Microwave signal processing method based on grating sensor composite channel

Abstract

The invention relates to the technical field, in particular to a grating sensor composite channel-based microwave signal processing method, which comprises the following steps of: capturing time sequence characteristics in a signal through a long-short-term memory network, accurately extracting each frequency component in the microwave signal during frequency spectrum calculation and time domain conversion; the defect that complex signal features cannot be effectively extracted in a traditional method is avoided, the signal is converted from a time domain to a frequency domain, frequency components are enhanced, the frequency, the amplitude and the phase of the signal can be finely adjusted stage by stage in time through a multi-stage recursive correction technology, the result after correction each time can be fed back to the next stage, and the accuracy of signal extraction is improved. According to the method, the quality of microwave signals is gradually improved, signal loss and interference can be remarkably reduced, unnecessary interference and errors in the signal transmission process are remarkably reduced through the intelligent optimization and dynamic correction mode, and the high efficiency and stability of the signal processing process are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing, and particularly to a microwave signal processing method based on a composite channel of a grating sensor. Background Art

[0002] In the technical field of signal processing, through the comprehensive application of mathematics, algorithms and electronic technology, operations such as filtering, denoising, modulation, demodulation, enhancement and spectrum analysis are performed on analog or digital signals from different sources to extract or restore the effective signal content carrying information.

[0003] A method that combines the signals sensed by a grating sensor with a microwave signal processing method. By constructing a composite channel and using the optical sensitivity characteristics of the grating sensor to sensitively capture and convert the changes in the microwave signal, high-precision optimization processing of the microwave signal is realized, aiming to improve the transmission accuracy of the microwave signal, reduce signal loss and interference, and ensure the stable and reliable transmission effect of the signal in a complex electromagnetic environment.

[0004] Traditional methods often rely on simple frequency-domain and time-domain processing. The processing process is single and lacks a fine-tuning mechanism. In a complex electromagnetic environment, signal interference is serious. Moreover, traditional methods usually use fixed frequency thresholds for processing, resulting in the inability to effectively eliminate noise and interference, the loss of important signal components, and the lack of a multi-stage recursive correction mechanism, making it impossible to finely adjust signal parameters according to environmental changes during the processing process, resulting in unstable signal quality, low processing efficiency, and affecting the reliability and stability of the system. Summary of the Invention

[0005] The purpose of the present invention is to provide a microwave signal processing method based on a composite channel of a grating sensor to solve the above problems.

[0006] The present invention is realized through the following technical solutions: A microwave signal processing method based on a composite channel of a grating sensor includes: Step 1: Based on the microwave signal obtained by the grating sensor, signal screening is performed. After the signal is low-pass filtered to remove noise with a frequency higher than the set threshold, the signal amplitude is normalized, and the signal is standardized through amplitude correction to generate an effective signal. Step 2: Based on the effective signal, a long short-term memory network is used to calculate the spectrum of the signal. At the same time, the signal is converted from the time domain to the frequency domain, each frequency component is extracted, and the main frequency range in the signal is determined. The frequency band is selected according to the component intensity and noise interference, and the amplitude of the frequency band is enhanced to generate an enhanced spectrum signal. Step 3: Based on the enhanced spectrum signal, adjust the frequency, amplitude, and phase of the signal stage by stage, perform multiple recursive corrections, detect the signal effect after each correction, adjust each parameter, further optimize the correction effect of each stage, and generate a recursively corrected signal; Step 4: Based on the recursively corrected signal, extract and analyze the time-domain characteristics of the signal, identify the interference sources in the signal, determine the frequency, intensity, and duration characteristics of the interference by comparing the signal change patterns frame by frame, and generate an interference recognition result; Step 5: Based on the interference recognition result, select an appropriate frequency adjustment rule, correct the gain control and frequency of the signal, and optimize the signal according to phase compensation and intensity compensation to generate an optimized signal.

[0007] As a further solution of the present invention, the specific steps for generating the effective signal are as follows: Perform low-pass filtering on the microwave signal obtained by the grating sensor, set a frequency threshold, and screen out the data points higher than the threshold. At the same time, set the amplitude to zero to remove high-frequency noise and generate a filtered signal; Based on the filtered signal, normalize the signal amplitude, calculate the limit amplitude value in the signal, divide the amplitude of each data point by the extreme value, and adjust the signal amplitude to a unified range to generate a normalized signal; Based on the normalized signal, correct the signal amplitude, traverse each signal point, compare it with the preset standard range, and at the same time adjust the signal amplitude to meet the set standard to generate an effective signal.

