Receiver frequency offset two-stage estimation method and device based on FFT + CIC-mSDFT

Abstract

The embodiment of the invention discloses a receiver frequency offset two-stage estimation method and device based on FFT + CIC-mSDFT, and the method comprises the steps: collecting a received digital signal, carrying out the preprocessing, carrying out the offset frequency rough estimation of a preprocessed signal through an FFT algorithm, and obtaining a first frequency corresponding to the maximum value of an amplitude frequency spectrum; performing offset frequency fine estimation on the first frequency by using a CIC-mSDFT algorithm to obtain a second frequency corresponding to the maximum amplitude-frequency spectrum value; and correcting the pre-processed signal by using the second frequency meeting the calibration condition to obtain a recovered signal. The CIC-mSDFT algorithm provided by the invention has an excellent recursive property, only one-time complex multiplication operation is needed for frequency spectrum estimation at each new input moment and updating of frequency spectrum information of a single spectral line, due to the fact that the recursive operation is extremely low in complexity, real-time frequency offset estimation and compensation can be achieved on a DSP, the CIC-mSDFT algorithm is suitable for high-order M-QAM modulation signals, and the accuracy of frequency offset estimation is improved. And when the frequency offset exceeds a monitoring range, an early warning characteristic is realized.

Description

Technical Field

[0001] The present application relates to the technical field of frequency deviation detection of coherent optical receivers, and in particular to a two-stage estimation method and device for receiver frequency deviation based on FFT+CIC-mSDFT. Background Art

[0002] With the rapid popularization of smart city infrastructure and the advancement of cloud computing, big data and mobile Internet technologies, the demand for optical communication systems with high capacity and long-distance transmission capabilities has increased significantly. The integration of coherent multi-level modulation formats with wavelength division multiplexing (WDM) systems is the key to achieving high-capacity and long-distance transmission optical communications. Advances in digital signal processing (DSP) technology have promoted the application of high-order modulation formats such as M-ary Quadrature Amplitude Modulation (M-QAM) in coherent optical communications, which significantly improves spectral efficiency and transmission rate.

[0003] The difference between the laser frequency of the transmitter and the local oscillator leads to frequency offset (FO) in homodyne (ideally, the local oscillator frequency is the same as the signal frequency, and the baseband signal is directly demodulated) and intradyne (ideally, there is a small frequency difference between the local oscillator frequency and the signal frequency) coherent optical detection receivers, causing the signal to rotate in the complex plane, affecting the subcarrier orthogonality of the signal, signal detection probability and signal stability, etc. Although the DSP-based carrier phase processing algorithm can tolerate frequency deviations within 100MHz, in actual applications, this frequency deviation can usually reach the GHz level, so accurate frequency offset estimation and compensation algorithms are required in coherent optical receivers.

[0004] The coherent optical receiver frequency measurement solutions that can be implemented based on DSP include:

[0005] 1) Data-aided Frequency Offset Estimation (DA-FOE) scheme: This scheme first performs polarization demultiplexing on the signal to separate signals of different polarization states, then compensates for signal distortion through equalization technology, and then performs preprocessing operations such as timing correction and mode alignment. Subsequently, the received signal is compared with the known signal using known auxiliary data, and the frequency offset is estimated through methods such as maximum likelihood estimation and minimum mean square error. This method can achieve frequency offset estimation in a noisy and interfering environment, but the auxiliary data processing resource consumption is extremely large, the multi-step processing complexity is high, and the algorithm will fail when there is no auxiliary data or the auxiliary data is interfered with.

[0006] 2) Non-data-assisted differential phase blind estimation scheme, which first removes the modulated (i.e., data-assisted information) by methods such as fourth power operation, and then multiplies the data at the next sampling moment by the complex conjugate of the previous moment to obtain the phase change amplitude of the carrier in a single sampling interval, and then estimates the frequency deviation. This method has low computational complexity, but is severely affected by the laser phase noise, and the estimation accuracy is low. Other differential phase (diff-FOE) methods based on equal time interval operations or average difference operations proposed later have reduced the impact of phase noise on accuracy to a certain extent, but require complex constellation operations for high-order modulated signals, and are not suitable for fast frequency deviation tracking.

