Large-scale broadband sparse array optimization design method and device considering mutual coupling effect

Abstract

The invention discloses a large-scale broadband sparse array optimization design method and device considering a mutual coupling effect, and belongs to the technical field of array antennas. The method comprises the following steps: constructing a Gaussian process regression model for geometric parameters and active directional diagrams of broadband sparse array units based on a mutual coupling effect between the broadband sparse array units; determining parameters according to a given array index, and setting an initial geometric layout of the broadband sparse array; based on the initial geometric layout, calculating an active directional diagram of each unit by using a Gaussian process regression model to obtain an array radiation directional diagram of the broadband sparse array; and based on the array radiation pattern, converting the broadband sparse array unit position optimization problem into a convex optimization problem, and performing iterative solution to obtain the optimal unit position of the broadband sparse array. According to the method, the iterative convex optimization algorithm and the broadband sparse array mutual coupling calculation model based on the Gaussian regression process are combined, optimization design of the large-scale broadband sparse array is achieved, universality is high, and the application range is wide.

Description

Technical Field

[0001] The present invention belongs to the technical field of array antennas, and in particular relates to a large-scale broadband sparse array optimization design method and device taking mutual coupling effects into consideration. Background Art

[0002] As a core component of various electronic equipment, the design of high-quality configuration of array antenna plays a vital role in realizing high-performance, low-cost and lightweight systems. In application fields such as radar systems and radio astronomy, array antennas are often required to have high target resolution and low manufacturing cost. To obtain a narrow beam and improve the spatial resolution of the antenna, one method is to increase the number of array elements in the array. Another method is to use a sparse array (also called an unequally spaced antenna array, or a non-periodic array). However, increasing the number of array elements not only increases the manufacturing cost of the array, but also increases the complexity of the system. Compared with a full array with uniform spacing of the same aperture, a sparse array not only has an almost equal main lobe beamwidth but also reduces system complexity and manufacturing cost. In addition, the sparse array uses equal amplitude feeding and can effectively avoid grating lobes.

[0003] At present, the prior art provides several optimization methods for sparse arrays, for example, in the literature "C. Bencivenni, MVIvashina, R. Maaskant, and J. Wettergren, Design of Maximally Sparse Antenna Arrays in the Presence of Mutual Coupling, IEEE Antennas and Wireless Propagation Letters, vol. 14, pp. 159-162, Nov. 2015" and the literature "X. Huang, Y. Liu, P. You, M. Zhang and QH Liu, Fast Linear Array Synthesis Including Coupling Effects Utilizing Iterative FFT via Least-Squares Active Element Pattern Expansion, IEEE Antennas and Wireless Propagation Letters, vol. 16, pp. 804-807, 2017. ", respectively proposed an iterative l1 norm optimization method combined with the moment method to optimize the design of sparse antenna arrays and an IFT-based hybrid algorithm that takes into account the mutual coupling effect in real time during the optimization synthesis process and the array synthesis in the mutual coupling factor to synthesize microstrip array antennas. The literature "Qi Wang, Jun Yang, Yi Tang, Renjing Gao, Shutian Liu, Optimization Method for Pattern Synthesis of Sparse Planar Arrays Considering Mutual Coupling and Nonoverlapping Constraint, IEEE Transactions on Antennas and Propagation, vol. 68, no. 8, pp. 6032-6038, 2020." proposed a sparse array optimization design method that handles array element constraints. This method achieves array sparseness by optimizing the array element position and rotation angle, and combines the MOM method to calculate the mutual coupling of the array. This type of method mainly takes into account the mutual coupling effect between array elements on the active element pattern (AEP) of the antenna unit by introducing full-wave simulation and other means.

[0004] However, in the optimization process of large-scale broadband sparse arrays, since the positions of antenna elements are constantly changing, it is difficult to consider the mutual coupling between units in real time during the optimization synthesis process, and thus it is impossible to extract the active radiation pattern of the units in the array. In addition, the array response is a complex exponential or trigonometric function of the element position and there are many optimization variables, so the large-scale element position optimization design is a constrained multivariable nonlinear optimization problem. In addition, with the increase in application requirements, the scale of array antennas is getting larger and larger, which makes the dimension of optimization variables increase sharply. Existing evolution-based algorithms (such as genetic algorithms, particle swarm optimization algorithms, and differential evolution algorithms, etc.) have difficulty in handling high-dimensional optimization problems, and are prone to fall into the "dimensionality disaster" and local optimal solutions.

