A method for simulating broadband low-frequency antenna impedance in a high-dynamic environment

Abstract

The present invention discloses a method for simulating the impedance of a broadband low-frequency antenna in a high-dynamic environment. The method includes the following steps: S1: obtaining the characteristic impedance of a low-frequency antenna within a desired frequency band and sampling a plurality of frequency points and their corresponding impedance characteristics; S2: respectively substituting the plurality of frequency points and their corresponding impedance characteristics into ADS software for simulation to obtain the capacitance, inductance, and resistance values in the antenna impedance simulation block diagram corresponding to each frequency point; S3: using the frequency value of each frequency point as input data for a BP neural network, and the capacitance, inductance, and resistance values corresponding to each frequency point as output data, and using the sum of the errors between the output data and the capacitance, inductance, and resistance values in the antenna impedance simulation block diagram at the corresponding frequency point as an indicator; S4: training the BP neural network until the indicators meet the requirements; and S5: embedding the trained BP neural network into an FPGA and connecting the various module units to form a broadband low-frequency antenna impedance simulator system. The present invention can quickly simulate the impedance characteristics of a real low-frequency antenna within a frequency range.

Description

Technical Field

[0001] The present invention relates to the technical field of low-frequency antenna communication equipment testing, and in particular to a method for simulating impedance of a broadband low-frequency antenna in a high-dynamic environment. Background Art

[0002] Testing antenna equipment is crucial to the stability of communication systems. Broadband, low-frequency antennas convert low-frequency current signals into electromagnetic waves and radiate them into space. However, low-frequency antennas are often bulky and require installation. The radiated electromagnetic waves can interfere with power supplies, computers, and other equipment used in on-site testing, significantly complicating the development and testing of antenna support circuits, such as impedance matching networks. Furthermore, due to size and mass constraints, broadband, low-frequency antennas often operate in an electrically small antenna state, causing drastic changes in antenna impedance. Existing antenna impedance simulation technology typically only simulates a single frequency point and a relatively simple operating environment. This technology cannot truly reflect the characteristic impedance of broadband, low-frequency antennas in highly dynamic environments and requires manual adjustment, making it difficult to rapidly simulate. Summary of the Invention

[0003] The purpose of the present invention is to overcome the above-mentioned technical deficiencies and propose a broadband low-frequency antenna impedance simulation method in a high-dynamic environment based on a BP neural network. This impedance simulation method can quickly simulate the impedance characteristics of a real low-frequency antenna within a frequency range, solving the problems of existing antenna impedance simulation technologies, such as narrow simulation bandwidth, high simulation operating frequency, low accuracy, simple simulation working environment, and the need for manual adjustment. In addition, this method does not radiate electromagnetic waves, but converts current signals into heat energy consumption. In the process of developing a broadband low-frequency antenna matching network, there is no need to set up a bulky antenna, which greatly reduces the R&D cost of the antenna system, facilitates development by R&D personnel, and improves development efficiency.

[0004] The present invention adopts the following technical solutions:

[0005] A method for simulating the impedance of a broadband low-frequency antenna in a high-dynamic environment comprises the following steps:

[0006] S1: Obtain the characteristic impedance of the low-frequency antenna within the required frequency band by actual measurement or simulation. The characteristic impedance includes the real part characteristic and the imaginary part characteristic of the antenna impedance, and obtain data of several frequency points within the required frequency band and the impedance characteristic data corresponding to each frequency point.

[0007] S2: Substitute the above-mentioned frequency point data and the impedance characteristic data corresponding to each frequency point into the ADS simulation software respectively, and build an antenna impedance simulation block diagram in ADS. When the frequency value of each frequency point is input, adjust the parameters of each component of the above-mentioned antenna impedance simulation block diagram so that the input impedance of the antenna impedance simulation block diagram at each frequency point is equal to the low-frequency antenna impedance characteristic data corresponding to each frequency point. Thus, the equivalent antenna impedance simulation block diagram corresponding to each frequency point of the actual low-frequency antenna and the capacitance value, inductance value, and resistance value in the antenna impedance simulation block diagram are obtained;

[0008] S3: Using the frequency values of each frequency point as input data of the BP neural network, and the capacitance, inductance, and resistance values corresponding to each frequency point as output data of the BP neural network, and using the sum of the errors between the output data and the capacitance, inductance, and resistance values in the antenna impedance simulation block diagram at the corresponding frequency point as an indicator;

