Antenna size optimization method, antenna and household electrical appliance product

Abstract

The invention discloses an antenna size optimization method, an antenna and a household electrical appliance product, and the method comprises the steps: carrying out the multi-time iteration of a size parameter combination of an ultrahigh frequency antenna, and evaluating the performance of the antenna corresponding to the size parameter combination obtained through each iteration through a pre-trained performance estimation model, so as to screen out the optimal antenna size of the ultrahigh frequency antenna. Electromagnetic simulation software does not need to be used in the antenna size automatic optimization process, the computing power cost and the time cost in the antenna design process are effectively reduced, and the risk of insufficient memory is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of antenna design, and in particular to an antenna size optimization method, an antenna and a household appliance product. Background Art

[0002] In recent years, in order to achieve automatic optimization of antenna size, antenna design based on optimization algorithms has become a research hotspot. However, practice has found that the optimization algorithm iteratively calls electromagnetic simulation software such as HFSS (High Frequency Structure Simulator) to evaluate the fitness of all individuals, which consumes a lot of computing power and time costs, and may even lead to insufficient memory. Summary of the invention

[0003] The purpose of the embodiments of the present invention is to provide an antenna size optimization method, an antenna and a household appliance product, which iterates multiple ultra-high frequency antenna size parameter combinations and evaluates the performance of the antenna corresponding to the size parameter combination obtained in each iteration through a pre-trained performance estimation model to screen out the optimal antenna size of the ultra-high frequency antenna. In the process of automatically optimizing the antenna size, there is no need to use electromagnetic simulation software, which effectively reduces the computing power cost and time cost in the antenna design process and reduces the risk of insufficient memory.

[0004] To achieve the above object, an embodiment of the present invention provides an antenna size optimization method, comprising:

[0005] Obtain the value range of each size parameter of the UHF antenna;

[0006] Within the value range of each size parameter, iterate the size parameter combination of the UHF antenna for multiple rounds, and use a pre-trained performance estimation model to calculate the performance index value of the UHF antenna at a preset frequency point corresponding to the size parameter combination value obtained after each round of iteration;

[0007] The size parameter combination value that optimizes the performance index value is used as the optimal antenna size of the UHF antenna.

[0008] As an improvement of the above scheme, within the value range of each size parameter, the size parameter combination of the UHF antenna is iterated for multiple rounds, and a pre-trained performance estimation model is used to calculate the performance index value of the UHF antenna at a preset frequency corresponding to the size parameter combination value obtained after each round of iteration, including:

[0009] Within the value range of each size parameter, a population including a preset number of individuals is initialized using a genetic algorithm encoding method, each individual represents a size parameter combination, and each size parameter combination includes a number of the size parameters;

[0010] In each round of iteration, based on a preset generation gap, a number of mothers are selected from the population for crossover mutation, and the individuals obtained by the crossover mutation are merged with the parent population;

[0011] The pre-trained performance estimation model is used to calculate the performance index value of each individual in the merged population at the preset frequency point, and several individuals with the best performance index value are selected as the population for a new round of iteration until the preset convergence condition is met and the iteration ends.

[0012] As an improvement of the above solution, the performance estimation model is a CNN-DualLSTM model; the use of the pre-trained performance estimation model to calculate the performance index value of each individual in the merged population at a preset frequency point includes:

[0013] Each individual in the merged population and the preset frequency point are input into the CNN-DualLSTM model, and the spatial feature information is extracted by the CNN unit and transmitted to the DualLSTM unit;

[0014] The DualLSTM unit extracts the electromagnetic characteristics of the spatial feature information and transmits it to the fully connected layer;

[0015] The fully connected layer integrates information of the electromagnetic characteristics and transmits the integrated information to the output layer;

[0016] The output layer processes the integrated information and outputs the performance index value of each individual in the combined population at the preset frequency point.

[0017] As an improvement of the above solution, the performance estimation model is trained by a pre-acquired training data set; the training data set is acquired in the following manner:

[0018] In the electromagnetic simulation software, based on the value range of each dimensional parameter, the each dimensional parameter is randomly adjusted to obtain several groups of reference dimensional combinations, and the performance index value of each reference dimensional combination at different reference frequencies is calculated, and the initial dimensional combination, the reference frequency and the corresponding performance index value are combined to obtain several groups of training data to form a training data set; wherein the reference frequency is greater than or equal to 860 MHz and less than or equal to 960 MHz.