[0005] Therefore, how to accurately consider the mutual coupling effect in the optimization process of sparse arrays and the optimal design of large-scale array antennas has become a core technical problem that needs to be solved urgently. Summary of the invention

[0006] In order to solve the above problems existing in the prior art, the present invention provides a large-scale broadband sparse array optimization design method and device considering the mutual coupling effect. The technical problem to be solved by the present invention is achieved by the following technical solutions:

[0007] In a first aspect, the present invention proposes a large-scale broadband sparse array optimization design method considering mutual coupling effects, comprising:

[0008] Step 1: Based on the mutual coupling effect between broadband sparse array units, a Gaussian process regression model is constructed for the geometric parameters and active pattern of broadband sparse array units;

[0009] Step 2: Determine the parameters according to the given array index and set the initial geometric layout of the broadband sparse array;

[0010] Step 3: Based on the initial geometric layout, the Gaussian process regression model is used to calculate the active pattern of each unit to obtain the array radiation pattern of the broadband sparse array;

[0011] Step 4: Based on the array radiation pattern of the broadband sparse array, the broadband sparse array unit position optimization problem is converted into a convex optimization problem, and iteratively solved to obtain the optimal unit position of the broadband sparse array.

[0012] In a second aspect, the present invention proposes a large-scale broadband sparse array optimization design device considering mutual coupling effects, which is used to implement the large-scale broadband sparse array optimization design method considering mutual coupling effects provided in the first aspect of the present invention, comprising:

[0013] A model building module is used to build a Gaussian process regression model for the geometric parameters of the broadband sparse array and the active pattern of the unit based on the mutual coupling effect between the broadband sparse array units;

[0014] An initialization module is used to determine parameters according to given array indicators and set the initial geometric layout of the broadband sparse array;

[0015] A calculation module is used to calculate the active pattern of each unit based on the initial geometric layout using a Gaussian process regression model to obtain the array radiation pattern of a broadband sparse array;

[0016] The optimization solution module is used to convert the broadband sparse array unit position optimization problem into a convex optimization problem based on the array radiation pattern of the broadband sparse array, and perform iterative solution to obtain the optimal unit position of the broadband sparse array.

[0017] Beneficial effects of the present invention:

[0018] The large-scale broadband sparse array optimization design method considering the mutual coupling effect provided by the present invention firstly constructs a Gaussian process regression model for the geometric parameters and active directional patterns of the broadband sparse array units based on the mutual coupling effect between the broadband sparse array units, and initializes the geometric layout of the broadband sparse array; then, based on the initial geometric layout, the active directional pattern of each unit is calculated using the Gaussian process regression model to obtain the array radiation directional pattern of the broadband sparse array; finally, the broadband sparse array unit position optimization problem is converted into a convex optimization problem, and iteratively solved to obtain the optimal unit position of the broadband sparse array. On the one hand, this method uses the sparse array element mutual coupling calculation model based on Gaussian regression process to obtain the active radiation pattern of the array unit, which effectively solves the problem of the inability to extract the active radiation pattern of the unit in the array due to the comprehensive mutual coupling effect of the large-scale broadband sparse array. On the other hand, the original large-scale array optimization problem is transformed into a convex problem for solution, which improves the calculation efficiency and avoids the problem that the existing evolution-based algorithm is prone to fall into the local optimal solution. By combining the iterative convex optimization algorithm with the broadband sparse array mutual coupling calculation model based on Gaussian regression process, the optimization design of large-scale broadband sparse array is realized, which has strong versatility and a wide range of applications.

