Four-channel global electromagnetic metasurface and design method and related device thereof

Abstract

This invention belongs to the field of electromagnetic materials and discloses a four-channel global electromagnetic metasurface, its design method, and related apparatus. The method includes: acquiring a phase profile of the four-channel global electromagnetic metasurface in a target frequency band; obtaining the electromagnetic frequency response curves of each unit of the four-channel global electromagnetic metasurface in the target frequency band based on the phase profile of the four-channel global electromagnetic metasurface in the target frequency band, and converting the electromagnetic frequency response curves of each unit in the target frequency band into two-dimensional images of each unit in the target frequency band; generating unit images of each unit using a pre-trained unit image generator based on the two-dimensional images of each unit in the target frequency band, and obtaining the unit structure of each unit based on the unit images. By converting the one-dimensional electromagnetic frequency response curves into two-dimensional images to solve the input-output dimension mismatch problem of the unit image generator, sufficient effective features can be extracted for accurate description and prediction, effectively improving the efficiency and stability of the design.

Description

Technical Field

[0001] This invention belongs to the field of electromagnetic materials and relates to a four-channel global electromagnetic metasurface, its design method, and related devices. Background Technology

[0002] Electromagnetic metasurfaces are artificially designed subwavelength planar units arranged periodically or aperiodically in a two-dimensional plane based on pre-defined electromagnetic frequency response curves. Electromagnetic metasurfaces are widely used in electronics, communications, and optics due to their ability to flexibly adjust various properties of electromagnetic waves, such as amplitude, phase, polarization, and frequency. In traditional electromagnetic metasurface design paradigms, researchers primarily rely on numerical simulation software, typically based on the finite element method, finite-difference time-domain method, or other numerical calculation methods. The design process generally begins with initially determining the geometry and material properties of the metasurface unit, then using simulation software to simulate the electromagnetic frequency response curves of the structure, obtaining electromagnetic parameters such as reflection coefficient, transmission coefficient, and absorptivity. The structural parameters are then repeatedly adjusted based on the simulation results until the design requirements are met. However, this trial-and-error method suffers from low efficiency, limited solution space, and long design cycles when dealing with complex targets.

[0003] In recent years, artificial intelligence technologies, represented by deep learning, have provided new methods for the design of electromagnetic metasurfaces. Deep learning possesses powerful nonlinear mapping and high-dimensional data fitting capabilities. Utilizing this characteristic, researchers can establish a direct mapping relationship between the structural parameters of electromagnetic metasurfaces and their electromagnetic frequency response curves, thereby rapidly predicting the electromagnetic frequency response curves under different structural parameters and significantly shortening the design time. Furthermore, deep learning can also derive the corresponding structural parameters from the target electromagnetic performance, simplifying the design process. Combined with optimization algorithms, it can efficiently search for the optimal design scheme in a high-dimensional parameter space, improving design efficiency and effectiveness. The application of deep learning in electromagnetic metasurface design provides new ideas and tools for solving many problems in traditional methods and is gradually changing the way research and applications are conducted in this field.

[0004] However, existing deep learning-based electromagnetic metasurface design methods often lack pre-screening of the training set, resulting in a large amount of low-quality datasets entering the training process. This low-quality data not only increases computational resource consumption but may also make model convergence difficult, reducing design efficiency and accuracy. Specifically, unscreened training datasets contain many unit structures unsuitable for metasurface design requirements. These structures perform poorly in terms of electromagnetic frequency response curves, such as failing to cover the required 2π full-phase range or 0–1 full-amplitude range, preventing the model from accurately learning key electromagnetic frequency response curve characteristics and thus affecting the final design performance. Simultaneously, the inverse electromagnetic metasurface design process faces the problem of input-output dimension mismatch. In existing methods, electromagnetic frequency response curves are typically one-dimensional, while electromagnetic metasurface unit structures are two-dimensional. This dimensional difference makes model training difficult. Because one-dimensional signals have limited feature representation capabilities, the model may not be able to extract enough effective features to accurately describe and predict the electromagnetic frequency response curve characteristics of the electromagnetic metasurface unit, leading to poor training results and an inefficient and unstable design process. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art, which is that the model may not be able to extract enough effective features to accurately describe and predict the electromagnetic frequency response curve characteristics of the electromagnetic metasurface unit, resulting in poor training effect and inefficient and unstable design process. The invention provides a four-channel global electromagnetic metasurface, its design method and related device.

[0006] To achieve the above objectives, the present invention employs the following technical solution:

[0007] In a first aspect, this invention provides a method for designing a four-channel global electromagnetic metasurface, comprising: acquiring a phase profile of the four-channel global electromagnetic metasurface in a target frequency band; obtaining electromagnetic frequency response curves of each unit of the four-channel global electromagnetic metasurface in the target frequency band based on the phase profile of the four-channel global electromagnetic metasurface in the target frequency band, and converting the electromagnetic frequency response curves of each unit in the target frequency band into two-dimensional images of each unit in the target frequency band; generating unit images of each unit using a pre-trained unit image generator based on the two-dimensional images of each unit in the target frequency band, and obtaining the unit structure of each unit based on the unit images of each unit; wherein... The pre-trained unit image generator is obtained as follows: acquire unit images of several unit structure samples and electromagnetic frequency response curves of several unit structure samples in the target frequency band, and convert the electromagnetic frequency response curves of several unit structure samples in the target frequency band into two-dimensional images of several unit structure samples in the target frequency band; based on the unit images of several unit structure samples and the two-dimensional images in the target frequency band, train a pre-set generative adversarial network model with the two-dimensional images in the target frequency band as input and the unit images as output, and use the generator of the trained generative adversarial network model as the pre-trained unit image generator.

