An array radar main lobe interference suppression method based on array virtual extension

Abstract

The application discloses an array radar main lobe interference suppression method based on array virtual expansion, estimates the non-circularity of a radar transmitting signal, adopts a conjugate mode to expand the steering vector of a radar receiving signal and an expected signal, and then realizes array virtual expansion; array elements are selected from the array after virtual expansion, and a waveform after interference suppression is obtained through beam forming based on the selected array elements; when the transmitting signal is a non-circular signal, the method has good main lobe and side lobe anti-interference performance, does not need to actually expand the array aperture, has strong robustness, and is suitable for different forms of array radars with different array element numbers.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of radar anti-jamming, and particularly relates to an array radar main lobe jamming suppression method based on array virtual extension. BACKGROUND

[0002] Adaptive beamforming technology is widely used in radar, sonar, communication and radar-communication integrated system. The technology uses sensor receiving data to adaptively point the main lobe peak of the directional diagram to the desired signal direction, and forms adaptive nulls in the interference direction and suppresses them, so as to achieve the maximum output signal-to-interference-and-noise ratio of the algorithm. However, with the increase of the number of electromagnetic devices, the electromagnetic environment becomes more and more complex. When the interference falls within the main lobe range of the directional diagram, the main lobe distortion, main lobe peak shift and side lobe level rise problems of the directional diagram generated by the classical adaptive beamforming technology occur, which greatly reduces the output signal-to-interference-and-noise ratio of the algorithm and the gain of the desired signal direction, greatly affecting the detection of the target by the radar.

[0003] In view of the problems caused by the main lobe interference, scholars have proposed different main lobe interference suppression algorithms. These algorithms can be roughly divided into the following four categories: the first category is the pre-processing directional diagram shape-preserving algorithm, such as the blocking matrix pre-processing (BMP) algorithm, which needs to accurately know the angle of the main lobe interference and will lose the system degrees of freedom; the eigen-projected pre-processing and covariance matrix pre-processing (EMP-CMR) algorithm will still have the problem of main lobe peak shift when the echo contains the desired signal, so this kind of algorithm can improve the main lobe distortion problem to some extent, but it does not improve the output signal-to-interference-and-noise ratio and the desired signal gain. The second category is the blind source separation (BSS) algorithm, such as the independent component analysis (ICA) algorithm, but this kind of algorithm has been proved to be equivalent to the classical adaptive beamforming technology, and although it can correctly separate the desired signal, it still faces the problem of main lobe interference. The third category is to use the extended array to suppress the main lobe interference, such as the distributed radar which converts the main lobe interference into side lobe interference for suppression, but this kind of algorithm needs a large space position, which is difficult to realize in practical application. The last category is to use the multi-domain joint method to suppress the side lobe interference through the spatial information and suppress the main lobe interference through other domain information, such as the space-polar joint method and the space-time joint interference suppression method, but this kind of algorithm needs to use additional receiving information, which increases the difficulty of designing the receiver.

[0004] In summary, the existing main lobe interference suppression algorithms mainly have the problems of main lobe peak shift, need a large space position or need additional receiving information, which leads to the unsatisfactory output signal-to-interference-and-noise ratio and desired signal gain, or difficulty in design and implementation. SUMMARY

[0005] Therefore, the application provides an array radar main lobe interference suppression method based on array virtual extension, which realizes interference suppression on the main lobe and the side lobe.

[0006] The application provides an array radar main lobe interference suppression method based on array virtual extension, which comprises the following steps:

[0007] The non-circularity of the radar transmitting signal is estimated; the radar receiving signal and the steering vector of the expected signal are extended in a conjugate manner based on the estimated non-circularity and the radar receiving signal, so as to realize virtual extension of the radar array; then, the extended array elements are selected by using a genetic algorithm, and the selected elements are used for beam forming, so as to obtain a waveform after interference suppression.