[0008] As a further solution of the present invention, the specific steps for generating the strong spectrum signal are as follows: Based on the effective signal, use a long short-term memory network. First, segment the signal according to a set time window, calculate the frequency components within each time window, weight each time period using a window function, and generate a frequency-domain signal using the frequency and amplitude information within each time window; Based on the frequency-domain signal, divide the spectrum into multiple frequency bands according to the frequency range, select the frequency bands that affect the transmission quality, calculate the boundaries of each frequency band, and at the same time extract all the frequency components within the frequency band to generate a key frequency band signal; Based on the key frequency band signal, perform gain adjustment on the signal of each selected frequency band. First, calculate the gain coefficient of the frequency band and apply it to the amplitudes of all frequency points within the frequency band to enhance the signal intensity of the frequency band and generate an enhanced spectrum signal.

[0009] As a further solution of the present invention, the long short-term memory network is calculated according to the formula: ; Where: represents a frequency-domain signal, is the frequency, is the time window, is the index of the time-domain signal, represents the th amplitude of the sampling point of the time-domain signal, is an integer index, represents the window function weighting coefficient, is the complex exponential term, is the base of the natural logarithm, is the imaginary unit, represents the frequency component in signal analysis, is the index of the time-domain signal, is the total number of sampling points of the signal.

[0010] Generating the key frequency band signal: The frequency-domain signal is divided into multiple sub-frequency bands according to the set frequency range, and then based on the correlation between the power spectral density, energy concentration, and transmission performance index within the frequency band, the frequency bands that affect the transmission quality are identified and defined as key frequency bands. For the key frequency bands, the complex spectral values of the frequency points within the frequency band are extracted, and the spectral components of the non-key frequency bands are set to zero, thus forming a spectral signal containing the frequency components of the key frequency bands, which is the key frequency band signal for subsequent gain adjustment or noise suppression processing.

[0011] As a further solution of the present invention, the specific steps for generating the recursive correction signal are as follows: Based on the enhanced spectral signal, first detect the signal frequency, obtain the actual frequency value of each frequency band, compare it with the target frequency, calculate the frequency error, and gradually adjust the frequency of each frequency band to generate a corrected frequency signal; Based on the corrected frequency signal, adjust the amplitude of each frequency band, measure the amplitude change of each frequency band, calculate the difference and adjust the amplitude of the corresponding frequency band, apply the correction amount to each frequency point, correct the signal strength of the frequency band, and generate a corrected amplitude signal; Based on the corrected amplitude signal, optimize the phase of the signal, detect the phase deviation, obtain the required phase adjustment amount, and then correct the phase of each signal according to the frequency component. After the phase adjustment is completed, a recursive correction signal is generated.

[0012] Correcting the phase according to the frequency component: Extract the current phase value of the frequency point according to the frequency component, compare it with the reference phase and the expected phase to obtain the phase error, calculate the correction factor according to this error, multiply the frequency component by the complex exponential factor in the frequency domain to achieve phase rotation compensation, and at the same time, it is necessary to maintain the consistency of the frequency-phase mapping relationship. After the phase correction, the spectral structure of the signal will not be damaged, and a phase-aligned frequency-domain signal is generated for subsequent time-domain reconstruction and control system input.

[0013] As a further solution of the present invention, the specific steps for generating the interference recognition result are as follows: Based on the recursive correction signal, time-domain feature extraction is performed. First, the signal is framed, and for each frame, the average amplitude, frequency change, signal peak value, and change rate information are extracted. At the same time, the change trend of each frame of the signal is compared to generate a time-domain feature signal; Based on the time-domain feature signal, a preliminary screening of the interference sources in the signal is performed. The abnormal fluctuations of each frame of the signal are monitored, a threshold is set and compared with each frame to identify the interference sources and generate an interference source candidate signal; Based on the interference source candidate signal, the interference sources are analyzed, the frequency range, amplitude, and duration of the interference signal are calculated, and the characteristics are analyzed at the same time to generate an interference recognition result.