[0019] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a flow chart of a large-scale broadband sparse array optimization design method considering mutual coupling effect provided by an embodiment of the present invention;

[0021] Figure 2 It is another flow chart of a large-scale broadband sparse array optimization design method considering mutual coupling effect provided by an embodiment of the present invention;

[0022] Figure 3 It is a schematic diagram of a process for constructing a Gaussian process regression model provided by an embodiment of the present invention;

[0023] Figure 4 It is a structural block diagram of a large-scale broadband sparse array optimization design device considering mutual coupling effects provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] The first aspect of the present invention provides a large-scale broadband sparse array optimization design method considering mutual coupling effects. Figure 1 and Figure 2 , Figure 1 is a flow chart of a large-scale broadband sparse array optimization design method considering mutual coupling effects provided by an embodiment of the present invention. Figure 2 FIG. 1 is another flow chart of a large-scale broadband sparse array optimization design method considering mutual coupling effect provided by an embodiment of the present invention. The method mainly includes the following steps:

[0026] Step 1: Based on the mutual coupling effect between broadband sparse array units, a Gaussian process regression model is constructed for the geometric parameters and active radiation pattern of broadband sparse array units.

[0027] Specifically, this embodiment can calculate the mutual coupling pattern data of the broadband sparse antenna array through full-wave simulation, and establish a Gaussian process regression model that can perform high-precision calculation of the electromagnetic response of the mutual coupling effect between the broadband sparse array elements. In other words, this embodiment establishes a broadband sparse array element mutual coupling calculation model based on Gaussian process regression.

[0028] See also Figure 3 , Figure 3 It is a schematic diagram of a process for constructing a Gaussian process regression model provided by an embodiment of the present invention.

[0029] Optionally, as an implementation manner, step 1 may be implemented according to the following sub-steps 1a-1c:

[0030] Step 1a: Select several groups of geometric layouts of broadband sparse arrays based on orthogonal experimental design.

[0031] Orthogonal experimental design is abbreviated as orthogonal design. It is a method of scientifically arranging and analyzing multi-factor experiments using orthogonal tables. It is one of the commonly used experimental analysis methods.

[0032] This embodiment uses orthogonal experimental design to design a geometric layout for M groups of broadband sparse arrays, and obtains a broadband sparse array with M groups of units.

[0033] Step 1b, obtaining mutual coupling pattern data corresponding to the geometric layout of several groups of broadband sparse arrays through full-wave electromagnetic simulation, and constructing a training data set; wherein the training data set includes the geometric parameters and active pattern of each unit of the broadband sparse array.

[0034] In this embodiment, the finite training sample data set constructed according to M observation values ​​can be denoted as D, which is expressed as:

[0035] D={(u m ,y m )|m=1,2,…,M};

[0036] In the formula, u m =[x r,m x r,m ...x r,m f m θ m ] T is the structural parameter of the input broadband sparse array, also known as the geometric parameter, x r,m x r,m ...x r,m is the position parameter of the mth group of sparse array units, f m and θ m Represent the values ​​of the frequency and angle of interest, respectively. For a broadband array, for f m Two side frequencies and the center frequency can be selected within the bandwidth, that is, three typical frequencies can be selected within the wide band; m =[E mag,m E pha,m ] T is the output electromagnetic response, that is, the active directional pattern of the unit, E mag,m and E pha,m They respectively represent the predicted calculation of the active amplitude and phase pattern of each unit of the broadband sparse array of the mth group.

[0037] Step 1c: construct a Gaussian process regression model that reflects the input-output mapping relationship using the training data set; wherein the Gaussian process regression model uses the geometric parameters of each unit of the broadband sparse array as input and the active directional pattern of each unit as output.

[0038] Specifically, the Gaussian regression process will learn the nonlinear mapping relationship between input u and output y based on the sample set, that is, the unit position of a set of sparse arrays and the values ​​of the active amplitude and phase pattern of each unit of the broadband sparse array. The regression model is expressed as:

[0039] y=f(u)+ε;

[0040] Where y represents the output active pattern, u represents the input geometric parameters, and ε represents the noise that obeys the Gaussian distribution.

[0041] f represents a multivariate Gaussian process, which is defined as any input variable that satisfies the Gaussian process Bayesian prior distribution, expressed as:

[0042] f~GP(m(u),k(u,u′));

[0043] Where GP represents Gaussian process, m(u) represents mean value, and k(u,u′) represents covariance function.