[0008] Optionally, the unit structure of the unit structure sample is a mirror-symmetric structure or a rotationally symmetric structure.

[0009] Optionally, obtaining the unit images of several unit structure samples includes: dividing each unit structure sample into several grids, setting each grid according to the method of metal as 1 and medium as 0, and forming a Boolean matrix to obtain a two-dimensional matrix of the unit structure sample and converting it into a binary image to obtain the unit image of each unit structure sample; obtaining the electromagnetic frequency response curves of several unit structure samples in the target frequency band includes: performing electromagnetic simulation on several unit structure samples in the target frequency band using simulation software to obtain the electromagnetic frequency response curves of several unit structure samples in the target frequency band.

[0010] Optionally, after obtaining the electromagnetic frequency response curves of several unit structure samples in the target frequency band, the unit structure samples with phase changes less than the phase change threshold in the target frequency band are removed based on the electromagnetic frequency response curves of each unit structure sample in the target frequency band.

[0011] Optionally, converting the electromagnetic frequency response curves of each unit in the target frequency band into two-dimensional images of each unit in the target frequency band includes: converting the electromagnetic frequency response curves of each unit in the target frequency band into two-dimensional images of each unit in the target frequency band using Gram angle difference field technology; converting the electromagnetic frequency response curves of several unit structure samples in the target frequency band into two-dimensional images of several unit structure samples in the target frequency band includes: converting the electromagnetic frequency response curves of several unit structure samples in the target frequency band into two-dimensional images of several unit structure samples in the target frequency band using Gram angle difference field technology.

[0012] Optionally, obtaining the phase profile of the four-channel global electromagnetic metasurface in the target frequency band includes: obtaining a holographic image of the x-polarized four-channel global electromagnetic metasurface, an orbital angular momentum vortex wave, and the target frequency band; obtaining a holographic image of the y-polarized four-channel global electromagnetic metasurface, an orbital angular momentum vortex wave, and the target frequency band; based on the holographic images of the x-polarized and y-polarized four-channel global electromagnetic metasurface, the orbital angular momentum vortex wave, and the target frequency band, and according to the diffraction propagation formula of the orbital angular momentum vortex wave and the electromagnetic wave, obtaining the ideal phase distribution map of the x-polarized four-channel global electromagnetic metasurface and the ideal phase distribution map of the y-polarized four-channel global electromagnetic metasurface; and arranging the ideal phase distribution maps of the x-polarized and y-polarized four-channel global electromagnetic metasurface in the horizontal and vertical directions in an alternating manner according to a preset step size to obtain the phase profile of the four-channel global electromagnetic metasurface in the target frequency band.

[0013] In a second aspect, the present invention provides a four-channel global electromagnetic metasurface design system, comprising: a data acquisition module for acquiring a phase profile of the four-channel global electromagnetic metasurface in a target frequency band; a two-dimensional conversion module for obtaining electromagnetic frequency response curves of each unit of the four-channel global electromagnetic metasurface in the target frequency band based on the phase profile of the four-channel global electromagnetic metasurface in the target frequency band, and converting the electromagnetic frequency response curves of each unit in the target frequency band into two-dimensional images of each unit in the target frequency band; and a structure determination module for generating unit images of each unit based on the two-dimensional images of each unit in the target frequency band using a pre-trained unit image generator, and determining the structure based on the unit images of each unit. The unit structure of each unit is obtained; wherein, the pre-trained unit image generator is obtained in the following way: acquiring unit images of several unit structure samples, and acquiring electromagnetic frequency response curves of several unit structure samples in the target frequency band, and converting the electromagnetic frequency response curves of several unit structure samples in the target frequency band into two-dimensional images of several unit structure samples in the target frequency band; based on the unit images of several unit structure samples and the two-dimensional images in the target frequency band, using the two-dimensional images in the target frequency band as input and the unit images as output, a pre-set generative adversarial network model is trained, and the generator of the trained generative adversarial network model is used as the pre-trained unit image generator.

[0014] In a third aspect, the present invention provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described four-channel global electromagnetic metasurface design method.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described four-channel global electromagnetic metasurface design method.

[0016] In a fifth aspect, the present invention provides a four-channel global electromagnetic metasurface, wherein the unit structure of each unit of the four-channel global electromagnetic metasurface is designed using the above-described four-channel global electromagnetic metasurface design method.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] This invention discloses a four-channel global electromagnetic metasurface design method. Based on the phase profile of the four-channel global electromagnetic metasurface in the target frequency band, the electromagnetic frequency response curves of each unit of the four-channel global electromagnetic metasurface in the target frequency band are obtained. Then, the electromagnetic frequency response curves of each unit in the target frequency band are converted into two-dimensional images of each unit in the target frequency band. Finally, based on the two-dimensional images of each unit in the target frequency band, a pre-trained unit image generator is used to generate unit images of each unit. Based on the unit images, the specific unit structure is obtained to realize the design of the four-channel global electromagnetic metasurface. Specifically, by converting one-dimensional electromagnetic frequency response curves into two-dimensional images to solve the input-output dimension mismatch problem of the unit image generator, the unit image generator can process these data in the manner of image recognition. In particular, it can effectively extract spatial features, patterns, and local relationships in convolution processing, improving the intuitiveness and effectiveness of the reverse design process. This overcomes the problem that training the unit image generator becomes difficult due to dimensional differences, enabling the unit image generator to extract enough effective features to accurately describe and predict the electromagnetic response characteristics of the electromagnetic metasurface unit, effectively improving the training effect, and thus making the design process of the four-channel global electromagnetic metasurface efficient and stable. Attached Figure Description

[0019] Figure 1 This is a flowchart of the four-channel global electromagnetic metasurface design method according to an embodiment of the present invention.