[0009] S4: Using the input data and output data described in S3 to train the BP neural network, the BP neural network model includes an input layer, a hidden layer, and an output layer, initializing the weight from the input layer to the hidden layer, the weight from the hidden layer to the output layer, the threshold from the input layer to the hidden layer, and the threshold from the hidden layer to the output layer; inputting the input data into the input layer, calculating the output data, calculating the index based on the output data and the capacitance, inductance, and resistance data in the corresponding frequency point antenna impedance simulation block diagram, and judging whether the index meets the requirements; if not, updating the weight from the input layer to the hidden layer, the weight from the hidden layer to the output layer, the threshold from the input layer to the hidden layer, and the threshold from the hidden layer to the output layer until the index meets the requirements;

[0010] S5: The BP neural network trained in S4 is embedded in the FPGA. When the signal source sends a signal of a certain frequency to the low-frequency antenna impedance simulator, the frequency meter detects the signal frequency on the transmission line in real time and sends the frequency value to the FPGA. This allows the component parameters required by the low-frequency antenna impedance simulator at the corresponding frequency to be quickly obtained. The FPGA then controls the electronic switches of the capacitor, inductor, and resistor matrices to achieve the component values required by the low-frequency antenna impedance simulator at each frequency.

[0011] The beneficial effects of the present invention are:

[0012] (1) The present invention can truly reflect the impedance characteristics of a broadband low-frequency antenna within a frequency range. The capacitance, inductance, and resistance values of the broadband low-frequency antenna impedance simulator change rapidly with frequency, so it has a high-precision fitting effect at each frequency point. R&D personnel can design and test low-frequency antenna matching networks at multiple frequency points without having to replace equipment, and there is no need to set up a bulky low-frequency antenna, thereby reducing R&D costs and improving R&D efficiency.

[0013] (2) During the implementation of the present invention, it is only necessary to monitor the operating frequency of the broadband low-frequency antenna impedance simulator and input the real-time operating frequency into the FPGA control system in which the trained BP neural network model has been embedded. This can quickly obtain the component parameters required by the antenna impedance simulator at the corresponding frequency, avoiding the complicated iterative calculation process of the traditional algorithm. The computing resource requirements of the FPGA control system are small, and the simulation speed and accuracy of the broadband low-frequency antenna impedance simulator in a high dynamic working environment are greatly improved.

[0014] (3) The present invention takes into account the problem of the time required for the parameters of each component of the antenna impedance simulator to change with frequency, that is, the dynamic response problem of the antenna impedance simulator. Therefore, a switch-type capacitor, inductor, and resistor matrix is introduced. The introduction of each component is controlled by an electronic switch, which greatly shortens the time required for the parameters of each component to change, enhances the dynamic response capability of the antenna impedance simulator, and improves the broadband low-frequency antenna impedance simulator's ability to simulate broadband low-frequency antenna impedance working in a complex environment, with higher simulation accuracy and speed.

[0015] (4) The present invention can be applied to high-precision, high-speed broadband simulation of low-frequency antenna impedance in a variety of high-dynamic working environments. By adjusting the training set of the BP neural network, different weight parameters can be obtained, making the present invention more universal. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a flow chart of the present invention;

[0017] Figure 2 The block diagram for antenna impedance simulation;

[0018] Figure 3 It is the BP neural network training graph;

[0019] Figure 4 It is a broadband low-frequency antenna impedance simulator system;

[0020] Figure 5 It is a digital impedance network. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. The specific embodiments described herein are only used to explain the technical solution of the present invention and are not intended to limit the present invention.

[0022] The present invention provides a method for simulating broadband low-frequency antenna impedance in a high dynamic environment. The process is as follows: Figure 1 As shown, the following steps are included:

[0023] S1: Obtain the characteristic impedance of the low-frequency antenna within the required frequency band by actual measurement or simulation. The characteristic impedance includes the real part characteristic and the imaginary part characteristic of the antenna impedance, and obtain data of several frequency points within the required frequency band and the impedance characteristic data corresponding to each frequency point.

