Millimeter wave filtering antenna and intelligent aided design method

Abstract

The application discloses a millimeter wave filtering antenna and an intelligent auxiliary design method, the antenna adopts a differential aperture coupling feeding mode to realize relatively wide impedance matching bandwidth and relatively high gain, low-frequency and high-frequency radiation zero points are respectively generated by adding a short-circuit patch and slotting a radiation patch, and out-of-band gain suppression is realized. A planar inverted F antenna and added parasitic patches are used to respectively improve the out-of-band suppression level of a low-frequency band and a high-frequency band. Deep learning is used on a machine learning algorithm, a proxy model of selected antenna structure parameters and corresponding performance parameters is established, the performance of an antenna with specific parameters is quickly predicted, a genetic algorithm is used to optimize the proxy model, an antenna structure corresponding to target performance parameters is quickly designed, and the application has the advantages of high efficiency, flexibility and scalability, helps to improve design efficiency and performance, reduces design cost, and the antenna structure has the characteristics of low loss, high gain and high reliability.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of communication antennas, in particular to a millimeter wave filter antenna and an intelligent auxiliary design method. BACKGROUND

[0002] Millimeter wave technology is an important branch of modern wireless communication field, which refers to the technology of using electromagnetic waves with wavelengths ranging from 1 to 10 millimeters (i.e. frequencies ranging from 30 GHz to 300 GHz) for communication. This frequency band has abundant spectrum resources and can provide larger communication bandwidth than traditional low frequency bands, which is crucial to meet the demand for high data transmission rate of 5G and future 6G mobile communication networks. The currently mainly concerned millimeter wave frequency bands include 28-30 GHz, 38 GHz, 45 GHz, 57-71 GHz, 71-76 GHz, 81-86 GHz, 100 GHz, etc. The core of 5G mobile communication module is its radio frequency front end, and the filter and antenna are the key components in the radio frequency front end. In the design of 5G millimeter wave radio frequency front end, the role of filter is particularly important, which is responsible for suppressing the out-of-band noise that may interfere with signal reception and transmission. With the continuous progress of wireless communication technology, the radio frequency front end system is developing towards higher integration, smaller size and higher performance. In the millimeter wave frequency band, the current trend is to integrate the antenna and the filter compactly to realize the integration of functions. Such integrated design not only can optimize the performance of the device, but also can reduce the size of the device, meeting the strict requirements of modern mobile communication devices on space and efficiency.

[0003] The most commonly used design method at present is the fusion method based on bandpass filter. The element to be designed by the fusion method is a filter antenna. The filter antenna is a new type of element that has both filtering and radiating functions. One of the most important characteristics of the filter antenna is that the filter and the antenna can no longer be clearly distinguished (i.e., the filter antenna simultaneously functions as a filter and an antenna). The filter antenna can replace the cascade connection between the filter and the antenna (another filter antenna design method), avoid additional insertion loss from their interface, and achieve a more compact size. Since the resonant structure is designed to resonate outside the passband, they have little effect on the antenna performance within the passband. These radiation zeros can be purposefully designed to control the impedance bandwidth, passband bandwidth, and frequency selectivity of the antenna. The main design challenge is to generate and control the radiation zeros. The advantages of this fusion method are low insertion loss and high radiation efficiency, and it can have a smaller size, but there is currently a lack of perfect and rigorous calculation formula support for modifying the antenna structure, and more experience of researchers and results from electromagnetic simulation software are needed.

[0004] According to the experience of the antenna shape, structure, material and other factors, simulation and debugging are continuously carried out, which is reliable but time-consuming and inefficient, and it is difficult to achieve the global optimal solution and meet the increasing demand. Moreover, the existing millimeter wave filtering antenna structure not only has high loss, but also has low gain, and is not reliable enough. SUMMARY

[0005] The purpose of the present application is to overcome the above-mentioned problems, and to provide a millimeter wave filtering antenna with low loss, high gain, and more reliable advantages.

[0006] Another purpose of the present application is to provide an intelligent auxiliary design method for a millimeter wave filtering antenna, which has the advantages of high efficiency, flexibility and scalability, and helps to improve the design efficiency and performance, and reduce the design cost.

[0007] The purpose of the present application is achieved by the following technical solutions:

[0008] A millimeter wave filtering antenna, comprising a patch, a dielectric substrate, a planar inverted F antenna metal layer, a metal ground plate, a rectangular feed slot, a microstrip feed line and a short circuit through hole;

[0009] The patch comprises a radiation patch, a driven patch, a parasitic patch and a short circuit patch;

[0010] The dielectric substrate comprises three and is stacked together from bottom to top, and the three dielectric substrates are respectively a first dielectric substrate, a second dielectric substrate and a third dielectric substrate;

[0011] The radiation patch, the parasitic patch and the planar inverted F antenna metal layer are printed on the upper surface of the third dielectric substrate; a plurality of slots are provided on the radiation patch; the driven patch and the short circuit patch are printed on the upper surface of the second dielectric substrate; the metal ground plate is printed on the upper surface of the first dielectric substrate; the rectangular feed slot is provided with a pair of and is etched symmetrically on the metal ground plate; the microstrip feed line is printed on the lower surface of the first dielectric substrate;

[0012] The short circuit through hole comprises a first short circuit through hole and a second short circuit through hole, the first short circuit through hole is provided through the second dielectric substrate, and the first short circuit through hole is connected between the short circuit patch and the metal ground plate; the second short circuit through hole is provided through the second dielectric substrate and the third dielectric substrate, and the second short circuit through hole is connected between the planar inverted F antenna metal layer and the metal ground plate.

