Deep learning-based tight-pitch antenna decoupling method and device, and storage medium

Abstract

The invention discloses a tight-pitch antenna decoupling method and device based on deep learning and a storage medium. The method comprises the following steps: acquiring tight-pitch antenna data; inputting the tight-spacing antenna data into the antenna decoupling structure prediction model to obtain decoupling structure parameters; performing tight-spacing antenna decoupling processing according to the decoupling structure parameters; the antenna decoupling structure prediction model is obtained through the following steps: obtaining rectangular microstrip patch antenna information; the patch length and the patch width are calculated; establishing a binary antenna model; performing electromagnetic field simulation processing by utilizing a preset script program and electromagnetic simulation software to obtain a simulation result data set; constructing an initial convolutional neural network model; and inputting the simulation result data set into the initial convolutional neural network model according to the mean square error, so that the initial convolutional neural network model is trained to obtain an antenna decoupling structure prediction model. According to the invention, antenna decoupling is realized, and efficiency and accuracy are improved. The method can be widely applied to the technical field of wireless communication.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to a method, device and storage medium for decoupling closely spaced antennas based on deep learning. Background Art

[0002] In the design process of MIMO multi-input multi-output system and array antenna architecture, there is a problem of electromagnetic coupling between adjacent antennas. Strong electromagnetic coupling between antennas will lead to problems such as reduced antenna gain, radiation pattern shift, and performance deterioration. With the advancement and development of modern communication methods, portable and miniaturized communication devices with closely spaced antennas have gradually become mainstream, which leads to stronger electromagnetic coupling between antennas and more serious shifts in antenna radiation patterns. Traditional decoupling methods use artificial design, adjustment and optimization to design the decoupling structure between antennas, which is time-consuming and inefficient. At the same time, the decoupling structure designed by machine learning and other methods still has the problem of large angle shift of the antenna pattern and low accuracy.

[0003] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the invention

[0004] The embodiments of the present invention provide a method, device and storage medium for decoupling closely spaced antennas based on deep learning, which effectively improves efficiency and accuracy.

[0005] On the one hand, an embodiment of the present invention provides a method for decoupling closely spaced antennas based on deep learning, comprising the following steps:

[0006] Acquire closely spaced antenna data;

[0007] Inputting the closely spaced antenna data into an antenna decoupling structure prediction model to obtain decoupling structure parameters;

[0008] Performing close-spaced antenna decoupling processing according to the decoupling structure parameters;

[0009] The antenna decoupling structure prediction model is obtained by the following steps:

[0010] Get rectangular microstrip patch antenna information;

[0011] Calculating the patch length and patch width according to the rectangular microstrip patch antenna information;

[0012] Establishing a binary antenna model according to the patch length and the patch width;

[0013] According to the binary antenna model, electromagnetic field simulation processing is performed using a preset script program and electromagnetic simulation software to obtain a simulation result data set, wherein the simulation result data set includes S21 coupling parameters and directional pattern parameters;

[0014] Build an initial convolutional neural network model;

[0015] According to the mean square error, the simulation result data set is input into the initial convolutional neural network model so that the initial convolutional neural network model is trained to obtain the antenna decoupling structure prediction model.

[0016] In some embodiments, the calculating the patch length and the patch width according to the rectangular microstrip patch antenna information includes:

[0017] Extracting the dielectric plate relative permittivity and dielectric plate thickness from the rectangular microstrip patch antenna information;

[0018] Calculating the patch width according to the speed of light, the working resonant frequency and the relative dielectric constant of the dielectric plate;

[0019] Calculating the length difference according to the patch width, the dielectric plate thickness and the dielectric plate relative dielectric constant;

[0020] The patch length is calculated according to the length difference, the speed of light, the working resonant frequency and the relative dielectric constant of the dielectric plate.

[0021] In some embodiments, establishing a binary antenna model according to the patch length and the patch width includes:

[0022] Extracting the dielectric plate thickness from the rectangular microstrip patch antenna information;

[0023] Set the antenna spacing;

[0024] The dual antenna model is established according to the antenna spacing, the patch length, the patch width, the dielectric plate thickness and the size of the radiation unit.

[0025] In some embodiments, the electromagnetic field simulation processing is performed according to the binary antenna model using a preset script program and electromagnetic simulation software to obtain a simulation result data set, including:

[0026] According to the binary antenna model, setting simulation model parameters;

[0027] Create electromagnetic simulation projects;

[0028] Initializing project settings for the electromagnetic simulation project;

[0029] Establishing a simulation model according to the simulation model parameters and the project settings;

[0030] The simulation model is used to perform simulation to obtain the simulation result data set.