[0019] As an improvement of the above scheme, in the CNN unit, the size of the convolution kernel is 1, the number of the convolution kernels is 32, the step is 1, and the filling method is SAME; the DualLSTM unit is an LSTM with a double-layer unit of 64; the activation function of the output layer is a sigmoid function.

[0020] As an improvement of the above solution, the antenna includes the UHF antenna and the HF antenna;

[0021] The value ranges of the various size parameters of the UHF antenna are obtained, including:

[0022] Obtaining size parameters of the high frequency antenna;

[0023] The value ranges of the various size parameters of the UHF antenna are determined according to the size parameters of the high frequency antenna, so that the UHF antenna does not overlap with the high frequency antenna or overlap itself.

[0024] As an improvement of the above solution, the preset frequency point is greater than or equal to 860 MHz and less than or equal to 960 MHz.

[0025] As an improvement to the above solution, the performance indicator value is a return loss characteristic value.

[0026] To achieve the above object, an embodiment of the present invention further provides an antenna, which includes an ultra-high frequency antenna and a high frequency antenna, and the size of the ultra-high frequency antenna is designed by the antenna optimization method described in any of the above embodiments.

[0027] To achieve the above objective, an embodiment of the present invention further provides a household appliance, wherein the household appliance includes the antenna described in the above embodiment.

[0028] Compared with the prior art, the antenna size optimization method, antenna and household appliance disclosed in the present invention obtain the value range of each size parameter of the UHF antenna; within the value range of each size parameter, perform multiple rounds of iterations on the size parameter combination, and use a pre-trained performance estimation model to calculate the performance index value of the UHF antenna at a preset frequency point corresponding to the size parameter combination value obtained after each round of iteration; and use the size parameter combination value that makes the performance index value optimal as the optimal antenna size of the UHF antenna. The antenna size optimization method, antenna and household appliance disclosed in the present invention perform multiple iterations on the size parameter combination of the UHF antenna and evaluate the performance of the antenna corresponding to the size parameter combination obtained in each iteration through a pre-trained performance estimation model, so as to screen out the optimal antenna size of the UHF antenna. In the process of automatically optimizing the antenna size, there is no need to use electromagnetic simulation software, which effectively reduces the computing power cost and time cost in the antenna design process and reduces the risk of insufficient memory. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 is a first flow chart of an antenna size optimization method provided by an embodiment of the present invention;

[0030] Figure 2 is a flow chart of an improved genetic algorithm provided by one embodiment of the present invention;

[0031] Figure 3 is a schematic diagram of an optimal antenna size of an ultra-high frequency antenna provided by an embodiment of the present invention;

[0032] Figure 4 It is a schematic diagram of a flow chart of calculating a performance index value provided by an embodiment of the present invention;

[0033] Figure 5 It is a schematic diagram of a process of obtaining a training data set provided by an embodiment of the present invention;

[0034] Figure 6 is a framework diagram of a performance estimation model provided by an embodiment of the present invention;

[0035] Figure 7 It is a structural schematic diagram of a high-frequency-ultra-high-frequency radio frequency identification tag antenna provided by one embodiment of the present invention;

[0036] Figure 8 It is a schematic diagram of the value ranges of various size parameters of a UHF antenna provided by an embodiment of the present invention;

[0037] Fig. 9 is a comparison chart of prediction errors and time consumption of various models provided by an embodiment of the present invention;

[0038] Fig.10 It is a schematic diagram of a process for determining a size range of an ultra-high frequency antenna provided by an embodiment of the present invention;

[0039] Fig.11 It is a structural schematic diagram of a high-frequency antenna provided by one embodiment of the present invention;

[0040] Fig.12 is a schematic diagram of the dimensions of a high-frequency antenna provided by an embodiment of the present invention;

[0041] Fig.13 It is a three-dimensional diagram of a refrigerator provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] See also Figure 1 , is a first flow chart of the antenna size optimization method provided by an embodiment of the present invention, comprising steps S11 to S13:

[0044] S11, obtaining the value range of each size parameter of the UHF antenna;

[0045] S12, within the value range of each size parameter, performing multiple rounds of iterations on the size parameter combination, and using a pre-trained performance estimation model to calculate the performance index value of the UHF antenna at a preset frequency point corresponding to the size parameter combination value obtained after each round of iteration;

[0046] S13, taking the size parameter combination value that optimizes the performance index value as the optimal antenna size of the UHF antenna.