[0013] In one preferred embodiment of the present application, the dielectric substrate is made of Ferro A6M-E material; the thickness of the dielectric substrate is a multiple of 0.094mm.

[0014] Furthermore, the thickness of the first dielectric substrate is 0.564 mm, the thickness of the second dielectric substrate is 0.47 mm, and the thickness of the third dielectric substrate is 0.094 mm.

[0015] In a preferred embodiment of the present invention, the size of the driving patch is 0.14λc×0.14λc, where λc is the wavelength at the center frequency of the passband.

[0016] In a preferred embodiment of the present invention, the two rectangular feeding slots are completely identical and have a length of 0.18λc, where λc is the wavelength at the center frequency of the passband.

[0017] In a preferred embodiment of the present invention, the size of the metal layer of the planar inverted-F antenna is 0.05λc×0.03λc.

[0018] In a preferred embodiment of the present invention, the diameters of the first short-circuit through hole and the second short-circuit through hole are both 0.1 mm.

[0019] An intelligent auxiliary design method for a millimeter wave filter antenna comprises the following steps:

[0020] The structural parameters that affect the antenna's radiation null and out-of-band suppression level are extracted, including the length of the short-circuit patch, the length of the metal layer of the planar inverted-F antenna, the width of the slot of the radiating patch, and the length of the parasitic patch.

[0021] The above four parameters are sampled exponentially;

[0022] The sampled parameters are combined in a simulation tool to perform full-wave simulation, obtaining S-parameters and gains for each group from 17 GHz to 37 GHz. The data are then combined to form a data set. The data set is then imported into MATLAB to construct a deep learning model.

[0023] After the model is established, it is optimized for the target;

[0024] By using the established proxy model, the genetic algorithm is used to optimize the antenna size parameters to obtain the optimal prediction value and its corresponding antenna parameter combination.

[0025] In a preferred embodiment of the present invention, the sampling ranges of the four parameters are:

[0026] {0.05-0.15}, {1-1.75}, {0.1-0.5}, {1-1.6}, 5 samples for each parameter, a total of 625 parameter combinations.

[0027] In a preferred embodiment of the present invention, building a deep learning model includes the following steps:

[0028] For data preprocessing, the first five columns of the data derived from the three-dimensional electromagnetic field simulation tool HFSS are four structure parameters and one frequency parameter, and the output is the S parameter;

[0029] x = _data(:, 1:5);

[0030] y = _data(:, 6);

[0031] The first five columns of the data list are set as input and put into the above x variable, and the sixth column of the data list is set as output and put into the above y variable;

[0032] X_norm = (X-mean(X)). / std(X);

[0033] y_norm = (y-mean(y)). / std(y);

[0034] The input and output data are normalized and put into new variables X_norm and y_norm respectively, and the data is normalized to make the characteristic values have the same scale;

[0035] cv = cvpartition(size(X_norm, 1), 'HoldOut', 0.20);

[0036] idxTrain = training(cv);

[0037] idxTest = test(cv);

[0038] XTrain = X_norm(idxTrain, :);

[0039] yTrain = y_norm(idxTrain, :);

[0040] XTest = X_norm(idxTest, :);

[0041] yTest = y_norm(idxTest, :);

[0042] The data set is divided into training set and test set using the cvpartition function, 20% of the data is reserved as test set, and the remaining 80% is used for training.

[0043] Further, after data preprocessing and division, the model is constructed by MATLAB:

[0044] layers = [

[0045] featureInputLayer(5,'Normalization','none','Name','input')

[0046] fullyConnectedLayer(64,'Name','fc1')

[0047] reluLayer('Name','relu1')

[0048] fullyConnectedLayer(32,'Name','fc2')

[0049] reluLayer('Name','relu2')

[0050] fullyConnectedLayer(1,'Name','output')

[0051] regressionLayer('Name','regression')];

[0052] First define the construction of the input layer featureInputLayer, the input data is four structure parameters and a frequency parameter, the input features are 5, and the layer is named input;

[0053] Add a fully connected layer fullyConnectedLayer after the input layer, the fully connected layer is used to learn the complex mapping of the input data, performs linear transformation, maps the input data to a new space, each neuron is connected to all neurons of the previous layer, and is calculated through weights and biases, the number of neurons is selected by experience and the prediction effect of the model, and the number of neurons in the layer is set to 64.