[0007] 3) Quartic + Fast Fourier Transform (FFT) spectrum analysis scheme, which first performs a quartic operation on the signal to remove some modulation information, then performs an N-point Fast Fourier Transform on the signal to obtain its spectrum information, and searches for the frequency corresponding to the peak in the spectrum to estimate the frequency deviation. This strategy is widely applicable to M-QAM high-order modulation, which greatly reduces the impact of phase noise on it, but the single-stage FFT method requires a large number of FFT points to achieve high accuracy, and increasing the number of FFT points will greatly increase the computational complexity.

[0008] 4) Other two-stage quartic spectrum analysis schemes: This type of scheme first performs a quartic operation on the signal, then performs a FFT with a smaller number of points, performs a peak search, and roughly estimates the frequency deviation value. Then, the Chrip-Z transform (CZT), zoom fast Fourier transform (Zoom-FFT) or all-phase FFT (apFFT) and other methods are used to perform a second-stage spectrum refinement operation near the frequency value roughly estimated in the first stage, and perform a peak search to achieve high-precision spectrum estimation. This type of strategy effectively reduces the complexity of a single estimation, but continuous frequency deviation estimation still requires a lot of resource consumption, and each frequency deviation estimation of this type of scheme is independent of the previous moment, and the entire spectrum needs to be recalculated, and the previous estimation results cannot be used, resulting in unnecessary waste of resources.

[0009] In summary, the advantage of the frequency offset estimation scheme based on carrier phase variation in the prior art is that the complexity of the frequency offset estimation part is relatively low, but for high-order modulated signals, extremely complex constellation data preprocessing is required, and the estimation accuracy and robustness are poor, and the estimation result is seriously affected by phase noise. While other existing frequency offset estimation schemes based on spectrum information ensure the estimation accuracy, the computational complexity and resource consumption are still large, and the estimation information at the previous moment is not reasonably used. Therefore, the resource consumption under the requirements of continuous frequency offset detection or frequency offset drift monitoring still needs to be further reduced. Summary of the invention

[0010] The present application provides a two-stage frequency offset estimation method and device for a receiver based on FFT+CIC-mSDFT to improve the accuracy of frequency offset estimation and reduce computational complexity. The specific scheme is as follows:

[0011] In a first aspect, the present application provides a two-stage receiver frequency offset estimation method based on FFT+CIC-mSDFT, including:

[0012] Collecting received digital signals;

[0013] Preprocessing the digital signal to eliminate the symbol phase and phase noise to obtain a preprocessed signal;

[0014] The FFT algorithm is used to roughly estimate the offset frequency of the preprocessed signal to obtain the first frequency corresponding to the maximum value of the amplitude spectrum;

[0015] The CIC-mSDFT algorithm is used to accurately estimate the offset frequency of the first frequency to obtain the second frequency corresponding to the maximum value of the amplitude spectrum;

[0016] The preprocessed signal is corrected using a second frequency that meets the calibration condition to eliminate the influence of the frequency deviation and obtain a restored signal.

[0017] Optionally, preprocessing the digital signal to eliminate the symbol phase and phase noise to obtain a preprocessed signal includes:

[0018] Performing quadratic processing on the digital signal to eliminate the symbol phase information therein;

[0019] The digital signal with symbol phase information eliminated is subjected to sliding average processing to remove phase noise therein to obtain a preprocessed signal.

[0020] Optionally, the digital signal is expressed as:

[0021] x(n)=exp{j[θ s (n)+ΔωnT O +θ L (n)+θ n (n)]};

[0022] In the formula, θ s (n) is the symbol phase information of the nth sampling signal, θ s (n) represents the modulation information; Δω is the angular frequency information of the frequency deviation to be estimated, T0 = 1 / f s is the sampling period, ΔωnT0=2πΔfnT0 is the phase error introduced by the frequency deviation, θ L (n) is the phase noise caused by the laser line width, θ n (n) is the phase noise caused by the spontaneous emission of the optical amplifier;

[0023] The digital signal with symbol phase information eliminated is expressed as:

[0024] x 4 (n) = exp{j[4ΔωnT O +4θ L (n)+4θ n (n)]}.

