A long-wave antenna radiation characteristic measurement system based on power amplification

Abstract

The application belongs to the technical field of magnetoelectric antenna control, and particularly relates to a long-wave antenna radiation characteristic measurement system based on power amplification, which comprises a reference characteristic parameter acquisition module, a rate correlation dynamic inverse model construction module, an adaptive weight updating module and a closed-loop driving and radiation correction module. A reference signal is output by a power amplifier, and a normalization coefficient and a reference change rate constant are obtained. A rate correlation correction factor is constructed based on the instantaneous change rate of an expected input voltage signal, and a pre-distortion control signal is generated in combination with a hysteresis operator. The total number of loop delay points is determined, the instantaneous error and the gradient vector after time alignment are calculated, the weight of the hysteresis operator is updated by using an energy normalization algorithm, and the pre-distortion control signal is input into the power amplifier to drive the magnetoelectric antenna. The application solves the problems of dynamic hysteresis distortion and thermal drift of the magnetoelectric antenna under high-frequency driving, and significantly improves the quality and stability of the radiation waveform.

Description

Technical Field

[0001] This invention relates to the field of magnetoelectric antenna control technology. More specifically, this invention relates to a long-wavelength antenna radiation characteristic measurement system based on power amplification. Background Technology

[0002] Magnetoelectric antenna technology utilizes the magneto-electro-mechanical coupling effect of magnetostrictive materials to efficiently convert electrical energy into magnetic energy, playing a crucial role in underwater cross-medium communication, geological exploration, and low-frequency electromagnetic compatibility testing. Compared to traditional coil antennas, magnetoelectric antennas based on magnetostrictive materials offer significant advantages such as small size, high energy density, and fast response speed. To generate an alternating magnetic field of specific frequency and amplitude in a target area, a wideband power amplifier is typically required to drive the magnetoelectric antenna. However, magnetostrictive materials are essentially ferromagnetic smart materials with significant hysteresis nonlinear characteristics. This means that the output magnetic field strength depends not only on the current input voltage but also on the material's historical magnetization state. If this nonlinearity is not corrected, it can lead to severe distortion of the radiated magnetic field waveform, significantly reducing the signal-to-noise ratio of the communication system or the reliability of test results.

[0003] For compensation of hysteresis nonlinearity, most existing mainstream technologies employ feedforward open-loop control using static hysteresis models based on phenomenological theory. Among these, the Prandtl-Ishlinskii (PI) model is widely used in the linearization control of piezoelectric ceramics and magnetostrictive actuators due to its ease of solving its analytical inverse model and low computational complexity. This type of method typically assumes that the input-output relationship of the material is quasi-static. It identifies a set of fixed weight parameters offline to fit the hysteresis loop of the material and then connects an inverse model in series in the controller to counteract the hysteresis effect, thereby attempting to achieve a linear correspondence between the input voltage and the output magnetic field.

[0004] However, in practical high-frequency driving applications of magnetoelectric antennas, the aforementioned static control method reveals serious limitations: the flipping process of magnetic domains within the magnetostrictive material is heavily influenced by eddy current losses and viscous damping, exhibiting strong rate-dependent dynamic characteristics; as the driving signal frequency or the rate of change of the signal increases, the energy loss within the material increases dramatically, macroscopically manifested as a significant increase in the area of ​​the hysteresis loop and a change in its shape; the static PI model can only characterize a fixed hysteresis path independent of frequency and cannot predict or compensate for this dynamic nonlinear error that changes drastically with the input speed; furthermore, in actual physical systems, there is a fixed transmission delay from the digital controller to the magnetic field radiation, and the antenna will experience thermal drift in the magnetostrictive coefficient due to self-heating under prolonged high-power excitation. Under the combined effect of these complex factors, the control strategy based on the static model cannot achieve effective waveform correction, leading to a significant decrease in control accuracy in the high-frequency band, and even causing system oscillations and divergence. Summary of the Invention

