Signal frequency offset compensation method and device, and electronic device

Abstract

This invention provides a signal frequency offset compensation method, apparatus, and electronic device, belonging to the field of optical communication technology. The method includes: acquiring the spectral structure of the transmitted signal; dividing the spectral structure to obtain multiple local symmetric units; acquiring the unbiased power spectrum of the local symmetric units; performing centroid calculation on the unbiased power spectrum to obtain a reference centroid position; acquiring the power spectrum of the received signal; performing centroid calculation and difference processing based on the power spectrum and the local symmetric units to obtain a local frequency offset prediction value; acquiring the pre-calibrated local frequency offset estimation error variance of the local symmetric units; performing weight calculation based on the local frequency offset estimation error variance to obtain a fusion weight; performing weighted fusion processing on each local frequency offset prediction value and its corresponding fusion weight to obtain a global frequency offset prediction value; and performing phase rotation on the received signal to obtain a frequency offset compensation signal. The above scheme can effectively achieve accurate correction of the received signal frequency offset.

Description

Technical Field

[0001] This invention relates to the field of optical communication technology, and more specifically to signal frequency offset compensation methods, devices, and electronic equipment. Background Technology

[0002] In digital subcarrier multiplexed optical communication systems, carrier frequency deviation is common during system operation because it is difficult to maintain a perfectly synchronized center frequency between the transmitting laser and the receiving local oscillator laser. Frequency deviation introduces time-varying phase fluctuations, leading to incorrect phase decisions. Therefore, accurate estimation and compensation of the frequency deviation are necessary. Existing technologies typically employ a method of estimating the overall frequency deviation across the entire received signal spectrum to obtain the value, and then compensate the signal based on this value.

[0003] However, this method of estimating the overall frequency offset of the entire signal spectrum is highly susceptible to interference from local spectral distortions or sudden noise in practical applications. When a local area of ​​the spectrum is subjected to strong noise interference, the overall estimated frequency offset value will have a large deviation, resulting in poor robustness. Secondly, in the actual spectrum structure, the degree of influence of noise and impairment varies in different frequency sub-intervals, and the reliability of their frequency offset estimates also varies. Existing overall estimation methods cannot distinguish the differences in the reliability of frequency offset estimates in each frequency sub-interval. The simple and crude overall processing masks the high-precision local frequency offset information, resulting in low frequency offset estimation accuracy, which in turn causes inaccurate subsequent phase rotation compensation and fails to effectively eliminate time-varying phase fluctuations.

[0004] Therefore, there is an urgent need for a high-precision frequency offset compensation method that can fully utilize local spectral features and distinguish the reliability of different local frequency offset estimates, so as to improve the robustness and accuracy of frequency offset estimation and achieve accurate phase compensation. Summary of the Invention

[0005] To address the aforementioned problems, this application provides a signal frequency offset compensation method to improve the accuracy and robustness of the global frequency offset prediction value, thereby achieving precise correction of the received signal frequency offset. The technical solution is as follows: In a first aspect, the present invention provides a signal frequency offset compensation method, comprising: The spectral structure of the transmitted signal is obtained, and the spectral structure is divided into multiple locally symmetric units, which are frequency sub-intervals with local symmetry characteristics. Obtain the unbiased power spectrum of the locally symmetric unit, perform centroid calculation on the unbiased power spectrum, and obtain the reference centroid position; The power spectrum of the received signal is obtained, and the centroid is calculated and the difference is processed based on the power spectrum and the local symmetric unit to obtain the local frequency offset prediction value. The local frequency offset estimation error variance of the local symmetric unit is obtained in advance, and the weight is calculated based on the local frequency offset estimation error variance to obtain the fusion weight. The local frequency offset prediction values ​​are weighted and fused with the corresponding fusion weights to obtain the global frequency offset prediction value. The received signal is then phase-rotated based on the global frequency offset prediction value to obtain the frequency offset compensation signal.

[0006] Combining the first aspect and the above implementation methods, in some possible implementation methods, the spectral structure is structurally partitioned to obtain multiple locally symmetric units, including: Obtain the baseband signal from the spectral structure; The adjacent subcarrier spectra of the baseband signal are overlapped and divided to obtain locally symmetrical units.

[0007] Combining the first aspect and the above implementation methods, in some possible implementation methods, the centroid of the unbiased power spectrum is calculated to obtain the reference centroid position, including: The reference centroid position is obtained by multiplying the frequency value of the frequency sampling point in the unbiased power spectrum with the power spectrum value, summing the products, and dividing by the sum of the power spectrum values.

[0008] Combining the first aspect and the above implementation methods, in some possible implementation methods, the local frequency offset prediction value is obtained by calculating the centroid and performing difference processing based on the power spectrum and local symmetric elements, including: The receiving centroid position is obtained by multiplying and summing the frequency values ​​of the power spectrum at the frequency sampling points within the local symmetric unit with the power spectrum value and then normalizing the result. The local frequency offset prediction value is obtained by subtracting the received centroid position from the reference centroid position.

[0009] Combining the first aspect and the above implementation methods, in some possible implementation methods, the local frequency offset estimation error variance of the local symmetric unit is obtained in advance, and the weights are calculated based on the local frequency offset estimation error variance to obtain the fusion weights, including: Obtain the known true frequency offset value and the local frequency offset estimation result during offline simulation and / or calibration. The known true frequency offset value is the preset true frequency offset value, and the local frequency offset estimation result is the value obtained by frequency offset estimation of the local symmetric element under the known true frequency offset value. The variance of the local frequency offset estimation error is obtained by summing the squares and taking the mean of the differences between the known true frequency offset value and the local frequency offset estimation result. The fusion weights are obtained by calculating the inverse and proportion of the variance of the local frequency offset estimation error.

[0010] Combining the first aspect and the above implementation methods, in some possible implementation methods, the difference between the known true frequency offset value and the local frequency offset estimation result is summed by squares and averaged to obtain the local frequency offset estimation error variance, including: The variance of the local frequency offset estimation error is obtained by summing the squares of the differences between the known true frequency offset value and the local frequency offset estimation result, dividing by the number of simulations, and then averaging the results.