[0025] Optionally, using an FFT algorithm to roughly estimate the offset frequency of the preprocessed signal to obtain a first frequency corresponding to the maximum value of the amplitude-spectrum ratio includes:

[0026] Calculating a first spectrum of the preprocessed signal using a predetermined number of FFT points;

[0027] In the first spectrum, the peak search is performed to find the first frequency f0 and the first index k corresponding to the maximum value of the spectrum.

[0028] Optionally, using a CIC-mSDFT algorithm to perform accurate frequency estimation on the first frequency to obtain a second frequency corresponding to the maximum value of the amplitude-spectrum ratio includes:

[0029] Perform a times frequency interpolation in the upper and lower spectrum lines of the first frequency f0, and use the CIC-mSDFT algorithm to calculate the second spectrum of the interpolation frequency point;

[0030] In the second spectrum, the second frequency Δf corresponding to the maximum value of the spectrum is searched by peak value. est and the corresponding second index k final .

[0031] Optionally, before correcting the preprocessed signal by using the second frequency to eliminate the influence of the frequency deviation to obtain the restored signal, the estimation method further includes:

[0032] Performing a recursive operation on the second frequency to confirm whether the second frequency meets the accuracy requirement;

[0033] If the second frequency does not meet the accuracy requirement, an early warning is issued and the second frequency is re-estimated;

[0034] If the second frequency meets the accuracy requirement, it is confirmed that the second frequency meets the calibration condition.

[0035] Optionally, performing a recursive operation on the second frequency to confirm whether the second frequency meets the accuracy requirement includes:

[0036] Two estimation boundaries are set at the upper and lower rth spectrum line positions of the second frequency at the current moment;

[0037] Using the recursive property of the CIC-mSDFT algorithm, calculate the third spectrum of the second frequency at the current moment and the fourth spectrum of the upper and lower spectral lines;

[0038] If the amplitude of the fourth spectrum at the current moment is higher than the corresponding estimated boundary, and the first amplitude is not less than θ times the second amplitude, then it is confirmed that the second frequency at the current moment meets the accuracy requirement, otherwise, it does not meet the requirement; wherein the first amplitude is the amplitude of the third spectrum at the current moment, the second amplitude is the amplitude of the third spectrum at the previous moment, and θ is a value between (0,1].

[0039] Optionally, before collecting the received digital signal, the estimation method further includes:

[0040] receiving an optical signal sent from a transmitting end;

[0041] filtering the received optical signal to obtain a filtered signal;

[0042] Mixing the filtered signal with the generated reference light to obtain mixed light;

[0043] Convert the mixed light into a current signal;

[0044] Convert the current signal into a digital signal.

[0045] Optionally, a two-stage estimation method of receiver frequency deviation based on FFT+CIC-mSDFT is applied to the receiving end, and the receiving end and the transmitting end transmit digital signals through a transmission path; the transmitting end maps and modulates the data source through 16 / 64QAM to obtain an I / Q modulated signal, and then generates a modulated optical signal through a laser; and transmits the modulated optical signal to the receiving end through the transmission path.

[0046] In a second aspect, the present application provides a two-stage frequency offset estimation device for a receiver based on FFT+CIC-mSDFT, comprising:

[0047] A signal acquisition module is configured to acquire received digital signals;

[0048] A preprocessing module is configured to preprocess the digital signal to eliminate the symbol phase and phase noise to obtain a preprocessed signal;

[0049] A frequency offset estimation module is configured to use an FFT algorithm to roughly estimate the frequency offset of the preprocessed signal to obtain a first frequency corresponding to a maximum value of the amplitude-spectrum ratio;

[0050] A frequency offset interpolation module is configured to use a CIC-mSDFT algorithm to perform an accurate frequency offset estimation on the first frequency to obtain a second frequency corresponding to a maximum value of the amplitude-spectrum ratio;

[0051] The signal recovery module is configured to correct the preprocessed signal using a second frequency that meets the calibration condition to eliminate the influence of the frequency deviation and obtain a recovered signal.