[0008] Further, the method further comprises:

[0009] It is assumed that there are M signal sources in the space incident to the radar array, the M signal sources include one expected signal source, one main lobe interference signal source and M-2 side lobe interference signal sources, wherein the angle between the direction in which the mth signal source is incident to the radar array and the normal of the array is θ m ; it is assumed that the positions of the signal sources and the radar array are far enough to meet the far field condition, and it is assumed that the leftmost element of the radar array is a reference element, i.e., the zero point of the coordinate axis, so that the radar receiving signal is represented as:

[0010]

[0011] wherein A(θ)=[a(θ0),...,a(θ M-1 )] is an array flow pattern matrix representing the response of the radar array to the space signal source; is a steering vector, representing the response of the radar array to the signal source with the direction of θ m ; S(t)=[s0(t),...,s M-1 (t)] T is a receiving signal vector, representing the signal complex envelope value of the space signal source at the tth moment, s m (t) representing the complex envelope of the mth signal source; N(t)=[n1(t),...,n N-1 (t)] T is an array receiving noise vector.

[0012] Further, the process of extending the radar receiving signal and the steering vector of the expected signal in a conjugate manner is:

[0013] When the expected signal of the radar receiving signal is a non-circular signal, the conjugate of the expected signal is decomposed as:

[0014]

[0015] Among them, the superscript [·] * This indicates the conjugate operation. Describes the orthogonal subspace of the desired signal. Let V be the variance of the desired radar signal received; therefore, the radar received signal vector of the virtual extended array is:

[0016]

[0017] in, IN(t) represents the desired signal steering vector of the virtual extended array, IN'(t) represents the interference plus noise vector of the virtual extended array, and IN'(t) represents the interference plus noise vector that includes the portion perpendicular to the desired signal after conjugation of the desired signal.

[0018] Furthermore, the method for selecting the expanded array elements using a genetic algorithm is as follows:

[0019] Step 4.1: Guide the desired signal vector of the virtual extended array. Search guide vector of virtual extended array Received signal of virtual extended array Let T be the number of iterations for the genetic algorithm and pop be the population size. size Initialize the initial population (pop0) and selection rate (μ). s Crossover rate μ c and the rate of variation μ m ;

[0020] Step 4.2: Calculate the current population pop. t-1 The fitness function;

[0021] Step 4.3: Use the roulette wheel method to obtain the parent population population (pop). A and offspring population pop B ;

[0022] Step 4.4: Pop the parent population using a single-point crossover method. A Cross over;

[0023] Step 4.5: Pop the parent population A pop of offspring population B The mutation is performed, and the two mutated populations are merged into the next generation population, pop. t ;

[0024] Step 4.6: Pop the current population t The individual with the worst fitness is replaced by the previous generation population pop. t-1 The individual with the best fitness;

[0025] Step 4.7: If the number of iterations has not reached T, execute step 4.2; otherwise, end this process and pop the individual with the highest fitness. T .

[0026] Furthermore, the method of beamforming using the selected array elements is as follows: adaptive beamforming is achieved by using the minimum variance distortionless response method.

[0027] Beneficial effects:

[0028] This invention estimates the non-circularity of the radar transmitted signal and then expands the steering vectors of the radar received signal and the desired signal using a conjugate method, thereby achieving virtual array expansion. Array elements are selected from the virtually expanded array, and beamforming is performed based on the selected array elements to obtain the waveform after interference suppression. When the transmitted signal is a non-circular signal, it has good anti-interference performance of the main lobe and side lobe, and at the same time, it does not require actual expansion of the array aperture. The method has strong robustness and is applicable to different types of array radars with different numbers of array elements. Attached Figure Description

[0029] Figure 1 The flowchart illustrates a method for suppressing main lobe interference in array radar based on array virtual extension, as provided by this invention.

[0030] Figure 2 This invention provides a virtual extended array topology structure established based on an array radar main lobe interference suppression method.

[0031] Figure 3 This invention provides an adaptive radiation pattern formed using an array radar main lobe interference suppression method based on array virtual extension and other different algorithms.

[0032] Figure 4 The amplified adaptive radiation pattern is generated by using the array radar main lobe interference suppression method based on array virtual extension provided by this invention and other different algorithms.

[0033] Figure 5 The graph shows the signal-to-interference-plus-noise ratio (SIR) versus the number of snapshots obtained using the array radar main lobe interference suppression method based on array virtual extension provided by this invention and other different algorithms.

[0034] Figure 6 The graph shows the variation of the desired signal directional gain with the number of snapshots, obtained by using the array radar main lobe interference suppression method based on array virtual extension provided by this invention and other different algorithms.

