Radio frequency microwave signal modulation method and system

Abstract

The invention relates to the technical field of frequency conversion, in particular to a radio frequency microwave signal modulation method and system, and the method comprises the following steps: detecting the ratio of the time domain peak interval to the frequency domain main sideband energy of a radio frequency signal, inputting the time domain peak interval and the frequency domain main sideband energy into a cross-correlation algorithm for synchronism analysis, extracting the periodic characteristics of a time domain peak sequence, and obtaining a frequency domain peak sequence; and generating an energy coupling factor based on a difference operation result of the frequency domain energy density gradient value and the time domain peak interval. According to the method, periodic characteristics are analyzed and extracted through the ratio of peak spacing to main sideband energy to generate an energy coupling factor, a main frequency point offset and an energy tilt vector are fused to generate a harmonic tendency parameter, an interference source path is positioned by combining a frequency trajectory vector, and disturbance levels are divided by clustering energy and harmonic parameters. And a conflict direction is pre-judged and stop band parameters are pre-loaded by grey prediction of a main sideband frequency difference value trend, and a closed-loop regulation and control system is constructed to improve the anti-drift capability and reduce the spectrum conflict risk.

Description

Technical Field

[0001] The present invention relates to the technical field of frequency conversion, and particularly to a method and system for modulating radio frequency and microwave signals. Background Art

[0002] The technical field of frequency conversion involves the transformation processing of radio frequency signals or microwave signals in terms of spectral position. The core content of this technical field includes the up-conversion and down-conversion operations of signals, and the conversion of signal frequencies is achieved through the collaborative action of a local oscillator signal and a mixer. In applications such as radar, communication, electronic countermeasure, and radio measurement, frequency conversion is a fundamental means to ensure the transmission, reception, and processing of system signals in a specified frequency band. Overall, this technical field also involves multiple aspects such as local oscillator frequency control, mixer path matching, noise suppression, and spectrum broadening, constituting a comprehensive system covering signal generation, frequency mapping, and multi-stage allocation.

[0003] Among them, the method for modulating radio frequency and microwave signals refers to a technical solution that realizes the orderly adjustment of signals in the time or frequency domain by controlling the amplitude, phase, or frequency characteristics of signals in the radio frequency or microwave frequency band. It mainly aims at the problem that the original baseband signal needs to be modulated in a specific communication link or radio frequency processing module to adapt to the transmission in a specific frequency band. Specifically, it adopts a method of realizing frequency synthesis based on a phase-locked loop structure, combines the output of a high-order analog mixer to generate a modulated signal, and completes the modulation process by precisely controlling the superposition sequence of the local oscillator frequency and the intermediate frequency signal. In addition, this method also combines a temperature compensation circuit to adjust the linear response of the signal amplitude, and limits the modulated output spectrum through a harmonic suppression filter, thereby completing the frequency conversion requirements of the entire modulation process.

[0004] The prior art relies on the combination of a phase-locked loop and a fixed filter to suppress interference, and does not establish a dynamic correlation mechanism for time-frequency domain parameters. For example, when the peak interval in the time domain of a signal and the energy distribution of the main sideband in the frequency domain show asynchronous changes, the existing scheme cannot identify the collaborative deviation of time-frequency domain characteristics, resulting in phase mismatch of the modulated signal. The parameters of the existing harmonic suppression filter are fixed, and the coupling relationship between the offset of the main frequency point and the harmonic energy distribution is not quantified. In the scenario of rapid signal frequency hopping, the residual harmonic energy exceeds the stopband range of the filter. The prior art does not construct a sequence of main sideband frequency trajectory vectors and cannot capture the continuity characteristics of adjacent cycle frequency offsets. When environmental noise causes cumulative drift of the main sideband frequency, the adjustment of the stopband boundary lags behind the actual interference change, resulting in energy leakage of the passband signal. The existing temperature compensation circuit only corrects the linear response of the amplitude and does not synchronously correlate the frequency offset trend with the energy tilt parameter, resulting in the lack of dynamic adaptability of interference suppression measures. Summary of the Invention

[0005] The purpose of the present invention is to solve the deficiencies existing in the prior art, and to propose a method and system for modulating radio frequency and microwave signals.

[0006] To achieve the above object, the present invention adopts the following technical solutions: A radio frequency microwave signal modulation method, comprising the following steps: S1: By detecting the ratio of the time-domain peak spacing to the frequency-domain main sideband energy of the radio frequency signal, inputting the two into the cross-correlation algorithm for synchronization analysis, extracting the periodic characteristics of the time-domain peak sequence, and based on the differential operation result of the frequency-domain energy density gradient value and the time-domain peak spacing, extracting the synchronization characteristics through the Pearson correlation coefficient under the differential window offset, and generating an energy coupling factor; S2: By calculating the offset of the main frequency point within the current modulation period, calculating the harmonic drift amount based on the phase difference of the fundamental frequency component of the discrete Fourier transform, and combining the normalization processing result of the energy tilt vector, calling the covariance matrix to perform weighted fusion on the harmonic drift amount and the energy tilt degree, outputting a harmonic influence tendency parameter, and simultaneously generating a main sideband frequency trajectory vector sequence based on the floating position of the center frequency of the main sideband within consecutive periods; S3: Based on the energy coupling factor and the harmonic influence tendency parameter, performing a point-by-point multiplication operation on the two and arranging them in ascending order of sample number as a 64-dimensional vector, using the K-means clustering algorithm to classify the multiplicative fusion result of the two, dividing the harmonic perturbation state level, generating a matrix control item, and triggering a corresponding passband suppression behavior decision instruction.

[0007] As a further solution of the present invention, the energy coupling factor is specifically a gradient difference, a peak interval difference, and a synchronization factor, the harmonic influence tendency parameter includes a weighted offset, an energy tilt coefficient, and a covariance weight, the main sideband frequency trajectory vector sequence is specifically a center frequency coordinate, a migration rate, and a period index, the harmonic perturbation state level is specifically a high-interference overlapping state, a low-interference isolated state, and a medium-interference transition state, and the matrix control item includes a passband compression instruction, a stopband migration instruction, and a peak suppression instruction.