[0011] Combining the first aspect and the above implementation methods, in some possible implementation methods, weight allocation is performed based on the variance of the local frequency offset estimation error to obtain fusion weights, including: The initial weights of each local symmetric element are determined based on the reciprocal of the sum of the local frequency offset estimation error variance and the zero-prevention constant. The initial weights of each local symmetric unit are normalized proportionally to obtain the fused weights.

[0012] Combining the first aspect and the above implementation methods, in some possible implementation methods, the received signal is phase-rotated based on the global frequency offset prediction value to obtain a frequency offset compensation signal, including: The received signal is multiplied by the negative phase exponent of the global frequency offset prediction value to obtain the frequency offset compensation signal.

[0013] Secondly, the present invention also provides a signal frequency offset compensation device, comprising: The data acquisition unit is used to acquire the spectral structure of the transmitted signal, divide the spectral structure into multiple locally symmetric units, and each locally symmetric unit is a frequency sub-interval with local symmetry characteristics. The data processing unit is used to obtain the unbiased power spectrum of the local symmetric unit, perform centroid calculation on the unbiased power spectrum, and obtain the reference centroid position. The frequency offset prediction unit is used to obtain the power spectrum of the received signal, and to perform centroid calculation and difference processing based on the power spectrum and the local symmetry unit to obtain the local frequency offset prediction value. The weight acquisition unit is used to acquire the local frequency offset estimation error variance of the local symmetric unit pre-calibrated, and to perform weight calculation processing based on the local frequency offset estimation error variance to obtain the fusion weight. The signal compensation unit is used to perform weighted fusion processing on each local frequency offset prediction value and the corresponding fusion weight to obtain the global frequency offset prediction value. Based on the global frequency offset prediction value, the phase of the received signal is rotated to obtain the frequency offset compensation signal.

[0014] Thirdly, the present invention also provides an electronic device, which includes a data acquisition unit, a memory, and a processor, wherein... The data acquisition unit is used to acquire the spectral structure of the transmitted signal; acquire the unbiased power spectrum of the local symmetric unit; acquire the power spectrum of the received signal; and acquire the variance of the local frequency offset estimation error pre-calibrated for the local symmetric unit. Memory, used to store executable program code; The processor is used to divide the spectral structure into multiple local symmetric units; to perform centroid calculation on the unbiased power spectrum to obtain the reference centroid position; to perform centroid calculation and difference processing based on the power spectrum and local symmetric units to obtain the local frequency offset prediction value; to perform weight calculation based on the variance of the local frequency offset estimation error to obtain the fusion weight; to perform weighted fusion processing to obtain the global frequency offset prediction value; and to perform phase rotation on the received signal based on the global frequency offset prediction value to obtain the frequency offset compensation signal.

[0015] Fourthly, the present invention also provides a computer-readable storage medium for storing a computer-readable program or instruction, which, when executed by a processor, can implement the steps of the signal frequency offset compensation method in any of the above implementations.

[0016] The beneficial effects of this invention are as follows: By acquiring the spectral structure of the transmitted signal and dividing the spectral structure into multiple local symmetric units, the defect of the overall frequency offset estimation being susceptible to local distortion interference is overcome; by acquiring the unbiased power spectrum of the local symmetric units and performing centroid calculation on the unbiased power spectrum to obtain the reference centroid position, and by acquiring the power spectrum of the received signal and performing centroid calculation and difference processing based on the power spectrum and the local symmetric units, a local frequency offset prediction value is obtained, realizing refined extraction of frequency offset in frequency sub-intervals; by acquiring the pre-calibrated local frequency offset estimation error variance of the local symmetric units and performing weight calculation based on the local frequency offset estimation error variance to obtain a fusion weight, effectively distinguishing the reliability differences of different local symmetric units; by performing weighted fusion processing on each local frequency offset prediction value and the corresponding fusion weight to obtain a global frequency offset prediction value, and by performing phase rotation on the received signal based on the global frequency offset prediction value to obtain a frequency offset compensation signal, the accuracy and robustness of the global frequency offset prediction value are improved, and accurate frequency offset compensation is achieved. Attached Figure Description To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating the signal frequency offset compensation method provided by the present invention. Figure 2 For the present invention Figure 1A schematic diagram of an embodiment of S103; Figure 3 For the present invention Figure 1 A schematic diagram of an embodiment of S104; Figure 4 For the present invention Figure 1 A schematic diagram of an embodiment of S104; Figure 5 This is a schematic diagram of the signal frequency offset compensation device provided by the present invention; Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0019] In the description of the embodiments of the present invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0020] The terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.

[0021] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0022] During signal transmission in an optical fiber link, a sudden burst of strong noise may occur, interfering with the spectrum of a certain frequency sub-interval. In this case, the traditional global frequency offset estimation will include all the spectrum data of the distorted area in the unified calculation. The distorted data directly deflects the overall frequency offset solution, resulting in a significant error in the final output global frequency offset estimate. The algorithm has poor robustness against interference.

[0023] Meanwhile, after the optical fiber transmission spectrum is divided into multiple non-overlapping frequency sub-intervals, the degree of influence of noise, nonlinearity, and polarization loss on each sub-interval is not uniform. That is, some sub-intervals are located in the flat region of the spectrum, with high signal-to-noise ratio and high reliability of frequency offset observation results; while the sub-intervals at the edge and in the in-band concave region have low signal-to-noise ratio and large frequency offset estimation errors. However, the traditional overall frequency offset estimation scheme uses indiscriminate weighted fusion calculation for all frequency sub-intervals, which cannot distinguish the differences in the reliability of frequency offset estimation results between different sub-intervals. The observation data of low-precision damaged sub-intervals will dilute and mask the high-precision frequency offset information output by high signal-to-noise ratio sub-intervals, directly reducing the overall frequency offset estimation accuracy.

[0024] The frequency offset estimation error will further cause the phase compensation operation at the receiver to be unable to completely offset the time-varying phase rotation introduced by the transmission. The residual phase fluctuation will continue to interfere with the demodulation process, ultimately causing a significant increase in the bit error rate of the received signal.

[0025] To address the aforementioned technical problems, this application provides a signal frequency offset compensation method. The subject of this method is a signal frequency offset compensation device or an electronic device equipped with a signal frequency offset compensation device. The following provides a detailed description. It should be noted that the order of description of the following embodiments is not intended to limit the preferred order of the embodiments.