[0005] To address the technical problem of severe distortion of the radiated magnetic field waveform and instability of the control system caused by the rate-dependent hysteresis characteristics of the magnetostrictive material, the transmission delay of the physical circuit, and the thermal drift effect during long-term operation of the aforementioned magnetoelectric antenna under dynamic high-frequency drive, which leads to the failure of the existing static control model, this invention provides a method for measuring the radiation characteristics of a long-wavelength antenna based on power amplification. The system includes the following modules: a reference characteristic parameter acquisition module, used to control the power amplifier to output a reference signal to drive the magnetoelectric antenna, and to obtain a normalization coefficient based on the collected peak input voltage and peak radiated magnetic field strength; and to obtain a reference rate of change constant based on the highest operating frequency of the magnetoelectric antenna and the maximum allowable output voltage amplitude of the power amplifier; and a rate-dependent dynamic inverse model construction module, used to collect the current... The expected input voltage signal at the previous moment is used to calculate the instantaneous rate of change of the signal using the differential principle; based on the absolute value of the instantaneous rate of change and the reference rate of change constant, a rate-related correction factor is constructed; based on the weight of the hysteresis operator, the rate-related correction factor is coupled as a time-varying gain to the hysteresis operator to generate a predistortion control signal; an adaptive weight update module is used to determine the total delay points of the system's hysteresis loop and to calculate the instantaneous error after time alignment and the corresponding gradient vector; based on the instantaneous error and the gradient vector, the weight of the hysteresis operator is updated using an energy normalization algorithm; a closed-loop drive and radiation correction module is used to input the predistortion control signal to a digital-to-analog converter to generate an analog voltage signal and drive the magnetoelectric antenna to realize closed-loop control of the magnetoelectric antenna radiation.

[0006] This invention constructs a dynamic inverse model that includes a rate-related correction factor and combines it with an adaptive weight update mechanism. This system can sense the rate of change of the input signal in real time and adjust the control strategy accordingly. This operation overcomes the shortcomings of traditional static models that cannot characterize the dynamic hysteresis characteristics of magnetostrictive materials under high-frequency drive due to eddy current losses and viscous damping. By coupling the rate-related correction factor into the hysteresis operator, the control model can automatically compensate for loop deformation as the signal frequency and amplitude change. At the same time, the adaptive update module can effectively offset the effects of physical loop transmission delay and thermal drift caused by long-term operation, thereby significantly reducing the waveform distortion of the radiated magnetic field under dynamic high-frequency conditions and improving the control accuracy and stability of the system.

[0007] Preferably, the normalization coefficient is the ratio of the peak value of the input voltage to the peak value of the radiated magnetic field strength; the reference rate of change constant is equal to the product of twice pi, the highest operating frequency of the magnetoelectric antenna design, and the maximum voltage amplitude allowed to be output by the power amplifier.

[0008] Preferably, the instantaneous change rate of the signal is the difference between the desired input voltage signal at the current moment and the desired input voltage signal at the previous moment, divided by the sampling period.

[0009] Preferably, the rate-related correction factor is calculated as follows: In the formula, For the current moment Rate-related correction factor; This is the dynamic sensitivity coefficient, with a value range of [1.0, 2.0]. It is a natural exponential function; For the current moment The instantaneous rate of change; To take the absolute value; The baseline rate of change constant; This is the static bias coefficient, with a value range of [0, 0.5].

[0010] This invention constructs a rate-related correction factor by utilizing the natural exponential function combined with the ratio of the instantaneous rate of change to the baseline rate of change constant. This factor can accurately simulate the nonlinear response of magnetostrictive materials under different driving rates. This calculation method nonlinearly maps the rate of change of the signal to a gain adjustment coefficient, reflecting the physical fact that as the frequency or rate of change increases, the internal damping effect of the material intensifies, leading to increased output attenuation or hysteresis. This provides a dynamic correction term for the static hysteresis operator that conforms to physical laws, enabling the model to accurately predict complex hysteresis behavior at high frequencies.

[0011] Preferably, the formula for calculating the predistortion control signal is: In the formula, For the current moment The predistortion control signal; The total number of hysteresis operators; The sequence number of the hysteresis operator; For the first The hysteresis operator at the current time The weights; For the current moment Rate-related correction factor; For the first The hysteresis operator at the current time The output status.