[0008] As a further solution of the present invention, the energy coupling factor is specifically a gradient difference, a peak interval difference, and a synchronization factor, the harmonic influence tendency parameter includes a weighted offset, an energy tilt coefficient, and a covariance weight, the main sideband frequency trajectory vector sequence is specifically a center frequency coordinate, a migration rate, and a period index, the harmonic perturbation state level is specifically a high-interference overlapping state, a low-interference isolated state, and a medium-interference transition state, and the matrix control item includes a passband compression instruction, a stopband migration instruction, and a peak suppression instruction.

[0009] As a further solution of the present invention, the energy coupling factor is specifically a gradient difference, a peak interval difference, and a synchronization factor. The harmonic influence tendency parameter includes a weighted offset, an energy tilt coefficient, and a covariance weight. The main sideband frequency trajectory vector sequence is specifically a center frequency coordinate, a migration rate, and a period index. The harmonic perturbation state level is specifically a high-interference overlapping state, a low-interference isolation state, and a medium-interference transition state. The matrix control term includes a passband compression instruction, a stopband migration instruction, and a peak suppression instruction.

[0010] As a further solution of the present invention, the steps for obtaining the main sideband frequency trajectory vector sequence are specifically as follows: S201: Based on the main frequency point data of the current modulation period, use the discrete Fourier transform to calculate the phase difference between the fundamental frequency component and the reference frequency, extract the fundamental frequency amplitude fluctuation amount, calculate the main frequency offset through linear interpolation, perform range normalization on the amplitude component of the energy tilt vector, compress the multi-components to the 0-1 interval, and generate a normalized energy tilt vector; S202: Map the main frequency offset to the harmonic drift amount dimension. Based on the diagonal element weight of the covariance matrix, use the product of the harmonic drift amount and the amplitude of the normalized energy tilt vector as the fusion factor, dynamically correct the fusion factor weight coefficient using the standard deviation, construct a linear combination of the harmonic drift amount and the energy tilt degree, and output the harmonic influence tendency parameter; S203: Extract the average value of the main sideband center frequencies of 10 consecutive modulation periods, use the harmonic influence tendency parameter as the correction factor to perform gain compensation on the difference between the average values of adjacent periods, perform trajectory fitting on the compensated discrete point sequence based on the cubic spline interpolation function, and set the first derivative of the cubic spline interpolation function to be continuous at adjacent nodes to generate the main sideband frequency trajectory vector sequence.

[0011] As a further solution of the present invention, the steps for obtaining the main sideband frequency trajectory vector sequence are specifically as follows: S201: Based on the main frequency point data of the current modulation period, use the discrete Fourier transform to calculate the phase difference between the fundamental frequency component and the reference frequency, extract the fundamental frequency amplitude fluctuation amount, calculate the main frequency offset through linear interpolation, perform range normalization on the amplitude component of the energy tilt vector, compress the multi-components to the 0-1 interval, and generate a normalized energy tilt vector; S202: Map the main frequency offset to the harmonic drift amount dimension. Based on the diagonal element weight of the covariance matrix, use the product of the harmonic drift amount and the amplitude of the normalized energy tilt vector as the fusion factor, dynamically correct the fusion factor weight coefficient using the standard deviation, construct a linear combination of the harmonic drift amount and the energy tilt degree, and output the harmonic influence tendency parameter; S203: Extract the average value of the center frequencies of the main sidebands in 10 consecutive modulation periods. Use the harmonic influence tendency parameter as a correction factor to perform gain compensation on the difference between the average values of adjacent periods. Perform trajectory fitting on the compensated discrete point sequence based on the cubic spline interpolation function, and set the first derivative of the cubic spline interpolation function to be continuous at adjacent nodes to generate the main sideband frequency trajectory vector sequence.

[0012] As a further solution of the present invention, the method further includes: S4: Through the main sideband frequency trajectory vector sequence, construct the difference between the frequency increments of adjacent periods into a frequency increasing amplitude sequence. Perform a first-order accumulation on the original sequence to generate a new sequence. Establish a grey differential equation between the accumulation sequence and the time variable to solve the trend deviation degree. Based on the prediction result and the passband suppression behavior decision instruction, perform a preloading operation on the stopband boundary parameter and the passband safety bandwidth.

[0013] As a further solution of the present invention, the frequency increasing amplitude sequence specifically includes positive increment, negative increment, and zero increment. The stopband boundary parameter specifically includes the upper cut-off frequency and the lower cut-off frequency. The passband safety bandwidth includes a high-frequency extension margin and a low-frequency buffer margin; The development coefficient and the grey action amount of the grey differential equation are solved by the least squares method, and the residual ratio threshold is set to 5%.

[0014] As a further solution of the present invention, the acquisition steps of S4 are specifically as follows: S401: Based on the main sideband frequency trajectory vector sequence, extract the frequency increments of the (n + 1)-th period and the n-th period within adjacent periods, calculate the absolute value of the difference between the two, arrange the difference sequence in chronological order, and divide each element in the difference sequence by the maximum value of the sequence to complete the normalization process to generate the frequency increment difference; S402: Invoke the grey prediction model. Use the frequency increment difference sequence as the input data, perform a first-order accumulation on the original sequence to generate a new sequence, establish a grey differential equation between the accumulation sequence and the time variable, solve the equation to obtain the development coefficient and the grey action amount, calculate the residual ratio between the predicted value and the original sequence, and output the change direction of the difference in the future period to obtain the trend deviation degree; S403: According to the positive or negative sign and the amplitude value of the trend deviation degree, combined with the band suppression intensity parameter in the passband suppression instruction matrix, calculate the step coefficient for the left or right shift of the stopband boundary, multiply the step coefficient by the passband safety bandwidth reference value, update the stopband boundary coordinates and the passband bandwidth value, and generate a preloading parameter group.