[0054] Add an activation layer reluLayer, set the activation function to introduce nonlinearity, the ReLU activation function provides the ability of nonlinear mapping for the neural network, and the result is transmitted to the subsequent layer through nonlinear transformation of the ReLU function;

[0055] Add a fully connected layer fullyConnectedLayer and an activation layer reluLayer to gradually reduce the dimension of the features and extract more abstract representations.

[0056] Define the output layer of the network for generating prediction results.

[0057] Add a regression layer regressionLayer to specify that the output layer of the network is suitable for a regression task, that is, to predict continuous values.

[0058] After the framework is established, the settings of each option are trained.

[0059] Further, the training algorithm is:

[0060] option = trainingOptions ('adam',...

[0061] 'MaxEpochs', 10000,...

[0062] 'MiniBatchSize', 32,...

[0063] 'InitialLearnRate', 1e-3,...

[0064] 'Plots', 'training-progress',...

[0065] 'Verbose', false);

[0066] The maximum number of iterations MaxEpochs during training is set, each Epoch represents that the entire training data set is traversed once, the number of samples in each small batch (Mini-batch) is set, and the initial learning rate is set.

[0067] Start training the model to obtain the proxy model, and after the completion of the proxy model, the performance of different antenna structure parameters at different frequencies is predicted.

[0068] In one preferred embodiment of the application, the genetic algorithm is used to optimize the size parameters of the antenna, including the following steps:

[0069] function fitness = proxyModelFunction (x, y, z, c);

[0070] params = [x, y, z, c];

[0071] S11 = predict (net, params);

[0072] s24 = S11 (:, 1);

[0073] s30 = S11 (:, 2);

[0074] fitness = s24 ^ 2 + s30 ^ 2;

[0075] end;

[0076] A fitness function is defined, and the S11 variable is accepted by the agent model with four structure parameters and returns the S parameter vector group of 24.2GHZ and 30GHZ, and then the vector group is split into two frequency point scalars;

[0077] options = optimoptions('ga',...

[0078] 'MaxGenerations',100,...

[0079] 'PopulationSize',20,...

[0080] 'EliteCount',2,...

[0081] 'CrossoverFraction',0.8,...

[0082] 'PlotFcn',@gaplotbestf,...

[0083] 'Display','iter');

[0084] ga is a genetic algorithm; the MaxGenerations option sets the maximum number of generations executed by the algorithm, and each generation corresponds to a complete population evaluation; the PopulationSize option sets the number of individuals in each generation; the EliteCount option specifies the number of optimal individuals retained to the next generation in each generation; the CrossoverFraction option sets the proportion of crossover operation; the genetic algorithm is run using the ga function.

[0085] [optimalParams,optimalSParam] = ga(fitness,4,...

[0086] [], [],[],[],...

[0087] L_norm_bounds,U_norm_bounds,...

[0088] [], options);

[0089] The fitness function, the number of parameters is 4, the parameter setting in the empty array is various constraints including linear and nonlinear, L_norm_bounds and U_norm_bounds are the upper and lower bounds of the optimization parameters this time, and finally the optimization result is run and waited.

[0090] Compared with the prior art, the present application has the following beneficial effects:

[0091] 1. The millimeter wave filtering antenna of the present application adopts a differential aperture coupling feeding mode to realize a relatively wide impedance matching bandwidth and a relatively high gain, and by adding a short-circuit patch and a slot in a radiation patch, low-frequency and high-frequency radiation zeros are respectively generated to realize out-of-band gain suppression; a planar inverted F antenna and an added parasitic patch are used to respectively improve the out-of-band suppression level of the low-frequency band and the high-frequency band.

[0092] 2. The present application adopts deep learning to establish a proxy model of selected antenna structure parameters and corresponding performance parameters, realizes fast prediction of the performance of a specific parameter structure antenna, uses a genetic algorithm to optimize the proxy model, and quickly designs an antenna structure corresponding to the target performance parameters, has the advantages of high efficiency, flexibility and scalability, and helps to improve design efficiency and performance and reduce design cost. BRIEF DESCRIPTION OF DRAWINGS

[0093] Figure 1 is a structural diagram of the millimeter wave filtering antenna of the present application.

[0094] Figure 2 is the reflection coefficient of the initial antenna design structure of the present application.

[0095] Figure 3 is the gain plot of the initial antenna design structure of the present application.

[0096] Figure 4 is the reflection coefficient of the optimized antenna structure of the present application.

[0097] Figure 5 is the gain plot of the optimized antenna structure of the present application.

[0098] Figure 6 is part of the data set of the present application. DETAILED DESCRIPTION

[0099] In order for those skilled in the art to have a good understanding of the technical solutions of the present application, the present application will be further described below in conjunction with the embodiments and the accompanying drawings, but the implementation of the present application is not limited thereto.