[0052] The innovative features of the embodiments of the present application include:

[0053] The embodiment of the present application discloses a two-stage frequency deviation estimation method and device for a receiver based on FFT+CIC-mSDFT, which collects received digital signals; preprocesses the digital signals to eliminate symbol phase and phase noise to obtain preprocessed signals; uses FFT algorithm to roughly estimate the frequency deviation of the preprocessed signals to obtain the first frequency corresponding to the maximum value of the amplitude spectrum; uses CIC-mSDFT algorithm to accurately estimate the frequency deviation of the first frequency to obtain the second frequency corresponding to the maximum value of the amplitude spectrum; uses the second frequency that meets the calibration conditions to correct the preprocessed signal to eliminate the influence of frequency deviation to obtain a restored signal. The CIC-mSDFT algorithm proposed in the present application has excellent recursive properties. For the spectrum estimation of each new input moment, updating the spectrum information of a single spectral line only requires one complex multiplication operation. Since the complexity of the recursive operation is extremely low, real-time frequency deviation estimation and compensation can be realized on DSP. It is suitable for high-order M-QAM modulated signals and has early warning characteristics when the frequency deviation exceeds the monitoring range. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art description are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without creative work.

[0055] Figure 1 is a schematic diagram of the principle of a coherent optical receiver;

[0056] Figure 2 A flowchart of a two-stage receiver frequency deviation estimation method based on FFT+CIC-mSDFT provided in an embodiment of the present application;

[0057] Figure 3 A detailed schematic diagram of a two-stage receiver frequency offset estimation method based on FFT+CIC-mSDFT provided in an embodiment of the present application;

[0058] Figure 4 A schematic diagram of the CIC-mSDFT algorithm provided in an embodiment of the present application;

[0059] Figure 5 A structural schematic diagram of a two-stage receiver frequency offset estimation device based on FFT+CIC-mSDFT provided in an embodiment of the present application. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0061] It should be noted that the terms "including" and "having" and any variations thereof in the embodiments of the present application and the accompanying drawings are intended to cover non-exclusive inclusions. For example, a process, method, device, product or equipment comprising a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or equipment.

[0062] The present application is dedicated to solving the problems in the current frequency offset detection scheme of coherent optical receivers, namely, the lack of a frequency offset estimation algorithm that can achieve low hardware resource occupancy and computational complexity while maintaining high estimation accuracy, and the independence of single estimation strategies, and the extremely high computational complexity of continuous frequency offset estimation and frequency offset drift monitoring, so as to realize a high-precision, low-complexity frequency offset detection and frequency offset drift monitoring strategy using recursive operations.

[0063] The embodiment of the present application discloses a two-stage frequency deviation estimation method for a receiver based on FFT+CIC-mSDFT, which collects received digital signals; preprocesses the digital signals to eliminate symbol phase and phase noise to obtain preprocessed signals; uses FFT algorithm to roughly estimate the frequency deviation of the preprocessed signals to obtain the first frequency corresponding to the maximum value of the amplitude spectrum; uses CIC-mSDFT algorithm to accurately estimate the frequency deviation of the first frequency to obtain the second frequency corresponding to the maximum value of the amplitude spectrum; uses the second frequency that meets the calibration conditions to correct the preprocessed signal to eliminate the influence of frequency deviation to obtain a restored signal. The CIC-mSDFT algorithm proposed in the present application has excellent recursive properties. For the spectrum estimation of each new input moment, it only takes one complex multiplication operation to update the spectrum information of a single spectral line. Since the complexity of the recursive operation is extremely low, it can realize real-time frequency deviation estimation and compensation on DSP. It is suitable for high-order M-QAM modulated signals and has early warning characteristics when the frequency deviation exceeds the monitoring range.