[0015] A radio frequency microwave signal modulation system, the radio frequency microwave signal modulation system is used to execute the above-mentioned radio frequency microwave signal modulation method, and the system includes: A time-frequency coupling analysis module, which is used to detect the ratio of the time-domain peak spacing to the frequency-domain main sideband energy of a radio frequency signal, input both of them into a cross-correlation algorithm for synchronization analysis, extract the periodic characteristics of the time-domain peak sequence, generate an energy coupling factor based on the differential operation result of the frequency-domain energy density gradient value and the time-domain peak spacing, and transfer the energy coupling factor to a perturbation classification control module; A harmonic trajectory modeling module, which is used to calculate the offset of the main frequency point within the current modulation period, combine the normalization processing result of the energy tilt vector, call a covariance matrix to perform weighted fusion on the harmonic drift amount and the energy tilt degree, output a harmonic influence tendency parameter, and simultaneously generate a main sideband frequency trajectory vector sequence based on the floating position of the main sideband center frequency within consecutive periods, transfer the harmonic influence tendency parameter to the perturbation classification control module, and transfer the main sideband frequency trajectory vector sequence to a frequency band pre-adjustment module; A perturbation classification control module, which is used to classify the multiplicative fusion result of the energy coupling factor and the harmonic influence tendency parameter by using a K-means clustering algorithm, divide the harmonic perturbation state level, generate a matrix control term and trigger a corresponding passband suppression behavior decision instruction, and transfer the matrix control term and the passband suppression behavior decision instruction to the frequency band pre-adjustment module; A frequency band pre-adjustment module, which is used to construct the difference between adjacent cycle frequency increments into a frequency increase amplitude sequence through the main sideband frequency trajectory vector sequence, input it into a grey prediction model for trend determination, and perform preloading operations on the stopband boundary parameters and the passband safety bandwidth based on the prediction result and the passband suppression behavior decision instruction.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In the present invention, through the cross-correlation analysis of the time-domain peak spacing and the frequency-domain main sideband energy ratio, the periodic characteristics are extracted and an energy coupling factor is generated, dynamically correlating the time-frequency domain parameters to solve the problem of signal synchronization deviation. The covariance matrix fusion of the main frequency point offset and the energy tilt vector generates a harmonic influence tendency parameter, quantifying the interference degree of harmonic drift on the main sideband energy distribution, and accurately positioning the evolution path of the interference source in combination with the main sideband frequency trajectory vector sequence. Based on the clustering algorithm, the energy coupling factor and the harmonic tendency parameter are classified to divide the perturbation level, realizing the adaptive matching of the interference suppression strategy. The grey prediction model is used to analyze the trend characteristics of the main sideband frequency increment difference sequence, predict the spectrum conflict direction and preload the stopband parameters, optimizing the passband suppression response speed. This method constructs a closed-loop control system for time-frequency joint analysis, harmonic dynamic modeling, interference state classification and frequency trend prediction, improving the anti-harmonic drift ability of the modulation signal in a complex interference environment and reducing the risk of spectrum resource conflict. Description of the Drawings

[0017] Figure 1Schematic diagram of the workflow of the present invention; Figure 2 Flowchart of the steps for obtaining the energy coupling factor of the present invention; Figure 3 Flowchart of the steps for obtaining the main sideband frequency trajectory vector sequence of the present invention; Figure 4 Flowchart of the steps for obtaining S3 of the present invention; Figure 5 Flowchart of the steps for obtaining S4 of the present invention. Detailed implementation manners

[0018] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0019] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. are based on the orientation or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention. In addition, in the description of the present invention, the meaning of "a plurality of" is two or more, unless otherwise specifically defined.

[0020] Embodiment 1: Please refer to Figure 1 , the present invention provides a technical solution: a radio frequency microwave signal modulation method, including the following steps: S1: By detecting the time-domain peak spacing and the frequency-domain main sideband energy ratio of the radio frequency signal, inputting the two into the cross-correlation algorithm for synchronization analysis, extracting the periodic characteristics of the time-domain peak sequence, and based on the differential operation result of the frequency-domain energy density gradient value and the time-domain peak spacing, extracting the synchronization characteristics through the Pearson correlation coefficient under the differential window offset, and generating an energy coupling factor; S2: By calculating the main frequency point offset within the current modulation period, calculating the harmonic drift amount based on the phase difference of the fundamental frequency components of the discrete Fourier transform, combining the normalization processing result of the energy tilt vector, calling the covariance matrix to perform weighted fusion on the harmonic drift amount and the energy tilt degree, outputting the harmonic influence tendency parameter, and at the same time generating the main sideband frequency trajectory vector sequence based on the floating position of the main sideband center frequency within consecutive periods; S3: Based on the energy coupling factor and the harmonic influence tendency parameter, perform a point-by-point multiplication operation on the two and arrange them in ascending order of sample numbers as a 64-dimensional vector. Use the K-means clustering algorithm to classify the multiplicative fusion result of the two, divide the harmonic disturbance state level, generate matrix control items, and trigger the corresponding passband suppression behavior decision instruction; S4: Through the main sideband frequency trajectory vector sequence, construct the difference in adjacent cycle frequency increments as a frequency increase amplitude sequence, perform a first-order accumulation on the original sequence to generate a new sequence, establish a grey differential equation between the accumulation sequence and the time variable to solve the trend deviation degree. Based on the prediction result and the passband suppression behavior decision instruction, perform a preloading operation on the stopband boundary parameter and the passband safety bandwidth.

[0021] The energy coupling factor specifically includes the gradient difference, peak interval difference, and synchronization factor. The harmonic influence tendency parameter includes the weighted offset, energy tilt coefficient, and covariance weight. The main sideband frequency trajectory vector sequence specifically includes the center frequency coordinate, migration rate, and period index. The harmonic disturbance state level specifically includes the high-interference overlap state, low-interference isolation state, and medium-interference transition state. The matrix control items include the passband compression instruction, stopband migration instruction, and peak suppression instruction. The frequency increase amplitude sequence specifically includes the positive increment, negative increment, and zero increment. The stopband boundary parameter specifically includes the upper cut-off frequency and the lower cut-off frequency. The passband safety bandwidth includes the high-frequency extension margin and the low-frequency buffer margin.

[0022] The differential window offset is the time-delay step amount of the sliding signal relative to the reference signal, and the step amount range is from 1 / 10 to 1 / 5 of the current modulation period; The harmonic drift is the frequency offset corresponding to the phase difference between the fundamental frequency component and the reference frequency. Extract the fundamental frequency amplitude fluctuation amount through the discrete Fourier transform and calculate it by linear interpolation; The development coefficient and grey action amount of the grey differential equation are solved by the least squares method, and the residual ratio threshold is set to 5%.