[0020] Figure 2 This is a schematic diagram of two-dimensional matrix transformation of a mirror-symmetric structure according to an embodiment of the present invention.

[0021] Figure 3 This is a schematic diagram of two-dimensional matrix transformation of a rotationally symmetric structure according to an embodiment of the present invention.

[0022] Figure 4 The electromagnetic response distribution of the unit structure sample of this invention at 8 GHz and different metal filling rates is shown in the figure.

[0023] Figure 5 The electromagnetic response distribution of the transmission amplitude of the unit structure sample in this embodiment of the invention at 8 GHz and different metal fill rates is shown.

[0024] Figure 6 The electromagnetic response distribution of the reflection phase of the unit structure sample in this embodiment of the invention is shown in the figure at 8 GHz and under different metal filling rates.

[0025] Figure 7 The electromagnetic response distribution of the transmission phase of the unit structure sample in this embodiment of the invention is shown in the figure at 8 GHz and at different metal fill rates.

[0026] Figure 8 This is a schematic diagram of a two-dimensional image of the electromagnetic frequency response curve of a unit structure sample according to an embodiment of the present invention after Gram angular difference field conversion.

[0027] Figure 9 This is a schematic diagram of the generative adversarial network model architecture according to an embodiment of the present invention.

[0028] Figure 10 This is a graph showing the loss curves of the producer and discriminator of the generative adversarial network model in an embodiment of the present invention as a function of the number of iterations during training.

[0029] Figure 11 This is a schematic diagram illustrating the functional implementation of the four-channel global electromagnetic metasurface in an embodiment of the present invention.

[0030] Figure 12 The figures (a), (b), (c), (d), (e), and (f) are phase profiles of the four-channel global electromagnetic metasurface in an embodiment of the present invention. Figures (a), (b), (c), (d), (e), and (f) are the upper phase distribution of the array at 8 GHz, the lower phase distribution at 8 GHz, the upper phase distribution at 12 GHz, the lower phase distribution at 12 GHz, the upper phase distribution of the array with an alternating arrangement, and the lower phase distribution of the array with an alternating arrangement, respectively.

[0031] Figure 13 This is the normalized radiation pattern of the first-order OAM wave of the four-channel global electromagnetic metasurface operating at 8 GHz, according to an embodiment of the present invention.

[0032] Figure 14 This is the normalized radiation pattern of the second-order OAM wave when the four-channel global electromagnetic metasurface of this invention is operating at 12 GHz.

[0033] Figure 15 The simulation diagram of the four-channel global electromagnetic metasurface at 8 GHz in an embodiment of the present invention is shown in Figure (a) and Figure (b), respectively, showing the amplitude and phase distribution of the first-order OAM near-field electric field.

[0034] Figure 16 The following is a simulation diagram of the four-channel global electromagnetic metasurface at 12 GHz according to an embodiment of the present invention. Figure (a) and Figure (b) are the amplitude and phase distribution diagrams of the near-field electric field of the second-order OAM, respectively.

[0035] Figure 17 The above are schematic diagrams of holographic patterns “AI” and “XD” simulated at different diffraction distances when the four-channel global electromagnetic metasurface operates at 8GHz and 12GHz, respectively, according to an embodiment of the present invention.

[0036] Figure 18 This is a block diagram of the four-channel global electromagnetic metasurface design system according to an embodiment of the present invention. Detailed Implementation

[0037] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0038] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0039] The present invention will now be described in further detail with reference to the accompanying drawings:

[0040] See Figure 1In one embodiment of the present invention, a four-channel global electromagnetic metasurface design method is provided, which can design a four-channel global electromagnetic metasurface that reflects or transmits orbital angular momentum vortex waves and specific holographic patterns under excitation of different frequencies and different linear polarization waves.

[0041] Specifically, the four-channel global electromagnetic metasurface design method of the present invention includes the following steps:

[0042] S1: Obtain the phase profile of the four-channel global electromagnetic metasurface in the target frequency band.

[0043] S2: Based on the phase profile of the four-channel global electromagnetic metasurface in the target frequency band, obtain the electromagnetic frequency response curves of each element of the four-channel global electromagnetic metasurface in the target frequency band, and convert the electromagnetic frequency response curves of each element in the target frequency band into two-dimensional images of each element in the target frequency band.

[0044] S3: Based on the two-dimensional image of each unit in the target frequency band, generate the unit image of each unit through a pre-trained unit image generator, and obtain the unit structure of each unit based on the unit image of each unit.

[0045] The pre-trained unit image generator is obtained as follows: acquiring unit images of several unit structure samples and electromagnetic frequency response curves of several unit structure samples in the target frequency band, and converting the electromagnetic frequency response curves of several unit structure samples in the target frequency band into two-dimensional images of several unit structure samples in the target frequency band; based on the unit images of several unit structure samples and the two-dimensional images in the target frequency band, using the two-dimensional images in the target frequency band as input and the unit images as output, training a pre-set generative adversarial network model, and using the generator of the trained generative adversarial network model as the pre-trained unit image generator.