[0100] Reference Figure 1The millimeter wave filtering antenna of the embodiment comprises a patch, a dielectric substrate, a planar inverted F antenna metal layer 1, a metal ground plate 2, a rectangular feed slot 3, a microstrip feed line 4 and a short circuit through hole; the patch comprises a radiation patch 5, a driven patch 6, a parasitic patch 7 and a short circuit patch 8; the dielectric substrate comprises three and is stacked together from bottom to top, and the three dielectric substrates are respectively a first dielectric substrate 9, a second dielectric substrate 10 and a third dielectric substrate 11; the radiation patch 5, the parasitic patch 7 and the planar inverted F antenna metal layer 1 are printed on the upper surface of the third dielectric substrate 11; a plurality of slots are arranged on the radiation patch 5; the driven patch 6 and the short circuit patch 8 are printed on the upper surface of the second dielectric substrate 10; the metal ground plate 2 is printed on the upper surface of the first dielectric substrate 9; the rectangular feed slot 3 is provided with a pair of and is symmetrically etched on the metal ground plate 2; the microstrip feed line 4 is printed on the lower surface of the first dielectric substrate 9; the short circuit through hole comprises a first short circuit through hole 12 and a second short circuit through hole 13, the first short circuit through hole 12 is provided through the second dielectric substrate 10, and the first short circuit through hole 12 is communicated between the short circuit patch 8 and the metal ground plate 2; the second short circuit through hole 13 is provided through the second dielectric substrate 10 and the third dielectric substrate 11, and the second short circuit through hole 13 is communicated between the planar inverted F antenna metal layer 1 and the metal ground plate 2.

[0101] Referring to Figure 1 The working principle of the millimeter wave filtering antenna of the embodiment is as follows:

[0102] At the frequency point where the low-frequency radiation zero point is located, the current with large current density is mainly concentrated on the surface of the short circuit patch 8, and the current density on the radiation patch 5 is small. In addition, the current on the short circuit patch 8 is opposite to the current on the radiation patch 5, and generates radiation opposite to the direction of the radiation patch 5. Therefore, the radiation from the opposite current produces a cancellation effect, forming a radiation zero point in the low-frequency band, and the current of the planar inverted F antenna is the same as the current direction of the radiation patch 5, which can better offset the reverse current generated by the short circuit patch 8 and the radiation patch 5, and bring better low-frequency band suppression effect. For the slots of the radiation patch 5 and the parasitic patch 7, at the frequency point where the low-frequency radiation zero point is located, the current direction of the parasitic patch 7 is opposite to that of the radiation patch 5, and a radiation zero point is generated in the high-frequency band.

[0103] Further, the dielectric substrate adopts Ferro A6M-E material; the thickness of the dielectric substrate is a multiple of 0.094 mm.

[0104] Further, the thickness of the first dielectric substrate 9 is 0.564 mm, the thickness of the second dielectric substrate 10 is 0.47 mm, and the thickness of the third dielectric substrate 11 is 0.094 mm.

[0105] Further, the size of the driving patch 6 is 0.14λc×0.14λc, λc being the wavelength at the center frequency of the passband.

[0106] Further, the two rectangular feed slots 3 are completely identical, with a length of 0.18λc, λc being the wavelength at the center frequency of the passband.

[0107] Further, the size of the planar inverted F antenna metal layer 1 is 0.05λc×0.03λc.

[0108] Further, the diameters of the first short-circuit through hole 12 and the second short-circuit through hole 13 are both 0.1 mm.

[0109] Specifically, the manufacturing process of the millimeter wave filtering antenna of the embodiment is as follows:

[0110] First, the antenna is modeled in the three-dimensional electromagnetic field simulation tool HFSS. The basic model is composed of a material layer at the bottom, which is ferroA6M, and a layer of metal floor GND on top, which is silver (sliver). There are two microstrip feed lines on the bottom of the material layer, which are symmetrical about the X-axis. Specifically, the structure of the microstrip feed line is designed on the dielectric substrate, and one end is coupled to the feed line through a slot. Compared with direct contact feeding, it has a wider bandwidth and is less likely to produce surface waves. The other end is connected to an external signal source or circuit, i.e., the ports are set on the left and right ends. The microstrip feed line is made of metal silver, and the shape chosen is rectangular. When current passes through the microstrip feed line, energy is coupled to the driving patch above the feed line, generating an electromagnetic field between the driving patch and the ground plate, which in turn radiates electromagnetic waves. This radiation is usually achieved through the gap between the patch and the surrounding ground plate, so two corresponding slots need to be dug in the GND to complete the energy radiation. Next, a dielectric layer is added between the ground and the driving chip, and the material chosen is ferroA6M. Next, a driving patch is added to the dielectric layer, and the material used is silver. Then, another layer of dielectric layer is added on top, and the material is still ferroA6M. The last step is to add a radiation chip to radiate energy. The optimization part follows, with a short-circuit patch connected to each side of the lower end of the driving patch. The short-circuit patch is composed of a short-circuit post and a patch on the short-circuit post, both made of metal silver. Connecting the short-circuit post to the GND achieves the short-circuit effect. Then, the radiation patch is slotted. It is important to note that the slotting here is divided into two parts: one part is for the middle of the patch, and the other part is for the edge of the patch. There are two slots in each part, for a total of four slots. Slotting can change the internal current distribution of the antenna and introduce new resonance points, which helps to expand the working bandwidth of the antenna and introduce radiation nulls. The number of slots can be tested, and opening a pair of slots can introduce a radiation null. Then, based on the smoothness of the S parameter, the number of slots is continuously optimized until it approaches the target parameter. The inverted planar F antenna is composed of a short-circuit post and a patch. The inverted planar F antenna metal layer is added to the third layer of dielectric substrate. The last part of the optimization is the addition of a parasitic patch (Parasitic Patch Antenna), which is a patch added around three sides of the radiation patch. Parasitic patch antenna is a design technique that uses electromagnetic coupling effect to enhance the performance of the antenna. It adds one or more radiation patches next to the antenna's radiation element (main radiation patch), which are not directly connected to the feed line, but interact with the main radiation patch through electromagnetic coupling. The addition of three parasitic patches changes the current distribution at high frequency and the radiation pattern in the high frequency stopband, reduces the gain of the antenna in the Z-axis direction, and thus strengthens the stopband suppression level, achieving better high-frequency filtering effect.