[0014] Analyze the interference source for the candidate signal segment, extract its main frequency components, calculate the frequency interval with dense energy distribution as the frequency range of the interference signal, select the average power as the interference amplitude index to measure the amplitude value in this interval, calculate the duration of the interference signal according to the number of frames within the signal duration and the time length of each frame, and further extract the feature vectors of the interference signal, including frequency stability, amplitude volatility, and modulation characteristics, to provide a basis for subsequent classification recognition and interference suppression strategy generation. The formed analysis result should include the interference frequency range, maximum amplitude, duration, and additional features, and the generated result supports the interference type judgment and system response decision.

[0015] As a further solution of the present invention, the specific steps for generating the optimized signal are as follows: Based on the interference recognition result, a gain control rule is selected, the gain requirements for each frequency band are calculated, the signal strength of each frequency band is measured, and then the amplitude of the corresponding frequency band is adjusted according to the required gain value to enhance the target frequency band and generate a gain control signal; Based on the gain control signal, frequency correction is performed, the frequency deviation of the signal is measured, and the frequency of each frequency band is adjusted one by one using the correction amount to generate a corrected frequency signal; Based on the corrected frequency signal, phase compensation is performed on the signal, the phase deviation of each frequency band is detected, the required adjustment amount is calculated, and the phase compensation is applied to each frequency component. The signal strength is adjusted according to the amplitude correction to generate an optimized signal.

[0016] Compared with the prior art, the present invention has the following advantages and beneficial effects: 1. In the present invention, the long short-term memory network is used to capture the timing characteristics in the signal. When calculating the spectrum and performing time-domain conversion, each frequency component in the microwave signal is accurately extracted, avoiding the defect that complex signal features cannot be effectively extracted in the traditional method; 2. In the present invention, by converting the signal from the time domain to the frequency domain and enhancing the frequency components, the multi-stage recursive correction technology can finely adjust the frequency, amplitude, and phase of the signal stage by stage. The result after each correction will be fed back to the next stage, gradually improving the quality of the microwave signal, and significantly reducing signal loss and interference. 3. In the present invention, through this intelligent optimization and dynamic correction method, the quality of the microwave signal during transmission has been greatly improved. It can identify and adapt to interference patterns in real time in an environment with multiple interference sources, which significantly reduces unnecessary interference and errors during signal transmission, and enables the efficiency and stability of the signal processing process. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, form a part of this application, and do not limit the embodiments of the present invention. In the drawings: Figure 1 It is a schematic diagram of the main steps of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] 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 in combination with embodiments and drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and do not limit the present invention. It should be noted that the present invention has been in the actual R & D and use stage.

[0019] Embodiment 1 Please refer to Figure 1 , the present invention provides a technical solution, a microwave signal processing method based on a grating sensor composite channel, including the following steps: Step 1: Based on the microwave signal obtained by the grating sensor, signal screening is performed. After the signal is low-pass filtered to remove noise with a frequency higher than the set threshold, the signal amplitude is normalized, and the signal is standardized through amplitude correction to generate an effective signal. Step 2: Based on the effective signal, the long short-term memory network is used to calculate the spectrum of the signal. At the same time, the signal is converted from the time domain to the frequency domain, each frequency component is extracted, and the main frequency range in the signal is determined. The frequency band is selected according to the component intensity and noise interference, and the amplitude of the frequency band is enhanced to generate an enhanced spectrum signal. Step 3: Based on the enhanced spectrum signal, the frequency, amplitude, and phase of the signal are adjusted stage by stage, and multiple recursive corrections are performed. The effect of the signal after each correction is detected and each parameter is adjusted to further optimize the correction effect of each stage to generate a recursively corrected signal. Step 4: Based on the recursively corrected signal, extract and analyze the time-domain features of the signal, identify the interference sources in the signal, determine the frequency, intensity, and duration characteristics of the interference by comparing the signal change patterns frame by frame, and generate an interference recognition result; Step 5: Based on the interference recognition result, select an appropriate frequency adjustment rule, correct the gain control and frequency of the signal, and optimize the signal according to phase compensation and intensity compensation to generate an optimized signal.