[0026] Please see Figure 1 , Figure 1 This is a schematic flowchart of the signal frequency offset compensation method provided in an embodiment of this application. Figure 1 As shown, the method in this application embodiment may include the following steps S101-S105: S101: Obtain the spectral structure of the transmitted signal, divide the spectral structure into multiple locally symmetric units.

[0027] In this embodiment, the transmitted signal refers to the transmitter signal generated by the transmitter under ideal conditions without carrier frequency deviation. It is formed by shifting and multiplexing multiple subcarrier signals through their corresponding center frequencies and is used to provide a local frequency-bias-free reference. The spectral structure refers to the power or amplitude distribution of the transmitted signal in the frequency domain, including the spectral shape of each subcarrier, its center frequency position, and the distribution of overlapping regions between adjacent subcarrier spectra. A locally symmetric unit refers to a frequency sub-interval whose spectral power distribution is geometrically symmetrical about the center position of the interval under frequency-bias-free conditions, obtained by dividing the overlapping region of adjacent subcarrier spectra.

[0028] In one feasible implementation, the transmitter first generates a binary pseudo-random sequence, forms an M-way symbol sequence to be loaded through Gray mapping, and performs upsampling and pulse shaping filtering respectively. The actual baud rate used can be, for example, 56 Gbaud.

[0029] Each subcarrier is shifted to its corresponding center frequency and multiplexed to form the transmitted signal. Under no frequency offset conditions, spectral analysis is performed on the generated transmitted signal to obtain its spectral structure. The spectra of adjacent subcarriers are overlapped and divided, with each unit exhibiting local symmetry. This division results in N locally symmetric units, denoted as . .

[0030] For example, the transmitted signal in the above steps The specific form of the formula can be:

[0031] Among them, formula parameters Representing the The baseband signal of each subcarrier, formula parameters Representing the The center frequency of each subcarrier, formula parameters Represents the sampling period, formula parameters This represents the total number of subcarriers.

[0032] S102, obtain the unbiased power spectrum of the local symmetric unit, perform centroid calculation on the unbiased power spectrum, and obtain the reference centroid position.

[0033] In this embodiment, the unbiased power spectrum refers to the power spectrum value of the transmitted signal at each frequency sampling point under unbiased conditions. Centroid calculation processing refers to the process of calculating the center frequency by weighting the frequency values ​​of the frequency sampling points using the power spectrum value as the weight. The reference centroid position refers to the frequency coordinates of the spectral centroid within a locally symmetrical unit under unbiased conditions.

[0034] In one feasible implementation, for the first For each locally symmetric element, the power spectrum of the transmitted signal under no-frequency-offset conditions is obtained as the no-frequency-offset power spectrum. The frequency values ​​of the frequency sampling points in the no-frequency-offset power spectrum are multiplied and summed, and then divided by the sum of the no-frequency-offset power spectrum values ​​for normalization. The reference centroid position of the locally symmetric element is then calculated.

[0035] For example, the reference centroid position in the above steps The specific form of the formula can be:

[0036] Among them, formula parameters Representing the Locally symmetric elements, formula parameters Represents the frequency sampling points within a locally symmetric unit, formula parameters Represents the frequency sampling point under no frequency offset condition The power spectral density at that location.

[0037] S103: Obtain the power spectrum of the received signal, and perform centroid calculation and difference processing based on the power spectrum and local symmetric units to obtain the local frequency offset prediction value.

[0038] In this embodiment, the received signal refers to the digital subcarrier multiplexed signal containing the frequency offset to be estimated, which is generated by the signal transmitter sending the transmitted signal to the optical fiber transmission link and then experiencing frequency offset due to factors such as the frequency difference between the lasers at the transmitting and receiving ends during optical fiber transmission. The power spectrum refers to the power spectrum distribution of the received signal at the frequency sampling point. The local frequency offset prediction value refers to the frequency offset difference between the receiving centroid position and the reference centroid position calculated within a single local symmetric cell.

[0039] In one feasible implementation, the transmitter sends the modulated transmit signal to the fiber optic transmission link. During fiber optic transmission, this signal experiences frequency offset due to factors such as inconsistencies in the frequencies of the local oscillator lasers at the transmitting and receiving ends. The receiver, in actual operation, receives this optical signal through the fiber optic transmission link and converts it into an electrical domain receive signal using a coherent reception method. A Fast Fourier Transform (FFT) is performed on the received signal to obtain its power spectrum. To reduce the impact of noise disturbances on the centroid position calculation, the power spectrum can be smoothed using a rectangular window, moving average window, or other smoothing filters. For the first... A locally symmetric unit, in the corresponding frequency range The receiving centroid position is calculated internally, and then subtracted from the reference centroid position to obtain the first centroid position. Local frequency offset prediction values ​​for a locally symmetric unit.

[0040] For example, the above-mentioned receiving centroid location The specific form of the formula can be:

[0041] Among them, formula parameters Represents the received signal at the frequency sampling point The power spectral density at that location.

[0042] For example, the local frequency offset prediction value in the above steps The specific form of the formula is:

[0043] Among them, formula parameters Representing the The receiving centroid position of each locally symmetric unit, formula parameters Representing the The reference centroid position of a locally symmetric unit.

[0044] S104: Obtain the local frequency offset estimation error variance of the local symmetric unit pre-calibrated, and perform weight calculation based on the local frequency offset estimation error variance to obtain the fusion weight.

[0045] In this embodiment, the local frequency offset estimation error variance refers to the statistical variance of the error obtained by multiple estimations of a local symmetric element under known true frequency offset conditions. Weight calculation refers to the process of allocating weights to each local symmetric element based on the magnitude of the local frequency offset estimation error variance. The fusion weight refers to the weight coefficient of each local symmetric element participating in the weighted fusion in the global frequency offset estimation.

[0046] In one feasible implementation, under the condition of known true frequency offset, the variance of the local frequency offset estimation error of each local symmetric element is statistically analyzed through offline simulation or calibration. A set of known true frequency offset values ​​is set, and the local frequency offset estimation results of the i-th local symmetric element under each known true frequency offset value are statistically analyzed. The sum and mean of the squares of the differences between the local frequency offset estimation results and the known true frequency offset values ​​are calculated to obtain the variance of the local frequency offset estimation error. The fusion weight is determined based on the variance of the local frequency offset estimation error of each local symmetric element; the smaller the variance of the local frequency offset estimation error, the larger the assigned fusion weight.

