A multi-band processing system for electronic communication

Abstract

The present invention relates to the field of information transmission technology, and specifically to a multi-band processing system for electronic communication. The system includes: a signal separation module, a frequency offset compensation module, an amplitude correction module, a phase correction module, and a dynamic feedback adjustment module. In the present invention, signal components are separated by performing band-pass filtering on multiple frequency bands, the short-time power spectral density is calculated to characterize the energy distribution, the normalization coefficient operation enhances the frequency domain equalization, the high-energy component screening reduces the noise interference, and the signal effectiveness is improved. The frequency offset compensation calculates the change rate based on adjacent phase data and eliminates abnormal fluctuations, and adjusts the time-domain coefficient to correct the frequency offset error. The amplitude correction relies on the short-time Fourier transform to screen the maximum amplitude to optimize the signal consistency. The phase adjustment fits the group delay error curve to optimize the compensation parameters. The dynamic feedback optimizes the strategy based on the signal-to-noise ratio trend to enhance the compensation accuracy, and improves the signal transmission quality and stability.

Description

Technical Field

[0001] The present invention relates to the technical field of information transmission, and particularly to a multi-band processing system for electronic communication. Background Art

[0002] The technical field of information transmission includes the design and implementation of wireless or wired transmission methods and related devices for information such as data, voice, and images. The core content of this technical field includes aspects such as signal modulation and demodulation, multiplexing, channel coding, waveform design, and transmission medium adaptation to ensure that data can be efficiently and stably transmitted from the sending end to the receiving end. Information transmission technology covers traditional analog communication, digital communication, optical communication, and wireless communication methods, and is widely used in fields such as radio and television, satellite communication, mobile communication, and the Internet. In the development process of modern communication technology, information transmission is gradually evolving towards high bandwidth, low latency, and multi-band compatibility to meet the data transmission requirements in different application scenarios.

[0003] Among them, a multi-band processing system for electronic communication refers to a system that can receive, convert, and process signals of different frequency bands. This system addresses the multi-band compatibility problem in the wireless communication environment, uses a radio frequency front-end circuit to filter and amplify signals of different frequency bands, generates the required local oscillator signals through a frequency synthesizer to complete mixing conversion, and uses a baseband processing unit to perform signal demodulation, error correction, and multi-channel data synthesis. The system can also combine spectrum sensing technology to achieve dynamic monitoring and frequency band allocation of signals to adapt to the multi-band parallel transmission requirements in a complex wireless communication environment.

[0004] The prior art lacks a dynamic equalization strategy for signal frequency band processing and has insufficient short-time power spectrum analysis, which easily causes unbalanced energy distribution, resulting in overloading or attenuation of signals in some frequency bands and affecting the transmission quality. The frequency offset correction relies on fixed compensation and fails to dynamically adjust based on adjacent frame data, leading to the accumulation of frequency offset errors and causing signal distortion. The amplitude control method does not fully consider the short-time power spectrum characteristics, and noise is easily introduced during the amplification process, and serious information loss occurs during attenuation, affecting the signal integrity. The phase adjustment uses static compensation and lacks adaptive adjustment for different signal characteristics, and the error accumulation exacerbates the phase distortion. The dynamic adjustment strategy is fixed and does not optimize parameters in combination with the real-time signal state, and the adjustment lag affects the system stability, which may lead to insufficient compensation or overcorrection, reducing the transmission efficiency. Summary of the Invention

[0005] The object of the present invention is to solve the deficiencies existing in the prior art, and to propose a multi-band processing system for electronic communication.

[0006] To achieve the above object, the present invention adopts the following technical solution: A multi-band processing system for electronic communication includes:

[0007] The signal separation module obtains the input signal, extracts multiple frequency bands, performs band-pass filtering, separates each frequency band component, calculates the short-time power spectral density, obtains the normalization coefficient according to the spectral energy, obtains the instantaneous power value, determines the frequency domain distribution, calculates the normalization ratio, verifies the signal equalization state, screens the high-energy components, and generates a multi-band normalized signal;

[0008] The frequency offset compensation module is based on the multi-band normalized signal, collects adjacent phases, calculates the phase change rate, statistically analyzes the continuous frame mean, eliminates abnormal fluctuations, compares the inter-frame frequency offset, obtains the compensation factor, adjusts the time-domain coefficient, and generates a time-domain offset compensation signal;

[0009] The amplitude correction module is based on the time-domain offset compensation signal, collects the short-time Fourier transform data, extracts the power spectral density, calculates the attenuation factor, analyzes the amplitude range, screens the maximum amplitude, normalizes the amplitude value, and generates a normalized amplitude signal;

[0010] The phase correction module is based on the normalized amplitude signal, collects the group delay data, extracts the phase trend, fits the error curve, judges the second-order change rate, selects the compensation parameter, adjusts the phase, and generates a phase compensation signal.

[0011] As a further solution of the present invention, the multi-band normalized signal includes a low-frequency normalized output, a medium-frequency normalized output, and a high-frequency normalized output. The time-domain offset compensation signal includes a phase stability value, a time-domain correction amount, and an inter-frame equalization index. The normalized amplitude signal includes an amplitude standard amount, an amplitude equalization index, and an amplitude reference value. The phase compensation signal includes a phase correction value, a trend balance amount, and an error correction index.

[0012] As a further solution of the present invention, the signal separation module includes:

[0013] The signal band-pass filtering sub-module obtains the input signal, extracts multiple preset frequency bands, calls the filter for band-pass filtering, adjusts the filtering parameters according to the frequency range, adjusts the filter order, calculates the amplitude spectrum of the filtered signal, compares the signal spectra before and after filtering, extracts the energy distribution of the multi-band signal, and obtains the multi-band signal components;

[0014] The short-time power spectrum calculation sub-module is based on the multi-band signal components, calculates the short-time power spectral density, obtains the signal power value within each time window, determines the spectral energy distribution, and uses the formula:

[0015] ;

[0016] Calculates to obtain the instantaneous power value, determines the energy proportion of the differentiated frequency bands, and obtains the spectral energy distribution;

[0017] Wherein, represents the instantaneous power value, Represents the signal component of the band, represents the center frequency of the band, represents the total center frequency for normalization calculation, a represents the sub - band index for normalization calculation, and b represents the sub - band index for normalization calculation;

[0018] The multi - band normalization sub - module calculates the normalization ratio based on the spectral energy distribution, adjusts the multi - band signal power, calculates the signal amplitude distribution after normalization, simultaneously compares the signal power difference before and after adjustment, screens the bands with high energy proportion, synthesizes the screened signals, and obtains the multi - band normalized signals.