[0031] In some embodiments, the process of constructing the initial convolutional neural network model includes:

[0032] Construct a first convolutional layer, where the convolution kernel size of the first convolutional layer is 32×32×32;

[0033] After the first convolutional layer, a first pooling layer is constructed, and the size of the first pooling layer is 16×16×32;

[0034] After the first pooling layer, a second convolutional layer is constructed, and the convolution kernel size of the second convolutional layer is 16×16×64;

[0035] After the second convolutional layer, a second pooling layer is constructed, and the size of the second pooling layer is 5×5×64;

[0036] After the second pooling layer, a third convolution layer is constructed, and the convolution kernel size of the third convolution layer is 5×5×128;

[0037] After the third convolutional layer, a third pooling layer is constructed, and the size of the third pooling layer is 2×2×128;

[0038] After the third pooling layer, construct a flattening layer;

[0039] After the flattening layer, a first fully connected layer is constructed, wherein the number of neurons in the first fully connected layer is 128;

[0040] After the first fully connected layer, a second fully connected layer is constructed, and the number of neurons in the second fully connected layer is 64;

[0041] After the second fully connected layer, a third fully connected layer is constructed, and the number of neurons in the third fully connected layer is 32;

[0042] After the third fully connected layer, an output layer is constructed, and the number of neurons in the output layer is 14.

[0043] In some embodiments, calculating the patch width according to the speed of light, the working resonant frequency and the relative dielectric constant of the dielectric plate includes:

[0044] According to the speed of light, the working resonant frequency and the relative dielectric constant of the dielectric plate, the patch width is calculated by the patch width calculation formula, and the patch width calculation formula is:

[0045]

[0046] Where W is the patch width, c is the speed of light, and f r is the working resonant frequency, ε reff is the relative dielectric constant of the dielectric plate.

[0047] In some embodiments, calculating the patch length according to the length difference, the speed of light, the operating resonant frequency, and the relative dielectric constant of the dielectric plate includes:

[0048] The patch length is calculated according to the length difference, the speed of light, the working resonant frequency and the relative dielectric constant of the dielectric plate using a patch length calculation formula, and the patch length calculation formula is:

[0049]

[0050] Where L is the length of the patch, c is the speed of light, and f r is the working resonant frequency, ε reff is the relative dielectric constant of the dielectric plate, and ΔL is the length difference.

[0051] On the other hand, an embodiment of the present invention provides a close-spaced antenna decoupling device based on deep learning, comprising:

[0052] The first module is used to obtain closely spaced antenna data;

[0053] The second module is used to input the close-spaced antenna data into an antenna decoupling structure prediction model to obtain decoupling structure parameters;

[0054] A third module is used to perform a decoupling process for closely spaced antennas according to the decoupling structure parameters;

[0055] The antenna decoupling structure prediction model is obtained by the following steps:

[0056] Get rectangular microstrip patch antenna information;

[0057] Calculating the patch length and patch width according to the rectangular microstrip patch antenna information;

[0058] Establishing a binary antenna model according to the patch length and the patch width;

[0059] According to the binary antenna model, electromagnetic field simulation processing is performed using a preset script program and electromagnetic simulation software to obtain a simulation result data set, wherein the simulation result data set includes S21 coupling parameters and directional pattern parameters;

[0060] Build an initial convolutional neural network model;

[0061] According to the mean square error, the simulation result data set is input into the initial convolutional neural network model so that the initial convolutional neural network model is trained to obtain the antenna decoupling structure prediction model.

[0062] In another aspect, an embodiment of the present invention provides a computer device, comprising:

[0063] at least one processor;

[0064] at least one memory for storing at least one program;

[0065] When the at least one program is executed by the at least one processor, the at least one processor implements the described method.

[0066] On the other hand, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described above is implemented.

[0067] The beneficial effects of the present invention are as follows:

[0068] The embodiment of the present invention first obtains the closely spaced antenna data, then inputs the closely spaced antenna data into the antenna decoupling structure prediction model to obtain the decoupling structure parameters, and finally performs closely spaced antenna decoupling processing according to the decoupling structure parameters, thereby realizing antenna decoupling and improving efficiency and accuracy. Among them, the antenna decoupling structure prediction model first obtains the rectangular microstrip patch antenna information, calculates the patch length and patch width, and then establishes a binary antenna model according to the patch length and patch width. According to the binary antenna model, a preset script program and electromagnetic simulation software are used to perform electromagnetic field simulation processing to obtain a simulation result data set, and then an initial convolutional neural network model is constructed. Finally, according to the mean square error, the simulation result data set is input into the initial convolutional neural network model so that the initial convolutional neural network model is trained, thereby improving the model accuracy.