[0111] The intelligent auxiliary design method of the millimeter wave filter antenna of this embodiment includes the following steps:

[0112] The structural parameters that affect the antenna's radiation null point and out-of-band suppression level are extracted, namely the length of the short-circuit patch, the length of the planar inverted-F antenna, the width of the slot for the radiation patch, and the length of the parasitic patch. The parameter sweep function in the 3D electromagnetic field simulation tool HFSS is used to perform exponential sampling. This sampling helps capture key characteristic changes over a wide frequency range, because physical phenomena (such as the propagation of electromagnetic waves) often follow exponential laws. The sampling ranges of the above four parameters are:

[0113] {0.05-0.15}{1-1.75}{0.1-0.5}{1-1.6}, 5 samples for each parameter, so there are a total of 625 parameter combinations, and then a full-wave simulation is performed in the simulation tool to obtain the S parameters and gain of each group from 17GHZ to 37GHZ, and the above data are used to form a data set. The data set is imported into MATLAB, and then the code is written to build a deep learning model. The first step is to preprocess the data. In the data structure exported from the three-dimensional electromagnetic field simulation tool HFSS, the first five columns are four structural parameters and one frequency parameter, and the output is S parameters, such as Figure 6 The following is an explanation of the code.

[0114] x = _data(:,1:5);

[0115] y = _data(:,6);

[0116] The first five columns of the data list are set as input and placed in the x variable, and the sixth column of the data list is placed as output in the y variable.

[0117] X_norm=(X-mean(X)). / std(X);

[0118] y_norm=(y-mean(y)). / std(y);

[0119] Normalize the input and output data and put them into new variables X_norm and y_norm respectively. Normalize the data so that the eigenvalues ​​have the same scale, which helps the deep learning model converge.

[0120] cv=cvpartition(size(X_norm,1),'HoldOut',0.20);

[0121] idxTrain = training(cv);

[0122] idxTest = test(cv);

[0123] XTrain = X_norm(idxTrain, :);

[0124] yTrain = y_norm(idxTrain, :);

[0125] XTest = X_norm(idxTest, :);

[0126] yTest = y_norm(idxTest, :);

[0127] The dataset is divided into training and test sets using the cvpartition function. 20% of the data is reserved as the test set, and the remaining 80% is used for training. cvpartition creates a split object, and the training, validation, and test methods return the indices of the training, validation, and test sets, respectively.

[0128] After data preprocessing and division, the model is built using MATLAB's deep learning toolbox.

[0129] layers = [

[0130] featureInputLayer(5, 'Normalization', 'none', 'Name', 'input')

[0131] fullyConnectedLayer(64, 'Name', 'fc1')

[0132] reluLayer('Name','relu1')

[0133] fullyConnectedLayer(32, 'Name', 'fc2')

[0134] reluLayer('Name','relu2')

[0135] fullyConnectedLayer(1, 'Name', 'output')

[0136] regressionLayer('Name','regression')] ;

[0137] First, define the input layer featureInputLayer, since the input data is four structural parameters and one frequency parameter, the input features are 5, and since the data has been normalized in the previous step, the Normalization here means not to normalize, and finally name this layer as input. Add a fully connected layer fullyConnectedLayer after the input layer, which is used to learn the complex mapping of the input data. It is a linear transformation that maps the input data to a new space. Each neuron is connected to all neurons in the previous layer and calculates the output through weights and biases. The number of neurons is selected based on experience and the prediction effect of the model. In this layer, the number of neurons is set to 64. Then add an activation layer reluLayer and set the activation function to introduce nonlinearity, which is the ability of the model to learn complex patterns. The ReLU activation function provides the ability of non-linear mapping for neural networks, enabling neural networks to learn and simulate complex function relationships. This layer will be applied to the output of the previous layer, which will be nonlinearly transformed by the ReLU function, and then the result will be passed to the subsequent layer. Add a fully connected layer fullyConnectedLayer and an activation layer reluLayer again to gradually reduce the dimensionality of the features and extract more abstract representations. Then define the output layer of the network, which is used to generate the prediction result. Finally, add a regression layer regressionLayer to specify that the output layer of the network is suitable for regression tasks, i.e., predicting continuous values. After the framework is established, the training options are set, such as the optimizer, the number of iterations, the learning rate, etc.