[0023] Please refer to Figure 2 , and the steps for obtaining the energy coupling factor are specifically as follows: S101: Detect the local amplitude extreme points of the RF signal time-domain waveform. Use the sliding window extreme value detection method to traverse the waveform data, mark the positions where the amplitude is higher than the adjacent sampling points before and after, record the time stamps corresponding to the extreme points, calculate the time intervals between adjacent extreme points, construct a time-domain peak interval sequence and normalize it to the same time axis, perform frequency point energy accumulation on the main sideband frequency range, and calculate the ratio with the total energy of the secondary sideband frequency points to generate the main sideband energy ratio; Detect the local amplitude extreme points of the radio frequency signal time-domain waveform. This process is achieved by traversing the collected waveform data using the sliding window extreme value detection method. Specifically, set a sliding window containing three consecutive sampling points. When the amplitude value of the window center point is greater than the amplitudes of its immediately adjacent previous and next sampling points, the center point is marked as a local amplitude maximum point, and the timestamp corresponding to this extreme point is recorded. For example, after sampling a radio frequency signal lasting 1000 nanoseconds (ns) at a rate of 10 Gs / s, 10000 amplitude data points are obtained. Traverse these data points to identify all local amplitude maximum points that meet the above conditions. Suppose 5 extreme points are identified within this 1000 ns segment, and their corresponding timestamps (unit: ns) are respectively: , , , , , Subsequently, calculate the time intervals between adjacent extreme points to obtain the sequence: ns, ns, ns, ns to construct the time-domain peak spacing sequence (unit: ns). To ensure the comparability of the peak spacing sequences obtained from different observation segments or different signal sources, it is necessary to normalize them to the same time axis. Here, normalization is performed by dividing each spacing value by the total duration of the observation segment (1000 ns) to obtain the normalized time-domain peak spacing sequence . Then, perform a Fourier transform on the original radio frequency signal to analyze its spectral characteristics. Set the center frequency of the signal to 2.45 GHz, and the main sideband range is defined as the center frequency , that is, from 2.44 GHz to 2.46 GHz. The setting of this range is based on the channel bandwidth standard of this wireless communication system. The secondary sideband range is defined as the frequency bands of 5 MHz on both sides of the main sideband, that is, from 2.435 GHz to 2.44 GHz and from 2.46 GHz to 2.465 GHz. Accumulate the signal energy (square of the amplitude) at all discrete frequency points within the main sideband range to obtain the total main sideband energy , for example, calculate to obtain watts (W). Similarly, accumulate the frequency point energies within the secondary sideband range to obtain the total secondary sideband energy , calculate to obtain W. Finally, calculate the ratio of the main sideband energy to the total secondary sideband energy to generate the main sideband energy ratio , and the calculation process is .

[0024] Table 1 Example of time-domain sampling, extreme points, and duration of radio frequency signals: ; As shown in Table 1, the sampled amplitudes of the radio frequency signals at specific timestamps are presented, and the identified local amplitude maximum points and their corresponding timestamps are marked. Meanwhile, the time intervals between adjacent extreme points calculated based on these timestamps are listed.

[0025] S102: Based on the ratio of the time-domain peak spacing sequence to the main sideband energy, normalize the sequence data to the same time axis. Using the main sideband energy ratio as the reference signal and the time-domain spacing sequence as the sliding signal, calculate the Pearson correlation coefficient of the two signals under different window offsets, extract the global maximum value of the correlation coefficient curve, and generate the synchronization correlation coefficient. Based on the normalized time-domain peak spacing sequence obtained in S101 and a series of main sideband energy ratios obtained in continuous observations. Here, the main sideband energy ratios are also calculated within the corresponding time periods of each element to form a sequence . The operation of normalizing these two sequence data with the same length to the same time axis has been completed in S101 by normalizing the time-domain peak spacing sequence. The main sideband energy ratio sequence itself is the value calculated within the corresponding normalized time period. Therefore, the two are already aligned in terms of time reference. Using the main sideband energy ratio sequence as the reference signal and the normalized time-domain peak spacing sequence as the sliding signal, calculate the Pearson correlation coefficient of the two signals under different window offsets. The different window offset is set to data points, which allows the sliding signal to move forward, not offset, or move backward by one data point relative to the reference signal for comparison. The calculation of the Pearson correlation coefficient invokes the standard formula. For the case where the offset is 0, , , calculate the mean value: ; ; Calculate sequence: ; Calculate sequence: ; ; ; ; ; For the offset of -1 ( shift one bit to the right, the number of comparison elements decreases) and +1 ( Similar calculations are also performed for a left shift by one bit. Assume that the calculated correlation coefficients are and , respectively. Then the correlation coefficient sequence is . Extract the global maximum value of the correlation coefficient curve from it. Here, the "global maximum value" refers to the maximum value in the algebraic sense. Therefore, is the largest, and this value is the generated synchronization correlation coefficient .

[0026] S103: Perform a point-by-point difference operation on the frequency-domain energy density distribution curve, calculate the energy difference between adjacent frequency points to generate an energy density gradient value, call the synchronization correlation coefficient, and input the absolute value and the energy density gradient value into a subtractor to calculate their difference, and superimpose the root mean square value of the time-domain peak distance sequence to generate an energy coupling factor.

[0027] For the frequency-domain energy density distribution curve, which is obtained from the result of the Fourier transform of the RF signal in S101 and represents the energy values at different frequency points, perform a point-by-point difference operation, calculate the energy difference between adjacent frequency points to generate an energy density gradient value. Assume that N equally spaced frequency points are selected near the main sideband, and their energy density values are , respectively. Each item in the energy density gradient sequence , for example, the obtained energy density gradient sequence (unit: W / Hz. For simplicity of representation, the unit is omitted in subsequent operations, but will be appended to the final result) is . Call the synchronization correlation coefficient generated in S102, calculate its absolute value , and perform a subtraction operation between this absolute value and each element in the energy density gradient value sequence. Specifically, ; ; ; ; ; ; Get the difference sequence . Next, calculate the root mean square value (RMS) of the original time-domain peak distance sequence ns constructed in S101; ns; To combine this value with the difference sequence (whose elements are derived from the energy density gradient, unit is For meaningful superposition (W / Hz and dimensionless correlation coefficient), it is necessary to perform a conversion or perform a conversion on the elements. Here, a conversion rule is set: in the evaluation of energy coupling, the contribution of time stability directly participates in the composition of the final factor with its root mean square value, while the influence of the difference sequence is reflected through its average value, and a weight is assigned to match its numerical range with that of time stability. Let be the average value of , and a weight coefficient is set. This coefficient is set based on system debugging experience and is used to balance the contributions of parameters from different sources. Then the energy coupling factor . Here, the unit of is ns, is an adjusted value, and finally is a comprehensive measure.