[0069] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0071] Figure 1 This is a flow chart of a method for decoupling closely spaced antennas based on deep learning according to an embodiment of the present invention;

[0072] Figure 2 A flowchart of obtaining an antenna decoupling structure prediction model according to an embodiment of the present invention;

[0073] Figure 3 A schematic diagram of a dual antenna model according to an embodiment of the present invention;

[0074] Figure 4 A schematic diagram of a joint simulation process according to an embodiment of the present invention;

[0075] Figure 5 A schematic diagram of the front side of a simulation model of a dual-element antenna array according to an embodiment of the present invention;

[0076] Figure 6 A schematic diagram of the back side of a simulation model of a dual-element antenna array according to an embodiment of the present invention;

[0077] Figure 7 A schematic diagram of a convolutional neural network model structure according to an embodiment of the present invention;

[0078] Figure 8 A schematic diagram of a comparison result of S parameters according to an embodiment of the present invention;

[0079] Fig. 9 A schematic diagram of a comparison result of a directional diagram according to an embodiment of the present invention;

[0080] Fig.10 A schematic diagram of an overall process of designing a decoupling structure according to an embodiment of the present invention;

[0081] Fig.11 This is a structural schematic diagram of a close-spaced antenna decoupling device based on deep learning according to an embodiment of the present invention;

[0082] Fig.12 The figure is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0083] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present application. They are only examples of devices and methods consistent with some aspects of the embodiments of the present application as detailed in the attached claims.

[0084] It is understood that the terms "first", "second", etc. used in this application can be used to describe various concepts in this article, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another concept. For example, without departing from the scope of the embodiment of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if" and "if" as used herein can be interpreted as "at the time of" or "when" or "in response to determination".

[0085] The terms "at least one", "multiple", "each", "any", etc. used in this application, at least one includes one, two or more, multiple includes two or more, each refers to each of the corresponding multiple, and any refers to any one of the multiple.

[0086] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0087] Before describing the embodiments of the present application in detail, some nouns and terms involved in the embodiments of the present application are first described. The nouns and terms involved in the embodiments of the present application are subject to the following explanations.

[0088] Antenna decoupling: refers to reducing or eliminating mutual interference between antenna units to improve the antenna's radiation efficiency and communication performance. In a multi-antenna system, mutual coupling between antenna units will lead to reduced antenna efficiency and distorted radiation patterns, thus affecting the performance of the entire system.

[0089] In the related technology, electromagnetic coupling between adjacent antennas is an inevitable problem in the design process of antenna arrays. Strong electromagnetic coupling between antennas will lead to problems such as reduced antenna gain, shifted radiation patterns, and deteriorated performance. With the advancement and development of modern communication methods, portable and miniaturized communication devices have gradually become the mainstream, and new requirements for antenna miniaturization have been put forward, which will lead to stronger electromagnetic coupling between antennas and more serious shifts in antenna radiation patterns. At that time, decoupling of tight spacing pattern correction will become an urgent problem to be solved. At the same time, with the rapid development of wireless communication technology, the field of antenna design is facing unprecedented performance challenges, especially the urgent need for high gain and excellent directivity, making it difficult for traditional single antennas to meet the complex requirements of the information age. In this context, array antennas have been widely adopted and applied in many fields as a solution. However, in MIMO multi-input and multi-output systems and array antenna architectures, the mutual coupling phenomenon between antenna units inevitably leads to distortion of the radiation pattern, especially in mobile devices that pursue extreme miniaturization. This problem is particularly prominent. Current research focuses on the decoupling technology of compact (spacing reduced to one twentieth of a wavelength) array antennas, striving to break through technical bottlenecks. Despite significant progress, the problem of pattern tilt still needs to be solved. Therefore, it is particularly important to explore and develop decoupling technology for compact microstrip array antennas, especially focusing on the precise correction and optimization of the pattern. This move can not only effectively improve the overall performance of the antenna system, but also has far-reaching significance for promoting the miniaturization and integration of array antennas and other engineering and technological advances. At the same time, deep learning, as an emerging force, is profoundly changing the traditional pattern of electromagnetic field modeling and simulation. Compared with traditional computationally intensive methods such as the method of moments and the finite element method, deep learning, with its powerful data processing capabilities, can extract efficient and accurate prediction models from massive data, significantly accelerate the process of electromagnetic field solution, and improve the calculation accuracy. In addition, deep learning is increasingly widely used in the field of electromagnetic optimization and design. It can not only automatically adjust design parameters to achieve the optimization of device performance, but also inject innovative design concepts and thinking methods into the field of electromagnetics, greatly improving design efficiency and flexibility. The traditional decoupling method uses artificial design adjustment and optimization to design the decoupling structure between antennas. The spacing between antennas is relatively large, time-consuming and inefficient. At the same time, the decoupling structure designed using machine learning and other methods still has the problem of large angle deviation of the antenna pattern and low accuracy.