[0138] option = trainingOptions('adam',...

[0139] 'MaxEpochs', 10000,...

[0140] 'MiniBatchSize', 32,...

[0141] 'InitialLearnRate', le-3,...

[0142] 'Plots', 'training-progress',...

[0143] 'Verbose', false);

[0144] 'adam' specifies the optimizer, an adaptive learning rate optimization algorithm that combines the advantages of AdaGrad and RMSProp, generally providing good performance and fast convergence. Sets the maximum number of iterations (MaxEpochs) during training. Each epoch represents a single pass through the entire training dataset. Setting this to 100 means the model will be trained a maximum of 100 times. Sets the number of samples in each mini-batch. Setting this to 32 means 32 samples will be used each time the model parameters are updated. The mini-batch size affects the training speed and memory usage of the model, as well as its generalization ability. Sets the initial learning rate. The learning rate determines the step size for parameter updates during optimization. Setting this to 0.001 is a common initial learning rate that helps the model converge quickly in the early stages of training. Next, begin training the model to obtain a surrogate model. Once the surrogate model is completed, performance predictions can be made for different antenna structure parameters at different frequencies.

[0145] After establishing the model, the next step was to optimize its target. Since the goal was to find structural parameters with a bandwidth below -10dB between 24.2GHz and 30GHz, and to improve accuracy and training speed, the proxy model needed to be adjusted. The model's input was changed to four structural parameters, and the outputs were S parameters at the 24.2GHz and 30GHz frequencies. The data partitioning and some neural network layer settings were adjusted, with the input features of the input layer, featureInputLayer, changed to 4 and the output features to 2. Due to the reduction in frequency input features, the data was significantly reduced. To make the model more accurate, the maximum number of iterations in the training options needed to be adjusted. After testing several different numbers of iterations, a setting of 5000 was found to be appropriate. Excessive iterations can lead to overfitting, learning rate decay, and decreased generalization ability. After adjusting the code, the model was retrained.

[0146] Using the established agent model, a genetic algorithm is employed to optimize the antenna's dimensional parameters, obtaining the optimal predicted value and its corresponding antenna parameter combination. In a genetic algorithm, the first thing to determine is the fitness function. This plays a central role in the algorithm, defining the objective of the problem. In optimization problems, the fitness function is typically designed to minimize or maximize a specific objective function. The higher the fitness value of an individual, the closer it is to the optimal solution.

[0147] functionfitness=proxyModelFunction(x,y,z,c)

[0148] params = [x, y, z, c];

[0149] S11 = predict(net, params);

[0150] s24 = S11(:, 1);

[0151] s30 = S11(:, 2);

[0152] fitness = s24^2 + s30^2;

[0153] end;

[0154] First define a fitness function, S11 variable is accepted by the agent model four structural parameters and return 24.2GHZ and 30GHZ S parameter vector group, and then the vector group is split into two frequency point scalar, genetic algorithm is the minimization problem, so we want the smaller the better fitness value

[0155] Since our goal is to make the two frequency bands of S parameters are zero, we can calculate the sum of the squares of S parameters as fitness function, here using the sum of squares because this can be considered at the same time two frequency bands of S parameters, and the square can amplify the error, so that the optimization process is more sensitive. Fitness function is written, it is to the genetic algorithm options configuration.

[0156] options = optimoptions('ga',...

[0157] 'MaxGenerations', 100,...

[0158] 'PopulationSize', 20,...

[0159] 'EliteCount', 2,...

[0160] 'CrossoverFraction', 0.8,...

[0161] 'PlotFcn', @gaplotbestf,...

[0162] 'Display', 'iter');

[0163] The first parameter 'ga' specifies the type of optimization algorithm, which is the genetic algorithm. The MaxGenerations option sets the maximum number of generations the algorithm will execute. Each generation corresponds to a complete population evaluation. The PopulationSize option sets the number of individuals (candidate solutions) in each generation. A larger population can provide more diversity but will increase the computational cost. The EliteCount option specifies the number of best individuals from the current population that are carried over to the next generation. These individuals are the highest fitness in the current population. The CrossoverFraction option sets the proportion of the new generation that is generated by the crossover operation, i.e., how many proportion of individuals in the new generation are generated by the crossover operation. Then the genetic algorithm is run using the ga function.

[0164] [optimalParams,optimalSParam] = ga(fitness, 4,...

[0165] [], [], [], [],...

[0166] L_norm_bounds, U_norm_bounds,...