[0035] Figure 7The graph shows the signal-to-interference-plus-noise ratio (SNR) versus main lobe interference angle obtained using the array radar main lobe interference suppression method based on array virtual extension provided by this invention and other different algorithms.

[0036] Figure 8 The graph shows the variation of the desired signal directional gain with the main lobe interference angle, obtained by using the array radar main lobe interference suppression method based on array virtual extension provided by this invention and other different algorithms.

[0037] Figure 9 The graph shows the output signal-to-interference-plus-noise ratio (SNR) versus input SNR obtained using the array radar main lobe interference suppression method based on array virtual extension provided by this invention and other different algorithms.

[0038] Figure 10 The graph shows the variation of the desired signal directional gain with the input signal-to-noise ratio obtained by using the array radar main lobe interference suppression method based on array virtual extension provided by this invention and other different algorithms.

[0039] Figure 11 This is a diagram of an outdoor experimental scenario.

[0040] Figure 12 This is a waveform diagram of the radar's received signal.

[0041] Figure 13 The waveform diagram is obtained after processing the array radar main lobe interference by using the array virtual extension-based main lobe interference suppression method provided by the present invention.

[0042] Figure 14 The waveform diagram is obtained after array processing using the array radar main lobe interference suppression method based on array virtual extension provided by the present invention. Detailed Implementation

[0043] The following examples illustrate the invention in detail.

[0044] Non-circular signals are a class of signals with second-order rotational variability. Utilizing this property, the array aperture can be virtually expanded without moving the existing array. Non-circular signals are commonly found in communication systems, such as binary phase-shift keying (BPSK) signals. With the integration of disciplines, non-circular signals are gradually demonstrating their role in radar and radar-communication integration. Initially used for super-resolution angle estimation to achieve higher-resolution angle estimation results, non-circular signals have also been introduced into the field of interference suppression. Therefore, this invention considers applying them to the main lobe interference suppression process in radar airspace anti-jamming.

[0045] This invention provides a method for suppressing main lobe interference in array radar based on array virtual expansion. The core idea is as follows: First, the non-circularity of the radar transmitted signal is estimated; then, for the radar received signal, the received signal is expanded in a conjugate manner, and the steering vector of the desired signal is also expanded, thereby achieving array virtual expansion without moving any array elements; subsequently, a genetic algorithm is used to select the expanded array elements, and beamforming is performed using the selected elements to obtain the waveform after interference suppression, which is beneficial for subsequent processing.

[0046] This invention provides a method for suppressing main lobe interference in array radar based on array virtual extension, such as... Figure 1 As shown, the specific steps include:

[0047] Step 1: Calculate the non-circularity of the radar transmitted signal.

[0048] Circular signals are common in radar waveforms, such as linear frequency modulated (LFM) signals. The first and second moments of a circular signal exhibit rotational invariance; that is, rotating the circular signal by any angle does not change its first and second moments.

[0049]

[0050] Where E[·] denotes the expectation operation, and in practical applications, maximum likelihood estimation is often used instead of the expectation operation; s t (t) represents the radar's transmitted signal, φ represents the angle of arbitrary rotation of the transmitted signal, [·] H This indicates the conjugate transpose operation, [·] T This indicates the transpose operation. For radar transmitted signals, the first two terms always satisfy the condition; however, for the third term, due to the characteristics of circular signals, its elliptic covariance matrix...

[0051] When the transmitted signal is non-circular, its elliptic covariance matrix... It contains information, and its elliptic covariance matrix can be written as:

[0052]

[0053] Where γ is the non-circularity of the transmitted signal. This formula can be used to estimate the non-circularity of the transmitted signal.

[0054] Non-circular signals are commonly found in communication systems. Binary Phase Shift Keying (BPSK) and Amplitude Shift Keying (ASK) are common signals with the highest non-circularity, meaning their non-circularity is equal to 1. Biscedastic Quadrature Phase Shift Keying (OQPSK) and Orthogonal Frequency Division Multiplexing (OFDM) are common signals with a non-circularity less than 1. With the development of radar-communication integration, the usage of these communication signals in radar systems has greatly increased.