[0090] In view of this, the embodiment of the present invention introduces a deep learning algorithm to address the problem of patch antenna pattern tilt when the spacing is tight at 0.05λ0, and combines a script program and electromagnetic simulation software to perform electromagnetic field simulation, exploring a comprehensive technology for achieving simultaneous decoupling and pattern correction.

[0091] The method for decoupling tightly spaced antennas based on deep learning provided in the embodiment of the present application relates to the field of wireless communication technology. The method for decoupling tightly spaced antennas based on deep learning provided in the embodiment of the present application can be applied to a terminal, can also be applied to a server, and can also be software running in a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and a vehicle-mounted terminal, etc., but is not limited to this; the server side can be configured as an independent physical server, or can be configured as a server cluster or a distributed system composed of multiple physical servers, and can also be configured to provide cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and cloud servers for basic cloud computing services such as big data and artificial intelligence platforms. The server can also be a node server in a blockchain network; the software can be an application that implements a method for decoupling tightly spaced antennas based on deep learning, etc., but is not limited to the above forms.

[0092] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0093] The following is a detailed explanation of the embodiments of the present application in conjunction with the accompanying drawings:

[0094] Figure 1 is an optional flow chart of a method for decoupling closely spaced antennas based on deep learning provided in an embodiment of the present application. Figure 1 The method may include but is not limited to steps S101 to S103.

[0095] Step S101, obtaining closely spaced antenna data;

[0096] Step S102, inputting the closely spaced antenna data into the antenna decoupling structure prediction model to obtain decoupling structure parameters;

[0097] Step S103: performing decoupling processing for closely spaced antennas according to decoupling structure parameters.

[0098] In some embodiments, closely spaced antenna data may be acquired first, and then input into an antenna decoupling structure prediction model to obtain decoupling structure parameters, and then closely spaced antenna decoupling processing may be performed based on the decoupling structure parameters, thereby correcting the radiation pattern, achieving antenna decoupling, and improving efficiency and accuracy.

[0099] In some embodiments, Figure 2 As shown, the specific implementation process of obtaining the antenna decoupling structure prediction model may include but is not limited to steps S201 to S206.

[0100] Step S201, obtaining rectangular microstrip patch antenna information;

[0101] Step S202, calculating the patch length and patch width according to the rectangular microstrip patch antenna information;

[0102] Step S203, establishing a binary antenna model according to the patch length and patch width;

[0103] Step S204: According to the binary antenna model, electromagnetic field simulation processing is performed using a preset script program and electromagnetic simulation software to obtain a simulation result data set, where the simulation result data set includes S21 coupling parameters and pattern parameters;

[0104] Step S205, constructing an initial convolutional neural network model;

[0105] Step S206: According to the mean square error, the simulation result data set is input into the initial convolutional neural network model so that the initial convolutional neural network model is trained to obtain the antenna decoupling structure prediction model.

[0106] In steps S201 to S206 shown in the embodiment of the present application, an antenna decoupling structure prediction model is obtained, thereby improving the model accuracy.

[0107] In step S201 of some embodiments, the rectangular microstrip patch antenna information can be obtained through the antenna database. The rectangular microstrip patch antenna information can also be obtained by other means, not limited thereto. It can be understood that by adding different metal strips between antennas or making grooves and seams in the floor, decoupling and pattern correction between antennas can be achieved. For the convenience of research, this embodiment selects parallel and side-by-side binary microstrip antennas operating at 5.8 GHz as the research object, and arranges the binary microstrip antennas along the H plane. The H plane of the antenna, also called the magnetic plane, refers to a plane parallel to the direction of the magnetic field (or perpendicular to the direction of the electric field) and containing the maximum radiation direction of the antenna.

[0108] In some embodiments, in step S202, the patch length and patch width are calculated according to the rectangular microstrip patch antenna information, which may include but is not limited to the following steps:

[0109] Extract the dielectric plate relative permittivity and dielectric plate thickness from the rectangular microstrip patch antenna information;

[0110] Calculate the patch width based on the speed of light, the working resonant frequency and the relative dielectric constant of the dielectric plate;

[0111] Calculate the length difference based on the patch width, dielectric plate thickness and dielectric plate relative dielectric constant;

[0112] The patch length is calculated based on the length difference, speed of light, operating resonant frequency and relative dielectric constant of the dielectric plate.