[0019] As a further solution of the present invention, the frequency offset compensation module includes:

[0020] The phase change rate calculation sub - module collects adjacent phases based on the multi - band normalized signal, calculates the phase change rate, extracts the phase change rate data of consecutive frames, and calculates the mean value to obtain the mean phase change rate;

[0021] The abnormal fluctuation elimination sub - module calls the mean phase change rate, statistics the change range of consecutive frames, eliminates the abnormal fluctuation data, calculates the inter - frame frequency offset, and obtains the frequency offset data after eliminating the abnormal fluctuations;

[0022] The time - domain compensation calculation sub - module calls the frequency offset data after eliminating the abnormal fluctuations, compares the inter - frame frequency offset, obtains the compensation factor, adjusts the time - domain coefficient, and uses the formula:

[0023] ;

[0024] Performs operations to obtain the time - domain offset compensation factor, adjusts the time - domain signal, and generates the time - domain offset compensation signal;

[0025] Wherein, represents the time - domain offset compensation factor, represents the frequency offset value of the frame, represents the weight parameter of the frame, represents the phase value of the frame, represents the time coefficient of the frame, represents the total number of frames,

[0026] As a further solution of the present invention, the amplitude correction module includes:

[0027] The short-time Fourier transform sub-module calculates the frequency components of the signal within a short-time window based on the time-domain offset compensation signal, intercepts the signal sequence using a sliding window, performs Fourier transform on the signal within each window, calculates the complex spectrum values after the transform, obtains the spectrum information on each time segment, and simultaneously calls the modulus value of the spectrum after the transform to generate short-time Fourier transform data;

[0028] The power spectral density extraction sub-module calls the short-time Fourier transform data, calculates the power values corresponding to multiple frequency components, normalizes the power spectra of all time segments, calculates the power density curve using integration, obtains the energy distribution, and screens the key power distributions within the characteristic frequency band range to generate power spectral density data;

[0029] The amplitude normalization sub-module calls the power spectral density data, calculates the attenuation factor of the signal within the characteristic frequency band range, using the formula:

[0030] ;

[0031] Calculates the weighted amplitude of each frequency component, extracts the amplitude range, screens the maximum amplitude value, and normalizes it to obtain the normalized amplitude signal;

[0032] Wherein, represents the amplitude value of the current frequency component, represents the attenuation factor, represents the frequency component corresponding attenuation factor, represents the frequency component within the characteristic frequency band range, represents the phase angle, represents the bias term associated with the frequency component, represents the total number of frequency components, represents the cosine value of the phase angle, f represents the frequency component index.

[0033] As a further solution of the present invention, the phase correction module includes:

[0034] The group delay acquisition sub-module acquires the group delay data of the signal at different frequencies based on the normalized amplitude signal, calculates the group delay values corresponding to multiple frequency points, and performs data sorting to screen out the noise interference data to obtain the effective group delay data;

[0035] The phase trend analysis sub-module calls the effective group delay data, extracts the phase change trend based on the group delay data of multiple frequency points, calculates the second-order change rate, analyzes the error curve, using the formula:

[0036] ;

[0037] Calculate the phase adjustment amount and obtain the phase adjustment parameter;

[0038] Wherein, represents the phase adjustment amount, represents the phase value of the th frequency point, represents the group delay value of the th frequency point, represents the adjustment coefficient of the th frequency point, N represents the total number of frequency points participating in the calculation, and g represents the frequency point index;

[0039] The phase correction calculation sub-module calls the phase adjustment parameter, selects a compensation strategy, performs phase adjustment on the normalized amplitude signal, and calculates and obtains a phase compensation signal.

[0040] As a further solution of the present invention, the system further includes:

[0041] The dynamic feedback adjustment module is based on the phase compensation signal, collects the offset compensation value and the amplitude coefficient, extracts the phase coefficient, calculates the signal-to-noise ratio trend, compares the past error gradient, determines the update step size, screens out the out-of-range parameters, and generates a dynamic optimization compensation signal;

[0042] The dynamic optimization compensation signal includes feedback correction parameters, error optimization amounts, and signal-to-noise balance indicators.

[0043] As a further solution of the present invention, the dynamic feedback adjustment module includes:

[0044] The offset compensation calculation sub-module obtains the phase compensation signal, collects the offset compensation value and the amplitude coefficient, calculates the normalized ratio between the offset compensation value and the amplitude coefficient, obtains the offset change rate by comparing with past data, and at the same time screens out the parameters with the offset change rate exceeding the limit. After removing the out-of-limit data, the offset compensation equilibrium value is calculated;

[0045] The signal-to-noise ratio trend analysis sub-module is based on the offset compensation equilibrium value, extracts the phase coefficient, calculates the current signal-to-noise ratio, calls the signal-to-noise ratio data, calculates the current signal-to-noise ratio change rate, and at the same time obtains the signal-to-noise ratio gradient trend based on the change rate, calculates the error amplitude according to the gradient trend, and generates the signal-to-noise ratio change trend;

[0046] The dynamic optimization compensation generation sub-module calls the signal-to-noise ratio change trend, compares the error gradient, calculates the optimized update step size, removes the out-of-range parameters, and uses the formula:

[0047] ;

[0048] Calculate the dynamic optimization coefficient, adjust the phase compensation signal, and generate a dynamic optimization compensation signal;

[0049] Among them, represents the calculation result of the dynamic optimization compensation signal, represents the adjustment coefficient, represents the derivative of the signal-to-noise ratio change rate with respect to time, represents the out-of-range error parameter, represents the compensation constant.