[0047] It is understandable that the size of the antenna is related to the wavelength it receives, and the wavelength is related to the frequency. Therefore, in order to ensure that the antenna can work effectively within a specific frequency range and has good performance and the ability to adapt to specific application requirements, the antenna size needs to be optimized when designing the antenna.

[0048] Specifically, in an implementation manner of the present invention, the value range of each size parameter of the UHF antenna is first determined, and the size is optimized within the range, which can shorten the optimization time. Within a limited value range, multiple rounds of iterations are performed on the size parameter combination of the UHF antenna and the performance index value of the UHF antenna at a preset frequency corresponding to the size parameter combination value after each round of iteration is calculated. The size parameter combination value that makes the performance index value optimal is selected as the optimal antenna size of the UHF antenna.

[0049] Furthermore, after obtaining the optimal antenna size, it also includes using HFSS simulation software to verify whether the performance indicator value of the optimal antenna size meets the requirements, that is, using electromagnetic simulation software to verify the optimal antenna size. If the requirements are met, the optimal antenna size is output; if the requirements are not met, the antenna size and the corresponding real performance indicator value (calculated by the electromagnetic simulation software) are added to the training set of the performance estimation model to correct the performance estimation model and recalculate the optimal antenna size.

[0050] Compared with the prior art, the embodiments of the present invention, when optimizing the antenna size, iterates the ultra-high frequency antenna size parameter combination multiple times and evaluates the performance of the antenna corresponding to the size parameter combination obtained in each iteration through a pre-trained performance estimation model, so as to screen out the optimal antenna size of the ultra-high frequency antenna. In the process of automatically optimizing the antenna size, there is no need to use electromagnetic simulation software, which effectively reduces the computing power cost and time cost in the antenna design process and reduces the risk of insufficient memory.

[0051] In a preferred embodiment, within the value range of each size parameter, multiple rounds of iterations are performed on the size parameter combination, and a pre-trained performance estimation model is used to calculate the performance index value of the ultra-high frequency antenna at a preset frequency point corresponding to the size parameter combination value obtained after each round of iteration, including:

[0052] Within the value range of each size parameter, a population including a preset number of individuals is initialized using a genetic algorithm encoding method, each individual represents a size parameter combination, and each size parameter combination includes a number of the size parameters;

[0053] In each round of iteration, based on a preset generation gap, a number of mothers are selected from the population for crossover mutation, and the individuals obtained by the crossover mutation are merged with the parent population;

[0054] The pre-trained performance estimation model is used to calculate the performance index value of each individual in the merged population at the preset frequency point, and several individuals with the best performance index value are selected as the population for a new round of iteration until the preset convergence condition is met and the iteration ends.

[0055] Specifically, CNN-DualLSTM is combined with genetic algorithm, and CNN-DualLSTM is used to assist genetic algorithm to optimize the size of UHF antenna. In this example, 10 sizes of UHF antenna in HF-UHF RFID tag are used as optimization variables, and S11 of UHF antenna at 915MHz is used as the fitness function of individuals. For example, according to the real integer coding rule, a population of 40 individuals is initialized. Each individual represents a size configuration of a set of UHF antennas. The parameters to be optimized are encoded in binary with a coding accuracy of 4. A total of 2 populations are generated, each containing 5 individuals with a generation gap of 0.9. Then, it is iterated to obtain the optimal antenna size.

[0056] Exemplarily, the present invention uses a genetic algorithm when calculating the optimal antenna size, see Figure 2 , Figure 2 : is a flow chart of an improved genetic algorithm provided by an embodiment of the present invention, including steps S14 to S21:

[0057] S14. Initialize the population using real integer coding rules.

[0058] S15. If the convergence condition is met, stop, otherwise continue to execute. The convergence condition of the algorithm in this example is the maximum number of evolutionary generations and the judgment threshold of target optimization stagnation, which are set to 25 and 1*10 respectively. -6 .