[0113] In some embodiments, the microstrip antenna used in this embodiment is an antenna formed by laying a conductive patch on top of a dielectric substrate with a ground plane. According to the different shapes of the conductive patch, the microstrip patch antenna can be divided into annular, rectangular and circular microstrip patch antennas, etc. The antenna designed in this embodiment is a rectangular microstrip patch antenna, which is fed by a microstrip line. In order to achieve impedance matching, a quarter impedance matching transformer is also designed. The size of the rectangular microstrip patch antenna is designed according to the electrical parameters of the selected medium (such as the relative dielectric constant and loss tangent of the dielectric plate). The relative dielectric constant and the thickness of the dielectric plate can be extracted from the rectangular microstrip patch antenna information, and then the patch width is calculated by the patch width calculation formula according to the speed of light, the working resonant frequency and the relative dielectric constant of the dielectric plate, where the patch width calculation formula is: Where W is the patch width, c is the speed of light, and f r is the working resonant frequency, ε reff is the relative dielectric constant of the dielectric plate. Then, the length difference is calculated based on the patch width, dielectric plate thickness and dielectric plate relative dielectric constant. The length difference is calculated as follows: In the formula, ΔL is the length difference, and h is the thickness of the dielectric plate. Finally, according to the length difference, light speed, working resonant frequency and relative dielectric constant of the dielectric plate, the patch length is calculated by the patch length calculation formula. The patch length calculation formula is: Where L is the patch length, c is the speed of light, and f r is the working resonant frequency, ε reff is the relative dielectric constant of the dielectric plate, and ΔL is the length difference. For example, according to the working resonant frequency of the antenna of 5.8 GHz, the thickness of the dielectric plate (known condition) is 1.6 mm, and the size of the entire dielectric plate of the binary antenna can be designed to be 46.00 mm × 31.00 mm (patch length × patch width).

[0114] In some embodiments, in step S203, a binary antenna model is established according to the patch length and the patch width, which may include but is not limited to the following steps:

[0115] Extract the dielectric plate thickness from the rectangular microstrip patch antenna information;

[0116] Set the antenna spacing;

[0117] A binary antenna model is established based on antenna spacing, patch length, patch width, dielectric plate thickness and radiation unit size.

[0118] In some embodiments, the dielectric plate thickness can be extracted from the rectangular microstrip patch antenna information, and then the antenna spacing can be set. Then, a binary antenna model can be established based on the antenna spacing, patch length, patch width, dielectric plate thickness, and radiation unit size. For example, the binary antenna model is as follows: Figure 3 As shown, the thickness of the extracted dielectric plate is 1.6mm, and the size of the entire dielectric plate is 46.00mm×31.00mm (patch length×patch width). The spacing between the two antennas can be set to 2.52mm (given the conditions, the specific spacing depends on the researcher's own considerations), which is about one twentieth of the wavelength, and a dual antenna model is obtained. At the same time, the size of the upper surface radiation unit of a single antenna is 12.80mm×11.33mm (length×width).

[0119] In some embodiments, in step S204, according to the binary antenna model, electromagnetic field simulation processing is performed using a preset script program and electromagnetic simulation software to obtain a simulation result data set, which may include but is not limited to the following steps:

[0120] According to the binary antenna model, set the simulation model parameters;

[0121] Create electromagnetic simulation projects;

[0122] Initialize the project settings for the electromagnetic simulation project;

[0123] Establish a simulation model according to simulation model parameters and project settings;

[0124] The simulation model is used to perform simulation and obtain a simulation result data set.

[0125] In some embodiments, the establishment of a prediction model requires a large amount of data training. However, in the process of acquiring a large data set, a lot of time and effort is still required. Therefore, a script program written in MATLAB can be used to communicate with CST Studio Suite (electromagnetic simulation software) through an API interface, and CST StudioSuite can be used to automatically perform electromagnetic field simulation, thereby realizing an automated modeling process and simulation, thereby greatly improving data acquisition efficiency. The preset script program can be set to a script program written in MATLAB, and the electromagnetic simulation software can be set to CST StudioSuite. The joint simulation process is as follows: Figure 4 As shown, in the electromagnetic simulation software, the simulation model parameters can be set according to the binary antenna model, and then the electromagnetic simulation project can be created, and then the project settings of the electromagnetic simulation project can be initialized, and the simulation model can be established according to the simulation model parameters and project settings, and finally the simulation model can be used for simulation to obtain the simulation result data set. Among them, the simulation result data set can include S21 decoupling parameters and directional pattern parameters. For example, the front view of the simulation model of the binary antenna array is as follows Figure 5 As shown in Figure 1, it mainly includes two patch antennas and a metal strip between the patch antennas. Figure 6 As shown, there is a strip-shaped defective ground structure between the patch antennas. It can be understood that decoupling can be performed using the metal strip and the defective ground structure.