[0012] This invention generates a predistortion control signal by multiplying the weights of the hysteresis operator with the rate-related correction factor in real time and accumulating the results. The system constructs a dynamic inverse model in the digital domain that is opposite to the physical magnetoelectric antenna characteristics. This operation actually pre-superimposes a distortion component that is equal in magnitude and opposite in direction to the dynamic hysteresis characteristics of the material into the driving signal. When this signal is applied to the antenna, it can precisely cancel the nonlinearity and rate-related hysteresis of the material itself, thereby forcing the final radiated magnetic field waveform to strictly follow the desired signal and achieve high-fidelity linearized output.

[0013] Preferably, the method for determining the output state of the hysteresis operator at the current moment includes: calculating the sum of the desired input voltage signal at the current moment and the threshold parameter of the hysteresis operator; selecting the minimum value of the sum and the output state of the hysteresis operator at the previous moment; calculating the difference between the desired input voltage signal at the current moment and the threshold parameter of the hysteresis operator; and selecting the maximum value of the difference and the minimum value as the output state.

[0014] Preferably, the instantaneous error is the product of the expected input voltage signal of the total delay points of the hysteresis loop minus the normalization coefficient and the actual radiated feedback signal.

[0015] Preferably, the gradient vector is the product of the rate-related correction factor of the total number of sampling points of the hysteresis loop and the output state of the hysteresis operator of the total number of sampling points of the hysteresis loop.

[0016] Preferably, updating the weights of the hysteresis operator using the energy normalization algorithm includes: In the formula, For the first The hysteresis operator at the next moment The weights; For the first The hysteresis operator at the current time The weights; This is the iteration step size factor, with a value range of [0.001, 0.05]. For the current moment The instantaneous error; , The first , The hysteresis operator at the current time The corresponding gradient vector; , For indexes of hysteresis operators; To prevent tiny positive numbers with a denominator of zero.

[0017] This invention utilizes an energy normalization algorithm combined with instantaneous error and gradient vector to update the weights of the hysteresis operator in real time, enabling the system to learn online and resist interference. The normalization process prevents the weight adjustment step size from getting out of control due to excessive gradient when the input signal amplitude is large, ensuring the numerical stability of the algorithm. At the same time, continuous iterative weight updates can automatically track and compensate for the magnetostriction coefficient drift caused by the heat generated by the long-term high-power operation of the magnetoelectric antenna, ensuring that the system can maintain an extremely high level of linearity control throughout the entire working cycle.

[0018] Preferably, controlling the power amplifier to output a reference signal to drive the magnetoelectric antenna includes: controlling the power amplifier to output a reference sine signal within the linear region to drive the magnetoelectric antenna, and acquiring the peak value of the input voltage and the peak value of the radiated magnetic field strength.

[0019] The beneficial effects of this invention are as follows:

[0020] This invention constructs a dynamic inverse model that includes a rate-related correction factor and combines it with an adaptive weight update mechanism. This system can sense the rate of change of the input signal in real time and adjust the control strategy accordingly. This operation overcomes the shortcomings of traditional static models that cannot characterize the dynamic hysteresis characteristics of magnetostrictive materials under high-frequency drive due to eddy current losses and viscous damping. By coupling the rate-related correction factor into the hysteresis operator, the control model can automatically compensate for loop deformation as the signal frequency and amplitude change. At the same time, the adaptive update module can effectively offset the effects of physical loop transmission delay and thermal drift caused by long-term operation, thereby significantly reducing the waveform distortion of the radiated magnetic field under dynamic high-frequency conditions and improving the control accuracy and stability of the system. Attached Figure Description

[0021] Figure 1 This is a schematic diagram illustrating a system block diagram of a long-wave antenna radiation characteristic measurement system based on power amplification according to the present invention.