[0064] The embodiments of the present application are described in detail below.

[0065] The present application provides a two-stage receiver frequency deviation estimation method based on FFT+CIC-mSDFT, which is applied to the receiving end. Figure 1 , Figure 1The figure is a schematic diagram of the principle of a coherent optical receiver. The receiving end and the transmitting end transmit digital signals through the transmission path; the transmitting end maps the data source through 16 / 64QAM and modulates the signal to obtain an I / Q modulated signal, and then generates a modulated optical signal through a laser; the modulated optical signal is transmitted to the receiving end through the transmission path. At the receiving end, the receiving end uses a local oscillator (LO) laser with a frequency close to the signal frequency of the transmitting end to generate reference light. The received optical signal and the reference light generated by the local oscillator are mixed on the photodetector to generate a current signal, which contains the amplitude and phase information of the original optical signal. Due to the slight difference between the frequency of the transmitter and the local oscillator, this will introduce a frequency offset (FO) in the coherent and intradyne detection receivers, resulting in a rotation of the signal in the complex plane. In order to recover the signal and perform effective data transmission, this frequency offset must be accurately estimated and compensated.

[0066] Figure 2 A flowchart of a two-stage receiver frequency offset estimation method based on FFT+CIC-mSDFT provided in an embodiment of the present application is shown. Figure 3 This is a detailed schematic diagram of a two-stage frequency offset estimation method for a receiver based on FFT+CIC-mSDFT provided in an embodiment of the present application. The purpose of this application is to improve the accuracy of frequency offset estimation and reduce computational complexity.

[0067] Combination Figure 2 and Figure 3 The present application provides a two-stage receiver frequency deviation estimation method based on FFT+CIC-mSDFT, comprising:

[0068] S1, collects the received digital signal;

[0069] The digital signal is expressed as:

[0070] x(n)=exp{j[n s (n)+ΔωnT O +θ L (n)+θ n (n)]};

[0071] In the formula, θ s (n) is the symbol phase information of the nth sampling signal, θ s (n) represents the modulation information; Δω is the angular frequency information of the frequency deviation to be estimated, T0 = 1 / f s is the sampling period, ΔωnT0=2πΔfnT0 is the phase error introduced by the frequency deviation, θ L (n) is the phase noise caused by the laser line width, θ n (n) is the phase noise caused by the spontaneous emission of the optical amplifier;

[0072] S2, preprocessing the digital signal to eliminate the symbol phase and phase noise to obtain a preprocessed signal;

[0073] The purpose of the preprocessing in this step is to eliminate the symbol phase information and phase noise, which specifically includes: performing quadratic processing on the digital signal to eliminate the symbol phase information therein; performing sliding average processing on the digital signal with the symbol phase information eliminated to remove the phase noise therein to obtain the preprocessed signal.

[0074] The digital signal with symbol phase information eliminated is expressed as:

[0075] x 4 (n) = exp{j[4ΔωnT O +4θ L (n)+4θ n (n)]}.

[0076] Phase noiseθ n (n) satisfies the Gaussian distribution and is a random variable with zero mean; θ L (n) approximately satisfies the Wiener process and is a slowly changing signal compared to the phase noise introduced by the frequency offset. Therefore, the phase noise can be removed by sliding average processing.

[0077] S3, using an FFT algorithm to roughly estimate the offset frequency of the preprocessed signal to obtain a first frequency corresponding to the maximum value of the amplitude spectrum;

[0078] S4, using the CIC-mSDFT algorithm to perform accurate frequency estimation on the first frequency to obtain a second frequency corresponding to the maximum value of the amplitude-spectrum ratio;

[0079] refer to Figure 4 , Figure 4 This is the schematic diagram of the CIC-mSDFT algorithm. Figure 4 The CIC-mSDFT algorithm is based on SDFT, which is improved into the mSDFT algorithm, and then improved into CIC-mSDFT. Figure 4 In the figure, (a) is the SDFT algorithm, (b) is the mSDFT algorithm, (c) is the improved CIC-mSDFT algorithm, and (d) is the implementation strategy of two CIC filters.