[0046] This invention discloses a four-channel global electromagnetic metasurface design method. Based on the phase profile of the four-channel global electromagnetic metasurface in the target frequency band, the electromagnetic frequency response curves of each unit of the four-channel global electromagnetic metasurface in the target frequency band are obtained. Then, the electromagnetic frequency response curves of each unit in the target frequency band are converted into two-dimensional images of each unit in the target frequency band. Finally, based on the two-dimensional images of each unit in the target frequency band, a pre-trained unit image generator is used to generate unit images of each unit. Based on the unit images, the specific unit structure is obtained to realize the design of the four-channel global electromagnetic metasurface. Specifically, by converting one-dimensional electromagnetic frequency response curves into two-dimensional images to solve the input-output dimension mismatch problem of the unit image generator, the unit image generator can process these data in the manner of image recognition. In particular, it can effectively extract spatial features, patterns, and local relationships in convolution processing, improving the intuitiveness and effectiveness of the reverse design process. This overcomes the problem that training the unit image generator becomes difficult due to dimensional differences, enabling the unit image generator to extract enough effective features to accurately describe and predict the electromagnetic response characteristics of the electromagnetic metasurface unit, effectively improving the training effect, and thus making the design process of the four-channel global electromagnetic metasurface efficient and stable.

[0047] In one possible implementation, obtaining a unit image of several unit structure samples includes: dividing each unit structure sample into several grids, setting each grid as 1 for metal and 0 for medium and forming a Boolean matrix, obtaining a two-dimensional matrix of the unit structure sample and converting it into a binary image, thereby obtaining a unit image of each unit structure sample.

[0048] Specifically, when acquiring unit images of several unit structure samples, these unit structure samples can be initially generated using MATLAB software.

[0049] This implementation uses a grid-like topology as the basic structure, dividing the unit structure into several grids. The attribute of each grid is randomly assigned as either metal or dielectric, represented by 1 and 0 respectively. In this way, the unit structure can be encoded into a two-dimensional matrix, which can then be converted into a binary image, i.e., a unit image, and saved into the dataset.

[0050] See Figure 2 and 3 To fully explore the design space of the structure while reducing data dimensionality and model complexity, and simultaneously exploring the design parameter space, the unit structure samples can be configured with either a mirror-symmetric structure or a rotationally symmetric structure. A mirror-symmetric structure exhibits symmetry about its central axis, while a rotationally symmetric structure remains unchanged after rotation by a certain angle. These two structural forms not only significantly reduce the number of parameters but also improve symmetry and stability, making the design more flexible and efficient.

[0051] In one possible implementation, taking the design of a four-channel global electromagnetic metasurface under two linear polarization excitations (x and y) as an example, the frequency range is 8–12 GHz. The electromagnetic metasurface consists of three metal layers and two F4B dielectric substrates. The dielectric substrates are all 1 mm thick, with a relative permittivity of 2.65 and a loss tangent of 0.001. The unit period p = 10 mm remains constant. To ensure that the two linear polarizations can operate to the maximum extent in the two different regions of transmission and reflection without coupling, the middle metal layer is a polarization gate, and the upper and lower layers are divided into 16×16 square grids with a side length of 0.5 mm. The attribute of each grid is randomly assigned by the program as metal (1) or dielectric (0), thereby encoding the structure of the entire electromagnetic metasurface into two Boolean matrices.

[0052] Optionally, obtaining the electromagnetic frequency response curves of several unit structure samples in the target frequency band includes: performing electromagnetic simulation on several unit structure samples in the target frequency band using simulation software to obtain the electromagnetic frequency response curves of several unit structure samples in the target frequency band.

[0053] Specifically, the generated unit structure can be electromagnetically simulated using commercial simulation software CST to obtain its electromagnetic frequency response curve. The electromagnetic frequency response includes the reflection and transmission coefficients within a specific frequency range, and the corresponding amplitude and phase characteristics are obtained.

[0054] Optionally, after obtaining the electromagnetic frequency response curves of several unit structure samples in the target frequency band, the unit structure samples with phase changes less than the phase change threshold in the target frequency band are removed based on the electromagnetic frequency response curves of each unit structure sample in the target frequency band.

[0055] Specifically, by analyzing the proportion of different metal regions in the unit structure, some meaningless unit structure samples are eliminated. For example, unit structure samples with small or no phase changes within the target frequency band have no practical significance for the target function, and therefore are not included in the training sample set. In this way, based on the CST simulation results, a set of high-quality samples with practical significance can be collected, laying the foundation for subsequent network training and ensuring the stability of the design.

[0056] This implementation analyzed the electromagnetic response at 8 GHz of 30,000 collected samples using four violin plots. See also Figure 4 and 5 , respectively, represent the reflection phase ∠R under x-polarization. xx And amplitude | R xx |; Figure 6 and Figure 7 These represent the transmission phase ∠T under y-polarization. yy And amplitude | T yyEach figure includes the different metal fill ratios (MFRs) for the elements in both mirror-symmetric and rotationally symmetric configurations. Figure 4 and Figure 5 It can be seen that both reflection and transmission amplitudes remain above 0.8. Except when the MFR is around 50%, the transmission amplitude |T yy |It has decreased somewhat, with some samples showing values ​​below 0.5. On the other hand, Figure 6 and Figure 7 The phase distribution shown also covers a 360° range. For example, for a mirror-symmetric element, when MFR equals 20%, the reflection phase ∠R... xx The phase values ​​are concentrated between 60° and 120°, while when the MFR increases to 70%, the phase values ​​are concentrated between -180° and -60°. Combining the different MFR distributions, it can be determined that the entire sample set, while ensuring high reflectance / transmission amplitude, basically covers the 2π range in phase.