[0024] S2: Substitute the above frequency data and the impedance characteristic data corresponding to each frequency into the ADS simulation software, and build the antenna impedance simulation block diagram in ADS. The antenna impedance simulation block diagram is as follows: Figure 2 As shown, when the frequency value of each frequency point is input, the parameters of each component of the antenna impedance simulation block diagram are adjusted so that the input impedance of the antenna impedance simulation block diagram of each frequency point is equal to the low-frequency antenna impedance characteristic data corresponding to each frequency point. Thus, the equivalent antenna impedance simulation block diagram corresponding to each frequency point of the actual low-frequency antenna and the capacitance value, inductance value, and resistance value in the antenna impedance simulation block diagram are obtained;

[0025] S3: Create a BP neural network: Use the frequency values of each frequency point mentioned above as the input data of the BP neural network, and the capacitance, inductance, and resistance values required for each frequency point as the output data of the BP neural network, and use the sum of the errors between the output data and the capacitance, inductance, and resistance values in the antenna impedance simulation block diagram of the corresponding frequency point as an indicator, so that the low-frequency antenna impedance simulator can represent the impedance characteristics of the actual low-frequency antenna within a certain frequency range, thereby improving the simulation accuracy of the impedance characteristics of the actual low-frequency antenna within a certain frequency range. In the existing antenna impedance simulator technology, the capacitance, inductance, and resistance values often cannot change with the change of the input signal frequency during the operation of the system. Attempting to use fixed capacitance, inductance, and resistance values to simulate the broadband impedance characteristics of the low-frequency antenna makes it difficult to ensure the simulation accuracy of the impedance change characteristics of the broadband low-frequency antenna working in a high dynamic environment;

[0026] S4: Using the input data and output data described in S3 to train the BP neural network, the BP neural network model includes an input layer, a hidden layer, and an output layer. The BP neural network training is as follows: Figure 3As shown, the weights from the input layer to the hidden layer, the weights from the hidden layer to the output layer, the thresholds from the input layer to the hidden layer, and the thresholds from the hidden layer to the output layer are initialized; the input data is input to the input layer, the output data is calculated, the indicators are calculated according to the output data and the capacitance, inductance, and resistance data in the corresponding frequency point antenna impedance simulation block diagram, and it is determined whether the indicators meet the requirements. If not, the weights from the input layer to the hidden layer, the weights from the hidden layer to the output layer, the thresholds from the input layer to the hidden layer, and the thresholds from the hidden layer to the output layer are updated until the indicators meet the requirements;

[0027] S5: The BP neural network trained in S4 is embedded in the FPGA. When the signal source sends a signal of a certain frequency to the low-frequency antenna impedance simulator, a frequency counter detects the signal frequency on the transmission line in real time and sends the frequency value to the FPGA. This allows the required component parameters of the low-frequency antenna impedance simulator to be quickly obtained at the corresponding frequency. The FPGA then controls the electronic switches of the capacitor, inductor, and resistor matrix to achieve the required component values of the low-frequency antenna impedance simulator at each frequency. Embedding the trained network model in the FPGA reduces the use of internal FPGA computing resources, avoids the complex iterative calculation process of traditional algorithms, and greatly improves the simulation speed of broadband low-frequency antenna impedance simulators in highly dynamic operating environments. Existing antenna impedance simulator technology requires obtaining capacitance, inductance, and resistance values in the simulation software. The input signal frequency of the antenna impedance simulator is often constantly changing, and each frequency change requires re-obtaining new capacitance, inductance, and resistance values in the simulation software. This makes the low-frequency antenna impedance simulation system very complex, time-consuming, and slow to simulate.

[0028] The antenna impedance simulator in the embodiment of the present invention refers to a broadband low-frequency antenna impedance simulator that can emit electromagnetic waves in the range of 30 kHz to 3 MHz.

[0029] In this embodiment of the present invention, it is first necessary to obtain the antenna impedance characteristics within the user's desired frequency band. In this embodiment, the desired frequency band is 30 kHz to 3 MHz. The antenna impedance characteristics include the real and imaginary parts of the antenna impedance. This can be accomplished by creating an antenna model using HFSS or CST software, performing a frequency sweep, and recording the antenna impedance characteristics. Alternatively, a network analyzer can be used to perform a frequency sweep on an actual low-frequency antenna and record the antenna impedance characteristics.