[0022] Figure 2 This is a schematic diagram illustrating the comparison of global temporal domain tracking effects;

[0023] Figure 3 This is a schematic diagram showing the enlarged details of signal tracking under high-frequency operating conditions;

[0024] Figure 4 This is a schematic diagram illustrating the comparison of instantaneous errors in closed-loop control;

[0025] Figure 5 This is a schematic diagram illustrating the comparison of input-output hysteresis characteristics under high-frequency drive. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0028] This invention provides a method for measuring the radiation characteristics of a long-wavelength antenna based on power amplification. For example... Figure 1 As shown, a method for measuring the radiation characteristics of a long-wave antenna based on power amplification includes a reference characteristic parameter acquisition module 100, a rate-correlation dynamic inverse model construction module 200, an adaptive weight update module 300, and a closed-loop driving and radiation correction module 400, which are described in detail below.

[0029] The reference characteristic parameter acquisition module 100 is used to control the power amplifier to output a reference signal to drive the magnetoelectric antenna. It obtains the normalization coefficient based on the peak value of the collected input voltage and the peak value of the radiated magnetic field strength, and obtains the reference rate of change constant based on the maximum operating frequency and maximum driving amplitude of the antenna.

[0030] It should be noted that since the driving voltage signal output by the power amplifier and the magnetic field signal radiated by the antenna belong to different physical dimensions and their numerical magnitudes differ greatly, direct subtraction cannot accurately reflect the control error of the system. In addition, when calculating the rate of change of the signal, high-frequency signals can cause the derivative value to be too large. If normalization is not performed, the result of subsequent exponential calculations will overflow. Therefore, before formal operation, this invention first obtains the normalization coefficient and the benchmark rate of change constant through benchmark operating condition tests to construct a unified measurement space for subsequent data processing.

[0031] Specifically, the reference sinusoidal signal output from the power amplifier within the linear region drives the magnetoelectric antenna to acquire the peak value of the input voltage. and the peak value of the received magnetic flux density Based on the antenna's maximum operating frequency and maximum driving amplitude Determine the normalization coefficients and the benchmark rate of change constant .

[0032] Among them, the normalization coefficient The formula for calculation is:

[0033]

[0034] In the formula, These are the normalization coefficients; This represents the peak value of the input voltage acquired under baseline operating conditions. This represents the peak value of the radiated magnetic field strength collected under the reference operating conditions.

[0035] The calculation formula is obtained through By linearly mapping magnetic signals of different physical dimensions to equivalent voltage signals, the feedback signal can be compared with the input signal on the same order of magnitude; when When it increases, The reduction ensures the numerical stability after mapping.

[0036] Among them, the benchmark rate of change constant The formula for calculation is:

[0037]

[0038] In the formula, The baseline rate of change constant; Pi; The highest operating frequency designed for magnetoelectric antennas; This represents the maximum voltage amplitude that the power amplifier is allowed to output.

[0039] The calculation formula is obtained through A physical upper limit is defined for the rate of signal change, serving as a reference benchmark for subsequent calculations of the relative rate of change; when or When it increases, This increases accordingly, ensuring the rationality of the normalized denominator.

[0040] It should be noted that by obtaining the normalization coefficients and the reference rate of change constant, a unified dimensional standard and physical boundary constraint are provided for the subsequent nonlinear control algorithm. The normalization operation eliminates the huge difference in magnitude between the input voltage and the radiated magnetic field strength, ensuring the numerical stability of the feedback error calculation. The reference rate of change constant sets the maximum dynamic range reference value of the system, enabling the algorithm to accurately measure the relative degree of change of the current signal, providing a reliable benchmark for accurately calculating the rate-related correction factor, and preventing calculation divergence caused by non-standardized parameters.

[0041] The rate-related dynamic inverse model construction module 200 is used to acquire the desired input voltage signal at the current moment, calculate the instantaneous rate of change of the signal using the differential principle, construct a rate-related correction factor based on the absolute value of the instantaneous rate of change and the reference rate of change constant, and couple the rate-related correction factor as a time-varying gain to the hysteresis operator based on the weight of the hysteresis operator to generate a predistortion control signal.