[0020] The specific steps to generate a valid signal are as follows: Based on the microwave signal obtained by the grating sensor, perform low-pass filtering, set a frequency threshold and filter out the data points higher than the threshold. At the same time, set the amplitude to zero to remove high-frequency noise and generate a filtered signal; Based on the filtered signal, normalize the signal amplitude, calculate the extreme amplitude value in the signal, divide the amplitude of each data point by the extreme value, and adjust the signal amplitude to a unified range to generate a normalized signal; Based on the normalized signal, correct the signal amplitude, traverse each signal point, compare it with the preset standard range, and at the same time adjust the signal amplitude to meet the set standard to generate a valid signal; Based on the microwave signal obtained by the grating sensor, use a low-pass filtering algorithm, set a frequency threshold, and perform low-pass filtering operations using the butter filter in the scipy library in Python. Set the frequency threshold to 5 GHz, generate the filter coefficients through the butter function, and apply the filtfilt function to filter the original signal to remove the frequency components higher than 5 GHz and generate a filtered signal. During the filtering process, compare each signal point with the set threshold in the frequency domain and set the amplitude of all frequency components higher than the threshold to zero; Based on the filtered signal, use a normalization algorithm, use the np.max function and np.min function in the NumPy library to calculate the extreme amplitude value of the signal, obtain the range of the signal amplitude by calculating the maximum and minimum values of the signal, and use np.divide to divide the amplitude value of each signal data point by the maximum amplitude value of the signal to adjust the signal amplitude to a unified range of 0 to 1 to generate a normalized signal; Based on the normalized signal, use an amplitude correction algorithm. First, through the set standard amplitude range, which is 0.2 to 0.8, use the if conditional statement in Python to traverse each signal data point to determine whether it falls within the set standard range. If not, correct the amplitude by linear interpolation. Use the interpolate.interp1d function in the scipy library to create an interpolation function and adjust the amplitude value of each data point to the preset range to generate a valid signal.

[0021] The specific steps to generate a strong spectrum signal are as follows: Based on the effective signal, using a long short-term memory network, first segment the signal according to a set time window, calculate the frequency components within each time window, weight each time period using a window function, and generate a frequency-domain signal using the frequency and amplitude information within each time window. Based on the frequency-domain signal, divide the spectrum into multiple frequency bands according to the frequency range, select the frequency bands that affect the transmission quality, calculate the boundaries of each frequency band, and at the same time extract all the frequency components within the frequency band to generate a key frequency band signal. Based on the key frequency band signal, perform gain adjustment on the signal of each selected frequency band. First, calculate the gain coefficient of the frequency band and apply it to the amplitudes of all frequency points within the frequency band to enhance the signal strength of the frequency band and generate an enhanced spectrum signal. Based on the effective signal, using a long short-term memory network, build a long short-term memory network model through the TensorFlow framework and the Keras library. Set the time window length to 50 ms, use the TimeDistributed layer to segment the input signal into time series, calculate the frequency components within each time period using the long short-term memory network layer, and perform weighted processing on each time period using the Hamming window function. The parameters of the window function are set as alpha = 0.54 and beta = 0.46 to generate a frequency-domain signal. By performing Fourier transform on the spectrum of each time window, extract the frequency and amplitude information within each time period to generate the frequency-domain data of each time window. Based on the frequency-domain signal, using a frequency division algorithm, set the frequency range from 0 Hz to 5 GHz. Use the fft.fftfreq function of the numpy library to segment the frequency-domain signal. Set the boundary of each frequency band to 1 GHz. Select the 0.5 GHz to 2 GHz frequency band that affects the transmission quality within the frequency range. Use the fft.fft function of numpy to extract all the frequency components within each frequency band, filter the signal of each frequency band according to the set frequency interval to generate a key frequency band signal, and screen and process the frequency components within each frequency band according to their influence to make the key frequency band signal extraction complete. Based on the key frequency band signal, using a gain adjustment algorithm, first calculate the gain coefficient of the frequency band. Calculate the amplitude average value of the selected frequency band through the np.mean function of numpy. Set the gain coefficient to 1.5. Apply the gain coefficient to the amplitude of each frequency point within the frequency band through the np.multiply function of numpy. Use the signal.resample function of the scipy library to enhance the amplitude of the frequency components and enhance the signal strength of the frequency band to generate an enhanced spectrum signal.

[0022] The long short-term memory network, according to the formula: ; Wherein: represents the frequency-domain signal, is the frequency, is the time window, is the index of the time-domain signal, represents the th amplitude of the sampling point of the time-domain signal, which is an integer index, represents the window function weighting coefficient, is the complex exponential term, is the base of the natural logarithm, is the imaginary unit, represents the frequency component in signal analysis, is the index of the time-domain signal, is the total number of sampling points of the signal; Execution process: Collect the microwave signal through the grating sensor to obtain the time-domain signal , after weighting by the window function , calculate the frequency-domain signal , which represents the complex spectrum of the signal at time and frequency , and at the same time perform the time-domain to frequency-domain conversion on the frequency basis function to obtain the frequency component at each moment and frequency point. After obtaining the frequency-domain signal, based on the spectrum , perform gain adjustment on the signal. The gain coefficient is dynamically calculated by analyzing each frequency band of the signal spectrum, adjusting the intensity of the frequency band and generating the enhanced spectrum , extract the phase information of each frequency band , compare it with the target phase, calculate the phase deviation, and at the same time calculate the phase adjustment amount of each frequency band, and add it to the original phase value of the frequency band to obtain the corrected phase information , and then combine the corrected frequency, amplitude and phase information to generate the recursively corrected microwave signal to improve the stability and accuracy of the signal.