[0030] The antenna impedance characteristic data described above is divided into frequency bands. This embodiment uses equally spaced frequency sampling. The sampling interval can be selected based on the user's actual needs. The smaller the sampling interval, the more data points are sampled, which improves the subsequent BP neural network training effect and increases the impedance simulator's accuracy in fitting the actual low-frequency antenna impedance. In this embodiment, a sampling interval of 6 kHz is selected to obtain several sampling frequency points and the corresponding low-frequency antenna impedance real and imaginary data at each frequency point.

[0031] Substitute the above-mentioned sampling frequency points and the low-frequency antenna impedance data corresponding to each frequency point into the ADS simulation software, and build an antenna impedance simulation block diagram in ADS. The antenna impedance simulation block diagram includes optimizable capacitors, optimizable inductors, and optimizable resistors. The antenna impedance simulation block diagram input impedance formula is:

[0032]

[0033] where Z in is the low-frequency antenna impedance data corresponding to the current frequency point, w is the frequency value corresponding to the current frequency point, R2 is the value of the resistor that can be optimized in the antenna impedance simulation block diagram, C2 is the value of the capacitor that can be optimized in the antenna impedance simulation block diagram, L2 is the value of the inductor that can be optimized in the antenna impedance simulation block diagram, and Z re is the real part of the low-frequency antenna impedance data corresponding to the current frequency point, Z im It is the imaginary part data of the low-frequency antenna impedance corresponding to the current frequency point.

[0034] By adjusting the parameters of each component of the antenna impedance simulation block diagram, the impedance characteristics of the antenna impedance simulation block diagram are made consistent with the low-frequency antenna impedance characteristics recorded by the user through simulation or actual measurement at the corresponding frequency point, and the parameters of each component of the impedance network corresponding to each frequency point are recorded.

[0035] In this embodiment, the BP neural network model includes an input layer, a hidden layer, and an output layer. The number of neurons in the input layer is 1, the number of neurons in the output layer is 3, and the number of neurons in the hidden layer can be set according to actual needs, for example, 10-15. The frequency value of each frequency point is used as the BP neural network input data, and the capacitance, inductance, and resistance values corresponding to each frequency point are used as the BP neural network output data. The sum of the average absolute values of the errors between the output data and the capacitance, inductance, and resistance values in the antenna impedance simulation block diagram at the corresponding frequency point is used as an indicator, so that the indicator is:

[0036]

[0037] Among them, E is the index, N is the total number of sampling data points, C M is the capacitance value in the output data of the Mth data point, L Mis the inductance value in the output data of the Mth data point, Z M is the resistance value in the output data of the Mth data point, C MT is the capacitance value in the antenna impedance simulation block diagram corresponding to the Mth data point, L MT is the inductance value in the antenna impedance simulation block diagram corresponding to the Mth data point, Z MT is the resistance value in the antenna impedance simulation block diagram corresponding to the Mth data point.

[0038] In selecting the indicator, the average value of the sum of squares of the errors between the output data and the capacitance, inductance, and resistance values in the antenna impedance simulation block diagram at the corresponding frequency point may also be used, that is, the indicator is:

[0039]

[0040] The error between the output data and the capacitance value, inductance value, and resistance value in the antenna impedance simulation block diagram at the corresponding frequency point is used as an indicator to ensure that the capacitance value, inductance value, and resistance value in the output data are closer to the capacitance value, inductance value, and resistance value in the antenna impedance simulation block diagram, thereby improving the fitting accuracy of the broadband low-frequency antenna impedance simulator.

[0041] In this embodiment, when determining the weight from the input layer to the hidden layer, the weight from the hidden layer to the output layer, the threshold from the input layer to the hidden layer, and the threshold from the hidden layer to the output layer in the BP neural network, a numerical value is randomly generated between 0 and 1, excluding 0 and 1, as the initial value. This method is used to set the weights and thresholds of the BP neural network, so that it can start training and can be gradually updated during the training process, and prevent the initial value from being too large and missing the optimal solution.

[0042] In this embodiment, the frequency values of the several sampling frequency points are input into one neuron of the input layer of the BP neural network, and then pass through the hidden layer and the output layer to finally obtain the required capacitance value, inductance value, and resistance value of the broadband low-frequency antenna impedance simulator at different frequencies.