[0126] In some embodiments, in step S205, the process of constructing the initial convolutional neural network model may include but is not limited to the following steps:

[0127] Construct the first convolutional layer, the convolution kernel size of the first convolutional layer is 32×32×32;

[0128] After the first convolutional layer, the first pooling layer is constructed, and the size of the first pooling layer is 16×16×32;

[0129] After the first pooling layer, the second convolutional layer is constructed, and the convolution kernel size of the second convolutional layer is 16×16×64;

[0130] After the second convolutional layer, the second pooling layer is constructed, and the size of the second pooling layer is 5×5×64;

[0131] After the second pooling layer, the third convolution layer is constructed, and the convolution kernel size of the third convolution layer is 5×5×128;

[0132] After the third convolutional layer, the third pooling layer is constructed, and the size of the third pooling layer is 2×2×128;

[0133] After the third pooling layer, a flattening layer is constructed;

[0134] After the flattening layer, the first fully connected layer is constructed. The number of neurons in the first fully connected layer is 128;

[0135] After the first fully connected layer, a second fully connected layer is constructed. The number of neurons in the second fully connected layer is 64.

[0136] After the second fully connected layer, the third fully connected layer is constructed. The number of neurons in the third fully connected layer is 32.

[0137] After the third fully connected layer, the output layer is constructed, and the number of neurons in the output layer is 14.

[0138] In some embodiments, the convolutional neural network model structure is as follows Figure 7 As shown, the first convolution layer and the first pooling layer can be constructed first, and then the second convolution layer and the second pooling layer can be constructed, and then the third convolution layer and the third pooling layer can be constructed. The convolution kernel size of the first convolution layer is 32×32×32, the size of the first pooling layer is 16×16×32, the size of the convolution kernel size of the second convolution layer is 16×16×64, the size of the second pooling layer is 5×5×64, the size of the convolution kernel size of the third convolution layer is 5×5×128, and the size of the third pooling layer is 2×2×128. Furthermore, there is a Dropout layer after each pooling layer, as well as the LeakyReLU activation function and the same padding setting. After the third pooling layer, a flattening layer is constructed, and the first fully connected layer, the second fully connected layer and the third fully connected layer are constructed, wherein the number of neurons in the first fully connected layer is 128, the number of neurons in the second fully connected layer is 64, and the number of neurons in the third fully connected layer is 32. Furthermore, each fully connected layer uses L2 regularization and LeakyReLU activation function. After the third fully connected layer, the output layer is constructed, where the number of neurons in the output layer is 14 and the ReLU activation function is used. Furthermore, the input image size of the convolutional neural network model is 32×32×1 (length×width×height).

[0139] In some embodiments, in step S206, the simulation result data set can be input into the initial convolutional neural network model according to the mean square error, so that the initial convolutional neural network model is trained to obtain the antenna decoupling structure prediction model. Exemplarily, the simulation result data set with an image size of 32×32×1 is input into the initial convolutional neural network model, and the mean square error MSE loss function is used to calculate the loss during the training process to perform parameter optimization, wherein the calculation formula of the mean square error is: In the formula, MSE is the mean square error, n is the number of samples, y is the true value, is the predicted value of the model. It can be understood that the predicted value output by the antenna decoupling structure prediction model is the relevant parameter of the antenna decoupling structure, which is used for mechanical energy antenna decoupling. Moreover, the model prediction value can be compared with the true value to verify the accuracy of the model. The true value is the actual decoupling structure parameter, which is obtained through actual simulation of CST. In order to verify the accuracy of the network design, the S21 decoupling coefficient corresponding to the decoupling structure parameter output by the antenna decoupling structure prediction model after training is compared with the S21 decoupling coefficient of the actual structure, and the S211 decoupling coefficient of the actual structure is obtained through simulation of CST software. On the premise of meeting the expected design, the antenna decoupling structure prediction model is used to design a decoupling structure that meets the actual required parameters. It can be understood that the CNN convolutional neural network can effectively extract the features of image, audio and other data through components such as convolutional layers, pooling layers, and activation functions. The decoupling structure is essentially a two-dimensional planar structure, which is equivalent to an image, so its features can be extracted through a convolutional neural network. In the comparison results between the true value and the predicted value, the comparison results of the S parameters are as follows Figure 8 The comparison results of the directional diagrams are shown in Fig. 9 As shown in the figure. From the prediction results, it can be seen that the predicted response (predicted value) of the antenna decoupling structure prediction model and the expected response (true value) are well matched in waveform, whether it is S parameters or radiation patterns, and the resonant frequency is also predicted more accurately. After using deep learning to design the decoupling structure, the tilted radiation pattern was corrected.