[0059] S16. Use CNN-DualLSTM to evaluate the fitness of the current population.

[0060] S17. Independently select several mothers from the current population.

[0061] S18. Perform crossover operation on the mother independently.

[0062] S19. Independently mutate the individuals after crossover.

[0063] S20. Merge the parent population and the population obtained by crossover mutation to obtain an expanded population.

[0064] S21. Sort the merged population according to fitness, select some individuals from high to low, obtain a new generation of population and return to S15.

[0065] like Figure 3 A schematic diagram of the optimal antenna size for a UHF antenna is shown. Figure 3 The numerical unit of each dimension parameter is mm. Under this dimension, the S11 of the UHF antenna drops from -10dB to -16.99dB. The dimension is output and the S11 of the UHF antenna under this dimension is -16.91dB obtained through HFSS simulation, which is very close to the -16.99dB predicted by CNN-DualLSTM. Because the S11 of the UHF antenna of this dimension has reached the engineering requirement of less than or equal to -10dB. Therefore, this dimension is the final dimension of the UHF antenna. At this time, the S11 of the HF antenna at 13.56MHz is -57.77dB. So far, the S11 of the HF-UHF RFID tag antenna at 13.56MHz and 915MHz has reached the engineering requirement of less than or equal to -10dB, and from their respective S11s, it can be seen that the resonant frequency of the dual-frequency antenna is also close to the expected target, and the design of the HF-UHF RFID tag antenna is completed.

[0066] The above-mentioned optimization of UHF antenna size based on CNN-DualLSTM assisted genetic algorithm only takes 29 seconds, plus 16 hours of simulation data collection and model training time, the total time is 16.008 hours. The optimization of UHF antenna size using HFSS assisted genetic algorithm requires 28.2 hours. The method described in the embodiment of the present invention takes only 56.8% of the time based on HFSS assisted genetic algorithm.

[0067] It is worth noting that the fitness is the performance index value of each individual at a preset frequency point. For example, the reflection coefficient S11 of the ultra-high frequency antenna at 915 MHz can be selected as the fitness function of the individual.

[0068] In a preferred embodiment, the performance estimation model is a CNN-DualLSTM model; see Figure 4The schematic diagram of the process of calculating the performance index value shown in the figure, wherein the performance index value of each individual in the merged population at the preset frequency point is calculated by using the pre-trained performance estimation model, comprising the following steps:

[0069] S22, input each individual in the merged population and the preset frequency point into the CNN-DualLSTM model, extract the spatial feature information through the CNN unit and transmit it to the DualLSTM unit;

[0070] S23, extracting the electromagnetic characteristics of the spatial feature information by the DualLSTM unit and transmitting it to the fully connected layer;

[0071] S24, the fully connected layer integrates the electromagnetic characteristics and transmits the integrated information to the output layer;

[0072] S25, the output layer processes the integrated information and outputs the performance index value of each individual in the merged population at the preset frequency point.

[0073] In a preferred embodiment, the performance estimation model is trained by a pre-acquired training data set; see Figure 5 The flowchart of obtaining a training data set is shown in FIG. 1 , wherein the training data set is obtained by the following steps:

[0074] S26. In electromagnetic simulation software, based on the value range of each dimension parameter, randomly adjust each dimension parameter to obtain a plurality of reference dimension combinations;

[0075] S27. Calculate the performance index value of each reference size combination at different reference frequencies, combine the initial size combination, the reference frequency and the corresponding performance index value, obtain several groups of training data, and form a training data set; wherein the reference frequency is greater than or equal to 860 MHz and less than or equal to 960 MHz.

[0076] It is worth noting that the performance estimation model is trained by a pre-acquired training data set, and the training data set is obtained by electromagnetic simulation software. Specifically, a plurality of reference size combinations and reference frequency points are input into the electromagnetic simulation software, and the corresponding performance indicators are output by the electromagnetic simulation software, thereby combining them into a training data set. Exemplarily, the performance indicator may be the reflection coefficient S11 of the antenna.

[0077] In a preferred embodiment, in the CNN unit, the size of the convolution kernel is 1, the number of convolution kernels is 32, the step is 1, and the padding mode is SAME; the DualLSTM unit is a LSTM with a double-layer unit of 64; the activation function of the output layer is a sigmoid function. For specific performance estimation models, see Figure 6 shown.

