identification, and comparing the identification results obtained by the K nearest neighbor method, the random forest method and the traditional support vector machine with the results obtained in step S50.

6. The identification method according to claim 5, characterized in that: Comparing the recognition obtained by the K nearest neighbor method, the random forest method and the traditional support vector machine with the result obtained in step S50 includes comparing the AUC value under the ROC curve.

Images