[0057] In one possible implementation, converting the electromagnetic frequency response curves of each unit in the target frequency band into two-dimensional images of each unit in the target frequency band includes: converting the electromagnetic frequency response curves of each unit in the target frequency band into two-dimensional images of each unit in the target frequency band using Gram angle difference field technology; converting the electromagnetic frequency response curves of several unit structure samples in the target frequency band into two-dimensional images of several unit structure samples in the target frequency band includes: converting the electromagnetic frequency response curves of several unit structure samples in the target frequency band into two-dimensional images of several unit structure samples in the target frequency band using Gram angle difference field technology.

[0058] The Gram Angular Difference Field (GADF) technique is used to convert the one-dimensional frequency response curve into a two-dimensional image, which is then paired with the corresponding unit image to form a labeled sample pair. This sample pair, in image form, contains both the spectral features of the electromagnetic frequency response curve and the corresponding physical structural information, thus ensuring a close correlation between the electromagnetic frequency response curve and the structural design. In this way, the subsequent network model can learn the mapping relationship between these data through image recognition, especially during convolution operations, effectively extracting spatial features and patterns from the data, thereby improving the accuracy and efficiency of the design. Specifically, after sample selection, the one-dimensional electromagnetic frequency response curve is converted into a two-dimensional image using GADF. The steps are as follows:

[0059] Given a one-dimensional sequence array x = {x1, x2, ..., xn} containing n values... n Normalize it using the following formula:

[0060]

[0061] The normalized sequence values ​​are then mapped to a polar coordinate system using the following formula, where their values ​​and indices are encoded as cosine angles. And radius r:

[0062]

[0063] Where N is a constant, used as a regularization factor for the polar coordinate system, and is set to 1 here. Based on the above equation, the value of the Gram angular difference field GADF can be obtained by the following equation:

[0064]

[0065] Where I is a unit row vector.

[0066] See Figure 8 The image demonstrates the effect of GADF transformation on the electromagnetic frequency response curve of one of the unit structure samples. It can be seen that the transformed image exhibits rich feature information. This feature information clearly reflects the amplitude and phase distribution in the original electromagnetic response through color changes, thus providing important input data for subsequent network training.

[0067] After collecting and processing the sample set, the model training phase begins. During training, a Generative Adversarial Network (GAN) is used for inverse design. A GAN consists of a generator and a discriminator. The generator first receives a noise vector and a 2D image transformed by GADF as input. Then, after multiple convolutional operations, it generates fake cell images. The discriminator receives these fake cell images and compares them with real cell images in the sample set. The network guides the parameter updates of both modules by calculating the loss functions of the generator and discriminator until the fake images generated by the generator become increasingly realistic, to the point that the discriminator cannot distinguish between them, ultimately achieving network convergence.

[0068] See Figure 9The diagram illustrates the overall structure of a GAN. First, a noise vector of size 128×128×1 and a 2D image obtained through GADF transformation are used as input to the generator. The noise vector introduces randomness, while the GADF-transformed 2D image provides initial feature information. The input data is first fused in the generator through an addition operation. Then, the fused data passes through the first convolutional layer with a kernel size k of 4, a number of kernels n of 64, and a stride s of 2, denoted as k4n64s2. After the convolution operation, the data passes through a batch normalization (BN) layer to standardize the feature maps, and then nonlinearity is introduced through the ReLU activation function. Subsequently, the data enters the downsampling process, passing through multiple convolutional and batch normalization layers. The specific parameters of each layer are as follows: the second convolutional layer has a kernel size of 4, a number of kernels of 128, and a stride of 2 (k4n128s2); the third convolutional layer has a kernel size of 4, a number of kernels of 256, and a stride of 2 (k4n256s2); the fourth convolutional layer has a kernel size of 4, a number of kernels of 512, and a stride of 2 (k4n512s2); and the fifth convolutional layer has a kernel size of 4, a number of kernels of 1024, and a stride of 2 (k4n1024s2). Each layer is processed using the ReLU activation function to progressively reduce the size of the feature map and extract higher-level features. Then, the network undergoes a symmetric transformation operation to enter the upsampling stage, and skip connections are used to directly pass the input features to the output stage, enhancing feature transfer. Finally, a deconvolution operation is performed to enlarge the feature map dimension to the target unit pattern (16×16×2), generating the final output image.

[0069] The generator inputs fake unit images, real unit images, and a 2D image into the discriminator. The discriminator first performs a series of convolution and downsampling operations on the input image to extract image features. The specific convolution parameters are: the first layer has a kernel size of 4, a number of kernels of 64, and a stride of 2 (k4n64s2); the second layer has a kernel size of 4, a number of kernels of 128, and a stride of 2 (k4n128s2); the third layer has a kernel size of 4, a number of kernels of 256, and a stride of 2 (k4n256s2); the fourth layer has a kernel size of 4, a number of kernels of 512, and a stride of 2 (k4n512s2). Each convolutional layer is processed by batch normalization and the ReLU activation function. After all convolutional operations are completed, the discriminator flattens the feature map into a one-dimensional vector, passes it through a Sigmoid activation function, and outputs the probability of the image being real or fake. The generator loss (L...) is calculated... G ) and discriminator loss (L D The network updates weights through backpropagation, thereby enabling the generation and discrimination of high-quality images.