[0042] It should be noted that this invention constructs a dynamic correction function that includes an input rate of change factor and couples it as a time-varying gain to the hysteresis operator to adjust the compensation strength of the model in real time.

[0043] Specifically, collect the current moment The expected input voltage signal Calculate the instantaneous rate of change of the signal. .

[0044] Among them, instantaneous rate of change The formula for calculation is:

[0045]

[0046] In the formula, For the current moment The instantaneous rate of change; For the current moment The desired input voltage signal; For the previous moment The desired input voltage signal; This refers to the sampling period of the digital control system.

[0047] This calculation formula utilizes the difference principle to extract the time-domain characteristics of the signal's rate of change; when the desired input voltage signal... Instantaneous rate of change during high-frequency oscillations or large-amplitude step changes The absolute value of will increase significantly, indicating that the magnetic material is in a high dynamic operating region at this time.

[0048] Furthermore, based on the instantaneous rate of change Construction rate related correction factor .

[0049] Among them, rate-related correction factors The formula for calculation is:

[0050]

[0051] In the formula, For the current moment Rate-related correction factor; The value is the dynamic sensitivity coefficient, which ranges from [1.0, 2.0]. In this embodiment, it is set to 1.5. In other embodiments, the implementer can adjust it according to the actual situation. It is a natural exponential function; For the current moment The instantaneous rate of change; To take the absolute value; The baseline rate of change constant; This is the static bias coefficient, with a value range of [0, 0.5]. In this embodiment, it is set to 0.2. In other embodiments, implementers can adjust it according to the actual situation.

[0052] The calculation formula utilizes the nonlinear growth characteristic of the exponential function to simulate the physical phenomenon of eddy current loss increasing rapidly with frequency; when the input signal changes gradually... When the value is small, the rate-related correction factor close to the base value When the input signal changes drastically... near At that time, the rate-related correction factor It increases exponentially; this relationship ensures that the model remains stable at low frequencies, while at high frequencies it can output a sufficiently large gain to compensate for the increased hysteresis loss.

[0053] Specifically, predistortion control signals are generated by combining hysteresis operators. .

[0054] Among them, the predistortion control signal The formula for calculation is:

[0055]

[0056]

[0057] In the formula, For the current moment The predistortion control signal; The total number of hysteresis operators; The sequence number of the hysteresis operator; For the first The hysteresis operator at the current time The weights; For the current moment Rate-related correction factor; For the first The hysteresis operator at the current time The output state; For the current moment The desired input voltage signal; For the first Threshold parameters for each hysteresis operator; For the first The hysteresis operator at the previous time... The output status.

[0058] It should be noted that the system is configured to start from a zero state, therefore, =0.

[0059] The calculation formula incorporates the dynamic characteristics at the physical level, namely the rate-related correction factor. Static hysteresis memory at the mathematical level, i.e., the output state of the hysteresis operator. To achieve multiplicative coupling; Responsible for based on threshold parameters This produces the basic hysteresis shape, while This serves as a dynamic magnification factor, amplifying or reducing the amplitude of the hysteresis shape in real time; when When the predistortion control signal increases with increasing signal frequency, The predistortion component contained therein is also enhanced, thereby physically offsetting the additional hysteresis caused by the antenna being driven at high frequencies.

[0060] It should be noted that by calculating the output state of the hysteresis operator based on the current input, threshold parameters, and the state at the previous time step, the memory effect and local extremum characteristics of magnetic domain flipping inside the magnetostrictive material are reproduced. This operator-based iterative calculation can accurately describe the multi-valued mapping and loop characteristics in the material's input-output relationship, forming the basic unit for describing complex hysteresis nonlinearity. This ensures that the control model can accurately fit the material's basic magnetization curve at the quasi-static level, providing an accurate static framework for further dynamic correction.

[0061] The adaptive weight update module 300 is used to determine the total delay points of the hysteresis loop in the system and to calculate the instantaneous error after time alignment and the corresponding gradient vector; based on the instantaneous error and gradient vector, the weights of the hysteresis operator are updated using an energy normalization algorithm.