[0167] [], options);

[0168] The fitness function is passed in, the number of parameters is 4, the parameter settings in the empty array are various constraints including linear and nonlinear, there are no constraints for this optimization target, so the empty array is set, L_norm_bounds and U_norm_bounds are the upper and lower bounds of the optimization parameters this time. Finally, the optimization result is run and waited.

[0169] Further, the reflection coefficient of the initial antenna design structure is as shown in Figure 2 The impedance matching bandwidth of S11 response less than -10 dB is 23.060 GHz-29.710 GHz, and the frequency range basically covers the N257, N258, and N261 frequency bands in the 5G millimeter wave communication frequency band, realizing good impedance matching. The gain response is as shown in Figure 3 The 5 dB bandwidth covers 23.607 GHz-29.086 GHz, and the highest gain in the passband is 5.5 dBi, and there is a radiation zero point at 19.3 GHz and 34.3 GHz, respectively, realizing a good out-of-band gain suppression level higher than 12.5 dB in the frequency range of 17-37 GHz. The overall size of the antenna is 0.75λc×0.54λc×0.1λc, realizing a compact size.

[0170] The established agent model using deep learning can directly input the structure parameters and predict the corresponding performance, and then the structure parameters are used for full-wave simulation according to the simulation. Then genetic algorithm is used for optimization, and the structure parameters with a bandwidth lower than-10dB from 24.2GHz to 30GHz are found. The optimized antenna reflection coefficient is shown in Figure 4 The impedance matching bandwidth with S11 response less than-10dB is 24.00-31.34, and the frequency range is wider than the initial result bandwidth and close to the target bandwidth. The optimized gain diagram is shown in Figure 5 The 5dB bandwidth covers 24.519GHz-31.150GHz, the frequency range is wider than the initial result bandwidth and close to the target bandwidth, and the out-of-band suppression at high frequency is more obvious.

[0171] The above is the preferred embodiment of the present application, but the embodiments of the present application are not limited by the above, any change, modification, substitution, combination, simplification made without departing from the spirit and principles of the present application should be an equivalent replacement method, and all are included in the protection scope of the present application.

Claims

1. A millimeter wave filtering antenna, characterized in that: It includes a patch, a dielectric substrate, a planar inverted-F antenna metal layer, a metal floor, a rectangular feed slot, a microstrip feed line, and a short-circuit through hole; The patches include radiation patches, driving patches, parasitic patches and short-circuit patches; The dielectric substrates include three and are stacked together in sequence from bottom to top, and the three dielectric substrates are respectively a first dielectric substrate, a second dielectric substrate and a third dielectric substrate; The radiation patch, parasitic patch, and planar inverted-F antenna metal layer are printed on the upper surface of the third dielectric substrate; The radiation patch is provided with a plurality of slots; the driving patch and the short-circuit patch are printed on the upper surface of the second dielectric substrate; The metal floor is printed on the upper surface of the first dielectric substrate; a pair of rectangular feed slots are provided and symmetrically etched on the metal floor; the microstrip feed line is printed on the lower surface of the first dielectric substrate; The short-circuit through hole includes a first short-circuit through hole and a second short-circuit through hole. The first short-circuit through hole is set through the second dielectric substrate, and the first short-circuit through hole is connected between the short-circuit patch and the metal floor; the second short-circuit through hole is set through the second dielectric substrate and the third dielectric substrate, and the second short-circuit through hole is connected between the metal layer of the planar inverted-F antenna and the metal floor.

2. The millimeter wave filtering antenna according to claim 1, characterized in that: The dielectric substrate is made of Ferro A6M-E material; the thickness of the dielectric substrate is a multiple of 0.094 mm; the thickness of the first dielectric substrate is 0.564 mm, the thickness of the second dielectric substrate is 0.47 mm, and the thickness of the third dielectric substrate is 0.094 mm.

3. The millimeter wave filtering antenna according to claim 1, characterized in that: The size of the driving patch is 0.14λc×0.14λc, and the size of the metal layer of the planar inverted-F antenna is 0.05λc×0.03λc, where λc is the wavelength at the center frequency of the passband.

4. The millimeter wave filtering antenna according to claim 1, wherein: The two rectangular feed slots are identical, with a length of 0.18λc, where λc is the wavelength at the center frequency of the passband.

5. An intelligent auxiliary design method for the millimeter wave filter antenna according to any one of claims 1 to 4, characterized in that: The following steps are involved: The structural parameters that affect the antenna's radiation null and out-of-band suppression level are extracted, including the length of the short-circuit patch, the length of the metal layer of the planar inverted-F antenna, the width of the slot of the radiating patch, and the length of the parasitic patch. The above four parameters are sampled exponentially; The sampled parameters are combined in a simulation tool to perform full-wave simulation, obtaining S-parameters and gains for each group from 17 GHz to 37 GHz. The data are then combined to form a data set. The data set is then imported into MATLAB to construct a deep learning model. After the model is established, it is optimized for the target; By using the established proxy model, the genetic algorithm is used to optimize the antenna size parameters to obtain the optimal prediction value and its corresponding antenna parameter combination.