[0028] Please refer to Figure 3 . The specific steps for obtaining the main sideband frequency trajectory vector sequence are as follows: S201: Based on the main frequency point data of the current modulation period, use the discrete Fourier transform to calculate the phase difference between the fundamental frequency component and the reference frequency, extract the fundamental frequency amplitude fluctuation amount, calculate the main frequency offset through linear interpolation, perform range normalization on the amplitude component of the energy tilt vector, compress the multi-components to the 0-1 interval, and generate a normalized energy tilt vector; Based on the main frequency point data of the current modulation period, in a certain radar system, the main frequency point measured by a high-precision spectrum analyzer within a modulation period is , the nominal reference frequency of the system is . Use the discrete Fourier transform (DFT) to perform a fine analysis on the narrowband signal near the main frequency point, and calculate that the complex spectrum value at is , its phase is . At the same time, refer to the phase at of the reference frequency (or a preset reference phase) , calculate the phase difference radians, and extract the short-term fluctuation amount of the fundamental frequency amplitude. For example, continuously monitor the fundamental frequency amplitudes of 10 pulses, and its standard deviation is V (relative to the average amplitude of 1V). Calculate the main frequency offset through linear interpolation. Here, the calculation is based on the phase difference, and the phase-frequency conversion coefficient is determined by system design parameters (such as the VCO tuning sensitivity). Let / , then the main frequency offset , Next, process the energy tilt vector obtained from the spectrum analysis. This vector characterizes the asymmetry of the energy distribution within the main sideband, and its amplitude component is (unit: dB). Perform range normalization on the amplitude component of this vector. First, find the minimum value dB and the maximum value dB. For each component apply ; ; ; ; Generate a normalized energy tilt vector ; S202: Map the main frequency offset to the harmonic drift dimension. Based on the diagonal element weights of the covariance matrix, take the product of the harmonic drift and the amplitude of the normalized energy tilt vector as the fusion factor, and dynamically correct the fusion factor weight coefficient using the standard deviation. Construct a linear combination of the harmonic drift and the energy tilt, and output the harmonic influence tendency parameter; Map the main frequency offset obtained in S201 to the harmonic drift dimension. The harmonic drift is defined as the ratio of the main frequency offset to a reference bandwidth. The reference bandwidth is set to 1% of the system operating bandwidth. If the system operating bandwidth is 50 MHz, then . Therefore, the harmonic drift , which is a dimensionless relative value. Based on the diagonal elements of a pre-established covariance matrix, these elements reflect the importance of the historical fluctuations of the components of the normalized energy tilt vector. Let the weights corresponding to the three components be . These weights are obtained through principal component analysis or variance analysis of a large amount of historical monitoring data and sum to 1. Take the product of the harmonic drift and the weighted average of the normalized energy tilt vector as the initial fusion factor, , The initial fusion factor , and use the standard deviation within the last N (e.g., N = 20) modulation periods to dynamically correct the weight coefficient of this fusion factor. If the calculated , set a reference standard deviation . When the actual standard deviation is lower than the reference, the system is considered more stable and a higher weight is assigned; otherwise, a lower weight is assigned. The weight correction coefficient The calculation rule is as follows: , where is a regulation factor, set to 2.0, then , the corrected fusion factor , construct a linear combination of the harmonic drift amount and the energy tilt , where and are coefficients optimized according to expert experience and system simulation to ensure the balanced contribution of both to the final parameters. Set and , ([[]] ), , this is output as the harmonic influence tendency parameter.

[0029] S203: Extract the average value of the center frequencies of the main sidebands in 10 consecutive modulation periods. Use the harmonic influence tendency parameter as a correction factor to perform gain compensation on the difference between the average values of adjacent periods. Based on the cubic spline interpolation function, perform trajectory fitting on the compensated discrete point sequence, and set the first-order derivative of the cubic spline interpolation function to be continuous at adjacent nodes to generate the main sideband frequency trajectory vector sequence.

[0030] Extract the average value of the center frequencies of the main sidebands in 10 consecutive modulation periods to obtain the sequence (unit: MHz). Use the harmonic influence tendency parameter obtained in step S202 as a correction factor to perform gain compensation on the difference between the average values of adjacent periods, and calculate the difference in the average frequencies of adjacent periods , MHz, MHz... (calculate all 9 differences) The compensated difference , where the gain compensation coefficient is a small regulation value, set according to the sensitivity of the system to harmonic influence. Here, set , for example, MHz, MHz, and calculate the compensated difference sequence in this way, and then reconstruct the compensated average frequency sequence , where , , to obtain the sequence: , , … (and so on until ) Based on this sequence containing 10 time points and their corresponding compensated frequency values For the discrete point sequence, a cubic spline interpolation function is used for trajectory fitting. Between each pair of adjacent data points and a cubic polynomial is constructed, and it is set that the cubic spline interpolation function has continuous first and second derivatives at each internal node . By solving a system of linear equations, the coefficients of all polynomials are determined, and finally a piecewise continuous and smooth function is obtained as the main sideband frequency trajectory vector sequence.

[0031] Please refer to Figure 4 , and the specific steps for obtaining S3 are as follows: S301: Based on the energy coupling factor and the harmonic influence tendency parameter, perform a point-by-point multiplication operation on the two. Multiply the numerical value of the energy coupling factor of each sample point by the numerical value of the corresponding harmonic influence tendency parameter, and arrange the results of the multiplication operation in ascending order of sample numbers as a 64-dimensional vector to establish a parameter fusion data group and obtain the harmonic fusion coefficient; Based on the energy coupling factor obtained in S103 and the harmonic influence tendency parameter obtained in S202 , perform a point-by-point multiplication operation on these two parameters. In practical applications, the system will continuously monitor and calculate a series of such parameter pairs. Here, we consider the data obtained in 64 consecutive monitoring windows (sample points), so as to obtain 64 energy coupling factor values and 64 harmonic influence tendency parameter values , where . Table 2 shows an example of the energy coupling factor and the harmonic influence tendency parameter (the first 5 samples): ; As shown in Table 2, the energy coupling factor, the harmonic influence tendency parameter, and their products of the first 5 samples are listed. For each sample point , multiply its corresponding by to obtain the product value . For example, for the first sample point , after performing this multiplication operation on all 64 sample points, 64 product values are obtained. Arrange these product values in ascending order of sample numbers to form a 64-dimensional vector (i.e., a one-dimensional array containing 64 elements) , and this vector is the established parameter fusion data group, also known as the harmonic fusion coefficient.