[0062] It should be noted that when a power amplifier drives a load, the signal undergoes a fixed physical time delay after amplification, transmission, radiation, and acquisition feedback. If the feedback signal at the current moment is directly compared with the reference signal at the current moment, the phase difference will be incorrectly identified as a nonlinear amplitude error, causing the algorithm to diverge. At the same time, after long-term operation, the magnetostriction coefficient of the antenna will drift due to the increase in temperature, and the fixed model parameters cannot adapt to this change. Therefore, this invention performs time alignment before error calculation and uses the normalized gradient descent algorithm to update the model weights online.

[0063] Specifically, the total delay points of the system's hysteresis loop are measured. And calculate the instantaneous error after time alignment. and the corresponding gradient vector .

[0064] Among them, instantaneous error The formula for calculation is:

[0065]

[0066] In the formula, For the current moment The instantaneous error; For the current moment Lag The expected input voltage signal at each sampling point; The number of physical loop delay sampling points in the system; These are the normalization coefficients; For the receiving coil at the current moment The actual radiation feedback signal collected.

[0067] It should be noted that by introducing the total delay points of the hysteresis loop to align the desired signal in time, and by using the normalization coefficient to align the feedback signal in amplitude, the calculated instantaneous error can truly reflect the nonlinear residual between the control algorithm and the physical system. After eliminating the interference of transmission delay and proportional difference, the error signal is purely caused by the uncompensated nonlinear characteristics of the material. This provides an accurate optimization target for the adaptive algorithm, prevents the controller from making erroneous advance or lag adjustments due to the inherent physical delay of the system, and ensures the effectiveness of closed-loop correction.

[0068] Wherein, gradient vector The formula for calculation is:

[0069]

[0070] In the formula, For the first The hysteresis operator at the current time The corresponding gradient vector; For the current moment Lag Rate-related correction factor for each sampling point; For the current moment Lag The hysteresis operator outputs the state at each sampling point.

[0071] This calculation formula utilizes historical data. With the actual radiation feedback signal at the current moment Matching eliminates the effects of transmission delay and ensures instantaneous error. It only reflects waveform distortion caused by nonlinearity and thermal drift; similarly, the gradient vector It also uses the historical state at the same time for calculation, ensuring mathematical causal consistency.

[0072] It should be noted that by calculating the product of the rate-related correction factor and the output state of the hysteresis operator as the gradient vector, the sensitivity direction of the influence of each hysteresis operator weight on the total output error is clarified. This operation, based on the chain rule, determines the direction and proportion that each weight should be adjusted to reduce the error at the current moment, providing the steepest descent search path for the adaptive algorithm. This ensures that in complex dynamic nonlinear environments, the weight update can quickly converge to the optimal solution, rather than oscillating in the wrong direction.

[0073] Furthermore, the weights for the next time step are updated using an energy normalization algorithm. .

[0074] Among them, the weight of the next time step The formula for calculation is:

[0075]

[0076] In the formula, For the first The hysteresis operator at the next moment The weights; For the first The hysteresis operator at the current time The weights; The iteration step size factor has a value range of [0.001, 0.05]. In this embodiment, it is set to 0.01. In other embodiments, implementers can adjust it according to the actual situation. For the current moment The instantaneous error; , The first , The hysteresis operator at the current time The corresponding gradient vector; For indexes of hysteresis operators; It is the sum of squares of the gradient vectors of all hysteresis operators; For the summation index of the hysteresis operator; To prevent tiny positive numbers with a denominator of zero, this embodiment is set as follows: .

[0077] The calculation formula is based on instantaneous error. The sign and magnitude of the gradient are used to adjust the weights in the opposite direction of the gradient; the energy normalization term in the denominator makes the adjustment of the weights no longer solely dependent on the magnitude of the signal amplitude; when the power amplifier outputs a large voltage, causing... When the value is large, the denominator increases accordingly, thus limiting the gain of weight updates and preventing the algorithm from oscillating or diverging under large signal driving conditions; if the antenna's radiation efficiency decreases due to heat, i.e. Reduce instantaneous error It will become a positive value, prompting the weights to... Increase, thereby enhancing the predistortion control signal The range of heat loss is automatically compensated.