[0050] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0051] In the present invention, during the signal separation process, multiple frequency bands are extracted and band-pass filtering is performed to clearly separate the signal components. The short-time power spectral density calculation accurately depicts the energy distribution, and the normalization coefficient operation enhances the frequency domain balance. The high-energy component screening reduces noise interference and improves the signal effectiveness. The frequency offset compensation calculates the change rate through adjacent phase data, combines the continuous frame mean to eliminate abnormal fluctuations, and adjusts the time-domain coefficient to correct the frequency offset error. The amplitude correction relies on the short-time Fourier transform to analyze the power spectrum, screens the maximum amplitude to optimize the signal amplitude consistency. The phase adjustment fits the error curve through the group delay data and adjusts the compensation parameter according to the change rate to reduce the cumulative error. The dynamic feedback adjustment combines the signal-to-noise ratio trend to optimize the compensation strategy, screens the out-of-range parameters according to the error gradient, and enhances the compensation accuracy and system stability. During the multi-band signal processing process, the synergistic effects of energy balance, frequency offset correction, amplitude adjustment, phase optimization, and dynamic compensation improve the signal transmission quality and stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 is the system flowchart of the present invention;

[0053] Figure 2 is the flowchart of the signal separation module of the present invention;

[0054] Figure 3 is the flowchart of the frequency offset compensation module of the present invention;

[0055] Figure 4 is the flowchart of the amplitude correction module of the present invention;

[0056] Figure 5 is the flowchart of the phase correction module of the present invention;

[0057] Figure 6 is the flowchart of the dynamic feedback adjustment module of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0058] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, 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.

[0059] 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. These are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of 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.

[0060] Embodiment 1

[0061] Please refer to Figure 1 , a multi-band processing system for electronic communication includes:

[0062] The signal separation module acquires the input signal, extracts multiple frequency bands, performs band-pass filtering, separates each frequency band component, calculates the short-time power spectral density, obtains the normalization coefficient according to the spectral energy, obtains the instantaneous power value, determines the frequency domain distribution, calculates the normalization ratio, verifies the signal equalization state, screens the high-energy components, and generates a multi-band normalized signal;

[0063] The frequency offset compensation module, based on the multi-band normalized signal, acquires the adjacent phases, calculates the phase change rate, statistically calculates the mean value of consecutive frames, eliminates abnormal fluctuations, compares the inter-frame frequency offset, obtains the compensation factor, adjusts the time-domain coefficient, and generates a time-domain offset compensation signal;

[0064] The amplitude correction module, based on the time-domain offset compensation signal, acquires the short-time Fourier transform data, extracts the power spectral density, calculates the attenuation factor, analyzes the amplitude range, screens the maximum amplitude, normalizes the amplitude value, and generates a normalized amplitude signal;

[0065] The phase correction module, based on the normalized amplitude signal, acquires the group delay data, extracts the phase trend, fits the error curve, judges the second-order change rate, selects the compensation parameter, adjusts the phase, and generates a phase compensation signal;

[0066] The dynamic feedback adjustment module, based on the phase compensation signal, acquires the offset compensation value and the amplitude coefficient, extracts the phase coefficient, calculates the signal-to-noise ratio trend, compares the previous error gradient, judges the update step size, screens the out-of-range parameters, and generates a dynamic optimization compensation signal.

[0067] The multi-band normalized signal includes a low-frequency normalized output, a medium-frequency normalized output, and a high-frequency normalized output. The time-domain offset compensation signal includes a phase stability value, a time-domain correction amount, and an inter-frame equalization index. The normalized amplitude signal includes an amplitude standard amount, an amplitude equalization index, and an amplitude reference value. The phase compensation signal includes a phase correction value, a trend balance amount, and an error correction index. The dynamic optimization compensation signal includes a feedback correction parameter, an error optimization amount, and a signal-to-noise balance index.

[0068] Please refer to Figure 2 , the signal separation module includes:

[0069] The signal band-pass filtering sub-module acquires the input signal, extracts multiple preset frequency bands, calls a filter for band-pass filtering, adjusts the filtering parameters according to the frequency range, adjusts the filter order, calculates the amplitude spectrum of the filtered signal, and compares the signal spectra before and after filtering, extracts the energy distribution of the multi-band signal, and obtains the multi-band signal components;

[0070] First, perform frequency-domain analysis on the input signal to determine its main energy distribution region, configure the filtering parameters according to different frequency ranges, set the bandwidth of each frequency band. For example, for the processing of voice signals, the frequency band can be set to 300Hz - 3.4kHz, and for biomedical signals (such as EEG signals), different frequency bands within the range of 0.5Hz - 100Hz can be selected. Then, call a filter for band-pass filtering. Taking the FIR filter as an example, the filter order needs to be set during design. The determination of the order is based on the trade-off between the filtering accuracy requirements of the target frequency band and the computational complexity. For example, if a steeper passband response is required, a higher order (such as 64) can be used, otherwise a lower order (such as 16) can be selected to reduce the computational burden. The filtered signal is subjected to Fourier transform to obtain the amplitude spectrum and compare it with the spectrum of the signal before filtering. By comparing the energy distribution of signals in different frequency bands, confirm whether the signal components meet the set filtering characteristics. If it is found that the energy of a certain frequency band signal accounts for a small proportion, the filtering parameters can be appropriately adjusted to optimize the frequency band division. For example, in the processing of EEG signals, if the energy of the low-frequency band (0.5Hz - 4Hz) is too low, the passband range of the filter can be appropriately adjusted to improve the energy capture ability of the low-frequency band. Finally, extract the multi-band signal components.