[0047] In one feasible implementation, the variance of the local frequency offset estimation error can not only be obtained through pre-calibration, but can also be dynamically updated during system operation to adapt to time-varying channels.

[0048] Specifically, a sliding time window can be established to obtain the historical local frequency offset prediction values ​​of each local symmetric unit within multiple consecutive time slots prior to the current moment. The statistical variance of the historical local frequency offset prediction values ​​within the sliding time window is then calculated, and this statistical variance is used as the local frequency offset estimation error variance of the corresponding local symmetric unit at the current moment. This dynamic variance update mechanism enables the fusion weights to adapt to the time-varying characteristics of channel conditions in real time. When a sudden channel change causes a certain sub-interval to deteriorate, the fusion weights are adaptively reduced; when the channel is stable, the fusion weights of high signal-to-noise ratio sub-intervals are increased, further improving the tracking accuracy and robustness of the global frequency offset prediction value in dynamic scenarios.

[0049] For example, the variance of the local frequency offset estimation error in the above steps The specific form of the formula can be:

[0050] Among them, formula parameters Represents the total number of known true frequency offset values, formula parameters The index representing the known true frequency offset value, formula parameters Representing the Given a known true frequency offset value, formula parameters Representing the The local symmetric unit in the th... Local frequency offset estimation results under known true frequency offset values.

[0051] For example, the fusion weights in the above steps The specific form of the formula can be:

[0052] Among them, formula parameters Representing the The variance of the local frequency offset estimation error for a locally symmetric element, and the formula parameters. Represents a positive constant to prevent the denominator from being zero; formula parameters. This represents the total number of locally symmetric units.

[0053] S105, weighted fusion of each local frequency offset prediction value with the corresponding fusion weight to obtain the global frequency offset prediction value, and phase rotation of the received signal based on the global frequency offset prediction value to obtain the frequency offset compensation signal.

[0054] In this embodiment, weighted fusion processing refers to the process of multiplying the local frequency offset prediction values ​​of each local symmetric unit by their corresponding fusion weights and then summing the results. The global frequency offset prediction value refers to the final system frequency offset estimate obtained after weighted fusion of the various local frequency offset prediction values. Phase rotation refers to the operation of reverse compensation for the phase offset of the received signal using the global frequency offset prediction value. The frequency offset compensation signal refers to the received signal after frequency offset is eliminated.

[0055] In one feasible implementation, the local frequency offset prediction values ​​of all local symmetric units are weighted and fused using calibrated fusion weights. The products of each local frequency offset prediction value and its corresponding fusion weight are summed to obtain the global frequency offset prediction value. The received signal is then phase-rotated based on this global frequency offset prediction value to compensate for the frequency offset, resulting in a frequency offset-compensated signal. After frequency offset compensation, subsequent digital signal processing (DSP) is performed, including polarization demultiplexing, subcarrier demultiplexing, resampling, dispersion compensation, matched filtering, adaptive equalization, carrier phase recovery, and demapping.

[0056] For example, the global frequency offset prediction value in the above steps The specific form of the formula can be:

[0057] Among them, formula parameters Representing the The fusion weights of local symmetric units, formula parameters Representing the Local frequency offset prediction values ​​for a locally symmetric element, formula parameters This represents the total number of locally symmetric units.

[0058] For example, the frequency offset compensation signal in the above steps The specific form of the formula can be:

[0059] Among them, formula parameters Represents the original received signal, formula parameters Represents the global frequency offset prediction value, formula parameters This represents the sampling period.

[0060] In summary, this application overcomes the vulnerability of overall frequency offset estimation to local distortion interference by obtaining the spectral structure of the transmitted signal, dividing the spectral structure into multiple local symmetric units, and acquiring the unbiased power spectrum of the local symmetric units. It then performs centroid calculation on the unbiased power spectrum to obtain the reference centroid position, and acquires the power spectrum of the received signal. Based on the power spectrum and the local symmetric units, it performs centroid calculation and difference processing to obtain local frequency offset prediction values, achieving refined extraction of frequency offset in frequency sub-intervals. Furthermore, it obtains the pre-calibrated local frequency offset estimation error variance of the local symmetric units, performs weight calculation based on the local frequency offset estimation error variance, and obtains fusion weights, effectively distinguishing the reliability differences of different local symmetric units. Finally, it performs weighted fusion processing on each local frequency offset prediction value and its corresponding fusion weight to obtain a global frequency offset prediction value. Based on the global frequency offset prediction value, it performs phase rotation on the received signal to obtain a frequency offset compensation signal, improving the accuracy and robustness of the global frequency offset prediction value and achieving accurate frequency offset compensation.

[0061] In one feasible implementation, during the process of dividing the spectrum structure to obtain multiple local symmetric units, the frequency band boundaries of each local symmetric unit can be adaptively determined based on the allocation and configuration information of the digital subcarriers in the transmitted signal and the distribution of power extrema points in the spectrum.

[0062] Specifically, by identifying power troughs in the spectral structure as candidate points for segmentation, and combining the bandwidth interval of the digital subcarriers to screen and merge the candidate points, it is ensured that each segmented local symmetric unit fully contains the corresponding subcarrier spectral features and maintains the symmetry of the structure, thereby enhancing the anti-interference capability of the local symmetric unit in complex spectral environments and improving the stability of subsequent centroid calculation processing.

[0063] Please see Figure 2 , Figure 2 For the present invention Figure 1 A schematic diagram of an embodiment of S103. (See attached diagram.) Figure 2 As shown, the method in this application embodiment may include the following steps S201-S205: S201, acquire the baseband signal in the spectrum structure.

[0064] In the embodiments of this application, the baseband signal refers to the modulated subcarrier signals that have not undergone spectrum shifting.

[0065] In one feasible implementation, the transmitter divides the modulated symbol sequence into M groups of sequences to be loaded, performs a 2x upsampling, and generates M subcarrier signals in the baseband as baseband signals, so that the baseband signals can be subsequently spectrum shifted to form digital subcarrier multiplexed signals.