[0078] The closed-loop drive and radiation correction module 400 is used to input the predistortion control signal to the digital-to-analog converter, generate an analog voltage signal and drive the magnetoelectric antenna to realize closed-loop control of the magnetoelectric antenna radiation.

[0079] It should be noted that the above calculation process needs to be executed in real time in a loop in a digital signal processor to form a complete closed-loop control; only by converting the calculated digital quantity into an analog quantity and applying it to the physical object can waveform correction be finally achieved.

[0080] Specifically, the predistortion control signal The signal is input to a digital-to-analog converter, which generates an analog voltage signal and inputs it to a power amplifier. The power amplifier then linearly amplifies the analog voltage signal to drive the magnetoelectric antenna.

[0081] At this point, the driving voltage applied across the antenna already includes the predistortion control signal. The antenna carries a reverse nonlinear characteristic; the positive physical hysteresis of the antenna itself and the reverse predistortion characteristic in the driving signal cancel each other out at the physical level, so that the magnetic field waveform radiated by the antenna at the end highly restores the desired input voltage signal. The waveform achieves high linearity signal radiation.

[0082] For example, Figure 2 This diagram illustrates the comparison of global time-domain tracking performance. The dashed lines represent the waveform curve of the desired input voltage signal set by the system. The output response curve under the existing static PI model control exhibits divergent large-amplitude oscillations, significantly deviating from the desired signal in subsequent time intervals. In contrast, the output response curve under the rate-correlation dynamic inverse model control described in this invention highly overlaps with the waveform curve of the desired input voltage signal and maintains stable amplitude. This demonstrates that under long-term frequency sweep conditions, this invention can effectively suppress oscillations caused by model mismatch and maintain stable tracking performance. Figure 3 This is a magnified schematic diagram showing the details of signal tracking under high-frequency operating conditions; the image is a cropped section. Figure 2 Waveform details in the mid-to-high frequency, high rate of change region; the dashed line represents the desired input voltage signal; for the curve corresponding to the output of the existing static model, the amplitude is significantly attenuated and the phase is severely lagging, indicating that it cannot compensate for eddy current losses at high frequencies; for the curve corresponding to the output of the dynamic model of this invention, the waveform is full and the phase is basically synchronized with the dashed line, indicating that the pass rate related correction factor of this invention accurately compensates for the gain attenuation of the physical system in the high frequency band.

[0083] further, Figure 4 This is a schematic diagram comparing the instantaneous error of closed-loop control. The curve corresponding to the instantaneous error under the existing static model control shows a trumpet-shaped divergence with violent amplitude fluctuations, indicating that the control accuracy decreases sharply as the frequency increases. The curve corresponding to the instantaneous error under the dynamic model control of this invention shows that after a brief fluctuation in the initial stage, it quickly converges and remains in a narrow band near the zero axis, proving that the adaptive weight update algorithm of this invention has extremely strong convergence and anti-interference ability, and can limit the tracking error to a very small range.

[0084] final, Figure 5 This diagram illustrates the comparison of input-output hysteresis characteristics under high-frequency drive. The desired input voltage is plotted on the horizontal axis, and the actual normalized output on the vertical axis. The diagonal dashed line represents the ideal linear reference relationship. The curves corresponding to the input-output relationship of the existing static model are scattered and irregularly distributed, forming a wide, ring-shaped set of points, indicating severe nonlinear hysteresis. In contrast, the curves corresponding to the input-output relationship of the dynamic model of this invention are closely distributed around the diagonal dashed line, forming an extremely narrow strip-shaped set of points. This demonstrates that the present invention successfully cancels out physical hysteresis through pre-distortion control, achieving a high degree of linearity between input and output.