[0070] In this embodiment, the PyTorch deep learning framework was used. The learning rate was 1e-4, and the optimizer used was PyTorch's default Adam optimizer. The network underwent 1500 iterations, with a batch size of 64 per iteration. The hardware used for the network was an Nvidia RTX 4080 graphics card. Figure 10 The diagram shows the loss curves for the generator and discriminator. It can be seen that in the early stages of training, the generator loss L... G And discriminator loss L D The values ​​were around 200 and -80 respectively. As training iterated, the generator gradually demonstrated its ability to generate high-quality samples, and its loss, along with the discriminator's loss, eventually stabilized around 80 and -30 respectively. The network then converged, and training ended.

[0071] After training, the unit image generator can be used to reverse engineer a four-channel global electromagnetic metasurface. The four-channel global electromagnetic metasurface in this embodiment achieves the following functions: Figure 11 As shown, under x-ray polarized plane wave excitation, the four-channel global electromagnetic metasurface projects a holographic image of the letter "AI" operating at 8 GHz and a second-order OAM vortex wave operating at 12 GHz in the backward space; under y-ray polarized plane wave excitation, the four-channel global electromagnetic metasurface projects a first-order OAM vortex wave operating at 8 GHz and a holographic image of the letter "XD" operating at 12 GHz in the forward space.

[0072] In one possible implementation, obtaining the phase profile of the four-channel global electromagnetic metasurface in the target frequency band includes: acquiring a holographic image of the x-polarized four-channel global electromagnetic metasurface, an orbital angular momentum vortex wave, and the target frequency band; acquiring a holographic image of the y-polarized four-channel global electromagnetic metasurface, an orbital angular momentum vortex wave, and the target frequency band; based on the holographic images of the x-polarized and y-polarized four-channel global electromagnetic metasurface, the orbital angular momentum vortex wave, and the target frequency band, and according to the diffraction propagation formulas of the orbital angular momentum vortex wave and the electromagnetic wave, obtaining the ideal phase distribution map of the x-polarized four-channel global electromagnetic metasurface and the ideal phase distribution map of the y-polarized four-channel global electromagnetic metasurface; and arranging the ideal phase distribution maps of the x-polarized and y-polarized four-channel global electromagnetic metasurface in the horizontal and vertical directions in an alternating manner according to a preset step size to obtain the phase profile of the four-channel global electromagnetic metasurface in the target frequency band.

[0073] Specifically, the design process for a four-channel global electromagnetic metasurface integrating OAM and holographic imaging is as follows: Figure 12 As shown. First, the total number of elements in the four-channel global electromagnetic metasurface array was determined to be 40×40, covering a total area of ​​400×400 mm. 2 Then, through theoretical calculations, the ideal phase distribution of the first-order OAM wave operating at 8 GHz and the hologram "AI" was obtained, as shown below. Figure 12 As shown in Figures (a) and (b).

[0074] For a vortex wave, its helical phase distribution can be obtained from the following equation:

[0075]

[0076] Where, x mn and y mn , where are the cell coordinates at the m-th row and n-th column, and l is the order of the OAM wave.

[0077] For holograms, the calculation process is as follows:

[0078] First, a random phase distribution φ(x,y) is initialized to form the wavefront function f(x,y)=exp(jφ(x,y)). Then, the wavefront function is transformed into the frequency domain using Fast Fourier Transform (FFT) to obtain g(x,y):

[0079]

[0080] in, and These represent the Fast Fourier Transform and the Inverse Fast Fourier Transform, respectively; λ is the wavelength; f x and f y is the spatial frequency component; z is the imaging distance.

[0081] In this embodiment, the imaging distance between the two holographic images is set to 180mm. In the frequency domain, the amplitude |G(x,y)| of the target holographic image is combined with the phase g(u,v) to form an intermediate function:

[0082]

[0083] Performing an inverse Fourier transform on this intermediate function to return it to the spatial domain yields the updated wavefront function:

[0084]

[0085] Replace the initial phase φ(x,y) with the phase of the updated wavefront function f′(x,y), and repeat the above process until f′(x,y) meets the requirements of the desired hologram.

[0086] Using the same method, the ideal phase distribution of the second-order OAM wave operating at 12 GHz and the hologram "XD" can be obtained, such as... Figure 12 As shown in Figures (c) and (d).

[0087] Finally, the units of the two frequency types are arranged alternately at intervals, and the corresponding upper and lower phases are also interleaved and merged as follows: Figure 12As shown in Figures (e) and (f), the cells at the two frequencies form a checkerboard-like alternating layout on the same array surface to obtain the final phase profile and achieve the best distribution effect.

[0088] Then, based on the phase profile diagram described above, the trained unit image generator sequentially generates the corresponding unit structure for each unit position in the array. Specifically, the two-dimensional image obtained by converting the theoretically calculated electromagnetic response (i.e., the electromagnetic frequency response curve) at each position through GADF is used as the input to the unit image generator. After the data undergoes multiple operation modules such as convolution, batch normalization, activation function, skip connections, and transpose, a unit image representing the unit structure is output. After the unit images at all positions are output, a full-wave simulation of the entire structure is performed in CST to verify its electromagnetic performance under dual-frequency dual-polarization excitation.

[0089] In another embodiment of the present invention, a four-channel global electromagnetic metasurface is provided, wherein the unit structure of each unit of the four-channel global electromagnetic metasurface is designed using the above-described four-channel global electromagnetic metasurface design method.