[0078] For example, see Figure 7 The HF-UHF RFID tag antenna shown is based on Figure 8 The figure shows the value range of each size parameter of the UHF antenna (the unit of each size parameter is mm). Within the value range of the size parameter, MATLAB is used to control HFSS to randomly adjust the size of the UHF antenna to calculate the antenna S11 to generate a total of 480 sets of antenna simulation data. The electromagnetic simulation software used in this embodiment is ANSYS HFSS 19.2, and the computer hardware configuration is Intel (R) Core (TM) i7-9750HF CPU @ 2.60GHz, 32G memory.

[0079] UHF Antenna S 11 The calculation formula is as follows:

[0080]

[0081] In the formula, Z a represents the impedance of the UHF antenna, Z c Represents the impedance of the chip, Z c * is the conjugate value of the chip impedance.

[0082] Construct a CNN-DualLSTM model, including an input layer, a feature extraction layer, and an output layer. The specific structure is as follows: the size of the convolution kernel in the CNN is 1, the number of convolution kernels is 32, the stride is 1, and the padding method is SAME. DualLSTM is an LSTM with 64 double-layer units. In addition, the model also includes two fully connected layers Dense with output sizes of 16 and 1, and the final output activation function is sigmoid.

[0083] Use the simulation data set to train CNN-DualLSTM until it achieves good prediction accuracy. The specific method is as follows:

[0084] From the simulation data set, within the working frequency band (860MHz-960MHz) of the UHF RFID tag antenna, the S of the antenna at 21 frequency points are sampled at a sampling interval of 4MHz. 11 , a total of 10080 sample data were collected.

[0085] The sample data was normalized and divided into training set and test set in a ratio of 8:2. During the model training process, the Adam optimizer was used to optimize the model parameters. The initial learning rate was set to 0.001 and the number of iterative training was 100 times.

[0086] The 10 dimension parameters and a sampling frequency of the UHF antenna are used as the input of CNN-DualLSTM. The input data is first extracted through CNN to extract the spatial features of the data, and then the feature information enters the DualLSTM to extract the electromagnetic characteristics of different structures. Subsequently, the feature information enters the fully connected layer with output sizes of 16 and 1, and finally the S of the UHF antenna is output through the sigmoid function. 11 .

[0087] In order to demonstrate the accuracy of the CNN-DualLSTM model provided by the embodiment of the present invention, CNN, LSTM, CNN-LSTM and BiLSTM models with the same training parameters as CNN-DualLSTM were established, and the accuracy of the CNN-DualLSTM model on the S of the UHF antenna in the HF-UHF RFID tag was verified. 11 Prediction accuracy. The mean absolute error (MAE), mean square error (MSE) and root mean square error (RMSE) evaluation indicators are used to quantify the prediction performance of each model. The error of each model and the time comparison with HFSS are shown in Fig. 9 The calculation method of the evaluation index is as follows:

[0088]

[0089]

[0090]

[0091] Among them, n represents the total number of samples in the test set, y prei represents the i-th predicted value, y truei represents the i-th true value.

[0092] Fig. 9 The prediction errors of CNN-DualLSTM, CNN-LSTM, LSTM, CNN, and BiLSTM models in the test set are shown, as well as the S of 100 UHF antennas calculated by HFSS. 11 The smaller the value of each evaluation index of the model, the better the predicted S 11 The closer the simulation value is to HFSS, the better the prediction performance of the model. Fig. 9 As shown in the figure, the MAE, MSE and RMSE of CNN-Dual LSTM are 0.0081, 0.00034 and 0.0184 respectively, which are better than the evaluation indicators of other models, indicating that the CNN-Dual LSTM model is effective for UHF antenna S 11 The prediction effect is better than that of CNN, LSTM, CNN-LSTM and BiLSTM models. From the perspective of computing time, CNN-DualLSTM can evaluate the S of 100 UHF antennas in less than 1 second.11 , while HFSS takes at least 4800 seconds. This shows that using CNN-Dual LSTM to evaluate the S performance of UHF antennas is 11 , which can greatly reduce time cost and computing power cost.