[0066] S202, the adjacent subcarrier spectra of the baseband signal are overlapped and divided to obtain locally symmetrical units.

[0067] In this embodiment, the subcarrier spectrum refers to the spectral distribution of each subcarrier signal after it has been shifted to its corresponding center frequency. Overlapping partitioning refers to the sharing of frequency regions between adjacent sub-intervals when dividing the frequency domain into intervals.

[0068] In one feasible implementation, the generated digital subcarrier multiplexed signal is subjected to spectral analysis. The adjacent subcarrier spectra of the baseband signal are overlapped and divided into groups, so that each unit has local symmetry characteristics, thereby dividing into N locally symmetric units.

[0069] S203, the frequency value of the frequency sampling point in the unbiased power spectrum is multiplied and summed with the power spectrum value, and then normalized by dividing by the sum of the power spectrum values ​​to obtain the reference centroid position.

[0070] In the embodiments of this application, a frequency sampling point refers to a discrete frequency sampling position in spectrum analysis.

[0071] In one feasible implementation, for the first For each locally symmetric unit, under no-frequency-offset conditions, the power spectrum of the transmitted signal is obtained as the no-frequency-offset power spectrum. The frequency values ​​of each frequency sampling point within the locally symmetric unit are then used as the frequency sampling points. With the corresponding unbiased power spectral density value After multiplying, the results are summed, and then the sum is divided by the sum of the unbiased power spectral values ​​of all frequency sampling points within the local symmetric element for normalization. Finally, the reference centroid position of the local symmetric element is calculated. .

[0072] S204, the frequency value of the power spectrum at the frequency sampling point in the local symmetric unit is multiplied and summed with the power spectrum value and then normalized to obtain the position of the receiving centroid.

[0073] In the embodiments of this application, the centroid location refers to the frequency coordinates of the centroid of the spectrum of the received signal within the local symmetric unit.

[0074] In one feasible implementation, a Fast Fourier Transform (FFT) is performed on the received signal to obtain its power spectrum. To reduce the impact of noise disturbances on the centroid position calculation, the power spectrum can be smoothed first using a rectangular window, a moving average window, or other smoothing filters. For the first... Each locally symmetric element samples the frequency values ​​of the power spectrum at each frequency sampling point within that element. With the corresponding power spectral value After multiplying, the results are summed, and then the sum is divided by the sum of the power spectral values ​​of all frequency sampling points within the local symmetry element for normalization. Finally, the receiving centroid position of the local symmetry element is calculated. .

[0075] S205, the received centroid position is subtracted from the reference centroid position to obtain the local frequency offset prediction value.

[0076] In the embodiments of this application, the subtraction process refers to the operation of calculating the algebraic difference between the receiving centroid position and the reference centroid position.

[0077] In one feasible implementation, for the first Each locally symmetric element will calculate the location of the receiving centroid. Compared with the pre-calculated reference centroid position Subtract the two numbers and calculate the difference; this difference is the nth... Local frequency offset prediction value corresponding to each local symmetric unit This reflects the frequency offset within the local symmetric unit, where the local frequency offset prediction value... For the specific calculation process, please refer to step S105 above, which will not be repeated here.

[0078] In summary, this application obtains locally symmetric units by overlapping the spectrum of adjacent subcarriers of the baseband signal. The reference centroid position and the received centroid position are calculated based on the unbiased power spectrum and the received power spectrum, respectively. The two are then subtracted to obtain the local frequency offset prediction value. This utilizes the spectral symmetry of the locally symmetric units to transform the frequency offset estimation into the calculation of the difference in centroid position, thereby reducing computational complexity and improving the reliability of the local frequency offset estimation.

[0079] Please see Figure 3 , Figure 3 For the present invention Figure 1 A schematic diagram of an embodiment of S104. (See attached diagram.) Figure 3 As shown, the method in this application embodiment may include the following steps S301-S303: S301, Obtain the known true frequency offset value and local frequency offset estimation results during offline simulation and / or calibration.

[0080] In this embodiment, offline simulation refers to the process of simulating a communication system using a computer to verify and statistically analyze parameters before the actual transmission system is put into operation. Calibration refers to the process of testing and calibrating on actual hardware by inputting known signals. Known true frequency offset refers to the accurate frequency offset value set manually during offline simulation or calibration. Local frequency offset estimation result refers to the estimated frequency offset value calculated by the centroid difference of a local symmetric element under the condition of a known true frequency offset value.

[0081] In one feasible implementation, during offline simulation and / or calibration, the transmitted signal is coherently received, and a set of known true frequency offset values ​​are set. ,in For each known true frequency offset value, the statistics for the first... Local frequency offset estimation results for each locally symmetric element at each known true frequency offset value .

[0082] S302, the difference between the known true frequency offset value and the local frequency offset estimation result is squared and summed and mean-valued to obtain the variance of the local frequency offset estimation error.

[0083] In the embodiments of this application, summation of squares refers to the calculation operation of squaring the error values ​​and then summing them. Mean value processing refers to the operation of dividing the result of summation of squares by the statistical frequency to obtain the average error level.

[0084] In one feasible implementation, for the first Each locally symmetric element will have a known true frequency offset value. With the corresponding local frequency offset estimation results Subtracting the values ​​yields the estimation error. Squaring and summing the estimation errors, then dividing by the total number of known true frequency offsets and averaging the sum, yields the result for the nth value. Variance of local frequency offset estimation error for a locally symmetric unit This is to quantify the frequency offset estimation accuracy of the local symmetric unit.

[0085] S303, based on the local frequency offset estimation error variance, performs reciprocal and proportional calculations to obtain the fusion weights.

[0086] In the embodiments of this application, the reciprocal and proportion calculation process refers to the operation of taking the reciprocal after adding a zero-prevention constant to the variance of the local frequency offset estimation error, and calculating the proportion of the reciprocal in the sum of the reciprocals of all local symmetric elements.

[0087] In one feasible implementation, the fusion weight is determined based on the variance of the local frequency offset estimation error of each local symmetric element. The smaller the variance of the local frequency offset estimation error, the higher the estimation accuracy of the local symmetric element, and the greater the fusion weight should be assigned. The reciprocal of the sum of the local frequency offset estimation error variance and the zero-prevention constant is calculated, and the proportion of this reciprocal to the sum of the reciprocals of all local symmetric elements is used as the fusion weight wi of that local symmetric element, thereby achieving adaptive weight allocation based on estimation accuracy.