[0080] SDFT algorithm:

[0081] The computational complexity of the discrete Fourier transform (DFT) algorithm is proportional to the square of the transform window length (number of points). As the window length increases, the computational complexity increases significantly. By taking advantage of the periodicity and symmetry of the spectrum, the computational complexity of the N-point fast Fourier transform (FFT) algorithm is greatly reduced through butterfly operations. However, this FFT algorithm calculates the entire spectrum and cannot isolate specific frequency bands for calculation. In addition, the N-point FFT calculations at different times are independent, and the results of previous calculations are not fully utilized.

[0082] The sliding discrete Fourier transform (SDFT) algorithm is implemented through recursive operations, making full use of the similarity between the N-point signal at the previous moment and the signal at the next moment. The SDFT algorithm can calculate a specific spectrum, thereby improving the efficiency of spectrum analysis and significantly reducing the computational complexity.

[0083]

[0084] W M =exp(j2π / M) is the complex rotation factor, q=n-M+1 is the position of the first sampling point in the frame sequence, X n (k) is the spectrum value of the kth frequency point at the Nth sampling time.

[0085] Therefore, the differential time domain equation of the SDFT algorithm is:

[0086]

[0087] The DFT result at the next time is obtained by recursion and phase shifting from the result at the previous time. Therefore, the SDFT algorithm only requires N real number addition operations and 1 complex number multiplication operation to implement the recursion.

[0088] Compared with FFT algorithm, SDFT has more advantages in real-time spectrum analysis. However, SDFT filter has Z-area poles on the unit circle. Therefore, the system poles move from the origin, which may lead to potential instability of the system. In addition, the rotation factor This can lead to round-off errors, which in turn can lead to cumulative errors in the recursive structure.

[0089] mSDFT algorithm:

[0090] The modulated SDFT algorithm (mSDFT) can be implemented by fully utilizing the modulation characteristics of DFT and moving the target spectrum line to (k=0). mSDFT avoids the cumulative error and potential instability caused by the complex rotation factor in the resonator feedback loop. The circular frequency shift characteristic of DFT is:

[0091]

[0092] So now multiply the input signal x(n) by the modulation sequence Get the new sequence y(n), and then calculate the amplitude spectrum of the new sequence at k=0:

[0093]

[0094] go through After phase correction, the accurate amplitude spectrum and phase spectrum of the sequence can be obtained. In the coherent FOE scenario, this application only needs the amplitude spectrum of the signal without phase correction, that is, |X n (k)|=|Y n (0)|. Compared with the SDFT algorithm, the mSDFT algorithm does not increase the computational complexity, avoids cumulative errors, and ensures the stability of coherent optical frequency deviation detection.

[0095] CIC-mSDFT algorithm:

[0096] because The present application can place the complex multiplication in the recursive stage in the feedforward part to avoid the clock rate limitation of the digital circuit. That is:

[0097]

[0098] The recursive structure of the new sequence is the same as that of the first-order CIC filter. Therefore, this application can summarize mSDFT into two steps: using the rotation factor The correction sequence is then used to calculate the frequency spectrum through a CIC filter without sampling rate changes. In addition, the sampling frequency of the coherent optical receiver is much higher than the FOE output requirement, and the FOD is also limited, so the estimation result can be downsampled. By combining the Noble transform, the present application can further reduce resource consumption.

[0099] F(Z R )(↓R)=(↓R)F(z);

[0100] Among them, F(Z R ) indicates that the frequency response of the original filter F(z) is stretched R times on the frequency axis, ↓R indicates the filtering operation stretched R times, and F(z) indicates the original filter.

[0101] Increasing the order (L) of the CIC filter can reduce spectral leakage and obtain a higher-order filter amplitude response. The CIC-mSDFT-based algorithm proposed in this application is used to estimate the frequency offset of the digital signal. This application first performs a rough frequency offset estimate through a fast Fourier transform (FFT) to determine the approximate range of the frequency offset. Within the frequency offset range determined by the FFT, the CIC-mSDFT algorithm is used to perform fine interpolation of the frequency offset points to improve the accuracy of the frequency offset estimation.