[0023] Generate the key frequency band signal: Divide the frequency-domain signal into multiple sub-frequency bands according to the set frequency range, and then identify the frequency bands that affect the transmission quality based on the correlation between the power spectral density, energy concentration and transmission performance index within the frequency band, and define them as key frequency bands. For the key frequency bands, extract the complex spectrum values of the frequency points within the frequency band, and set the spectrum components of the non-key frequency bands to zero, so as to form a spectrum signal containing the frequency components of the key frequency bands, which is the key frequency band signal for subsequent gain adjustment or noise suppression processing.

[0024] The specific steps to generate the recursively corrected signal are as follows: Based on the enhanced spectrum signal, first detect the signal frequency, obtain the actual frequency value of each frequency band, compare it with the target frequency, calculate the frequency error, and gradually adjust the frequency of each frequency band to generate a corrected frequency signal; Based on the corrected frequency signal, adjust the amplitude of each frequency band, measure the amplitude change of each frequency band, calculate the difference and adjust the amplitude of the corresponding frequency band, apply the correction amount to each frequency point, correct the signal intensity of the frequency band, and generate a corrected amplitude signal; Based on the corrected amplitude signal, optimize the phase of the signal, detect the phase deviation, obtain the required phase adjustment amount, and then correct the phase of each signal according to the frequency components. After the phase adjustment is completed, a recursive corrected signal is generated; Based on the enhanced spectrum signal, adopt a frequency detection algorithm. Calculate the actual frequency value of each frequency band through the fft.fftfreq function in the numpy library, and use the signal.find_peaks function in the scipy library to detect the peaks of the frequency-domain signal, so as to obtain the actual frequency of each frequency band. Compare it with the set target frequency, calculate the frequency error, and calculate the difference between the error value and the target frequency through the np.subtract function in the numpy library. Gradually adjust the frequency of each frequency band, use the signal.resample function in the scipy library to adjust the frequency, apply the frequency correction amount, and generate a corrected frequency signal. By adjusting the frequency components of each frequency band one by one, the frequency correction is finally completed; Based on the corrected frequency signal, adopt an amplitude adjustment algorithm. Use the np.abs function of numpy to measure the amplitude change of each frequency band, calculate the corrected amplitude difference, and adjust the amplitude of each frequency band through the np.multiply function of numpy. Set the correction amount to 1.2, apply the correction amount to the amplitudes of all frequency points within each frequency band, and use the signal.resample function in the scipy library to resample the corrected signal to correct the amplitude of each frequency band; Based on the corrected amplitude signal, adopt a phase optimization algorithm. Use the np.angle function of numpy to obtain the phase information of each frequency band signal and calculate the phase deviation of each signal. Set the target phase to 0, use the signal.angle function in the scipy library to optimize the phase of each frequency band, obtain the required phase adjustment amount, apply the adjustment amount to the phase of each frequency point through the np.add function of numpy, and use the signal.resample function in the scipy library to sample the corrected phase to complete the phase correction of each signal so that the phase of the signal meets the target requirements.

[0025] Phase correction based on frequency components extracts the current phase value of the frequency points according to the frequency components, compares it with the reference phase and the expected phase to obtain the phase error, calculates the correction factor based on this error, multiplies the frequency component by a complex exponential factor in the frequency domain to achieve phase rotation compensation, and at the same time, the consistency of the frequency-phase mapping relationship needs to be maintained. After phase correction, the signal spectrum structure will not be damaged, and a phase-aligned frequency-domain signal is generated for subsequent time-domain reconstruction and control system input.