[0088] In summary, this application obtains known true frequency offset values ​​and local frequency offset estimation results during offline simulation and / or calibration, calculates the difference between the two, and obtains the local frequency offset estimation error variance by summing the squares and taking the mean. Then, based on the variance, it performs reciprocal and proportional calculations to obtain the fusion weights, thereby enabling local symmetric units with high estimation accuracy to obtain greater weights, providing a reliable weight basis for subsequent high-precision global frequency offset estimation.

[0089] Please see Figure 4 , Figure 4 For the present invention Figure 1 A schematic diagram of an embodiment of S104. (See attached diagram.) Figure 4 As shown, the method in this application embodiment may include the following steps S401-S403: S401, the sum of squares of the difference between the known true frequency offset value and the local frequency offset estimation result, and the mean value is obtained by dividing by the number of simulations.

[0090] In the embodiments of this application, the number of simulations refers to the total number of known true frequency offset values ​​set during offline simulation or calibration.

[0091] In one feasible implementation, during offline simulation or calibration, a set of known true frequency offset values ​​is set. ,in , That is, the number of simulations. For the th For each locally symmetric element, the local frequency offset estimation results are statistically analyzed under each known true frequency offset value. The differences between the known true frequency offset value and the local frequency offset estimation result are squared, summed, and then the sum is divided by the number of simulations. After averaging, the variance of the local frequency offset estimation error is calculated. .

[0092] For example, the variance of the local frequency offset estimation error in the above steps The specific form of the formula can be:

[0093] Among them, formula parameters Represents the number of simulations, formula parameters The index representing the known true frequency offset value, formula parameters Representing the Given a known true frequency offset value, formula parameters Representing the The local symmetric unit in the th... Local frequency offset estimation results under known true frequency offset values.

[0094] S402, the initial weights of each local symmetric unit are determined based on the reciprocal of the sum of the local frequency offset estimation error variance and the preset zero-prevention constant.

[0095] In this embodiment, the zero-prevention constant refers to a positive constant preset to prevent the denominator from being zero. The initial weight refers to the weight value of each local symmetric unit before normalized proportional allocation.

[0096] In one feasible implementation, to prevent the denominator from becoming zero due to excessively small variance in the local frequency offset estimation error, a preset zero-prevention constant is introduced. Calculate the variance of the local frequency offset estimation error. With zero constant Sum of the sums, then take the reciprocal of the sum, and determine the reciprocal as the first value. The initial weights of a locally symmetric unit.

[0097] S403, normalize the initial weights of each local symmetric unit by proportional allocation to obtain the fused weights.

[0098] In the embodiments of this application, the normalized proportional allocation process refers to the process of dividing each initial weight by the sum of the initial weights of all local symmetric units so that the final sum of weights is 1.

[0099] In one feasible implementation, the sum of the initial weights of all locally symmetric elements is calculated, and the weights of the first element are then... The initial weight of each local symmetric unit is divided by the sum of those initial weights, and then normalized proportionally allocated to ensure that the sum of the weights of all local symmetric units is 1, thus obtaining the first local symmetric unit. The final fusion weights of the local symmetric units .

[0100] For example, the fusion weights in the above steps The specific form of the formula can be:

[0101] Among them, formula parameters Representing the The variance of the local frequency offset estimation error for a locally symmetric element, and the formula parameters. Represents a zero-prevention constant, formula parameters Representing the The variance of the local frequency offset estimation error for a locally symmetric element, and the formula parameters. This represents the total number of locally symmetric units.

[0102] In summary, this application obtains the local frequency offset estimation error variance by summing the squared differences and dividing by the number of simulations, calculates the initial weights using a zero-prevention constant, and then performs normalized proportional allocation processing on the initial weights to obtain the fusion weights. This ensures the mathematical rigor and system stability of the weight allocation, avoids the anomaly of zero denominator, and ensures the accuracy of the weighted fusion results.

[0103] In one feasible implementation, when performing the step of phase rotation of the received signal based on the global frequency offset prediction value to obtain a frequency offset compensation signal, the following is also specifically performed: The received signal is multiplied by the negative phase exponent of the global frequency offset prediction value to obtain the frequency offset compensation signal.

[0104] In this embodiment, the negative phase exponent refers to a complex exponential function that uses the negative value of the global frequency offset prediction as the phase angle. The multiplication process refers to the operation of multiplying the time-domain received signal with the complex exponential function point by point.

[0105] In one feasible implementation, after obtaining the global frequency offset prediction value... Then, the original received signal is analyzed based on this global frequency offset prediction value. Perform frequency offset compensation. Convert the original received signal... Negative phase index of global frequency offset prediction The multiplication process is performed to cancel out the phase rotation introduced by the frequency offset in the received signal, thereby obtaining the frequency offset compensation signal. .

[0106] After frequency offset compensation is completed, the frequency offset compensation signal is then subjected to subsequent digital signal processing (DSP), including polarization demultiplexing, subcarrier demultiplexing, resampling, dispersion compensation, matched filtering, adaptive equalization, carrier phase recovery and demapping, etc.

[0107] For example, the frequency offset compensation signal in the above process The specific form of the formula is:

[0108] Among them, formula parameters Represents the original received signal, formula parameters Represents the global frequency offset prediction value, formula parameters Represents discrete-time index, formula parameters This represents the sampling period.

[0109] In summary, this application performs reverse compensation for the frequency offset phase rotation of the received signal directly in the time domain by multiplying the received signal with the negative phase exponent of the global frequency offset prediction value. This method has low computational complexity, accurate compensation effect, and effectively restores the frequency accuracy of the signal.