[0032] S302: Invoke the K-means clustering algorithm, set the initial number of centroids to 3, use the harmonic fusion coefficient vector as the input data, initialize the centroids as 3 randomly selected sample vectors, calculate the Euclidean distances between the multiple sample vectors and the centroid vectors, classify the samples into the centroid cluster with the closest distance, update the centroids as the mean values of the samples within the cluster, iterate until the centroid offset is less than 0.1%, output the sample classification numbers, and generate the perturbed clustering labels. Invoke the K-means clustering algorithm, set the initial number of centroids to 3, and use the harmonic fusion coefficient sequence generated in step S301 as the input data (each is regarded as a one-dimensional data point), and the method of initializing the centroids is to randomly select 3 different values from as the initial centroids , , . For each data point in the dataset , calculate its Euclidean distances from the 3 centroids (in the one-dimensional case, it is the absolute value of the difference), and classify into the cluster where the centroid with the closest distance is located. For example, for , its distances from the three centroids are , , respectively. Since has the smallest distance from , is classified into cluster 1. After classifying all 64 data points, update the centroid of each cluster. The new centroid is the arithmetic mean of all data points within the cluster. This assignment and update process is iterated continuously, and the iteration stops when the offset of all centroids between two consecutive iterations is less than 0.1% of the centroid value, that is, for each centroid

[0033] S303: Invoke the K-means clustering algorithm, set the initial number of centroids to 3, use the harmonic fusion coefficient vector as the input data, initialize the centroids as 3 randomly selected sample vectors, calculate the Euclidean distances between the multiple sample vectors and the centroid vectors, classify the samples into the centroid cluster with the closest distance, update the centroids as the mean values of the samples within the cluster, iterate until the centroid offset is less than 0.1%, output the sample classification numbers, and generate the perturbed clustering labels.

[0034] The description of this step is exactly the same as that of S302. Its purpose is to perform K-means clustering to generate perturbed clustering labels. Therefore, for its specific execution process, data processing method, and calculation example, refer to the detailed description in S302. Similarly, using the harmonic fusion coefficient sequence generated by S301 as the input, set the number of clusters to 3. By randomly selecting the initial centroids and iteratively calculating distances, assigning samples, and updating the centroids until the convergence criterion of a centroid offset less than 0.1% is achieved, finally output the clustering labels of each sample.

[0035] Please refer to Figure 5 , and the specific steps for obtaining S4 are as follows: S401: Based on the main sideband frequency trajectory vector sequence, extract the frequency increment between the (n + 1)-th period and the n-th period within adjacent periods, calculate the absolute value of the difference between them, arrange the difference sequence in chronological order, divide each element in the difference sequence by the maximum value of the sequence to complete the normalization process, and generate the frequency increment difference; The description of this step is exactly the same as that of S302. Its purpose is to perform K-means clustering to generate perturbed clustering labels. Therefore, for its specific execution process, data processing method, and calculation example, refer to the detailed description in S302. Similarly, using the harmonic fusion coefficient sequence generated by S301 as the input, set the number of clusters to 3. By randomly selecting the initial centroids and iteratively calculating distances, assigning samples, and updating the centroids until the convergence criterion of a centroid offset less than 0.1% is achieved, finally output the clustering labels of each sample.

[0036] S402: Call the grey prediction model. Using the frequency increment difference sequence as the input data, perform a first-order accumulation on the original sequence to generate a new sequence, establish a grey differential equation between the accumulated sequence and the time variable, solve the equation to obtain the development coefficient and grey action quantity, calculate the residual ratio between the predicted value and the original sequence, output the change direction of the difference in the future period, and obtain the trend offset degree; Call the GM(1,1) grey prediction model, using the frequency increment difference sequence generated by S401 as the original sequence , where , perform a first-order accumulation on to generate the (1-AGO) sequence : ; ; ; ; That is , establish the grey differential equation , where ; ; ; ; Solve the development coefficient by the least squares method and the grey action quantity , construct the matrices and , calculate , ; ; ; ; Therefore , , the time response sequence is , predict the value of the first future period , where . ; ; This predicted value is the obtained trend deviation degree

[0037] S403: According to the positive or negative sign and amplitude value of the trend deviation degree, combined with the frequency band suppression intensity parameter in the passband suppression instruction matrix, calculate the step coefficient for the left or right shift of the stopband boundary, multiply the step coefficient by the passband safety bandwidth reference value, update the stopband boundary coordinates and passband bandwidth values, and generate a preloaded parameter group

[0038] According to the trend deviation degree obtained in S402 , which is positive with an amplitude of 0.9211, combined with the frequency band suppression intensity parameter in a passband suppression instruction matrix (shown in Table 3 below), Table 3 Passband Suppression Instruction Reference Table: ; As shown in Table 3, this table sets corresponding frequency band suppression intensity parameters according to different intervals of the trend deviation degree. Since falls within interval, look up the table to obtain the frequency band suppression intensity parameter , according to and the sign (positive) and amplitude of the trend deviation degree, calculate the step coefficient for adjusting the stopband boundary, set , the passband safety bandwidth reference value is set to according to the system requirements, then the specific adjustment step , a positive trend offset indicates the possibility of the passband extending towards higher frequencies or being unstable. It is necessary to further shift the stopband boundary on the high-frequency side to the right (increase its starting frequency) to ensure the passband performance. If the current lower boundary of the passband , the upper boundary of the passband , then the passband bandwidth , the start of the stopband on the high-frequency side is at , the new lower boundary of the stopband on the high-frequency side . At the same time, to maintain the communication quality, the effective bandwidth value of the passband is also updated. The new upper boundary of the passband , the new passband bandwidth . These updated stopband boundary coordinates (such as the new starting frequency of the stopband on the high-frequency side ) and the updated values of the upper boundary and bandwidth of the passband ( and ) together constitute the generated preloading parameter group.