[0043] The output of the hidden layer can be calculated according to the following formula:

[0044] α h =f(V 1h x1-γ 1h ),h=1,2,3...q

[0045] Among them, α h is the hidden layer output of the hth neuron in the hidden layer, q is the number of neurons in the hidden layer, V 1h is the weight from the first neuron in the input layer to the hth neuron in the hidden layer, x1 is the input data of the first neuron in the input layer, γ 1his the threshold from the first neuron in the input layer to the hth neuron in the hidden layer, f(*) is the activation function, and the activation function is x is the independent variable.

[0046] The output of the output layer can be calculated according to the following formula:

[0047]

[0048] Among them, y j is the output of the jth neuron in the output layer, w hj is the weight from the hth neuron in the hidden layer to the jth neuron in the output layer, θ hj is the threshold from the hth neuron in the hidden layer to the jth neuron in the output layer.

[0049] The indicator is calculated according to the indicator calculation model described above. If the calculated indicator is less than or equal to the expected error, it means that the indicator meets the requirements. Otherwise, it means that the indicator does not meet the requirements. The expected error can be selected according to actual needs.

[0050] Based on the above indicators, determine whether the calculated indicators meet the requirements. If not, it is necessary to update the weights from the input layer to the hidden layer, the weights from the hidden layer to the output layer, the thresholds from the input layer to the hidden layer, and the thresholds from the hidden layer to the output layer. If the requirements are met, the judgment ends. Among them, the weights from the input layer to the hidden layer can be updated according to the following formula:

[0051]

[0052]

[0053] where ΔV 1h is the weight change from the first neuron in the input layer to the hth neuron in the hidden layer, μ is the preset learning rate, E is the index, and y j is the output of the jth neuron in the output layer, α h is the hidden layer output of the hth neuron in the hidden layer, V 1h is the weight from the first neuron in the input layer to the hth neuron in the hidden layer, is the updated weight from the first neuron in the input layer to the hth neuron in the hidden layer, and q is the number of neurons in the hidden layer.

[0054] The weights from the hidden layer to the output layer can be updated according to the following formula:

[0055]

[0056]

[0057] where Δwhj is the weight change from the hth neuron in the hidden layer to the jth neuron in the output layer, w hj is the weight from the hth neuron in the hidden layer to the jth neuron in the output layer, is the updated weight from the hth neuron in the hidden layer to the jth neuron in the output layer.

[0058] The threshold for updating the input layer to the hidden layer can be based on the following formula:

[0059]

[0060]

[0061] where Δγ 1h is the threshold change from the first neuron in the input layer to the hth neuron in the hidden layer, γ 1h The threshold from the first neuron in the input layer to the hth neuron in the hidden layer, The updated threshold from the first neuron in the input layer to the hth neuron in the hidden layer.

[0062] The threshold for updating the hidden layer to the output layer can be based on the following formula:

[0063]

[0064]

[0065] where Δθ hj is the threshold change from the hth neuron in the hidden layer to the jth neuron in the output layer, θ hj is the threshold from the hth neuron in the hidden layer to the jth neuron in the output layer, is the updated threshold from the hth neuron in the hidden layer to the jth neuron in the output layer.

[0066] In this embodiment, by judging whether the indicators meet the requirements, it is decided whether it is necessary to update the weights from the input layer to the hidden layer, the weights from the hidden layer to the output layer, the thresholds from the input layer to the hidden layer, and the thresholds from the hidden layer to the output layer, and the updated weights and thresholds are substituted into the next iteration process. The weights and thresholds in the network are continuously updated through back propagation, so that the error indicators become smaller and smaller, and the output data becomes more accurate.

[0067] In this embodiment, the trained BP neural network model is embedded into the FPGA and combined with various functional modules to form a broadband low-frequency antenna impedance simulator system. Figure 4 As shown, the functional modules include a signal source, a coupler, a frequency detector, and a digital impedance network. The digital impedance network is as follows: Figure 5As shown, it includes a series inductance matrix, a parallel capacitance matrix, and a parallel resistance matrix. The inductance matrix includes 8 series inductors and 8 electronic switches in parallel with them. The inductance value increases by 2 times in sequence. The inductance value range of the series inductance matrix can reach 0~255L min , L min The inductance value of the inductor with the smallest inductance in the inductance matrix. The capacitance matrix contains 8 parallel capacitors and 8 electronic switches connected in series with them. The capacitance value increases by 2 times in sequence. The capacitance range of the parallel capacitance matrix can reach 0~255C min , C min The capacitance value of the smallest capacitor in the capacitance matrix. The resistance matrix contains 8 series resistors and 8 electronic switches connected in parallel. The resistance values increase by 2 times. The resistance range of the parallel resistor matrix can reach 0~255Z min , Z min The resistance value of the resistor with the smallest resistance in the resistor matrix. The use of a digital impedance network greatly reduces the time required for the capacitance, inductance, and resistance values to change with frequency, thereby improving the dynamic response capability of the entire system. Existing antenna impedance simulator technology typically uses fixed or sliding adjustable capacitors, inductors, and resistors. When the frequency of the antenna impedance simulator's input signal changes, manual adjustment of the capacitance, inductance, and resistance values is often required. This operation is cumbersome and time-consuming, and cannot meet the impedance simulation requirements for broadband, low-frequency antennas in dynamic environments.