[0026] The specific steps to generate the interference recognition result are as follows: Based on the recursively corrected signal, time-domain feature extraction is performed. First, the signal is framed. For each frame, the average amplitude, frequency change, signal peak, and change rate information are extracted. At the same time, the change trend of each frame of the signal is compared to generate a time-domain feature signal. Based on the time-domain feature signal, the interference sources in the signal are preliminarily screened. The abnormal fluctuations of each frame of the signal are monitored, a threshold is set and compared with each frame to identify the interference sources and generate an interference source candidate signal. Based on the interference source candidate signal, the interference sources are analyzed, the frequency range, amplitude, and duration of the interference signal are calculated, and the characteristics are analyzed at the same time to generate the interference recognition result. Based on the recursively corrected signal, a time-domain feature extraction algorithm is used. First, the np.split function of numpy is used to frame the signal, with the length of each frame set to 100 ms and an overlap of 50 ms between frames. The np.mean function of numpy is used to calculate the average amplitude of each frame of the signal, the np.diff function of numpy is used to calculate the frequency change within each frame of the signal, the signal.find_peaks function of scipy is used to detect the signal peak, the np.gradient function of numpy is used to calculate the change rate of each frame of the signal, and the change trend of each frame of the signal is compared to generate a time-domain feature signal. Based on the time-domain feature signal, an interference source screening algorithm is adopted. The threshold for abnormal fluctuations of each frame of the signal is set to 3 standard deviations. The np.std function of numpy is used to calculate the standard deviation of each frame of the signal and compare it with the set threshold. The np.where function of numpy is used to detect the frames with fluctuations exceeding the threshold, identify the possible interference sources, the signal.find_peaks function of scipy is used to detect the peak of the fluctuations of each frame of the signal, and the interference sources are preliminarily screened through the fluctuation amplitude and frequency characteristics to form a set of interference source candidate signals. Based on the candidate signals of the interference source, an interference source analysis algorithm is adopted. First, the np.fft.fft function of numpy is used to calculate the frequency range of each candidate signal. The amplitude range of the interference signal is analyzed through the np.max function and np.min function of numpy. The frequency range of each interference signal is calculated. The signal.find_peaks function of scipy is used to analyze the duration of the interference signal. Combining the amplitude and frequency information, the characteristic differences of each interference signal are calculated through the np.subtract function and np.add function of numpy, and the key characteristic parameters of each interference source are output.

[0027] Analyze the candidate signal segment of the interference source to extract its main frequency components, calculate the frequency interval with dense energy distribution as the frequency range of the interference signal, select the average power as the interference amplitude index to measure the amplitude value in this interval, calculate the duration of the interference signal according to the number of frames within the signal duration and the time length of each frame, and further extract the feature vectors of the interference signal, including frequency stability, amplitude volatility, and modulation characteristics, providing a basis for subsequent classification recognition and interference suppression strategy generation. The formed analysis results should include the interference frequency range, maximum amplitude, duration, and additional features, and the generated results support the judgment of the interference type and the decision-making of the system response.

[0028] The specific steps to generate the optimized signal are as follows: Based on the interference recognition result, select the gain control rule, calculate the gain requirement of each frequency band, measure the signal strength of each frequency band, and then adjust the amplitude of the corresponding frequency band according to the required gain value to enhance the target frequency band and generate the gain control signal; Based on the gain control signal, perform frequency correction, measure the frequency deviation of the signal, and use the correction amount to adjust the frequency of each frequency band one by one to generate the corrected frequency signal; Based on the corrected frequency signal, perform phase compensation on the signal, detect the phase deviation of each frequency band, calculate the required adjustment amount, and apply the phase compensation to each frequency component. Adjust the signal strength according to the amplitude correction to generate the optimized signal; Based on the interference recognition result, adopt the gain control rule. First, use the np.abs function of the numpy library to measure the signal strength of each frequency band, calculate the amplitude value of each frequency band, obtain the average amplitude of each frequency band through the np.mean function of numpy, calculate the gain requirement of each frequency band through the np.subtract function of numpy, set the target gain value to 1.5, and apply the gain value to the amplitude of each frequency band through the np.multiply function of numpy according to the required gain value to enhance the signal strength of the target frequency band and generate the gain control signal. This gain value is dynamically adjusted according to the signal strength of each frequency band to make the amplitude of the signal meet the preset requirements; Based on the gain control signal, frequency correction is performed. First, the np.angle function of numpy is used to measure the phase information of the signal. The frequency deviation of each frequency band is calculated through the np.diff function of numpy. The frequency correction amount is set to 0.05 Hz. The frequencies of each frequency band are adjusted one by one using the np.add function of numpy, and the correction amount is applied to each frequency band to adjust the frequency of the frequency band and generate a corrected frequency signal. During this process, by comparing the target frequency and the actual frequency of each frequency band, the frequency deviation correction is gradually performed to make the signal frequency match the target value; Based on the corrected frequency signal, phase compensation is performed on the signal. First, the np.angle function of numpy is used to detect the phase deviation of each frequency band, and the phase difference of each frequency band is calculated. The phase correction amount is set to π / 4. The phase compensation amount is applied to the phase of each frequency band through the np.add function of numpy, and the amplitude of the signal is adjusted using the np.multiply function of numpy. The amplitude of the signal generated after phase compensation is readjusted to the set value, and finally an optimized signal is obtained.