[0055] Step 2: Assume there are M signal sources incident on the radar array in space. The M signal sources include one desired signal source, one main lobe interference signal source, and M-2 sidelobe interference signal sources. The angle between the direction of the m-th signal source incident on the radar array and the normal of the array is θ. m Assuming the signal source is far enough from the radar array to satisfy the far-field condition, and assuming the leftmost element of the radar array is the reference element (i.e., the zero point of the coordinate axis), then the received signal model of the radar array can be represented as follows:

[0056]

[0057] Among them, A(θ)=[a(θ0),...,a(θ M-1 The array manifold matrix () represents the array's response to a spatial information source. Called the steering vector, it represents the direction of the array pair θ. m The response of the source; the received source vector S(t) = [s0(t),...,s...,sM-1] M-1 (t)] T s represents the complex envelope value of the spatial source signal at time t. m (t) represents the complex envelope of the m-th source; N(t) = [n1(t),...,n...,nN-1] N-1 (t)] T This is the array received noise vector. For ease of discussion, parameters with a subscript of 0 are considered as the desired signal, and those with a subscript of 1 are considered as the main lobe interference signal.

[0058] The signal modeling method of this invention is applicable to array models with various structures, including one-dimensional linear arrays and two-dimensional planar arrays. Taking a one-dimensional uniform linear array as an example, when constructing the signal model, consider a linear uniform array composed of N identical omnidirectional antenna elements arranged at equal intervals, with the element spacing being half the wavelength, i.e., d = λ / 2. Simultaneously, M signals (including one desired signal, one main lobe interference, and M-2 sidelobe interferences) are incident on this array in space. The angle between the m-th signal source incident on the array and the array's normal is θ. m Assuming the signal source and array are far enough apart to satisfy the far-field condition, and that the leftmost element of the array is the reference element (i.e., the zero point of the coordinate axis), the received noise is spatially white Gaussian noise with a mean of zero and a variance of...

[0059] Step 3: Expand the original radar array through conjugate expansion.

[0060] When the received desired signal is a non-circular signal, the conjugate of the desired signal can be decomposed into:

[0061]

[0062] Among them, the superscript [·] * This indicates the conjugate operation. Describes the orthogonal subspace of the desired signal. Let V be the variance of the received desired signal. Therefore, the received signal of the virtual extended array can be written as:

[0063]

[0064] Where IN(t) represents the interference plus noise vector of the virtual extended array. By decomposing the desired signal conjugate, the received signal vector of the virtual extended array can be rewritten as:

[0065]

[0066] in, Let IN'(t) be the desired signal steering vector of the virtual extended array, representing the interference plus noise vector that includes the portion perpendicular to the desired signal after conjugation. This vector is uncorrelated with the desired signal. The topology of the virtual extended array is as follows: Figure 2 As shown.

[0067] Step 4: Use a genetic algorithm to select the appropriate array elements from the virtual extended array to obtain the selected radar array.

[0068] Genetic algorithm (GA) is used to select array elements in the virtual extended array. To select elements more efficiently, a 2N×1 dimensional binary encoded sequence is used as the individual in the genetic algorithm, and the main lobe width is used as the fitness function. To keep the overall array aperture constant, elements on both sides of the virtual extended array must be selected. The flowchart of the genetic algorithm is shown in Table 1.

[0069] Table 1. Genetic Algorithm Element Selection Process

[0070]

[0071] Using the selection vector obtained in the genetic algorithm, the corresponding array elements can be selected from the virtual extended array to obtain the selected radar array. The corresponding received signal vector and desired signal steering vector can be expressed as follows:

[0072] Step 5: Perform adaptive beamforming on the selected radar array obtained in Step 4 to obtain the signal after interference suppression processing.

[0073] The Minimum Variance Distortionless Response (MVDR) method is a classic adaptive beamforming method. This method can accurately determine the adaptive weight vector. Its optimization problem can be expressed as:

[0074]

[0075] in, This is the covariance matrix of the selected radar array.