[0093] In a preferred embodiment, see Fig.10 The schematic diagram of the process of determining the size range of the UHF antenna shown, the antenna includes the UHF antenna and the high frequency antenna;

[0094] The value ranges of the various size parameters of the UHF antenna are obtained, including:

[0095] S28, obtaining the size parameters of the high frequency antenna;

[0096] S29, determining the value ranges of various size parameters of the UHF antenna according to the size parameters of the high frequency antenna, so that the UHF antenna does not overlap with the high frequency antenna or overlap itself.

[0097] Specifically, construct Figure 7 The initial structure of the HF-UHF RFID tag antenna shown in the figure, the HF antenna (high frequency antenna) is a coil antenna, and the UHF antenna (ultra-high frequency antenna) is a bent dipole antenna with an impedance matching loop. The number of turns and size of the coil antenna can be determined by the improved Wheeler formula for calculating the planar spiral inductance.

[0098] Improved Wheeler formula:

[0099]

[0100]

[0101]

[0102] In the above formula, L is the inductance, K is 1 , K 2 are the coefficients of the improved Wheeler expression, which are 2.34 and 2.75 respectively, μ 0 is the vacuum magnetic permeability, n is the number of turns of the coil, d avg is the average diameter of the coil, ρ represents the filling rate of the coil, do ut and d in Respectively represent the outer diameter and inner diameter of the coil antenna.

[0103] It is worth mentioning that the design of the antenna is determined by the chip used in the antenna. There is a capacitor in the chip. The capacitive reactance of the capacitor needs to be balanced by the inductive reactance provided by the antenna to make it conjugate matching, that is, the real part is equal and the imaginary part is opposite. Inductive reactance is inductance, and there is a mutual calculation relationship between capacitive reactance and inductive reactance, so you need to know the inductance first, and then design the antenna size so that the antenna parameters produce a reasonable inductance. The inductance of the antenna is closely related to the size of the antenna.

[0104] The HF antenna is easy to match the impedance of the dual-frequency RFID chip (EM4425 produced by EM Microelectronics) used, and is less affected by the size change of the UHF antenna. Therefore, the embodiment of the present invention first determines the size of the HF antenna and optimizes the size of the UHF antenna based on the existing HF antenna.

[0105] It is understandable that since both the UHF antenna and the HF antenna transmit and receive wireless signals when working, when the two antennas are in contact or overlapped, electromagnetic coupling may occur between the two antennas, thereby interfering with each other's work, resulting in signal attenuation, distortion or loss. Furthermore, since the HF antenna is easy to match the impedance of the dual-frequency RFID chip (EM4425 produced by EM Microelectronics) used, and is less affected by the size change of the UHF antenna, the present invention first determines the size of the HF antenna, and then determines the value range of each size parameter of the UHF antenna according to the size parameters of the HF antenna.

[0106] For example, see Figure 7 The structural diagram of the HF-UHF RFID tag antenna shown in FIG. 1 is a symmetrical structure, which mainly includes a substrate, a chip, a HF antenna and an UHF antenna. The chip, the HF antenna and the UHF antenna are arranged on the substrate. a1 represents the width of the matching loop and the bent dipole trace, b5 represents the width of the radiation patch, and a1, a2, a3, a4, a5, b1, b2, b3, b4, and b5 are the size parameters of the UHF antenna. Fig.11 and Fig.12 , Fig.12 This is a schematic diagram of the dimensions of the high-frequency antenna. Fig.11 It is a structural diagram of the high-frequency antenna, including the upper and lower layers of the high-frequency antenna. The tag is double-layered and identified on both sides, one is the front layer and the other is the back layer. s1, m, dout1, dout2, dout3, dout4, din1, din2, k, l, g, c1, d, and e1 are the size parameters of the high-frequency antenna.

[0107] In order to prevent the UHF antenna from overlapping with the HF antenna or itself, its dimensional parameters should meet the following constraints:

[0108]

[0109] Combined with the above constraints, the value range of each size parameter of the UHF antenna is determined through experiments as follows: Figure 8 shown.

[0110] It is worth noting that the structure of the tag antenna in the embodiment of the present invention is not limited to Figure 7 and 11 The specific structure can be designed according to actual conditions.

[0111] In a preferred implementation, the preset frequency is greater than or equal to 860 MHz and less than or equal to 960 MHz.