[0029] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above description is only the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A microwave signal processing method based on a composite channel of a grating sensor, characterized in that Microwave signal processing for the composite channel of a grating sensor includes the following steps: Step 1: Based on the microwave signal obtained by the grating sensor, signal screening is performed. After the signal is low-pass filtered to remove noise with a frequency higher than the set threshold, the signal amplitude is then normalized, and the signal is standardized through amplitude correction to generate a valid signal; Step 2: Based on the valid signal, a long short-term memory network is used to calculate the spectrum of the signal. At the same time, the signal is converted from the time domain to the frequency domain, each frequency component is extracted, and the main frequency range in the signal is determined. According to the component intensity and noise interference, a frequency band is selected, and the amplitude of the frequency band is enhanced to generate an enhanced spectrum signal; Step 3: Based on the enhanced spectrum signal, the frequency, amplitude, and phase of the signal are adjusted stage by stage, and multiple recursive corrections are performed. The signal effect after each correction is detected and each parameter is adjusted to further optimize the correction effect at each stage to generate a recursively corrected signal; Step 4: Based on the recursively corrected signal, the time-domain characteristics of the signal are extracted and analyzed to judge the interference source in the signal. By comparing the signal change patterns frame by frame, the frequency, intensity, and duration characteristics of the interference are determined to generate an interference recognition result; Step 5: Based on the interference recognition result, an appropriate frequency adjustment rule is selected to correct the gain control and frequency of the signal, and the signal is optimized according to phase compensation and intensity compensation to generate an optimized signal.

2. The microwave signal processing method based on a grating sensor composite channel according to claim 1, wherein The specific steps for generating the valid signal are as follows: Based on the microwave signal obtained by the grating sensor, low-pass filtering is performed. A frequency threshold is set and data points higher than the threshold are screened out. At the same time, the amplitude is set to zero to remove high-frequency noise and generate a filtered signal; Based on the filtered signal, the signal amplitude is normalized. The limit amplitude value in the signal is calculated, and the amplitude of each data point is divided by the extreme value, and the signal amplitude is adjusted to a unified range to generate a normalized signal; Based on the normalized signal, the signal amplitude is corrected. Each signal point is traversed and compared with the preset standard range. At the same time, the signal amplitude is adjusted so that the signal amplitude meets the set standard to generate a valid signal.

3. The microwave signal processing method based on a grating sensor composite channel according to claim 1, wherein The specific steps for generating the enhanced spectrum signal are as follows: Based on the valid signal, a long short-term memory network is used. First, the signal is segmented according to a set time window, the frequency components within each time window are calculated, a window function is used to weight each time period, and a frequency-domain signal is generated using the frequency and amplitude information within each time window; Based on the frequency-domain signal, the spectrum is divided into multiple frequency bands according to the frequency range. The frequency bands that affect the transmission quality are selected, and the boundaries of each frequency band are calculated. At the same time, all frequency components within the frequency band are extracted to generate a key frequency band signal; Based on the key frequency band signal, the gain of the signal in each selected frequency band is adjusted. First, the gain coefficient of the frequency band is calculated and applied to the amplitude of all frequency points within the frequency band to enhance the signal intensity of the frequency band and generate an enhanced spectrum signal.

4. The microwave signal processing method based on a grating sensor composite channel according to claim 1, wherein The long short-term memory network, according to the formula: ； Wherein: represents a frequency-domain signal, is the frequency, is the time window, is the index of the time-domain signal, represents the amplitude of the th sampling point of the time-domain signal, is an integer index, represents the window function weighting coefficient, is the complex exponential term, is the base of the natural logarithm, is the imaginary unit, is the total number of sampling points of the signal.