[0102] S5, correcting the preprocessed signal using a second frequency that meets the calibration condition to eliminate the influence of the frequency deviation and obtain a restored signal.

[0103] This step corrects the received digital signal based on the estimated frequency deviation information (second frequency), eliminates the influence of the frequency deviation, and restores the original form of the signal. Finally, the restored signal is demodulated to extract the transmitted data information.

[0104] See also Figure 3 The estimation method of this application is mainly divided into two stages. The first stage is the FFT stage, i.e., the rough estimation stage, and the second stage is the CIC-mSDFT stage, i.e., the precise estimation stage. Figure 3 In the rough estimation stage, the preprocessed signal is first roughly estimated in terms of offset frequency, and the spectrum line is moved to k0, where k0 is the serial number corresponding to the roughly estimated spectrum line, corresponding to the amplitude of the spectrum A0. Figure 3 In the equation, k0=argmax(|·|), In the second stage, CIC-mSDFT is used to perform accurate estimation of the offset frequency, and a-fold interpolation is performed during the accurate estimation process, expressed as a(k0 - 1), a(k0 + 1). Figure 3 In the figure, the interpolation of the upper spectrum line is expressed as USL, and the interpolation of the lower spectrum line is expressed as LSL. The maximum amplitude is found to obtain the second frequency. LSL <k<USL,k∈Z,M=aN。

[0105] The embodiment of the present application discloses a two-stage frequency deviation estimation method and device for a receiver based on FFT+CIC-mSDFT, which collects received digital signals; preprocesses the digital signals to eliminate symbol phase and phase noise to obtain preprocessed signals; uses FFT algorithm to roughly estimate the frequency deviation of the preprocessed signals to obtain the first frequency corresponding to the maximum value of the amplitude spectrum; uses CIC-mSDFT algorithm to accurately estimate the frequency deviation of the first frequency to obtain the second frequency corresponding to the maximum value of the amplitude spectrum; uses the second frequency that meets the calibration conditions to correct the preprocessed signal to eliminate the influence of frequency deviation to obtain a restored signal. The CIC-mSDFT algorithm proposed in the present application has excellent recursive properties. For the spectrum estimation of each new input moment, updating the spectrum information of a single spectral line only requires one complex multiplication operation. Since the complexity of the recursive operation is extremely low, real-time frequency deviation estimation and compensation can be realized on DSP. It is suitable for high-order M-QAM modulated signals and has early warning characteristics when the frequency deviation exceeds the monitoring range.

[0106] In a specific implementation of the present application, the preprocessed signal is roughly estimated by using an FFT algorithm to obtain a first frequency corresponding to the maximum value of the amplitude spectrum, which includes:

[0107] Calculating a first spectrum of the preprocessed signal using a predetermined number of FFT points;

[0108] The predetermined FFT point number may be a smaller number of FFT points, specifically 512 points. When the FFT point number is 512 points, the square root error can reach 10 11 , which is more than an order of magnitude better than other two-stage solutions with 1024 FFT points in the first stage.

[0109] In the first spectrum, the peak search is performed to find the first frequency f0 and the first index k corresponding to the maximum value of the spectrum.

[0110] In a specific implementation of the present application, the CIC-mSDFT algorithm is used to perform accurate frequency estimation on the first frequency, and the second frequency corresponding to the maximum value of the amplitude spectrum is obtained, which includes:

[0111] Perform a times frequency interpolation in the upper and lower spectrum lines of the first frequency f0, and use the CIC-mSDFT algorithm to calculate the second spectrum of the interpolation frequency point;

[0112] Among them, the a-fold frequency interpolation is expressed as k∈ (a(k0 - 1), a(k0 + 1)), k0 is the serial number corresponding to the roughly estimated spectrum line, and k is the index of the kth spectrum line.