[0044] It is understandable that after constructing the Gaussian process regression model, that is, after step 1c, it also includes:

[0045] Step 1d: Based on any set of input data, perform error judgment on the output of the Gaussian process regression model and the output obtained according to the full-wave electromagnetic simulation. If the error between the two is less than the preset threshold, output the current Gaussian process regression model; otherwise, add sample data and return to step 1b to update the training data set.

[0046] Specifically, any input data with the same specifications as the training data set can be selected and input into the constructed Gaussian process regression model to obtain an active pattern. At the same time, the input data is used to perform full-wave electromagnetic simulation to obtain another active pattern. The two patterns are compared. If the error between the two is less than the preset threshold, that is, the output obtained by the Gaussian process regression model is close enough to the output obtained by simulation, it means that the constructed Gaussian process regression model has good accuracy. Therefore, it is directly output as the best model.

[0047] If the error between the two is greater than or equal to the preset threshold, that is, the output obtained by the Gaussian process regression model is far from the output obtained by the simulation, indicating that the accuracy of the constructed Gaussian process regression model is low and needs to be rebuilt. Then return to the step of building a training data set, increase sample data, and rebuild the Gaussian process regression model until the threshold requirement is met.

[0048] The established Gaussian process regression model can directly output accurate electromagnetic responses in a very short time according to the structural parameters and can be reused, that is, the Gaussian regression process model quickly and accurately obtains the AEP of each element in the array. Therefore, it can be combined with the optimization algorithm and used in the optimization synthesis of broadband sparse arrays considering mutual coupling effects to reduce the design time.

[0049] Different from the previous optimization methods of sparse arrays that take mutual coupling effects into consideration by combining active unit radiation patterns or full-wave simulation with optimization algorithms, the sparse array element mutual coupling calculation model based on Gaussian regression process proposed in the present invention can effectively solve the comprehensive mutual coupling effect problem of large-scale broadband sparse arrays. This method has significant advantages such as strong versatility and wide applicability.

[0050] Step 2: Determine the parameters according to the given array indicators and set the initial geometric layout of the broadband sparse array.

[0051] Specifically, for a large-scale broadband sparse array with an array plane located in the XOY plane, the number of its units is M*N, and a geometric layout can be initialized arbitrarily for the broadband sparse array. Generally speaking, the initial geometric layout is a uniform array.

[0052] Step 3: Based on the initial geometric layout, the Gaussian process regression model is used to calculate the active pattern of each unit to obtain the array radiation pattern of the broadband sparse array.

[0053] For a large-scale broadband sparse array consisting of M*N units, the Gaussian process regression model is used to calculate the active pattern of each unit and obtain the array radiation pattern Its expression is:

[0054]

[0055] In the formula, θ and denote the elevation angle and azimuth angle respectively, w(m,n) denotes the complex weighting coefficient of the (m,n)th antenna element, represents the unit radiation pattern of the (m,n)th antenna unit. The unit active radiation pattern includes the mutual coupling factors between different units of the array, and can also be simulated using full-wave electromagnetic simulation software. (x m ,y n ) represents the unit position of the (m,n)th antenna unit, k1 represents the wave number.

[0056] Step 4: Based on the array radiation pattern of the broadband sparse array, the broadband sparse array unit position optimization problem is converted into a convex optimization problem, and iteratively solved to obtain the optimal unit position of the broadband sparse array.

[0057] In general, the problem of optimizing the location of broadband sparse array elements can be described as:

[0058]

[0059] In the formula, represents any set of elevation angles and azimuth angles in the sidelobe region, ρ0 represents the suppression coefficient of the sidelobe rise, Θ sllrepresents the sidelobe area, Indicates the direction of maximum radiation of the main beam.

[0060] However, the optimization described above is non-convex, mainly because the expression for the cell position is in the exponential position, It is a nonlinear function of (x, y) about the unit position. At this time, it is impossible to directly use convex optimization methods to effectively solve the original optimization problem in polynomial time.

[0061] In this regard, this embodiment intends to transform the original non-convex optimization problem into a convex optimization problem by deriving an equivalent transformation form of the problem and constructing new convex constraints.