5. The microwave signal processing method based on a grating sensor composite channel according to claim 3, wherein generating the key frequency band signal divides the frequency domain signal into multiple sub - frequency bands according to a set frequency range, and then, based on the correlation between the power spectral density, energy concentration degree and transmission performance index within the frequency band, identifies the frequency bands that affect the transmission quality, which are defined as key frequency bands. For the key frequency bands, the complex spectral values of the frequency points within the frequency band are extracted, and the spectral components of the non - key frequency bands are set to zero, thereby forming a spectral signal containing the frequency components of the key frequency bands, which is the key frequency band signal for subsequent gain adjustment or noise suppression processing.

6. The microwave signal processing method based on a grating sensor composite channel according to claim 1, wherein, The specific steps for generating the recursive correction signal are as follows: Based on the enhanced spectral signal, first detect the signal frequency, obtain the actual frequency value of each frequency band, compare it with the target frequency, calculate the frequency error, and gradually adjust the frequency of each frequency band to generate a corrected frequency signal; Based on the corrected frequency signal, adjust the amplitude of each frequency band, measure the amplitude change of each frequency band, calculate the difference and adjust the amplitude of the corresponding frequency band, apply the correction amount to each frequency point, correct the signal intensity of the frequency band, and generate a corrected amplitude signal; Based on the corrected amplitude signal, optimize the phase of the signal, detect the phase deviation, obtain the required phase adjustment amount, and then correct the phase of each signal according to the frequency components. After the phase adjustment is completed, a recursive correction signal is generated.

7. The microwave signal processing method based on a grating sensor composite channel according to claim 6, wherein correcting the phase according to the frequency components extracts the current phase value of the frequency point according to the frequency components, compares it with the reference phase and the expected phase to obtain the phase error, calculates the correction factor according to this error, and realizes phase rotation compensation by multiplying the frequency component by a complex exponential factor in the frequency domain. At the same time, it is necessary to maintain the consistency of the frequency - phase mapping relationship. After the phase correction, the spectral structure of the signal will not be damaged, and a phase - aligned frequency domain signal is generated for subsequent time - domain reconstruction and control system input.

8. The microwave signal processing method based on a grating sensor composite channel according to claim 1, wherein The specific steps for generating the interference identification result are as follows: Based on the recursive correction signal, perform time - domain feature extraction. First, perform frame - by - frame processing on the signal, extract information such as the average amplitude, frequency change, signal peak, and change rate for each frame, and at the same time compare the change trends of the signals in each frame to generate a time - domain feature signal; Based on the time - domain feature signal, preliminarily screen the interference sources in the signal, monitor the abnormal fluctuations of the signals in each frame, set a threshold and compare each frame to identify the interference sources and generate an interference source candidate signal; Based on the interference source candidate signal, analyze the interference source, calculate the frequency range, amplitude, and duration of the interference signal, and at the same time analyze the characteristics to generate an interference identification result.

9. The microwave signal processing method based on the composite channel of the grating sensor according to claim 8, wherein for the candidate signal segment affected by the interference source, its main frequency components are extracted, the frequency range with intensive energy distribution is calculated as the frequency range of the interference signal, the average power is selected as the interference amplitude index to measure the amplitude value within this range, the duration of the interference signal is calculated according to the number of frames within the signal duration and the time length of each frame, and the feature vector of the interference signal is further extracted, including frequency stability, amplitude volatility and modulation characteristics, so as to provide a basis for subsequent classification recognition and the generation of interference suppression strategies. The analysis result should include the interference frequency range, the maximum amplitude, the duration and additional features, and the generated result supports the judgment of the interference type and the decision-making of the system response.

10. The microwave signal processing method based on a grating sensor composite channel according to claim 1, characterized in that The specific steps for generating the optimized signal are as follows: Based on the interference recognition result, select the gain control rule, calculate the gain requirement for each frequency band, measure the signal strength of each frequency band, and then adjust the amplitude of the corresponding frequency band according to the required gain value to enhance the target frequency band and generate the gain control signal; Based on the gain control signal, perform frequency correction, measure the frequency deviation of the signal, and use the correction amount to adjust the frequency of each frequency band one by one to generate the corrected frequency signal; Based on the corrected frequency signal, perform phase compensation on the signal, detect the phase deviation of each frequency band, calculate the required adjustment amount, apply the phase compensation to each frequency component, and adjust the signal strength according to the amplitude correction to generate the optimized signal.

Images