[0071] The short-time power spectrum calculation sub-module calculates the short-time power spectral density based on the multi-band signal components, obtains the signal power value within each time window, determines the spectral energy distribution, using the formula:

[0072] ;

[0073] Perform operations to obtain the instantaneous power value, determine the energy proportion of the differentiated frequency bands, and obtain the spectral energy distribution;

[0074] where represents the instantaneous power value, Represents the signal component of the frequency band, represents the center frequency of the frequency band, represents the number of sub - frequency bands, represents the sum of the center frequencies for normalization calculation, a represents the sub - frequency band index, and b represents the sub - frequency band index for normalization calculation;

[0075] First, set the short - time window length. For example, a 25 - ms window is often used in speech signal processing, and a 500 - ms window can be set in EEG signal analysis. Then, perform power calculation on the signal within each window. Obtain the signal energy of each time window by summing the squares, and calculate the instantaneous power based on the spectral energy distribution. The formula is:

[0076] ;

[0077] where, is the signal component of the frequency band, represents the center frequency of this frequency band, is the total number of sub - frequency bands. During calculation, sum the squares of the signals of each frequency band and perform weighted normalization. Taking a practical example, assume a signal is divided into 3 frequency bands, corresponding to center frequencies 100 Hz, 500 Hz, and 1000 Hz respectively, and the signal component energies are 0.2, 0.8, and 1.5 respectively. Then the calculation is as follows:

[0078] ;

[0079] The energy calculation of each frequency band is as follows:

[0080] ;

[0081] ;

[0082] ;

[0083] The instantaneous power is obtained. If this value is within a certain set threshold (for example, the preset threshold 1.5), it is determined that the signal power within this window is high, otherwise the signal is considered weak. Through this calculation, determine the energy proportion of different frequency bands for subsequent signal screening.

[0084] The multi - band normalization sub - module calculates the normalization ratio based on the spectral energy distribution, adjusts the multi - band signal power, calculates the signal amplitude distribution after normalization, and at the same time compares the signal power difference before and after adjustment, screens the frequency bands with high energy proportion, synthesizes the screened signals, and obtains the multi - band normalized signal.

[0085] First, set the normalization target value. For example, set the maximum energy component not to exceed 1.0 to avoid excessive amplification in certain frequency bands affecting the balance of the overall signal. When calculating the normalization ratio, take the current maximum energy value as the reference, and calculate the normalization coefficient for each frequency band:

[0086] ;

[0087] Suppose the maximum component value of a certain signal energy is 2.0, then the normalization ratio:

[0088] ;

[0089] Then, multiply the signal components of all frequency bands by the normalization ratio to ensure that the adjusted energy meets the preset range. Calculate and compare the signal power differences before and after adjustment, screen the frequency bands with a relatively high energy proportion, and set the screening threshold , for example, set it to 0.4. Then the signal energy in the low-frequency band is low and is discarded, only the mid-high frequency band signals are retained, and finally the screened signals are synthesized to obtain the multi-band normalized signal.

[0090] Table 1 Multi-band signal energy normalization results

[0091]

[0092] As shown in Table 1, by normalizing and adjusting the energies of different frequency bands to maintain the overall balance of the signal. On this basis, screen the frequency bands with a high energy proportion, and finally synthesize the normalized signal to meet the requirements of subsequent signal processing.

[0093] Please refer to Figure 3 , the frequency offset compensation module includes:

[0094] The phase change rate calculation sub-module is based on the multi-band normalized signal, collects adjacent phases, calculates the phase change rate, extracts the phase change rate data of consecutive frames, and calculates the mean value to obtain the mean phase change rate;

[0095] First, obtain the phase values of adjacent frames. In the specific implementation process, extract consecutive multi-frame data, calculate the phase difference frame by frame. Suppose the phase value of the th frame is , then the phase change rate can be expressed as , where is obtained through frequency domain analysis. For the signals of different frequency bands, normalization processing is performed so that their phase values can be calculated under the same reference. By calculating the phase difference, the distribution of the phase change rate can be obtained. In the calculation process, the window sliding method is adopted. Each group of windows contains 10-frame signal data (which can be adjusted according to the specific application scenario), and the mean value of the phase change rate within the window is calculated and defined as , where Represents the number of window frames. During the calculation of the mean value, in order to reduce noise interference, abnormal data can be filtered out before calculating the mean value, and a change rate threshold is set. If the phase change rate of a certain frame exceeds the set threshold (for example, more than 3 times the standard deviation of the mean value), then the data of this frame is excluded. Finally, the corrected mean value of the phase change rate is obtained. This result shows that by calculating the mean value of the phase change rate, the average phase change trend between frames can be obtained, providing basic data for subsequent frequency offset calculation, making the subsequent calculation based on stable data rather than the instantaneous phase change affected by noise.

[0096] The abnormal fluctuation elimination sub-module calls the mean value of the phase change rate, counts the change range of consecutive frames, eliminates abnormal fluctuation data, calculates the inter-frame frequency offset, and obtains the frequency offset data after eliminating abnormal fluctuations.

[0097] First, perform a distribution analysis on the phase change rates of all frames and calculate the change range. If this range exceeds the set upper and lower limits, then further screen the abnormal fluctuation data. Adopt the median method of removing extreme values to eliminate the extreme value frames that exceed the set upper and lower limits. The elimination method uses dynamic threshold setting, such as setting an abnormal threshold. where and represent the mean value and standard deviation of the change rate within the normal range respectively. Finally, calculate the inter-frame frequency offset, and the formula is expressed as where is the frame period. During the calculation process, 1000-frame data is used for statistical analysis. The example is as follows:

[0098] Table 2 Data table for frequency offset calculation

[0099]

[0100] As shown in Table 2, through the calculation of the inter-frame frequency offset, stable frequency offset data is obtained after eliminating abnormal fluctuations, avoiding the influence of extreme values on the calculation accuracy. This result shows that the data after eliminating abnormal fluctuations can effectively reduce the interference of extreme values on the frequency offset calculation, making the finally obtained frequency offset data more stable, ensuring the accuracy of frequency offset compensation, and improving the reliability of subsequent time-domain compensation calculation.

[0101] The time-domain compensation calculation sub-module calls the frequency offset data after eliminating abnormal fluctuations, compares the inter-frame frequency offset, obtains the compensation factor, adjusts the time-domain coefficient, and uses the formula:

[0102] ;

[0103] Calculate through operations to obtain the time-domain offset compensation factor, adjust the time-domain signal, and generate the time-domain offset compensation signal.