[0039] A radio frequency and microwave signal modulation system. A radio frequency and microwave signal modulation system is used to execute the above-mentioned radio frequency and microwave signal modulation method. The system includes: A time-frequency coupling analysis module, which is used to detect the ratio of the time-domain peak spacing to the frequency-domain main sideband energy of the radio frequency signal, input both into the cross-correlation algorithm for synchronization analysis, extract the periodic characteristics of the time-domain peak sequence, generate an energy coupling factor based on the differential operation result of the frequency-domain energy density gradient value and the time-domain peak spacing, and transfer the energy coupling factor to the perturbation hierarchical control module; A harmonic trajectory modeling module, which is used to calculate the offset of the main frequency point within the current modulation period, combine the normalized processing result of the energy tilt vector, call the covariance matrix to perform weighted fusion on the harmonic drift amount and the energy tilt degree, output the harmonic influence tendency parameter, and at the same time generate a main sideband frequency trajectory vector sequence based on the floating position of the main sideband center frequency within consecutive periods, transfer the harmonic influence tendency parameter to the perturbation hierarchical control module, and transfer the main sideband frequency trajectory vector sequence to the frequency band pre-adjustment module; A perturbation hierarchical control module, which is used to classify the multiplicative fusion result of the energy coupling factor and the harmonic influence tendency parameter by using the K-means clustering algorithm, divide the harmonic perturbation state level, generate a matrix control term and trigger the corresponding passband suppression behavior decision instruction, and transfer the matrix control term and the passband suppression behavior decision instruction to the frequency band pre-adjustment module; A frequency band pre-adjustment module, which is used to construct the difference between the frequency increments of adjacent periods into a frequency increase amplitude sequence through the main sideband frequency trajectory vector sequence, input it into the grey prediction model for trend determination, and perform preloading operations on the stopband boundary parameters and the passband safety bandwidth based on the prediction result and the passband suppression behavior decision instruction.

[0040] The above are only the preferred embodiments of the present invention and do not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A method for modulating radio frequency and microwave signals, characterized in that It includes the following steps: S1: By detecting the time-domain peak spacing and the frequency-domain main sideband energy ratio of the radio frequency signal, inputting the two into the cross-correlation algorithm for synchronization analysis, extracting the periodic characteristics of the time-domain peak sequence, based on the differential operation result of the frequency-domain energy density gradient value and the time-domain peak spacing, extracting the synchronization characteristics through the Pearson correlation coefficient under the differential window offset, and generating an energy coupling factor; S2: By calculating the main frequency point offset within the current modulation period, calculating the harmonic drift amount based on the phase difference of the fundamental frequency component of the discrete Fourier transform, combining the normalization processing result of the energy tilt vector, calling the covariance matrix to perform weighted fusion on the harmonic drift amount and the energy tilt degree, outputting the harmonic influence tendency parameter, and simultaneously generating a main sideband frequency trajectory vector sequence based on the floating position of the main sideband center frequency within consecutive periods; S3: Based on the energy coupling factor and the harmonic influence tendency parameter, performing a point-by-point multiplication operation on the two and arranging them in ascending order of sample number as a 64-dimensional vector, using the K-means clustering algorithm to classify the multiplicative fusion result of the two, dividing the harmonic perturbation state level, generating a matrix control item and triggering a corresponding passband suppression behavior decision instruction.

2. The radio frequency microwave signal modulation method according to claim 1, characterized in that, The energy coupling factor is specifically the gradient difference value, the peak interval difference value, and the synchronization factor. The harmonic influence tendency parameter includes the weighted offset, the energy tilt coefficient, and the covariance weight. The main sideband frequency trajectory vector sequence is specifically the center frequency coordinate, the migration rate, and the period index. The harmonic perturbation state level is specifically the high-interference overlapping state, the low-interference isolated state, and the medium-interference transition state. The matrix control item includes the passband compression instruction, the stopband migration instruction, and the peak suppression instruction.

3. The radio frequency microwave signal modulation method according to claim 2, wherein The differential window offset is the time delay step amount of the sliding signal relative to the reference signal, and the step amount range is from 1 / 10 to 1 / 5 of the current modulation period; The harmonic drift amount is the frequency offset corresponding to the phase difference between the fundamental frequency component and the reference frequency, and is obtained by extracting the fundamental frequency amplitude fluctuation amount through the discrete Fourier transform and calculating it with linear interpolation.

4. A radio frequency microwave signal modulation method according to claim 3, characterized in that The specific steps for obtaining the energy coupling factor are as follows: S101: Detect the local amplitude extreme points of the time-domain waveform of the radio frequency signal, traverse the waveform data using the sliding window extreme value detection method, mark the positions where the amplitude is higher than the adjacent sampling points before and after, record the time stamps corresponding to the extreme points, calculate the time intervals between adjacent extreme points, construct a time-domain peak spacing sequence and normalize it to the same time axis, accumulate the frequency point energies in the main sideband range of the frequency domain, and calculate the ratio with the total energy of the sub-sideband frequency points to generate the main sideband energy ratio; S102: Based on the time-domain peak spacing sequence and the main sideband energy ratio, normalize the sequence data to the same time axis, use the main sideband energy ratio as the reference signal and the time-domain spacing sequence as the sliding signal, calculate the Pearson correlation coefficient of the two signals under the differential window offset, extract the global maximum value of the correlation coefficient curve, and generate the synchronization correlation coefficient; S103: Perform point-by-point difference operation on the frequency-domain energy density distribution curve, calculate the energy difference between adjacent frequency points to generate the energy density gradient value, call the synchronization correlation coefficient, input the absolute value and the energy density gradient value into a subtractor to calculate the difference between the two, and superimpose the root mean square value of the time-domain peak spacing sequence to generate the energy coupling factor.