[0140] In some embodiments, the overall process of designing a decoupling structure is as follows: Fig.10 As shown, the binary antenna with a decoupling structure can be modeled first, and then MATLAB can be used to control the CST simulation software for automatic modeling and simulation. A certain amount of simulation structure data can then be input into the convolutional neural network for training to obtain a prediction model. Finally, the accuracy of the prediction model can be verified to design the decoupling structure.

[0141] In some embodiments, this embodiment can decouple closely spaced antennas and correct the directional pattern, taking both into account. This design not only reduces the area size of the antenna, but also reduces the weight and material usage, thereby reducing the processing and manufacturing costs. The lightweight and portable design achieved in the application enables the new technology to be widely used in multiple related fields such as airborne, satellite-borne, shipborne, and vehicle-borne. This embodiment utilizes deep learning to perform closely spaced decoupling and correct the directional pattern, and realizes its functions in combination with metal strips and defective ground. A prediction model based on a deep convolutional neural network is constructed to achieve the design of a decoupling structure.

[0142] The beneficial effects of implementing the embodiments of the present invention include: the embodiments of the present invention first obtain the closely spaced antenna data, then input the closely spaced antenna data into the antenna decoupling structure prediction model, obtain the decoupling structure parameters, and finally perform closely spaced antenna decoupling processing according to the decoupling structure parameters, thereby achieving antenna decoupling and improving efficiency and accuracy. Among them, the antenna decoupling structure prediction model first obtains the rectangular microstrip patch antenna information, calculates the patch length and patch width, and then establishes a binary antenna model according to the patch length and patch width. According to the binary antenna model, a preset script program and electromagnetic simulation software are used to perform electromagnetic field simulation processing to obtain a simulation result data set, and then an initial convolutional neural network model is constructed. Finally, according to the mean square error, the simulation result data set is input into the initial convolutional neural network model so that the initial convolutional neural network model is trained, thereby improving the model accuracy.

[0143] like Fig.11 As shown, the embodiment of the present invention also provides a close-spaced antenna decoupling device based on deep learning, including:

[0144] The first module 801 is used to obtain closely spaced antenna data;

[0145] The second module 802 is used to input the closely spaced antenna data into the antenna decoupling structure prediction model to obtain the decoupling structure parameters;

[0146] The third module 803 is used to perform decoupling processing of closely spaced antennas according to decoupling structure parameters;

[0147] The antenna decoupling structure prediction model is obtained through the following steps:

[0148] Get rectangular microstrip patch antenna information;

[0149] According to the rectangular microstrip patch antenna information, calculate the patch length and patch width;

[0150] A binary antenna model is established based on the patch length and patch width;

[0151] According to the binary antenna model, the electromagnetic field simulation is performed using a preset script program and electromagnetic simulation software to obtain a simulation result data set, which includes S21 coupling parameters and directional pattern parameters;

[0152] Build an initial convolutional neural network model;

[0153] According to the mean square error, the simulation result data set is input into the initial convolutional neural network model so that the initial convolutional neural network model is trained to obtain the antenna decoupling structure prediction model.

[0154] The contents of the above method embodiments are all applicable to the present device embodiments. The functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0155] like Fig.12 As shown, an embodiment of the present invention further provides a computer device, including:

[0156] at least one processor 901;

[0157] At least one memory 902, used to store at least one program;

[0158] When at least one program is executed by at least one processor, the at least one processor implements Figure 1 The method shown.

[0159] The contents of the above method embodiments are all applicable to the present device embodiments. The functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0160] The embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, Figure 1 The method shown.

[0161] The contents of the above method embodiments are all applicable to the present storage medium embodiments. The functions specifically implemented by the present storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0162] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the rights of the present invention.

Claims

1. A method for decoupling closely spaced antennas based on deep learning, characterized in that: The following steps are involved: Acquire closely spaced antenna data; Inputting the closely spaced antenna data into an antenna decoupling structure prediction model to obtain decoupling structure parameters; Performing close-spaced antenna decoupling processing according to the decoupling structure parameters; The antenna decoupling structure prediction model is obtained by the following steps: Get rectangular microstrip patch antenna information; Calculating the patch length and patch width according to the rectangular microstrip patch antenna information; Establishing a binary antenna model according to the patch length and the patch width; According to the binary antenna model, electromagnetic field simulation processing is performed using a preset script program and electromagnetic simulation software to obtain a simulation result data set, wherein the simulation result data set includes S21 coupling parameters and directional pattern parameters; Build an initial convolutional neural network model; According to the mean square error, the simulation result data set is input into the initial convolutional neural network model so that the initial convolutional neural network model is trained to obtain the antenna decoupling structure prediction model.