[0062] Optionally, as an implementation manner, step 4 may be implemented according to the following sub-steps 4a-4e:

[0063] Step 4a: In the current iteration process, the array radiation pattern is approximated by a first-order Taylor expansion to obtain an approximate array radiation pattern.

[0064] Specifically, assuming that the unit position of the sparse array at the kth iteration is (x k ,y k ), given a set of initial unit positions (x 0 ,y 0 ), satisfying the unit position constraint; in order to transform the nonlinear function about the unit position into a convex optimization problem, first define in the kth iteration:

[0065]

[0066] Where m = 1, 2, .., M, n = 1, 2, .., N, ε m and δ n are the changes in the unit position in the x-axis direction and the y-axis direction respectively.

[0067] In the kth iteration, a first-order Taylor expansion is used to approximate e jγ ≈1+jγ, the original non-convex optimization problem is equivalently transformed to obtain a series of convex sub-optimization problems about the unit position for iterative solution.

[0068] The array radiation pattern is obtained by first-order Taylor expansion The approximate array radiation pattern is expressed as:

[0069]

[0070] In the formula, Indicates that during the kth iteration The first-order Taylor expansion approximation of is: represents the unit position of the (m,n)th antenna unit in the k-1th iteration process, ε m and δ n They represent the micro-winding amount of the (m,n)th antenna unit position in the x-axis direction and the y-axis direction respectively.

[0071] Generally speaking, when |uε is satisfied m |<<1 and |vδ n |<<1, yes A good approximation.

[0072] Step 4b: Based on the approximate array radiation pattern, the unit position perturbation is introduced to convert the broadband sparse array unit position optimization problem into a convex optimization problem.

[0073] Optionally, this embodiment introduces an M-dimensional unit position perturbation and an N-dimensional unit position perturbation The original optimization problem is transformed into a standard convex optimization problem form, and a new convex constraint condition is constructed so that the receiving intensity at angles other than the desired beam pointing is less than the receiving intensity in the desired beam direction, to ensure that the maximum radiation direction is the desired beam direction.

[0074] Specifically, after converting the k-th large-scale sparse array solution problem into a convex optimization problem, it can be expressed as follows:

[0075]

[0076] In the formula, represents any set of pitch angles and azimuth angles, ρ0 represents the suppression coefficient of the sidelobe lift, Θ sll represents the sidelobe area, represents the maximum radiation direction of the main beam, and α represents a preset positive constant.

[0077] Step 4c: Iteratively solve the convex optimization problem to obtain the current broadband sparse array unit position.

[0078] Step 4d: Update the broadband sparse array unit position according to the current broadband sparse array unit position.

[0079] Among them, the update formula of the broadband sparse array unit position is:

[0080]

[0081] In the formula, and Respectively represent the unit position of the (m,n)th antenna unit in the kth and k-1th iterations, and They respectively represent the perturbations of the (m,n)th antenna unit position in the x-axis direction and the y-axis direction during the k-th iteration.

[0082] Step 4e: determine whether the iteration stop condition is met. If so, output the current broadband sparse array unit position as the optimal solution. Otherwise, return to step 4a for the next iteration.

[0083] Generally speaking, when the side lobes of a large-scale broadband sparse array no longer change, it can be considered that the condition for stopping the iteration is met.

[0084] It should be noted that for the iterative convex optimization algorithm, in addition to using the initialized uniform array unit position as the initial value, an evolutionary algorithm or an iterative IFT (Iterative Fourier Transform) algorithm can also be used to optimize the array without considering mutual coupling, and the optimization result can be used as the initial value of the algorithm to make the algorithm converge faster.

[0085] The large-scale broadband sparse array optimization design method considering the mutual coupling effect provided by the present invention firstly constructs a Gaussian process regression model for the geometric parameters and active directional patterns of the broadband sparse array units based on the mutual coupling effect between the broadband sparse array units, and initializes the geometric layout of the broadband sparse array; then, based on the initial geometric layout, the active directional pattern of each unit is calculated using the Gaussian process regression model to obtain the array radiation directional pattern of the broadband sparse array; finally, the broadband sparse array unit position optimization problem is converted into a convex optimization problem, and iteratively solved to obtain the optimal unit position of the broadband sparse array. On the one hand, this method uses the sparse array element mutual coupling calculation model based on Gaussian regression process to obtain the active radiation pattern of the array unit, which effectively solves the problem of the inability to extract the active radiation pattern of the unit in the array due to the comprehensive mutual coupling effect of the large-scale broadband sparse array. On the other hand, the original large-scale array optimization problem is transformed into a convex problem for solution, which improves the calculation efficiency and avoids the problem that the existing evolution-based algorithm is prone to fall into the local optimal solution. By combining the iterative convex optimization algorithm with the broadband sparse array mutual coupling calculation model based on Gaussian regression process, the optimization design of large-scale broadband sparse array is realized, which has strong versatility and a wide range of applications.