[0110] In one feasible implementation, after obtaining the frequency offset compensation signal, it is necessary to perform subsequent digital signal processing on the frequency offset compensation signal. The digital signal processing includes, but is not limited to, resampling, subcarrier demultiplexing, matched filtering, dispersion compensation, clock recovery, adaptive equalization, carrier phase recovery and demapping. The specific process is as follows: The frequency offset compensation signal is resampled to ensure that subsequent digital signal processing operates at a uniform sampling rate or meets the sampling rate requirements of various algorithms. The resampled signal is then demultiplexed with subcarriers, breaking down the broadband multiplexed signal containing multiple frequency components back into individual baseband subcarrier signals. Each individual baseband subcarrier signal undergoes matched filtering, using conjugate matching with the transmitter's pulse shaping filter to maximize the signal-to-noise ratio at the sampling points. Dispersion compensation is then applied to the matched-filtered signal to counteract pulse broadening and waveform distortion caused by differences in the transmission speeds of different frequency components due to fiber chromatic dispersion.

[0111] Furthermore, clock recovery is performed on the dispersion-compensated signal to compensate for frequency deviation and phase jitter of the sampling clock at the transmitting and receiving ends, eliminating timing errors. Based on this, adaptive equalization is performed on the clock-recovered signal, using an adaptive algorithm to dynamically track and compensate for residual linear and nonlinear channel impairments and polarization mode dispersion. Simultaneously, crosstalk between two orthogonal polarization state signals caused by random changes in polarization state during fiber transmission is eliminated, separating the mixed polarization signal into independent polarization channel signals.

[0112] Subsequently, carrier phase recovery is performed on the equalized signal to estimate and eliminate phase rotation and jump caused by phase noise and residual frequency offset of the transceiver laser, so as to accurately restore the constellation point position of the signal. Finally, the signal with the correct constellation point is demapped, and the complex symbols are inversely mapped into the corresponding binary bit sequence according to the preset modulation format rules, thereby completing the entire process of signal recovery and demodulation at the receiving end.

[0113] The following will combine Figure 5This application provides a detailed description of the signal frequency offset compensation device provided in its embodiments. It should be noted that... Figure 5 The signal frequency offset compensation device 500 in the present application is used to perform the functions described herein. Figures 1-4 The methods shown in the embodiments are for illustrative purposes only, illustrating the parts relevant to the embodiments of this application. For specific technical details not disclosed, please refer to this application. Figures 1-4 In the embodiment shown, the signal frequency offset compensation device 500 may include a first data acquisition unit 501, a data processing unit 502, a frequency offset prediction unit 503, a weight acquisition unit 504, and a signal compensation unit 505, as detailed below: The data acquisition unit 501 is used to acquire the spectral structure of the transmitted signal, divide the spectral structure into multiple locally symmetric units, and the locally symmetric units are frequency sub-intervals with local symmetric characteristics. The data processing unit 502 is used to obtain the unbiased power spectrum of the local symmetric unit, perform centroid calculation on the unbiased power spectrum, and obtain the reference centroid position. The frequency offset prediction unit 503 is used to acquire the power spectrum of the received signal, and to perform centroid calculation and difference processing based on the power spectrum and the local symmetry unit to obtain the local frequency offset prediction value. The weight acquisition unit 504 is used to acquire the local frequency offset estimation error variance of the local symmetric unit pre-calibrated, and perform weight calculation processing based on the local frequency offset estimation error variance to obtain the fused weights. The signal compensation unit 505 is used to perform weighted fusion processing on each local frequency offset prediction value and the corresponding fusion weight to obtain the global frequency offset prediction value, and to perform phase rotation on the received signal based on the global frequency offset prediction value to obtain the frequency offset compensation signal.

[0114] The signal frequency offset compensation device 500 provided in the above embodiments can realize the technical solutions described in the above signal frequency offset compensation method embodiments. The specific implementation principles of each module or unit can be found in the corresponding content in the above signal frequency offset compensation method embodiments, and will not be repeated here.

[0115] like Figure 6 As shown, the present invention also provides an electronic device 600. The electronic device 600 includes a processor 601, a memory 602, a display 603, and a data acquisition device 604. Figure 6 Only some components of the electronic device 600 are shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead.

[0116] In some embodiments, processor 601 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 602 or process data, such as the signal frequency offset compensation method of the present invention.

[0117] In some embodiments, processor 601 may be a single server or a group of servers. The server group may be centralized or distributed. In some embodiments, processor 601 may be local or remote. In some embodiments, processor 601 may be implemented on a cloud platform. In one embodiment, the cloud platform may include a private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, intranet, multi-cloud, etc., or any combination thereof.

[0118] In some embodiments, memory 602 may be an internal storage unit of electronic device 600, such as a hard disk or memory of electronic device 600. In other embodiments, memory 602 may also be an external storage device of electronic device 600, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on electronic device 600.

[0119] Furthermore, the memory 602 may include both internal storage units of the electronic device 600 and external storage devices. The memory 602 is used to store application software and various types of data installed on the electronic device 600.

[0120] In some embodiments, display 603 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 603 is used to display information from electronic device 600 and to display a visual user interface. Data acquisition unit 604 is used to acquire the spectral structure of the transmitted signal; acquire the unbiased power spectrum of the locally symmetric unit; acquire the power spectrum of the received signal; and acquire the variance of the pre-calibrated local frequency offset estimation error of the locally symmetric unit. Components 601-604 of electronic device 600 communicate with each other via a system bus.

[0121] In one embodiment, when the processor 601 executes the signal frequency offset compensation program in the memory 602, the following steps can be implemented: The spectral structure of the transmitted signal is obtained, and the spectral structure is divided into multiple locally symmetric units, which are frequency sub-intervals with local symmetry characteristics. Obtain the unbiased power spectrum of the locally symmetric unit, perform centroid calculation on the unbiased power spectrum, and obtain the reference centroid position; The power spectrum of the received signal is obtained, and the centroid is calculated and the difference is processed based on the power spectrum and the local symmetric unit to obtain the local frequency offset prediction value. The local frequency offset estimation error variance of the local symmetric unit is obtained in advance, and the weight is calculated based on the local frequency offset estimation error variance to obtain the fusion weight. The local frequency offset prediction values ​​are weighted and fused with the corresponding fusion weights to obtain the global frequency offset prediction value. The received signal is then phase-rotated based on the global frequency offset prediction value to obtain the frequency offset compensation signal.

[0122] It should be understood that when the processor 601 executes the signal frequency offset compensation program in the memory 602, in addition to the functions mentioned above, it can also perform other functions, as can be found in the description of the corresponding method embodiments above.