Claims

1. A system for measuring the radiation characteristics of a long-wavelength antenna based on power amplification, characterized in that, include: The reference characteristic parameter acquisition module is used to control the power amplifier to output a reference signal to drive the magnetoelectric antenna, and to obtain the normalization coefficient based on the collected peak input voltage and peak radiated magnetic field strength. The reference rate of change constant is obtained based on the highest operating frequency of the magnetoelectric antenna design and the maximum allowable voltage amplitude of the power amplifier output. The rate-correlation dynamic inverse model construction module is used to acquire the desired input voltage signal at the current moment and calculate the instantaneous rate of change of the signal using the difference principle: the difference between the desired input voltage signal at the current moment and the desired input voltage signal at the previous moment is divided by the sampling period; based on the absolute value of the instantaneous rate of change and the reference rate of change constant, a rate-correlation correction factor is constructed. ; For the current moment Rate-related correction factor; This is the dynamic sensitivity coefficient, with a value range of [1.0, 2.0]. It is a natural exponential function; For the current moment The instantaneous rate of change; To take the absolute value; The baseline rate of change constant; This is the static bias coefficient, with a value range of [0, 0.5]. Based on the weights of the hysteresis operator, the rate-related correction factor is coupled as a time-varying gain to the hysteresis operator to generate a predistortion control signal. ; For the current moment The predistortion control signal; The total number of hysteresis operators; The sequence number of the hysteresis operator; For the first The hysteresis operator at the current time The weights; For the first The hysteresis operator at the current time The output state; An adaptive weight update module is used to determine the total delay points of the system's hysteresis loop and to calculate the instantaneous error after time alignment and the corresponding gradient vector; based on the instantaneous error and the gradient vector, the weights of the hysteresis operator are updated using an energy normalization algorithm. The closed-loop drive and radiation correction module is used to input the predistortion control signal to the digital-to-analog converter, generate an analog voltage signal and drive the magnetoelectric antenna to realize closed-loop control of the magnetoelectric antenna radiation.

2. The long-wavelength antenna radiation characteristic measurement system based on power amplification according to claim 1, characterized in that, The normalization coefficient is the ratio of the peak value of the input voltage to the peak value of the radiated magnetic field strength; The reference rate of change constant is equal to the product of twice pi, the highest operating frequency of the magnetoelectric antenna design, and the maximum voltage amplitude allowed to be output by the power amplifier.

3. The long-wavelength antenna radiation characteristic measurement system based on power amplification according to claim 1, characterized in that, The method for determining the output state of the hysteresis operator at the current time includes: Calculate the sum of the desired input voltage signal at the current moment and the threshold parameter of the hysteresis operator; select the minimum value of the sum and the output state of the hysteresis operator at the previous moment; Calculate the difference between the desired input voltage signal at the current moment and the threshold parameter of the hysteresis operator; select the maximum value between the difference and the minimum value as the output state.

4. The long-wavelength antenna radiation characteristic measurement system based on power amplification according to claim 1, characterized in that, The instantaneous error is the product of the expected input voltage signal of the total delay points of the hysteresis loop minus the normalization coefficient and the actual radiation feedback signal.

5. The long-wavelength antenna radiation characteristic measurement system based on power amplification according to claim 1, characterized in that, The gradient vector is the product of the rate-related correction factor of the total number of sampling points of the hysteresis loop and the output state of the hysteresis operator of the total number of sampling points of the hysteresis loop.

6. The long-wavelength antenna radiation characteristic measurement system based on power amplification according to claim 1, characterized in that, The method of updating the weights of the hysteresis operator using the energy normalization algorithm includes: ； In the formula, For the first The hysteresis operator at the next moment The weights; For the first The hysteresis operator at the current time The weights; This is the iteration step size factor, with a value range of [0.001, 0.05]. For the current moment The instantaneous error; , The first , The hysteresis operator at the current time The corresponding gradient vector; , For indexes of hysteresis operators; To prevent tiny positive numbers with a denominator of zero.

7. The long-wavelength antenna radiation characteristic measurement system based on power amplification according to claim 1, characterized in that, The control power amplifier outputs a reference signal to drive the magnetoelectric antenna, including: The power amplifier outputs a reference sinusoidal signal within the linear region to drive the magnetoelectric antenna, and collects the peak value of the input voltage and the peak value of the radiated magnetic field strength.

Images