2. The method according to claim 1, characterized in that The step of calculating the patch length and patch width according to the rectangular microstrip patch antenna information includes: Extracting the dielectric plate relative permittivity and dielectric plate thickness from the rectangular microstrip patch antenna information; Calculating the patch width according to the speed of light, the working resonant frequency and the relative dielectric constant of the dielectric plate; Calculating the length difference according to the patch width, the dielectric plate thickness and the dielectric plate relative dielectric constant; The patch length is calculated according to the length difference, the speed of light, the working resonant frequency and the relative dielectric constant of the dielectric plate.

3. The method according to claim 1, characterized in that The step of establishing a binary antenna model according to the patch length and the patch width includes: Extracting the dielectric plate thickness from the rectangular microstrip patch antenna information; Set the antenna spacing; The dual antenna model is established according to the antenna spacing, the patch length, the patch width, the dielectric plate thickness and the size of the radiation unit.

4. The method according to claim 1, characterized in that: According to the binary antenna model, electromagnetic field simulation processing is performed using a preset script program and electromagnetic simulation software to obtain a simulation result data set, including: According to the binary antenna model, setting simulation model parameters; Create electromagnetic simulation projects; Initializing project settings for the electromagnetic simulation project; Establishing a simulation model according to the simulation model parameters and the project settings; The simulation model is used to perform simulation to obtain the simulation result data set.

5. The method according to claim 1, characterized in that The construction process of the initial convolutional neural network model includes: Construct a first convolutional layer, where the convolution kernel size of the first convolutional layer is 32×32×32; After the first convolutional layer, a first pooling layer is constructed, and the size of the first pooling layer is 16×16×32; After the first pooling layer, a second convolutional layer is constructed, and the convolution kernel size of the second convolutional layer is 16×16×64; After the second convolutional layer, a second pooling layer is constructed, and the size of the second pooling layer is 5×5×64; After the second pooling layer, a third convolution layer is constructed, and the convolution kernel size of the third convolution layer is 5×5×128; After the third convolutional layer, a third pooling layer is constructed, and the size of the third pooling layer is 2×2×128; After the third pooling layer, construct a flattening layer; After the flattening layer, a first fully connected layer is constructed, wherein the number of neurons in the first fully connected layer is 128; After the first fully connected layer, a second fully connected layer is constructed, and the number of neurons in the second fully connected layer is 64; After the second fully connected layer, a third fully connected layer is constructed, and the number of neurons in the third fully connected layer is 32; After the third fully connected layer, an output layer is constructed, and the number of neurons in the output layer is 14.

6. The method according to claim 2, characterized in that The step of calculating the patch width according to the light speed, the working resonant frequency and the relative dielectric constant of the dielectric plate includes: According to the speed of light, the working resonant frequency and the relative dielectric constant of the dielectric plate, the patch width is calculated by the patch width calculation formula, and the patch width calculation formula is: Where W is the patch width, c is the speed of light, and f r is the working resonant frequency, ε reff is the relative dielectric constant of the dielectric plate.

7. The method according to claim 2, characterized in that The step of calculating the patch length according to the length difference, the speed of light, the operating resonant frequency and the relative dielectric constant of the dielectric plate includes: The patch length is calculated according to the length difference, the speed of light, the working resonant frequency and the relative dielectric constant of the dielectric plate using a patch length calculation formula, and the patch length calculation formula is: Where L is the length of the patch, c is the speed of light, and f r is the working resonant frequency, ε reff is the relative dielectric constant of the dielectric plate, and ΔL is the length difference.

8. A close-spaced antenna decoupling device based on deep learning, characterized in that: include: The first module is used to obtain closely spaced antenna data; The second module is used to input the close-spaced antenna data into an antenna decoupling structure prediction model to obtain decoupling structure parameters; A third module is used to perform a decoupling process for closely spaced antennas according to the decoupling structure parameters; The antenna decoupling structure prediction model is obtained by the following steps: Get rectangular microstrip patch antenna information; Calculating the patch length and patch width according to the rectangular microstrip patch antenna information; Establishing a binary antenna model according to the patch length and the patch width; According to the binary antenna model, electromagnetic field simulation processing is performed using a preset script program and electromagnetic simulation software to obtain a simulation result data set, wherein the simulation result data set includes S21 coupling parameters and directional pattern parameters; Build an initial convolutional neural network model; According to the mean square error, the simulation result data set is input into the initial convolutional neural network model so that the initial convolutional neural network model is trained to obtain the antenna decoupling structure prediction model.

9. A computer device, characterized in that: include: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images