[0112] It is worth noting that the value range of the preset frequency point (860MHz to 960MHz) is set according to the working frequency band of the UHF antenna.

[0113] In a preferred implementation, the performance indicator value is a return loss characteristic value. Preferably, the performance indicator value adopts the S11 parameter (return loss characteristic value) among the S parameters of the antenna.

[0114] It is worth noting that the performance indicator value is not limited to the above-mentioned S11, but may also be other S parameters of the antenna, such as S22 parameter, S12 parameter, S21 parameter, etc., which are not limited here.

[0115] Specifically, the overall process of the antenna size optimization method provided by the embodiment of the present invention is as follows: constructing the initial structure of the HF-UHF RFID tag antenna, determining the size of the HF antenna, and defining the value range of each size parameter of the UHF antenna; using MATLAB-HFSS-API joint simulation to generate a data set of the HF-UHF RFID tag antenna according to the value range of the size parameter of the UHF antenna; establishing a CNN-DualLSTM model; using the simulation data set to train the CNN-DualLSTM until a good prediction accuracy is achieved; using the trained CNN-DualLSTM as an individual fitness calculator in the genetic algorithm optimization process; and using the genetic algorithm to automatically complete the optimization of the UHF antenna size in the HF-UHF RFID tag.

[0116] Compared with the prior art, the antenna size optimization method disclosed in the embodiment of the present invention obtains the value range of each size parameter of the UHF antenna; within the value range of each size parameter, iterates the size parameter combination for multiple rounds, and uses a pre-trained performance estimation model to calculate the performance index value of the UHF antenna at a preset frequency point corresponding to the size parameter combination value obtained after each round of iteration; and uses the size parameter combination value that makes the performance index value optimal as the optimal antenna size of the UHF antenna. The antenna size optimization method disclosed in the present invention iterates the size parameter combination of the UHF antenna for multiple times and evaluates the performance of the antenna corresponding to the size parameter combination obtained in each iteration through a pre-trained performance estimation model, so as to screen out the optimal antenna size of the UHF antenna. In the process of automatically optimizing the antenna size, there is no need to use electromagnetic simulation software, which effectively reduces the computing power cost and time cost in the antenna design process and reduces the risk of insufficient memory.

[0117] An embodiment of the present invention further provides an antenna, comprising an ultra-high frequency antenna and a high frequency antenna, wherein the size of the ultra-high frequency antenna is designed by the antenna optimization method described in any of the above embodiments.

[0118] An embodiment of the present invention further provides a household appliance product, which includes the antenna described in the above embodiment.

[0119] Optionally, the household appliance may be a refrigerator, an air conditioner or a television, etc., which is not limited here.

[0120] See also Fig.13 , Fig.13 : is a three-dimensional diagram of a refrigerator provided by an embodiment of the present invention. The refrigerator of this embodiment is approximately rectangular in shape, and includes a box body 100 defining a storage space and a plurality of door bodies 200 provided at the opening of the box body 100. The box body 100 is provided with a chamber, wherein the chamber includes a component storage chamber for placing components in the refrigerator, such as a compressor, etc., and also includes a storage space for storing food, medicine, etc.; the storage space can be divided into a plurality of storage rooms (compartments), and the storage rooms can be configured as a refrigeration room, a freezing room, a temperature-changing room (also called a fresh-keeping room) according to different uses, and each storage room corresponds to one or more door bodies, for example, in Figure 1 The storage room at the upper part is provided with a double-opening door body. The door body can be pivotally arranged at the opening of the box body, and can also be opened in a drawer-like manner to realize drawer-like storage.

[0121] Compared with the prior art, the antenna and household appliance products of the embodiments of the present invention obtain the value range of each size parameter of the UHF antenna; within the value range of each size parameter, perform multiple rounds of iterations on the size parameter combination, and use a pre-trained performance estimation model to calculate the performance index value of the UHF antenna at a preset frequency point corresponding to the size parameter combination value obtained after each round of iteration; and use the size parameter combination value that makes the performance index value optimal as the optimal antenna size of the UHF antenna. The antenna size optimization method, antenna and household appliance products disclosed in the present invention perform multiple iterations of the UHF antenna size parameter combination and evaluate the performance of the antenna corresponding to the size parameter combination obtained in each iteration through a pre-trained performance estimation model, so as to screen out the optimal antenna size of the UHF antenna. In the process of automatically optimizing the antenna size, there is no need to use electromagnetic simulation software, which effectively reduces the computing power cost and time cost in the antenna design process and reduces the risk of insufficient memory.