[0104] Among them, represents the time-domain offset compensation factor, represents the frequency offset value of the th frame, represents the phase value of the th frame, represents the signal amplitude of the th frame, represents the

[0105] total number of frames, represents the number of frames for phase change calculation,

[0106] ;

[0107] where the weight parameter can be set according to the frame amplitude size. For example, if is set, then the higher the amplitude of the frame, the higher the weight. In addition, to further improve the compensation accuracy, the inter-frame phase accumulation amount is introduced:

[0108] ;

[0109] Calculate the normalized amplitude square sum of the signal:

[0110] ;

[0111] Finally, the time-domain offset compensation factor is calculated as:

[0112] ;

[0113] where the time coefficient is calculated based on the inter-frame stability. For example, , where represents the time normalization coefficient. Assuming the sampling frequency of a certain frame is 1000 Hz, the calculation is as follows:

[0114] ;

[0115] The calculation result is:

[0116] ;

[0117] ;

[0118] Finally, a time-domain compensation factor is obtained. , the time-domain signal is adjusted to generate a time-domain offset compensation signal. This result indicates that the calculated time-domain compensation factor shows the comprehensive influence degree of the inter-frame frequency offset. The larger this value is, the greater the change in the inter-frame frequency offset, and stronger time-domain compensation is required. When is in a smaller range (such as close to 1), it indicates that the frequency offset is small and the system runs stably. By calculating this compensation factor, it can be directly used to adjust the time-domain characteristics of the signal, thereby compensating for the frequency offset error and ensuring the signal synchronization accuracy.

[0119] Please refer to Figure 4 , the amplitude correction module includes:

[0120] Based on the time-domain offset compensation signal, the short-time Fourier transform sub-module calculates the frequency components of the signal within a short-time window. It uses a sliding window to intercept the signal sequence, performs Fourier transform on the signal within each window, calculates the complex spectrum values after the transform, obtains the spectrum information on each time segment, and at the same time calls the modulus value of the spectrum after the transform to generate short-time Fourier transform data;

[0121] First, the input signal is sampled, and the time-domain signal is segmented according to a fixed time window. The length of each window is set to , and the interval between each window is set to . The window uses a Hanning window to reduce spectrum leakage. After intercepting the signal, discrete Fourier transform (DFT) is performed on the signal within each window. For sampling points within a window, the signal is expressed as , and its transformation formula is:

[0122] ;

[0123] Among them, is the spectrum value after the transform, takes the value of 1024 to ensure sufficient frequency resolution. The DFT operation of each window generates a set of complex spectrum values, and each spectrum value contains amplitude and phase information. For example, for the signal , the spectrum amplitudes obtained after the transform are approximately 2 and 1.5 at 50 Hz and 120 Hz respectively. Through the repeated calculation of the sliding window, the spectrum change of the signal on the entire time axis is obtained, and the complete short-time Fourier transform data is obtained, and the spectrum information of each time segment is recorded, as shown in Table 3.

[0124] Table 3 Short-time Fourier transform data

[0125]

[0126] As shown in Table 3, the spectral amplitudes within different time windows vary slightly, indicating that the instantaneous frequency components of the signal can be reflected by the short-time Fourier transform data, which can be used for subsequent power spectral density calculations. This result shows that the short-time Fourier transform can accurately obtain the frequency components of the signal on the time axis and provide a data basis for power spectrum calculations. These variations in spectral amplitudes can be used to analyze the energy distribution characteristics of the signal for subsequent power spectral density calculations and normalization processing.

[0127] The power spectral density extraction sub-module calls the short-time Fourier transform data, calculates the power values corresponding to multiple frequency components, normalizes the power spectra of all time segments, uses integration to calculate the power density curve, obtains the energy distribution, and screens the key power distributions within the characteristic frequency band range to generate power spectral density data;

[0128] Call the short-time Fourier transform data and calculate the power values corresponding to multiple frequency components. The power calculation formula for each frequency component is:

[0129] ;

[0130] where is the power value at frequency , is the amplitude of this frequency component, is the window length. Let . For the 50 Hz frequency point, its power calculation is as follows:

[0131] ;

[0132] Calculate for all frequency points and perform normalization processing. The normalization calculation method is as follows:

[0133] ;

[0134] where is the total power of all frequency components, is the total number of frequency components to ensure that the power distribution is between . For the 50 Hz frequency point, its normalized power calculation is as follows:

[0135] ;

[0136] After calculating the power spectra of all time segments, use integration to calculate the power density curve. The calculation method is:

[0137] ;

[0138] where is the target characteristic frequency band range. Let the characteristic frequency band be , the integral value is calculated to obtain the energy distribution within the characteristic frequency band, and the frequency range with the highest energy density is selected to obtain the power spectral density data. This result indicates that the power spectral density calculation can effectively extract the energy variation within the characteristic frequency band and provide the power data for amplitude normalization processing. This normalized power distribution data determines the main energy contribution part of the signal characteristic frequency band and can be used for further processing of the signal amplitude.

[0139] The amplitude normalization sub-module calls the power spectral density data to calculate the attenuation factor of the signal within the characteristic frequency band range, using the formula:

[0140] ;

[0141] Calculate the weighted amplitude of each frequency component, extract the amplitude range, select the maximum amplitude value, and perform normalization to obtain the normalized amplitude signal;

[0142] Among them, represents the amplitude value of the current frequency component, represents the attenuation factor, represents the frequency component corresponding attenuation factor, represents the frequency component within the characteristic frequency band range, represents the phase angle, represents the bias term associated with the frequency component, represents the total number of frequency components, represents the cosine value of the phase angle, and f represents the frequency component index.

[0143] The specific calculation method is as follows:

[0144] ;

[0145] Among them, represents the attenuation coefficient of the frequency component. Let , represents the amplitude of each frequency component, is the corresponding frequency component, is the phase angle, is the bias term. Suppose the frequency components and amplitudes within a certain time window are shown in Table 4:

[0146] Table 4 Frequency Components and Amplitudes

[0147]

[0148] Calculate the amplitude value after attenuation:

[0149] ;

[0150] Calculate the final normalized amplitude:

[0151] ;

[0152] The maximum amplitude value in the characteristic frequency band is screened and normalized to obtain the normalized amplitude signal. The result shows that the normalized amplitude signal has removed the power attenuation effect in the characteristic frequency band, making the contribution of the signal on different frequency components more balanced. At the same time, the signal adjusted by amplitude normalization can be used for subsequent signal analysis and classification processing to improve the analysis accuracy.