6. The intelligent assisted design method according to claim 5, characterized in that: The sampling ranges of the four parameters are: {0.05-0.15}, {1-1.75}, {0.1-0.5}, {1-1.6}, 5 samples for each parameter, a total of 625 parameter combinations.

7. The intelligent assisted design method according to claim 5, characterized in that: Building a deep learning model involves the following steps: For data preprocessing, the first five columns of the data exported from the three-dimensional electromagnetic field simulation tool HFSS are four structural parameters and one frequency parameter, which are output as S parameters; x = _data(:,1:5); y = _data(:,6); Take the first five columns of the data list as input and put them into the above x variable, and take the sixth column of the data list as output and put it into the above y variable; X_norm = (X-mean(X)). / std(X); y_norm = (y-mean(y)). / std(y); Normalize the input and output data and put them into new variables X_norm and y_norm respectively. Normalize the data so that the eigenvalues ​​have the same scale. cv = cvpartition(size(X_norm, 1), 'HoldOut', 0.20); idxTrain = training(cv); idxTest = test(cv); XTrain = X_norm(idxTrain, :); yTrain = y_norm(idxTrain, :); XTest = X_norm(idxTest,:); yTest = y_norm(idxTest,:); The cvpartition function is used to divide the dataset into training and test sets. 20% of the data is reserved as the test set, and the remaining 80% is used for training.

8. The intelligent assisted design method according to claim 7, characterized in that: After data preprocessing and division are completed, the model is built using MATLAB: layers = [ featureInputLayer(5, 'Normalization', 'none', 'Name', 'input') fullyConnectedLayer(64, 'Name', 'fc1') reluLayer('Name', 'relu1') fullyConnectedLayer(32, 'Name', 'fc2') reluLayer('Name', 'relu2') fullyConnectedLayer(1, 'Name', 'output') regressionLayer('Name', 'regression')]; First, define and construct the input layer featureInputLayer. The input data is four structure parameters and one frequency parameter. The input feature is 5. Name this layer input. Add a fully connected layer after the input layer. The fully connected layer is used to learn the complex mapping of input data, perform linear transformations, and map the input data to a new space. Each neuron is connected to all neurons in the previous layer and calculations are performed through weights and biases. The number of neurons should be set based on experience and the prediction effect of the model. The number of neurons in this layer is set to 64. Add an activation layer reluLayer and set the activation function to introduce nonlinearity. The ReLU activation function provides the neural network with the ability of nonlinear mapping. The ReLU function is used to perform nonlinear transformation on it and pass the result to the subsequent layers. Add a fully connected layer fullyConnectedLayer and an activation layer reluLayer to gradually reduce the dimension of the features and extract more abstract representations; Define the output layer of the network to generate prediction results; Add a regression layer regressionLayer to specify that the output layer of the network is suitable for regression tasks, that is, predicting continuous values; After the framework is established, train the settings of various options.

9. The intelligent assisted design method according to claim 8, characterized in that: The training algorithm is: option= trainingOptions('adam',... 'MaxEpochs',10000,... 'MiniBatchSize',32,... 'InitialLearnRate',1e-3,... 'Plots','training-progress',... 'Verbose',false ); Set the maximum number of iterations (MaxEpochs) during training, where each Epoch represents one pass through the entire training dataset; set the number of samples in each mini-batch; and set the initial learning rate. The model is trained to obtain a proxy model. After the proxy model is completed, the performance of different antenna structure parameters at different frequencies is predicted.

10. The intelligent assisted design method according to claim 9, characterized in that: The optimization of antenna size parameters using genetic algorithm includes the following steps: function fitness = proxyModelFunction(x,y,z,c); params = [x,y,z,c]; S11 = predict(net,params); s24=S11(:, 1); s30=S11(:, 2); fitness = s24^2 + s30^2; end; Define a fit function. The S11 variable is a proxy model that accepts four structural parameters and returns a 24.2GHz and 30GHz S-parameter vector group. The vector group is then split into two frequency point scalars. options = optimoptions('ga', ... 'MaxGenerations', 100, ... 'PopulationSize', 20, ... 'EliteCount', 2, ... 'CrossoverFraction', 0.8, ... 'PlotFcn', @gaplotbestf, ... 'Display', 'iter'); ga is a genetic algorithm; the MaxGenerations option sets the maximum number of generations executed by the algorithm, with each generation corresponding to a complete population evaluation; the PopulationSize option sets the number of individuals in each generation; the EliteCount option specifies the optimal number of individuals in each generation to be retained in the next generation; the CrossoverFraction option sets the ratio of crossover operations; use the ga function to run the genetic algorithm; [optimalParams, optimalSParam] = ga(fitness, 4, ... [], [], [], [], ... L_norm_bounds, U_norm_bounds, ... [], options); Pass in the fitness function and the number of parameters as 4. The parameters in the empty array are set as various linear and nonlinear constraints. L_norm_bounds and U_norm_bounds are the upper and lower bounds of the optimization parameters respectively. Finally, run and wait for the optimization results.

Images