[0122] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for optimizing antenna size, characterized in that: include: Obtain the value range of each size parameter of the UHF antenna; Within the value range of each size parameter, iterate the size parameter combination of the UHF antenna for multiple rounds, and use a pre-trained performance estimation model to calculate the performance index value of the UHF antenna at a preset frequency point corresponding to the size parameter combination value obtained after each round of iteration; The size parameter combination value that makes the performance index value optimal is screened out to serve as the optimal antenna size of the UHF antenna.

2. The antenna size optimization method according to claim 1, characterized in that: The step of iterating the size parameter combination of the UHF antenna for multiple rounds within the value range of each size parameter, and using a pre-trained performance estimation model to calculate the performance index value of the UHF antenna at a preset frequency corresponding to the size parameter combination value obtained after each round of iteration, includes: Within the value range of each size parameter, a population including a preset number of individuals is initialized using a genetic algorithm encoding method, each individual represents a size parameter combination, and each size parameter combination includes a number of the size parameters; In each round of iteration, based on a preset generation gap, a number of mothers are selected from the population for crossover mutation, and the individuals obtained by the crossover mutation are merged with the parent population; The pre-trained performance estimation model is used to calculate the performance index value of each individual in the merged population at the preset frequency point, and several individuals with the best performance index value are selected as the population for a new round of iteration until the preset convergence condition is met and the iteration ends.

3. The antenna size optimization method according to claim 2, characterized in that: The performance estimation model is a CNN-DualLSTM model; the performance estimation model trained in advance is used to calculate the performance index value of each individual in the merged population at a preset frequency point, including: Each individual in the merged population is input into the CNN-DualLSTM model with the preset frequency points, and the spatial feature information is extracted by the CNN unit and transmitted to the DualLSTM unit; The DualLSTM unit extracts the electromagnetic characteristics of the spatial feature information and transmits it to the fully connected layer; The fully connected layer integrates information of the electromagnetic characteristics and transmits the integrated information to the output layer; The output layer processes the integrated information and outputs the performance index value of each individual in the combined population at the preset frequency point.

4. The antenna size optimization method according to claim 2 or 3, characterized in that: The performance estimation model is trained by a pre-acquired training data set; the training data set is acquired in the following manner: In the electromagnetic simulation software, based on the value range of each dimensional parameter, the each dimensional parameter is randomly adjusted to obtain several groups of reference dimensional combinations, and the performance index value of each reference dimensional combination at different reference frequencies is calculated, and the initial dimensional combination, the reference frequency and the corresponding performance index value are combined to obtain several groups of training data to form a training data set; wherein the reference frequency is greater than or equal to 860 MHz and less than or equal to 960 MHz.

5. The antenna size optimization method according to claim 3, characterized in that: In the CNN unit, the size of the convolution kernel is 1, the number of convolution kernels is 32, the step is 1, and the filling method is SAME; the DualLSTM unit is an LSTM with a double-layer unit of 64; the activation function of the output layer is a sigmoid function.

6. The antenna size optimization method according to claim 1, characterized in that: The antenna includes the UHF antenna and the HF antenna; The value ranges of the various size parameters of the UHF antenna are obtained, including: Obtaining size parameters of the high frequency antenna; The value ranges of the various size parameters of the UHF antenna are determined according to the size parameters of the high frequency antenna, so that the UHF antenna does not overlap with the high frequency antenna or overlap itself.

7. The antenna size optimization method according to claim 1, characterized in that: The preset frequency point is greater than or equal to 860 MHz and less than or equal to 960 MHz.

8. The antenna size optimization method according to claim 1, characterized in that: The performance indicator value is a return loss characteristic value.

9. An antenna, characterized in that: The antenna comprises an ultra-high frequency antenna and a high frequency antenna, and the size of the ultra-high frequency antenna is designed by the antenna optimization method according to any one of claims 1 to 8.

10. A household appliance, characterized in that: Comprising the antenna as claimed in claim 9.

Images