[0153] See also Figure 5 , the phase correction module includes:

[0154] The group delay acquisition submodule collects the group delay data of the signal at the differentiated frequency based on the normalized amplitude signal, calculates the group delay values ​​corresponding to multiple frequency points, and sorts the data to filter out the noise interference data to obtain the effective group delay data;

[0155] First, for the input normalized amplitude signal, the group delay data is collected at different frequency points. The specific method is to collect the group delay data of the set frequency point set. Measure each frequency point one by one The corresponding group delay It can be calculated by phase response, and the specific calculation method is:

[0156] ;

[0157] in, is the phase value at this frequency point, The system excites multiple frequency points with signals, measures the corresponding phase response by scanning, and calculates the group delay value. In practical applications, a linear frequency modulation pulse signal can be used to excite the system, and the phase changes at different frequency points can be recorded to obtain the corresponding group delay data. Then, for the collected group delay data, a data preprocessing method is used to eliminate abnormal data. First, the mean value of the group delay data is calculated. and standard deviation , then filter the data and set the threshold range , remove the data beyond this range to reduce the impact of noise, and finally obtain the effective group delay data set , as shown in Table 5.

[0158] Table 5 Group delay data collection and screening results

[0159]

[0160] The results show that after data screening, the abnormal group delay data with large deviations (such as 7.8ns at the 2.0GHz frequency point) are removed, and the effective group delay data can more accurately reflect the delay characteristics of the system at different frequency points, thereby providing more reliable input data for subsequent phase trend analysis.

[0161] The phase trend analysis submodule calls the effective group delay data, extracts the phase change trend based on the group delay data of multiple frequency points, calculates the second-order change rate, and analyzes the error curve using the formula:

[0162] ;

[0163] Calculate the phase adjustment amount and obtain the phase adjustment parameter;

[0164] in, represents the phase adjustment amount, Representative The phase value at the frequency point, Representative The group delay value at the frequency point, Representative The adjustment coefficient of the frequency point, represents the total number of frequency points involved in the calculation, and g represents the frequency point index;

[0165] First, the phase change trend is extracted based on the group delay data of multiple frequency points. The corresponding method is to calculate the phase increment of adjacent frequency points. , and further calculate the second-order rate of change , calculated as follows:

[0166] ;

[0167] in, For the The phase value at each frequency point, is the group delay value at this frequency point, is the adjustment coefficient, which is used to measure the impact of this frequency point on the overall trend. Or adaptively adjust based on signal quality. If the signal-to-noise ratio of a certain frequency point is low, reduce its weight, such as setting To reduce its impact, set Perform calculations: Frequency points , , , , corresponding to the group delay ns, ns, ns, ns

[0168] Substituting into the formula:

[0169] ;

[0170] ;

[0171] The result shows that the calculated phase adjustment amount reflects the overall offset of the phase change trend. The larger the value, the higher the degree of non-uniformity of the phase change. This value will be used as a key adjustment parameter in the subsequent phase compensation process to correct the phase deviation and improve the signal consistency.

[0172] The phase correction calculation sub-module calls the phase adjustment parameter, selects the compensation strategy, adjusts the phase of the normalized amplitude signal, and calculates and obtains the phase compensation signal.

[0173] To adjust the phase of the normalized amplitude signal, the specific operation is to compensate the phase of the input signal:

[0174] ;

[0175] where is the compensation factor, which is set according to the system performance. For example, for high-precision applications, is set, while for general scenarios, is set, and the compensation calculation is performed using : - - - - ;

[0176] The adjusted phase compensation signal is , that is, .

[0177] This result shows that after the phase compensation calculation, the original phase data is corrected, making the phase change smoother, reducing the non-linear phase error caused by the group delay change. The adjusted phase value will provide a more stable phase response during the signal transmission and processing, ensuring the stable system performance.

[0178] Please refer to Figure 6 , the dynamic feedback adjustment module includes:

[0179] The offset compensation calculation sub-module obtains the phase compensation signal, collects the offset compensation value and the amplitude coefficient, calculates the normalized ratio between the offset compensation value and the amplitude coefficient, obtains the offset change rate by comparing with the previous data, and at the same time screens the parameters with the offset change rate exceeding the limit. After removing the out-of-limit data, the offset compensation equilibrium value is calculated;

[0180] First, receive the phase compensation signal, and obtain the measured offset compensation value and amplitude coefficient in the current system. The offset compensation value can be obtained by collecting the signal amplitude in the time domain through a high-precision sampler, while the amplitude coefficient is determined by calculating the ratio of the maximum value to the minimum value of the signal. Then, perform a normalization calculation on the offset compensation value and the amplitude coefficient. The calculation method of the normalization ratio is to divide the offset compensation value by the amplitude coefficient, so as to ensure that signal offsets at different amplitudes can be compared on the same scale, thereby improving the stability of the calculation. For example, if the current offset compensation value is 0.03V and the amplitude coefficient is 1.2V, the normalization ratio is calculated as follows:

[0181] ;

[0182] Subsequently, the system extracts the past normalization ratio sequence from the historical data and calculates its change rate. The calculation method of the change rate is the difference between the current normalization ratio and the normalization ratio at the previous moment, divided by the normalization ratio at the previous moment. For example, if the normalization ratio at the previous moment is 0.028, the change rate is calculated as follows:

[0183] ;

[0184] After obtaining the change rate, the system further screens out the out-of-limit parameters, that is, determines whether the change rate exceeds the set threshold. Assuming that the threshold range is set to ±0.1, the current change rate of -0.107 exceeds the allowable range, so this data will be excluded. After excluding the out-of-limit data, the system calculates the average value of the remaining data to obtain the offset compensation equilibrium value. Assuming that the remaining normalization ratios after excluding the abnormal values are , the equilibrium value is calculated as follows:

[0185] ;

[0186] This result indicates that the offset compensation value of the current system remains relatively stable within the historical data range, and after the normalization ratio calculation, its equilibrium value still maintains at the level of 0.025. This means that in this step, through the screening of historical data and the calculation of the normalization ratio, a relatively reliable offset compensation reference value is obtained, providing a basis for the subsequent signal-to-noise ratio trend analysis.