[0086] Based on the same inventive concept, the second aspect of the present invention also provides a large-scale broadband sparse array optimization design device considering the mutual coupling effect. Figure 4 , Figure 4 : is a structural block diagram of a large-scale broadband sparse array optimization design device considering mutual coupling effect provided by an embodiment of the present invention. The device mainly includes:

[0087] A model building module, used for building a Gaussian process regression model for the geometric parameters and active pattern of the broadband sparse array unit based on the mutual coupling effect between the broadband sparse array units;

[0088] An initialization module is used to determine parameters according to given array indicators and set the initial geometric layout of the broadband sparse array;

[0089] A calculation module is used to calculate the active pattern of each unit based on the initial geometric layout using a Gaussian process regression model to obtain the array radiation pattern of a broadband sparse array;

[0090] The optimization solution module is used to convert the broadband sparse array unit position optimization problem into a convex optimization problem based on the array radiation pattern of the broadband sparse array, and perform iterative solution to obtain the optimal unit position of the broadband sparse array.

[0091] As for the device embodiment provided in the second aspect of the present invention, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts may refer to the partial description of the method embodiment.

[0092] It should be noted that the method provided in the embodiment of the present invention can also be applied to electronic devices. Specifically, the electronic device can be: a desktop computer, a portable computer, an intelligent mobile terminal, a server, etc. This is not limited here, and any electronic device that can implement the present invention belongs to the protection scope of the present invention.

[0093] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, devices (equipment), or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware, which are collectively referred to as "modules" or "systems" herein. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The computer program is stored / distributed in a suitable medium, provided together with other hardware or as a part of the hardware, or may adopt other distribution forms, such as through the Internet or other wired or wireless telecommunication systems.

[0094] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the protection scope of the present invention.

Claims

1. A large-scale broadband sparse array optimization design method considering mutual coupling effects, characterized in that: include: Step 1: Based on the mutual coupling effect between broadband sparse array units, a Gaussian process regression model is constructed for the geometric parameters and active pattern of broadband sparse array units; Step 2: Determine the parameters according to the given array index and set the initial geometric layout of the broadband sparse array; Step 3: Based on the initial geometric layout, the Gaussian process regression model is used to calculate the active pattern of each unit to obtain the array radiation pattern of the broadband sparse array; Step 4: Based on the array radiation pattern of the broadband sparse array, the broadband sparse array unit position optimization problem is converted into a convex optimization problem, and iteratively solved to obtain the optimal unit position of the broadband sparse array.

2. The large-scale broadband sparse array optimization design method considering mutual coupling effect according to claim 1 is characterized in that: Step 1 specifically includes: Step 1a, selecting several groups of geometric layouts of broadband sparse arrays based on orthogonal experimental design; Step 1b, obtaining the mutual coupling pattern data corresponding to the geometric layout of the plurality of groups of broadband sparse arrays through full-wave electromagnetic simulation, and constructing a training data set; wherein the training data set includes the geometric parameters and active pattern of each unit of the broadband sparse array; Step 1c, using the training data set to construct a Gaussian process regression model that reflects the input-output mapping relationship; wherein the Gaussian process regression model uses the geometric parameters of each unit of the broadband sparse array as input and the active directional pattern of each unit as output.