[0068] In this embodiment, the signal source can transmit a signal within a frequency range. The signal enters the coupler through the transmission line, is output to a frequency detector through the coupling end of the coupler for real-time frequency calculation, and is output to the digital impedance network through the output end of the coupler. The frequency detector transmits the real-time frequency calculation result to the FPGA. The trained BP neural network in the FPGA maps the required capacitance value, inductance value, and resistance value at the frequency point. The FPGA performs an 8-bit binary equivalent replacement on the capacitance value, inductance value, and resistance value to control the 8 electronic switches of the inductance matrix, the 8 electronic switches of the capacitance matrix, and the 8 electronic switches of the resistance matrix in the digital impedance network, so that the input impedance of the digital impedance network at the corresponding frequency point is approximately equal to the impedance characteristics of the actual low-frequency antenna.

[0069] The above description merely represents the preferred embodiments of the present invention, and while the description is relatively detailed and specific, it should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make various modifications, improvements, and substitutions without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A method for simulating broadband low-frequency antenna impedance in a high-dynamic environment, characterized in that: The following steps are involved: S1: Obtain the characteristic impedance of the low-frequency antenna within the required frequency band by actual measurement or simulation. The characteristic impedance includes the real part characteristic and the imaginary part characteristic of the antenna impedance, and obtain data of several frequency points within the required frequency band and the impedance characteristic data corresponding to each frequency point. S2: Substitute the above-mentioned frequency point data and the impedance characteristic data corresponding to each frequency point into the ADS simulation software respectively, and build an antenna impedance simulation block diagram in ADS. When the frequency value of each frequency point is input, adjust the parameters of each component of the above-mentioned antenna impedance simulation block diagram so that the input impedance of the antenna impedance simulation block diagram at each frequency point is equal to the low-frequency antenna impedance characteristic data corresponding to each frequency point. Thus, the equivalent antenna impedance simulation block diagram corresponding to each frequency point of the actual low-frequency antenna and the capacitance value, inductance value, and resistance value in the antenna impedance simulation block diagram are obtained; S3: Using the frequency values of each frequency point as input data of the BP neural network, and the capacitance, inductance, and resistance values corresponding to each frequency point as output data of the BP neural network, and using the sum of the errors between the output data and the capacitance, inductance, and resistance values in the antenna impedance simulation block diagram at the corresponding frequency point as an indicator; S4: Using the input data and output data described in S3 to train the BP neural network, the BP neural network model includes an input layer, a hidden layer, and an output layer, initializing the weight from the input layer to the hidden layer, the weight from the hidden layer to the output layer, the threshold from the input layer to the hidden layer, and the threshold from the hidden layer to the output layer; inputting the input data into the input layer, calculating the output data, calculating the index based on the output data and the capacitance, inductance, and resistance data in the corresponding frequency point antenna impedance simulation block diagram, and judging whether the index meets the requirements; if not, updating the weight from the input layer to the hidden layer, the weight from the hidden layer to the output layer, the threshold from the input layer to the hidden layer, and the threshold from the hidden layer to the output layer until the index meets the requirements; S5: The BP neural network trained in S4 is embedded in the FPGA. When the signal source sends a signal of a certain frequency to the low-frequency antenna impedance simulator, the frequency meter detects the signal frequency on the transmission line in real time and sends the frequency value to the FPGA to obtain the component parameters required by the low-frequency antenna impedance simulator at the corresponding frequency. The FPGA then controls the electronic switches of the capacitor, inductor, and resistor matrix to achieve the component values required by the low-frequency antenna impedance simulator at each frequency.

Images