[0187] The signal-to-noise ratio trend analysis sub-module extracts the phase coefficient based on the offset compensation equilibrium value, calculates the current signal-to-noise ratio, calls the signal-to-noise ratio data, calculates the change rate of the current signal-to-noise ratio, and at the same time obtains the signal-to-noise ratio gradient trend based on the change rate, and calculates the error amplitude according to the gradient trend to generate the signal-to-noise ratio change trend;

[0188] The specific method is to perform a Fourier transform on the signal, obtain its phase information, and calculate the offset angle relative to the reference signal. For example, if the calculated phase offset angle is 5 degrees, the phase coefficient can be expressed as and , namely:

[0189] ;

[0190] ;

[0191] Subsequently, based on the phase coefficient, the current signal-to-noise ratio is calculated. The method for calculating the signal-to-noise ratio is the ratio of the effective signal power to the noise power, namely:

[0192] ;

[0193] Assuming the signal power is 2.5 mW and the noise power is 0.1 mW, the signal-to-noise ratio is calculated as follows:

[0194] ;

[0195] Then, the system calls the historical signal-to-noise ratio data to calculate the current change rate of the signal-to-noise ratio. The calculation method of the change rate is the difference between the current signal-to-noise ratio and the signal-to-noise ratio at the previous moment, and then divided by the signal-to-noise ratio at the previous moment. Assuming the signal-to-noise ratio at the previous moment is 13.8 dB, the change rate is calculated as follows:

[0196] ;

[0197] Finally, based on the change rate, the signal-to-noise ratio gradient trend is calculated. The calculation method of the gradient trend is the derivative of the change rate with respect to time, namely:

[0198] ;

[0199] Assuming , then

[0200] ;

[0201] This result indicates that the change rate of the current signal-to-noise ratio per second is relatively low, meaning that the change of the current signal quality of the system is relatively stable. In this step, the calculated signal-to-noise ratio gradient trend provides input parameters for subsequent dynamic optimization compensation, ensuring that the compensation adjustment can accurately track the change trend of the signal quality.

[0202] The dynamic optimization compensation generation sub-module calls the signal-to-noise ratio change trend, compares the error gradient, calculates the optimization update step size, eliminates out-of-range parameters, and uses the formula:

[0203] ;

[0204] Calculate the dynamic optimization coefficient, adjust the phase compensation signal, and generate the dynamic optimization compensation signal;

[0205] Among them, Represents the calculation result of the dynamic optimization compensation signal, Represents the adjustment coefficient, Represents the derivative of the signal-to-noise ratio change rate with respect to time, Represents the out-of-range error parameter, Represents the compensation constant.

[0206] Based on the signal-to-noise ratio change trend, calculate the optimized update step size, eliminate the out-of-range parameters, and the calculation formula for the optimized update step size is:

[0207] ;

[0208] Among them, the adjustment coefficient , the maximum signal-to-noise ratio change rate , the out-of-range error parameter , the compensation constant , substitute into the calculation to get:

[0209] ;

[0210] This result indicates that the update step size of the current optimized compensation signal is 3.0, which means that in this step, the system can dynamically adjust the phase compensation signal by calculating the update step size to match the current signal-to-noise ratio change trend, ensure the stability of the signal quality, and the calculated update step size will be used for the generation of the final dynamic optimization compensation signal and continuously adjust the compensation amount to keep the signal-to-noise ratio stable within the expected range.

[0211] The above is only the preferred embodiment of the present invention, and it does not limit the present invention in other forms. Any person skilled in the art may use the disclosed technical content 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 content of the present invention's technical solution, 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 present invention's technical solution.

Claims

1. A multi-band processing system for electronic communications, characterized in that: The system comprises: The signal separation module obtains the input signal, extracts multiple frequency bands, performs bandpass filtering, separates the components of each frequency band, calculates the short-time power spectrum density, finds the normalization coefficient based on the spectrum energy, obtains the instantaneous power value, determines the frequency domain distribution, calculates the normalization ratio, verifies the signal balance state, filters the high-energy components, and generates a multi-band normalized signal; The signal separation module comprises: The signal bandpass filtering submodule obtains the input signal, extracts multiple preset frequency bands, calls the filter to perform bandpass filtering, adjusts the filtering parameters according to the frequency range, adjusts the filter order, calculates the amplitude spectrum of the filtered signal, and compares the signal spectrum before and after filtering, extracts the energy distribution of the multi-band signal, and obtains the multi-band signal components; The short-time power spectrum calculation submodule calculates the short-time power spectrum density based on the multi-band signal components, obtains the signal power value in each time window, and determines the spectrum energy distribution using the formula: ； Obtain instantaneous power value through calculation, determine the energy proportion of differentiated frequency bands, and obtain spectrum energy distribution; in, Represents the instantaneous power value, Representative The signal components of the frequency band, Representative The center frequency of the band, Represents the number of frequency bands, represents the sum of the center frequencies used for normalization calculation, a represents the sub-band index, and b represents the sub-band index used for normalization calculation; The multi-band normalization submodule calculates the normalization ratio based on the spectrum energy distribution, adjusts the multi-band signal power, and calculates the normalized signal amplitude distribution, and compares the signal power difference before and after the adjustment, selects the frequency band with high energy proportion, synthesizes the selected signal, and obtains the multi-band normalized signal; The frequency offset compensation module collects adjacent phases, calculates the phase change rate, calculates the mean of continuous frames, removes abnormal fluctuations, compares the frequency offset between frames, obtains the compensation factor, adjusts the time domain coefficient, and generates a time domain offset compensation signal based on the multi-band normalized signal; The amplitude correction module collects short-time Fourier transform data based on the time domain offset compensation signal, extracts power spectrum density, calculates attenuation factor, analyzes amplitude range, selects maximum amplitude, normalizes amplitude value, and generates normalized amplitude signal; Based on the normalized amplitude signal, the phase correction module collects group delay data, extracts phase trend, fits the error curve, determines the second-order change rate, selects compensation parameters, adjusts the phase, and generates a phase compensation signal.