3. The large-scale broadband sparse array optimization design method considering mutual coupling effect according to claim 2 is characterized in that: In step 1c, the constructed Gaussian process regression model is expressed as: y=f(u)+ε; In the formula, y represents the output active pattern, u represents the input geometric parameters, ε represents the noise that obeys the Gaussian distribution, and f represents the multivariate Gaussian process, which is defined as any input variable satisfies the Gaussian process Bayesian prior distribution. The expression is: f~GP(m(u),k(u,u′)); Where GP represents Gaussian process, m(u) represents mean value, and k(u,u′) represents covariance function.

4. The large-scale broadband sparse array optimization design method considering mutual coupling effect according to claim 2 is characterized in that: After step 1c, the method further includes: Step 1d: Based on any set of input data, perform error judgment on the output of the Gaussian process regression model and the output obtained according to full-wave electromagnetic simulation. If the error between the two is less than a preset threshold, output the current Gaussian process regression model; otherwise, add sample data and return to step 1b to update the training data set.

5. The large-scale broadband sparse array optimization design method considering mutual coupling effect according to claim 1 is characterized in that: In step 3, for a large-scale broadband sparse array consisting of M*N units, the array radiation pattern is expressed as The expression is: In the formula, θ and denote the azimuth and elevation angles respectively, w(m,n) denotes the complex weighting coefficient of the (m,n)th antenna element, represents the unit radiation pattern of the (m,n)th antenna unit, (x m ,y n ) represents the unit position of the (m,n)th antenna unit, k1 represents the wave number.

6. The large-scale broadband sparse array optimization design method considering mutual coupling effect according to claim 5 is characterized in that: Step 4 specifically includes: Step 4a, in the current iteration process, performing a first-order Taylor expansion approximation on the array radiation pattern of the broadband sparse array to obtain an approximate array radiation pattern; Step 4b, based on the approximate array radiation pattern, introduce the unit position micro-winding amount, and transform the broadband sparse array unit position optimization problem into a convex optimization problem; Step 4c, iteratively solving the convex optimization problem to obtain the current broadband sparse array unit position; Step 4d, updating the broadband sparse array unit position according to the current broadband sparse array unit position; Step 4e: determine whether the iteration stop condition is met. If so, output the current broadband sparse array unit position as the optimal solution. Otherwise, return to step 4a for the next iteration.

7. The large-scale broadband sparse array optimization design method considering mutual coupling effect according to claim 6 is characterized in that: In step 4a, the approximate array radiation pattern is expressed as: In the formula, Indicates that during the kth iteration The first-order Taylor expansion approximation of is: represents the unit position of the (m,n)th antenna unit in the k-1th iteration process, ε m and δ n They represent the perturbations of the (m,n)th antenna unit position in the x-axis direction and the y-axis direction respectively.

8. The large-scale broadband sparse array optimization design method considering mutual coupling effect according to claim 7 is characterized in that: In step 4b, the convex optimization problem is expressed as: In the formula, represents any set of elevation angles and azimuth angles in the sidelobe region, ρ0 represents the suppression coefficient of the sidelobe lift, Θ sll represents the sidelobe area, represents the maximum radiation direction of the main beam, and α represents a preset positive constant.

9. The large-scale broadband sparse array optimization design method considering mutual coupling effect according to claim 6, characterized in that: In step 4d, the updating formula of the broadband sparse array unit position is: In the formula, and Respectively represent the unit position of the (m,n)th antenna unit in the kth and k-1th iterations, and They respectively represent the perturbations of the (m,n)th antenna unit position in the x-axis direction and the y-axis direction during the k-th iteration.

10. A large-scale broadband sparse array optimization design device considering mutual coupling effect, used to implement the large-scale broadband sparse array optimization design method considering mutual coupling effect as described in any one of claims 1 to 9, characterized in that: include: A model building module, used for building a Gaussian process regression model for the geometric parameters and active pattern of the broadband sparse array unit based on the mutual coupling effect between the broadband sparse array units; An initialization module is used to determine parameters according to given array indicators and set the initial geometric layout of the broadband sparse array; A calculation module, used to calculate the active pattern of each unit based on the initial geometric layout using the Gaussian process regression model to obtain the array radiation pattern of the broadband sparse array; The optimization solution module is used to convert the broadband sparse array unit position optimization problem into a convex optimization problem based on the array radiation pattern of the broadband sparse array, and perform iterative solution to obtain the optimal unit position of the broadband sparse array.

Images