5. A radio frequency microwave signal modulation method according to claim 4, characterized in that The specific steps for obtaining the main sideband frequency trajectory vector sequence are as follows: S201: Based on the main frequency point data of the current modulation period, use the discrete Fourier transform to calculate the phase difference between the fundamental frequency component and the reference frequency, extract the fundamental frequency amplitude fluctuation amount, calculate the main frequency offset through linear interpolation, perform range normalization on the amplitude component of the energy tilt vector, compress multiple components to the 0-1 interval, and generate the normalized energy tilt vector. S202: Map the main frequency offset to the harmonic drift amount dimension, use the product of the harmonic drift amount and the amplitude of the normalized energy tilt vector as the fusion factor based on the diagonal element weights of the covariance matrix, dynamically correct the fusion factor weight coefficient using the standard deviation, construct a linear combination of the harmonic drift amount and the energy tilt degree, and output the harmonic influence tendency parameter. S203: Extract the average value of the main sideband center frequencies of 10 consecutive modulation periods, use the harmonic influence tendency parameter as the correction factor to perform gain compensation on the difference between the average values of adjacent periods, perform trajectory fitting on the compensated discrete point sequence based on the cubic spline interpolation function, and set the first derivative of the cubic spline interpolation function to be continuous at adjacent nodes to generate the main sideband frequency trajectory vector sequence.

6. The radio frequency microwave signal modulation method according to claim 5, wherein The specific steps for obtaining S3 are as follows: S301: Based on the energy coupling factor and the harmonic influence tendency parameter, perform point-by-point multiplication operation on them, multiply the energy coupling factor value of each sample point by the corresponding harmonic influence tendency parameter value, arrange the multiplication operation results in ascending order of sample numbers as a 64-dimensional vector, establish a parameter fusion data group, and obtain the harmonic fusion coefficient. S302: Call the K-means clustering algorithm, set the initial number of centroids to 3, use the harmonic fusion coefficient vector as the input data, initialize the centroids as 3 randomly selected sample vectors, calculate the Euclidean distances between multiple sample vectors and the centroid vectors, classify the samples into the centroid cluster with the closest distance, update the centroids as the mean values of the samples within the cluster, iterate until the centroid offset is less than 0.1%, output the sample classification numbers, and generate the perturbation clustering labels. S303: Call the K-means clustering algorithm, set the initial number of centroids to 3, use the harmonic fusion coefficient vector as the input data, initialize the centroids as 3 randomly selected sample vectors, calculate the Euclidean distances between multiple sample vectors and the centroid vectors, classify the samples into the centroid cluster with the closest distance, update the centroids as the mean values of the samples within the cluster, iterate until the centroid offset is less than 0.1%, output the sample classification numbers, and generate the perturbation clustering labels.

7. A radio frequency microwave signal modulation method according to claim 6, characterized in that, The method further includes: S4: Based on the main sideband frequency trajectory vector sequence, construct the difference between adjacent cycle frequency increments into a frequency increasing amplitude sequence, perform first-order accumulation on the original sequence to generate a new sequence, establish a grey differential equation between the accumulation sequence and the time variable to solve the trend deviation degree, and perform preloading operations on the stopband boundary parameters and the passband safety bandwidth based on the prediction result and the passband suppression behavior decision instruction.

8. A radio frequency microwave signal modulation method according to claim 7, characterized in that, The frequency increasing amplitude sequence specifically includes positive increments, negative increments, and zero increments. The stopband boundary parameters specifically include the upper cut-off frequency and the lower cut-off frequency. The passband safety bandwidth includes a high-frequency extension margin and a low-frequency buffer margin. The development coefficient and grey action quantity of the grey differential equation are solved by the least squares method, and the residual ratio threshold is set to 5%.

9. A radio frequency microwave signal modulation method according to claim 8, characterized in that, The acquisition steps of S4 are specifically as follows: S401: Based on the main sideband frequency trajectory vector sequence, extract the frequency increments of the (n + 1)-th cycle and the n-th cycle within adjacent cycles, calculate the absolute value of the difference between the two, arrange the difference sequence in chronological order, divide each element in the difference sequence by the maximum value of the sequence to complete the normalization process, and generate a frequency increasing difference. S402: Invoke the grey prediction model, use the frequency increasing difference sequence as the input data, perform first-order accumulation on the original sequence to generate a new sequence, establish a grey differential equation between the accumulation sequence and the time variable, solve the equation to obtain the development coefficient and grey action quantity, calculate the residual ratio between the predicted value and the original sequence, and output the change direction of the difference in the future cycle to obtain the trend deviation degree. S403: According to the positive / negative sign and amplitude value of the trend deviation degree, combined with the frequency band suppression intensity parameters in the passband suppression instruction matrix, calculate the step coefficient for the left or right shift of the stopband boundary, multiply the step coefficient by the passband safety bandwidth reference value, update the stopband boundary coordinates and the passband bandwidth value, and generate a preloading parameter group.

10. A radio frequency microwave signal modulation system, characterized in that, According to the radio frequency and microwave signal modulation method according to any one of claims 1 - 9, the system includes: A time-frequency coupling analysis module, configured to synchronously analyze the two by detecting the time-domain peak spacing and the frequency-domain main sideband energy ratio of a radio frequency signal, input them into a cross-correlation algorithm, extract the periodic characteristics of the time-domain peak sequence, generate an energy coupling factor based on the differential operation result of the frequency-domain energy density gradient value and the time-domain peak spacing, and transfer the energy coupling factor to the disturbance hierarchical control module. A harmonic trajectory modeling module, configured to calculate the main frequency point offset within the current modulation cycle, combine the normalization processing result of the energy tilt vector, call the covariance matrix to perform weighted fusion on the harmonic drift amount and the energy tilt degree, output a harmonic influence tendency parameter, and at the same time generate a main sideband frequency trajectory vector sequence based on the floating position of the main sideband center frequency within consecutive cycles, transfer the harmonic influence tendency parameter to the disturbance hierarchical control module, and transfer the main sideband frequency trajectory vector sequence to the frequency band pre-adjustment module. The disturbance hierarchical control module is used to classify the multiplicative fusion result of the two by using the K-means clustering algorithm through the energy coupling factor and the harmonic influence tendency parameter, divide the harmonic disturbance state level, generate a matrix control term and trigger a corresponding passband suppression behavior decision instruction, and transmit the matrix control term and the passband suppression behavior decision instruction to the frequency band pre-adjustment module; The frequency band pre-adjustment module is used to construct the adjacent cycle frequency increment difference into a frequency increasing amplitude sequence through the main sideband frequency trajectory vector sequence, input it into the grey prediction model for trend determination, and perform preloading operations on the stopband boundary parameter and the passband safety bandwidth based on the prediction result and the passband suppression behavior decision instruction.

Images