2. The multi-band processing system for electronic communication according to claim 1, characterized in that: The multi-band normalized signal includes a low-frequency normalized output, a medium-frequency normalized output, and a high-frequency normalized output; the time domain offset compensation signal includes a phase smoothness value, a time domain correction amount, and an inter-frame equalization index; the normalized amplitude signal includes an amplitude standard amount, an amplitude equalization index, and an amplitude reference value; the phase compensation signal includes a phase correction value, a trend balance amount, and an error correction index.

3. The multi-band processing system for electronic communication according to claim 2, characterized in that: The frequency offset compensation module comprises: The phase change rate calculation submodule collects adjacent phases based on the multi-band normalized signal, calculates the phase change rate, extracts the phase change rate data of continuous frames, and calculates the mean to obtain the phase change rate mean; The abnormal fluctuation elimination submodule calls the phase change rate mean, counts the change range of continuous frames, eliminates abnormal fluctuation data, calculates the inter-frame frequency deviation, and obtains the frequency deviation data after eliminating the abnormal fluctuation; The time domain compensation calculation submodule calls the frequency deviation data after removing abnormal fluctuations, compares the frequency deviation between frames, obtains the compensation factor, and adjusts the time domain coefficient using the formula: ； Obtaining a time domain offset compensation factor by operation, adjusting the time domain signal, and generating a time domain offset compensation signal; in, represents the time domain offset compensation factor, Representative The frequency deviation value of the frame, Representative The weight parameters of the frame, Representative The phase value of the frame, Representative The signal amplitude of the frame, Representative The time coefficient of the frame, Represents the total number of frames. represents the frame number of phase change calculation, Represents the frame number of signal amplitude and time coefficient statistics, and c, d, and e represent frame indexes.

4. The multi-band processing system for electronic communication according to claim 3, characterized in that: The amplitude correction module comprises: The short-time Fourier transform submodule calculates the frequency component of the signal in the short-time window based on the time domain offset compensation signal, intercepts the signal sequence using a sliding window, performs Fourier transform on the signal in each window, calculates the complex spectrum value after the transform, obtains the spectrum information on each time segment, and calls the modulus value of the transformed spectrum at the same time to generate short-time Fourier transform data; The power spectrum density extraction submodule calls the short-time Fourier transform data, calculates the power values ​​corresponding to the multi-frequency components, normalizes the power spectra of all time segments, calculates the power density curve by integration, obtains the energy distribution, and screens the key power distribution within the characteristic frequency band to generate power spectrum density data; The amplitude normalization submodule calls the power spectrum density data to calculate the attenuation factor of the signal within the characteristic frequency band using the formula: ； Calculate the weighted amplitude of each frequency component, extract the amplitude range, filter the maximum amplitude value, and perform normalization to obtain a normalized amplitude signal; in, Represents the amplitude value of the current frequency component, represents the attenuation factor, , represents the phase angle, represents the bias term associated with the frequency component, represents the total number of frequency components, represents the cosine value of the phase angle, and f represents the frequency component index.

5. The multi-band processing system for electronic communication according to claim 4, characterized in that: The phase correction module comprises: The group delay acquisition submodule collects the group delay data of the signal at the differentiated frequency based on the normalized amplitude signal, calculates the group delay values ​​corresponding to multiple frequency points, and sorts the data to filter out the noise interference data to obtain the effective group delay data; The phase trend analysis submodule calls the effective group delay data, extracts the phase change trend based on the group delay data of multiple frequency points, calculates the second-order change rate, analyzes the error curve, and uses the formula: ； Calculate the phase adjustment amount and obtain the phase adjustment parameter; in, represents the phase adjustment amount, Representative The phase value at the frequency point, Representative The group delay value at the frequency point, Representative The adjustment coefficient of the frequency point, represents the total number of frequency points involved in the calculation, and g represents the frequency point index; The phase correction calculation submodule calls the phase adjustment parameter, selects a compensation strategy, performs phase adjustment on the normalized amplitude signal, and calculates and obtains a phase compensation signal.

6. The multi-band processing system for electronic communication according to claim 5, characterized in that: The system further comprises: The dynamic feedback adjustment module collects the offset compensation value and the amplitude coefficient based on the phase compensation signal, extracts the phase coefficient, calculates the signal-to-noise ratio trend, compares the error gradient in the past, determines the update step size, filters out-of-range parameters, and generates a dynamic optimization compensation signal; The dynamic optimization compensation signal includes feedback correction parameters, error optimization amount, and signal-to-noise balance index.

7. The multi-band processing system for electronic communication according to claim 6, characterized in that: The dynamic feedback adjustment module comprises: The offset compensation calculation submodule obtains the phase compensation signal, collects the offset compensation value and the amplitude coefficient, calculates the normalized ratio between the offset compensation value and the amplitude coefficient, compares the past data to obtain the offset change rate, and simultaneously screens the offset change rate exceeding the limit parameter, removes the exceeding limit data, and calculates the offset compensation equilibrium value; The signal-to-noise ratio trend analysis submodule extracts the phase coefficient based on the offset compensation equalization value, calculates the current signal-to-noise ratio, calls the signal-to-noise ratio data, calculates the current signal-to-noise ratio change rate, obtains the signal-to-noise ratio gradient trend based on the change rate, calculates the error amplitude according to the gradient trend, and generates the signal-to-noise ratio change trend; The dynamic optimization compensation generation submodule calls the signal-to-noise ratio change trend, compares the error gradient, calculates the optimization update step, removes out-of-range parameters, and uses the formula: ； Calculating a dynamic optimization coefficient, adjusting the phase compensation signal, and generating a dynamic optimization compensation signal; in, represents the calculation result of the dynamic optimization compensation signal, represents the adjustment factor, represents the derivative of the rate of change of the signal-to-noise ratio with respect to time, represents the out-of-range error parameter, Represents the compensation constant.

Images