[0076] Solving the above optimization problem yields the adaptive weight vector:

[0077]

[0078] Therefore, the signal processed by the adaptive beamforming algorithm can be calculated as follows:

[0079]

[0080] The adaptive radiation pattern is the response of the array, which is calculated using the following formula:

[0081]

[0082] In summary, the above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0083] Example 1

[0084] To compare the performance of the algorithm of this invention with existing main lobe interference suppression algorithms, a simulation experiment was designed for comparative analysis. The simulation parameters are shown in Table 2.

[0085] Table 2 Simulation Parameters

[0086]

[0087] The desired signal is set as a binary phase shift keying (BPSK) signal with a non-circularity of 1. During the simulation, it is assumed that the desired signal, interference, and noise are uncorrelated. Comparison methods include the MVDR method without any array manipulation, the main lobe interference suppression method based on eigenprojection preprocessing and covariance matrix reconstruction (EMP-CMR), and a main lobe interference suppression method for moving end elements (INM). The results of comparing the virtual extended array and the MVDR method of the selected array in this invention using these three methods are as follows.

[0088] Figure 3 and Figure 4 The adaptive radiation patterns generated by different algorithms are shown. It can be seen that all methods can form nulls at the interference locations. Calculations show that the proposed method outperforms existing methods in terms of output signal-to-interference-plus-noise ratio (SINR) and desired signal directional gain, as shown in Table 3.

[0089] Table 3. Calculation results of indicators using different methods

[0090]

[0091] As shown in Table 3, both arrays proposed in this invention achieve a 7dB higher output signal-to-interference-plus-noise ratio (SINR) than existing algorithms, and both arrays also achieve a 5dB higher gain in the desired signal direction. A comparison of the two arrays reveals that the selected array maintains the highest gain in the desired signal direction, while the virtual extended array maintains the highest output SINR.

[0092] Figure 5 and Figure 6 The figures show the output signal-to-interference-plus-noise ratio (SIR) and the desired signal directional gain as a function of the number of snapshots for different algorithms. Figure 5 As can be seen, when the number of snapshots is greater than 16, the output signal-to-interference-plus-noise ratio (SNR) of the proposed method is much higher than that of existing methods. The reason why the virtual extended array has a slower convergence speed is that it inherently has more array elements, so it can have a higher output SNR when the method reaches convergence. At the same time, when all algorithms reach convergence, the proposed method has a much higher gain in the desired signal direction than existing methods, and the selected array is superior.

[0093] Figure 7 and Figure 8 The figures show the output signal-to-interference-plus-noise ratio (SIR) and the desired signal directional gain as a function of the main lobe interference angle for different algorithms. Figure 7 As can be seen, the proposed method maintains a high output signal-to-interference-plus-noise ratio (SNR) as the main lobe interference gradually approaches the desired signal in angle. When the main lobe interference is sufficiently close to the desired signal angle, the output SNR of MVDR and INM is low because nulls are generated in incorrect locations. Simultaneously, it can be observed that when the main lobe interference angle approaches the desired signal, the gain of the EMP-CMR algorithm in the desired signal direction drops rapidly. This means that EMP-CMR primarily suppresses main lobe interference rather than preserving the desired signal, while other algorithms primarily aim to preserve the desired signal. It is worth noting that when the angle between the main lobe interference and the desired signal is less than 0.2°, the gain of MVDR and INM in the desired signal direction is even higher, because both methods fail at this point.

[0094] Figure 9 and Figure 10 The figures show the output signal-to-interference-plus-noise ratio (SINNR) and the desired signal directional gain as a function of the input SINNR for different algorithms. It can be seen that the method proposed in this invention outperforms existing methods in both output SINNR and desired signal directional gain.

[0095] Example 2

[0096] To verify the proposed array radar main lobe interference suppression method based on array virtual expansion, measured data obtained from actual radar and jammers were used for interference suppression processing. A uniform linear array of 16 Ku-band elements was used in the experimental test, with the element spacing being half the transmitted wavelength. Figure 11 As shown. The radar array is composed of four uniformly weighted subarrays. The experiment was conducted on an outdoor rooftop, as shown in the experimental setup. Figure 7 As shown. During the actual measurement, noise suppression interference with a bandwidth of 20MHz and an interference-to-noise ratio of 24.58dB was generated by a jammer and propagated through the antenna, incident on the radar at a direction of 3.2°. Subsequently, a BPSK signal generated by MATLAB simulation was added to the radar received signal, with a signal-to-noise ratio of 0dB, and incident on the radar at a direction of 0°.