[0123] Furthermore, this embodiment of the invention does not specifically limit the type of electronic device 600 mentioned. Electronic device 600 can be a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, laptop computer, or other portable electronic device. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic device can also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the invention, electronic device 600 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0124] Accordingly, this application also provides a computer-readable storage medium for storing a computer-readable program or instruction. When the program or instruction is executed by a processor, it can implement the steps or functions of the signal frequency offset compensation method provided in the above-described method embodiments.

[0125] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.), and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0126] The signal frequency offset compensation method, apparatus, and electronic device provided by the present invention have been described in detail above. Specific examples have been used to illustrate the principle and implementation of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation and application scope based on the idea of ​​the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method of signal frequency offset compensation, the method comprising: include: The spectral structure of the transmitted signal is obtained, and the spectral structure is divided into multiple locally symmetric units, wherein each locally symmetric unit is a frequency sub-interval with local symmetry characteristics. Obtain the unbiased power spectrum of the local symmetric unit, and perform centroid calculation on the unbiased power spectrum to obtain the reference centroid position; The power spectrum of the received signal is obtained, and the centroid is calculated and the difference is processed based on the power spectrum and the local symmetric unit to obtain the local frequency offset prediction value. The local frequency offset estimation error variance of the local symmetric unit is obtained in advance, and the weight is calculated based on the local frequency offset estimation error variance to obtain the fusion weight. The local frequency offset prediction values ​​are weighted and fused with the corresponding fusion weights to obtain the global frequency offset prediction value. The received signal is then phase-rotated based on the global frequency offset prediction value to obtain the frequency offset compensation signal.

2. The method of claim 1, wherein, The process of dividing the spectral structure to obtain multiple locally symmetric units includes: Obtain the baseband signal in the aforementioned spectral structure; The local symmetric unit is obtained by overlapping and dividing the spectrum of adjacent subcarriers of the baseband signal.

3. The method of claim 1, wherein, The process of calculating the centroid of the unbiased power spectrum to obtain the reference centroid position includes: The reference centroid position is obtained by multiplying the frequency value of the frequency sampling point in the unbiased power spectrum with the power spectrum value, summing the products, and dividing by the sum of the power spectrum values.

4. The method of claim 1, wherein, The step of calculating the centroid and performing difference processing based on the power spectrum and the local symmetric unit to obtain the local frequency offset prediction value includes: The frequency value and power spectrum value of the power spectrum at the frequency sampling point within the local symmetric unit are multiplied, summed, and normalized to obtain the receiving centroid position. The local frequency offset prediction value is obtained by subtracting the received centroid position from the reference centroid position.

5. The method of claim 1, wherein, The step of obtaining the pre-calibrated local frequency offset estimation error variance of the local symmetric unit, and performing weight calculation based on the local frequency offset estimation error variance to obtain the fusion weights includes: Acquire the known true frequency offset value and the local frequency offset estimation result during offline simulation and / or calibration process. The known true frequency offset value is a preset true frequency offset value, and the local frequency offset estimation result is the value obtained by the local symmetric unit performing frequency offset estimation under the known true frequency offset value. The variance of the local frequency offset estimation error is obtained by summing the squares and taking the mean of the difference between the known true frequency offset value and the local frequency offset estimation result. The fusion weights are obtained by calculating the reciprocal and proportion of the variance of the local frequency offset estimation error.

6. The method according to claim 5, characterized in that, The step of summing the squares and taking the mean of the differences between the known true frequency offset value and the local frequency offset estimation result to obtain the variance of the local frequency offset estimation error includes: The variance of the local frequency offset estimation error is obtained by summing the squares of the differences between the known true frequency offset value and the local frequency offset estimation result, dividing by the number of simulations, and then averaging the sums.

7. The method of claim 5, wherein, The process of calculating the inverse and proportion of the variance of the local frequency offset estimation error to obtain the fusion weights includes: The initial weights of each local symmetric unit are determined based on the reciprocal of the sum of the local frequency offset estimation error variance and the preset zero-prevention constant. The initial weights of each local symmetric unit are normalized proportionally to obtain the fused weights.

8. The method of claim 1, wherein, The step of performing phase rotation on the received signal based on the global frequency offset prediction value to obtain a frequency offset compensation signal includes: The received signal is multiplied by the negative phase exponent of the global frequency offset prediction value to obtain the frequency offset compensation signal.

9. A signal frequency offset compensation apparatus, characterized by, The device includes: The data acquisition unit is used to acquire the spectral structure of the transmitted signal, divide the spectral structure into multiple locally symmetric units, wherein the locally symmetric units are frequency sub-intervals with local symmetry characteristics. The data processing unit is used to obtain the unbiased power spectrum of the local symmetric unit, perform centroid calculation on the unbiased power spectrum, and obtain the reference centroid position. The frequency offset prediction unit is used to acquire the power spectrum of the received signal, and to perform centroid calculation and difference processing based on the power spectrum and the local symmetry unit to obtain the local frequency offset prediction value. The weight acquisition unit is used to acquire the local frequency offset estimation error variance pre-calibrated by the local symmetric unit, and perform weight calculation processing based on the local frequency offset estimation error variance to obtain the fusion weight; The signal compensation unit is used to perform weighted fusion processing on each local frequency offset prediction value and the corresponding fusion weight to obtain a global frequency offset prediction value, and to perform phase rotation on the received signal based on the global frequency offset prediction value to obtain a frequency offset compensation signal.

10. An electronic device, comprising: The electronic device includes a data acquisition unit, a memory, and a processor, wherein, The data acquisition unit is used to acquire the spectral structure of the transmitted signal; acquire the unbiased power spectrum of the local symmetric unit; acquire the power spectrum of the received signal; and acquire the local frequency offset estimation error variance of the local symmetric unit pre-calibrated. The memory is used to store executable program code; The processor is configured to: divide the spectral structure into multiple locally symmetric units; perform centroid calculation on the unbiased power spectrum to obtain a reference centroid position; perform centroid calculation and difference processing based on the power spectrum and the locally symmetric units to obtain a local frequency offset prediction value; perform weight calculation based on the variance of the local frequency offset estimation error to obtain a fusion weight; perform weighted fusion processing to obtain a global frequency offset prediction value; and perform phase rotation on the received signal based on the global frequency offset prediction value to obtain a frequency offset compensation signal.

Images