[0090] The invention will be further described below with reference to simulation results:

[0091] like Figure 13 and Figure 14 As shown, the normalized radiation patterns of the OAM vortex waves at 8 GHz and 12 GHz are represented in polar coordinates. The amplitude null depth at 8 GHz is -14 dB, corresponding to an angle of 2°; the amplitude null depth at 12 GHz is -22 dB, corresponding to an angle of -5°.

[0092] The amplitude and phase distributions of the electric field of the first-order OAM vortex wave in near-field simulation are shown in the figure. Figure 15 Figure (a) and Figure 15 In Figure (b), it can be seen that the amplitude distribution at 8 GHz has zero energy at its center, and the phase exhibits a first-order spiral pattern; the amplitude and phase distributions of the second-order OAM vortex wave electric field from near-field simulation are shown in Figure (b). Figure 16 Figure (a) and Figure 16 In Figure (b), the amplitude corresponding to 12 GHz also approximately exhibits hollow characteristics, and the phase follows a second-order spiral pattern. Therefore, the designed four-channel global electromagnetic metasurface can generate distinctive OAM vortex electromagnetic waves with different modes at different frequencies.

[0093] like Figure 17As shown, holographic images at different diffraction distances were extracted through near-field simulation. The backward region (z<0) represents the letter "AI" operating at 8 GHz, and the forward region (z>0) represents the letter "XD" operating at 12 GHz. It can be seen that at the theoretical value z = ±180 mm, the two holographic images can be clearly distinguished. At positions deviating from the theoretical values ​​of z = ±160 mm and z = ±200 mm, the image quality decreases slightly, but the overall letter outlines are still clearly discernible. Therefore, the designed four-channel global electromagnetic metasurface can generate different holographic images at different frequencies.

[0094] The four-channel global electromagnetic metasurface design method of this invention has the following characteristics:

[0095] 1) Data pre-screening improves model accuracy: By manually screening the metal proportion of the cells, it is ensured that the generated training dataset covers the entire 2π phase range. This pre-screening process effectively eliminates low-quality data, reduces the waste of computing resources, enables the deep learning model to converge faster, and accurately learns key electromagnetic response characteristics during training, thereby improving design efficiency and accuracy.

[0096] 2) Addressing the Dimensionality Mismatch: By introducing the Gram angular difference field into the generative adversarial network (GAN), the dimensionality mismatch between the electromagnetic response and the unit structure is resolved by transforming the one-dimensional electromagnetic response curve into a two-dimensional image. The model can process this data in an image recognition manner, especially in convolutional processing, where it can effectively extract spatial features, patterns, and local relationships, improving the intuitiveness and effectiveness of the reverse engineering process.

[0097] 3) Saves computational resources: Data pre-screening and visualization effectively reduce the consumption of computational resources during training. High-quality datasets and efficient model training processes shorten the design cycle, reduce costs, and significantly improve resource utilization.

[0098] 4) Enhanced model generalization ability: During model training, the pre-selected high-quality dataset and effective feature extraction methods enable the model to have stronger generalization ability. The model can adapt to more diverse design requirements and application scenarios, improving its flexibility and adaptability in practical applications.

[0099] The following are embodiments of the apparatus of the present invention, which can be used to execute embodiments of the method of the present invention. For details not disclosed in the apparatus embodiments, please refer to the embodiments of the method of the present invention.

[0100] See Figure 18In another embodiment of the present invention, a four-channel global electromagnetic metasurface design system is provided, which can be used to implement the above-mentioned four-channel global electromagnetic metasurface design method. Specifically, the four-channel global electromagnetic metasurface design system includes a data acquisition module, a two-dimensional conversion module, and a structure determination module.

[0101] The system comprises the following modules: a data acquisition module for acquiring the phase profile of the four-channel global electromagnetic metasurface in the target frequency band; a two-dimensional conversion module for obtaining the electromagnetic frequency response curves of each element of the four-channel global electromagnetic metasurface in the target frequency band based on the phase profile of the four-channel global electromagnetic metasurface in the target frequency band, and converting the electromagnetic frequency response curves of each element in the target frequency band into two-dimensional images of each element in the target frequency band; and a structure determination module for generating element images of each element using a pre-trained element image generator based on the two-dimensional images of each element in the target frequency band, and obtaining the element structure of each element based on the element images. The pre-trained unit image generator is obtained as follows: acquire unit images of several unit structure samples and electromagnetic frequency response curves of several unit structure samples in the target frequency band, and convert the electromagnetic frequency response curves of several unit structure samples in the target frequency band into two-dimensional images of several unit structure samples in the target frequency band; based on the unit images of several unit structure samples and the two-dimensional images in the target frequency band, train a pre-defined generative adversarial network model with the two-dimensional images in the target frequency band as input and the unit images as output, and use the generator of the trained generative adversarial network model as the pre-trained unit image generator.

[0102] All relevant content of each step involved in the aforementioned embodiments of the four-channel global electromagnetic metasurface design method can be referenced to the functional description of the corresponding functional module of the four-channel global electromagnetic metasurface design system in the embodiments of the present invention, and will not be repeated here.

[0103] The module division in this embodiment of the invention is illustrative and represents only one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional modules in the various embodiments of the invention can be integrated into a single processor, exist as separate physical entities, or be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0104] In another embodiment of the present invention, a computer device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used for the operation of a four-channel global electromagnetic metasurface design method.

[0105] In another embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the four-channel global electromagnetic metasurface design method in the above embodiments.