[0097] Figure 12 and Figure 13 The figure shows the radar received waveform and the received waveform after processing using the method proposed in this invention. It can be seen from the figure that before processing by the proposed method, the desired signal waveform is submerged in interference and cannot be correctly detected. After processing by the virtual extended array and the selected array, the desired signal can be distinguished, and each symbol in the BPSK can be accurately distinguished. Therefore, the algorithm can be used to accurately handle main lobe interference and is beneficial for subsequent processing. The processed signal is as follows: Figure 14 As shown in Table 4, the output signal-to-interference-plus-noise ratio and desired signal directional gain for the two methods are as follows.

[0098] Table 4 shows the index calculation results of the proposed method.

[0099]

[0100] In summary, the above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for suppressing main lobe interference in array radar based on array virtual extension, characterized in that, Includes the following steps: The non-circularity of the radar transmitted signal is estimated. Based on the estimated non-circularity and the radar received signal, the steering vectors of the radar received signal and the desired signal are extended in a conjugate manner to realize the virtual extension of the radar array. Then, a genetic algorithm is used to select the array elements after the extension, and the selected array elements are used for beamforming to obtain the waveform after interference suppression. The process of expanding the steering vectors of the radar received signal and the desired signal using a conjugate method is as follows: When the desired signal received by the radar is a non-circular signal, the conjugate of the desired signal is decomposed into: Among them, superscript This indicates the conjugate operation. Describes the orthogonal subspace of the desired signal. Let V be the variance of the desired radar signal received; therefore, the radar received signal vector of the virtual extended array is: in, The desired signal steering vector for the virtual extended array. This represents the interference plus noise vector of the virtual extended array. This represents the interference plus noise vector that includes the portion perpendicular to the desired signal after conjugating the desired signal.

2. The array radar main lobe interference suppression method according to claim 1, characterized in that, Also includes: Assuming there is One signal source incident radar array, The signal source includes a desired signal source, a main lobe interference signal source, and... There are several sidelobe interference signal sources, among which the first... The angle between the direction of the incident signal source on the radar array and the normal of the array is: Assuming the signal source and radar array are far enough apart to satisfy the far-field condition, and assuming the leftmost element of the radar array is the reference element (i.e., the zero point of the coordinate axis), then the radar received signal is represented as: in, The array manifold matrix represents the radar array's response to a spatial information source; The steering vector represents the direction of the radar array. The response from the source; For receiving the source vector, it represents the first... The complex envelope value of the spatial source signal at time. Indicates the first The complex envelope of a single information source; The array receives noise vector; parameters with subscript 0 are considered as the desired signal, and those with subscript 1 are considered as the main lobe interference signal.

3. The array radar main lobe interference suppression method according to claim 1, characterized in that, Using a 2N×1 dimensional binary encoded sequence as an individual in the genetic algorithm, and the main lobe width as the fitness function, the genetic algorithm selects the elements of the expanded array as follows: Step 4.1: Guide the desired signal vector of the virtual extended array. Search guide vector of virtual extended array The received signal vector of the virtual extended array Let be the input, and let the number of iterations of the genetic algorithm be . Population size is Initialize the initial population. Selection rate Cross rate and variability ; Step 4.2: Calculate the current population The fitness function; Step 4.3: Obtain the parent population using the roulette wheel method. and offspring population ; Step 4.4: Use single-point crossover to cross the parent population. Cross over; Step 4.5: Transfer the parent population With offspring population The mutation is performed, and the two mutated populations are merged into the next generation population. ; Step 4.6: Population The individual with the worst fitness was replaced by the population. The individual with the best fitness; Step 4.7: Determine if the current iteration number t has reached T. If not, set t = t + 1 and return to step 4.2; otherwise, end this process and obtain the individual with the highest fitness. Using the selection vector obtained in the genetic algorithm, the corresponding array elements are selected from the virtual extended array.

4. The array radar main lobe interference suppression method according to claim 1, characterized in that, The method of beamforming using the selected array elements is as follows: adaptive beamforming is achieved by using the minimum variance distortionless response method.

Images