[0106] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0107] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0108] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0109] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A four-channel global electromagnetic metasurface design method, characterized in that, include: Obtain the phase profile of the four-channel global electromagnetic metasurface in the target frequency band; Based on the phase profile of the four-channel global electromagnetic metasurface in the target frequency band, the electromagnetic frequency response curves of each element of the four-channel global electromagnetic metasurface in the target frequency band are obtained, and the electromagnetic frequency response curves of each element in the target frequency band are converted into two-dimensional images of each element in the target frequency band. Based on the two-dimensional images of each unit in the target frequency band, a pre-trained unit image generator is used to generate unit images of each unit, and the unit structure of each unit is obtained based on the unit images of each unit. The pre-trained unit image generator is obtained in the following way: Acquire unit images of several unit structure samples and electromagnetic frequency response curves of several unit structure samples in the target frequency band, and convert the electromagnetic frequency response curves of several unit structure samples in the target frequency band into two-dimensional images of several unit structure samples in the target frequency band; based on the unit images of several unit structure samples and the two-dimensional images in the target frequency band, use the two-dimensional images in the target frequency band as input and the unit images as output to train a pre-set generative adversarial network model, and use the generator of the trained generative adversarial network model as a pre-trained unit image generator.

2. The four-channel global electromagnetic metasurface design method according to claim 1, characterized in that, The unit structure of the sample is a mirror-symmetric structure or a rotationally symmetric structure.

3. The four-channel global electromagnetic metasurface design method according to claim 1, characterized in that, Obtaining unit images of several unit structure samples includes: dividing each unit structure sample into several grids, setting each grid as 1 for metal and 0 for medium and forming a Boolean matrix to obtain a two-dimensional matrix of the unit structure sample and converting it into a binary image to obtain the unit image of each unit structure sample. The process of obtaining the electromagnetic frequency response curves of several unit structure samples in the target frequency band includes: performing electromagnetic simulation on several unit structure samples in the target frequency band using simulation software to obtain the electromagnetic frequency response curves of several unit structure samples in the target frequency band.

4. The four-channel global electromagnetic metasurface design method according to claim 1, characterized in that, After obtaining the electromagnetic frequency response curves of several unit structure samples in the target frequency band, the unit structure samples with phase changes less than the phase change threshold in the target frequency band are removed based on the electromagnetic frequency response curves of each unit structure sample in the target frequency band.

5. The four-channel global electromagnetic metasurface design method according to claim 1, characterized in that, The step of converting the electromagnetic frequency response curves of each unit in the target frequency band into two-dimensional images of each unit in the target frequency band includes: converting the electromagnetic frequency response curves of each unit in the target frequency band into two-dimensional images of each unit in the target frequency band using Gram angle difference field technology; the step of converting the electromagnetic frequency response curves of several unit structure samples in the target frequency band into two-dimensional images of several unit structure samples in the target frequency band includes: converting the electromagnetic frequency response curves of several unit structure samples in the target frequency band into two-dimensional images of several unit structure samples in the target frequency band using Gram angle difference field technology.

6. The four-channel global electromagnetic metasurface design method according to claim 1, characterized in that, The process of obtaining the phase profile of the four-channel global electromagnetic metasurface in the target frequency band includes: Acquire holographic images of x-polarized, orbital angular momentum vortex waves and target frequency bands of a four-channel global electromagnetic metasurface; and holographic images of y-polarized, orbital angular momentum vortex waves and target frequency bands. Based on the holographic images of x-polarization and y-polarization, the orbital angular momentum vortex wave and the target frequency band, the ideal phase distribution map of x-polarization and the ideal phase distribution map of y-polarization are obtained according to the diffraction propagation formula of orbital angular momentum vortex wave and electromagnetic wave. The ideal phase distribution maps of x-polarization and y-polarization are arranged alternately in the horizontal and vertical directions according to a preset step size to obtain the phase profile map of the four-channel global electromagnetic metasurface in the target frequency band.

7. A four-channel global electromagnetic metasurface design system, characterized in that, include: The data acquisition module is used to acquire the phase profile of the four-channel global electromagnetic metasurface in the target frequency band; The two-dimensional conversion module is used to obtain the electromagnetic frequency response curves of each unit of the four-channel global electromagnetic metasurface in the target frequency band based on the phase profile diagram of the four-channel global electromagnetic metasurface in the target frequency band, and convert the electromagnetic frequency response curves of each unit in the target frequency band into two-dimensional images of each unit in the target frequency band. The structure determination module is used to generate unit images of each unit based on the two-dimensional images of each unit in the target frequency band using a pre-trained unit image generator, and to obtain the unit structure of each unit based on the unit images; wherein, the pre-trained unit image generator is obtained in the following manner: Acquire unit images of several unit structure samples and electromagnetic frequency response curves of several unit structure samples in the target frequency band, and convert the electromagnetic frequency response curves of several unit structure samples in the target frequency band into two-dimensional images of several unit structure samples in the target frequency band; based on the unit images of several unit structure samples and the two-dimensional images in the target frequency band, use the two-dimensional images in the target frequency band as input and the unit images as output to train a pre-set generative adversarial network model, and use the generator of the trained generative adversarial network model as a pre-trained unit image generator.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the four-channel global electromagnetic metasurface design method as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the four-channel global electromagnetic metasurface design method as described in any one of claims 1 to 6.

10. A four-channel global electromagnetic metasurface, characterized in that, The unit structure of each cell of the four-channel global electromagnetic metasurface is designed using the four-channel global electromagnetic metasurface design method described in any one of claims 1 to 6.

Images