[0113] In the second spectrum, the second frequency Δf corresponding to the maximum value of the spectrum is searched by peak value. est and the corresponding second index k final .

[0114] The frequency interpolation density in the precise estimation stage of the present application can be customized according to different frequency deviation ranges and accuracy requirements, and thanks to the characteristics of the mSDFT algorithm, changes in the frequency interpolation density will not affect the computational complexity of the algorithm in the subsequent continuous monitoring stage.

[0115] In the frequency offset fine estimation stage, a minimum of only three complex multiplication operations are required for each operation. Compared with several existing strategies such as FFT+CZT, FFT+Zoom-FFT and apFFT, the computational complexity of the frequency offset fine estimation stage is reduced by 93%, 86.9% and 75% respectively.

[0116] In a specific implementation of the present application, before correcting the preprocessed signal using the second frequency to eliminate the influence of the frequency deviation to obtain the restored signal, the estimation method further includes:

[0117] Performing a recursive operation on the second frequency to confirm whether the second frequency meets the accuracy requirement;

[0118] If the second frequency does not meet the accuracy requirement, an early warning is issued and the second frequency is re-estimated;

[0119] If the second frequency meets the accuracy requirement, it is confirmed that the second frequency meets the calibration condition.

[0120] In a specific implementation manner of the present application, performing a recursive operation on the second frequency to confirm whether the second frequency meets the accuracy requirement includes:

[0121] Set two estimation boundaries at the upper and lower rth spectrum line positions of the second frequency at the current moment;

[0122] Using the recursive property of the CIC-mSDFT algorithm, calculate the third spectrum of the second frequency at the current moment and the fourth spectrum of the upper and lower spectral lines;

[0123] If the amplitude of the fourth spectrum at the current moment is higher than the corresponding estimated boundary, and the first amplitude is not less than θ times the second amplitude, then it is confirmed that the second frequency at the current moment meets the accuracy requirement, otherwise, it does not meet the requirement; wherein the first amplitude is the amplitude of the third spectrum at the current moment, the second amplitude is the amplitude of the third spectrum at the previous moment, and θ is a value between (0,1].

[0124] Continue to refer Figure 3 , using the recursive property of CIC-mSDFT to continuously calculate its amplitude spectrum. If the amplitude of the estimated spectrum line is always higher than the two spectrum lines on the left and right, and not lower than θ times the amplitude estimated at the previous moment, θ is an estimated value between (0,1], then the frequency deviation estimation is considered accurate. If it does not meet the requirements, it is considered that the frequency deviation drifts out of the trajectory range, and an alarm is issued and re-estimated.

[0125] In the continuous monitoring and detection stage, that is, continuously monitoring whether the second frequency meets the calibration conditions, a minimum of only three complex multiplication operations are required for each operation. Compared with several existing strategies such as FFT+CZT, FFT+Zoom-FFT and apFFT, the computational complexity of this application is reduced by 93%, 86.9% and 75% respectively.

[0126] In a specific implementation of the present application, before collecting the received digital signal, the estimation method further includes:

[0127] receiving an optical signal sent from a transmitting end;

[0128] filtering the received optical signal to obtain a filtered signal;

[0129] Mixing the filtered signal with the generated reference light to obtain mixed light;

[0130] Convert the mixed light into a current signal;

[0131] Convert the current signal into a digital signal.

[0132] Continue to refer Figure 1, the receiving end includes a local oscillator (LO) laser, an optical filter, an optical mixer, a photodetector and an ADC module; the optical filter is used to receive the digital signal sent from the transmitting end and filter it; the local oscillator is used to generate reference light; the optical mixer mixes the filtered signal with the generated reference light; the photodetector converts the mixed light into a current signal; the current signal contains the original signal information. The ADC module converts the current signal into a digital signal; after the receiving end of the present application receives the optical signal, it is filtered by the optical filter, the reference light is generated by the LO laser, and then the optical mixer completes the mixing of the reference light and the filtered signal, the photodetector completes the conversion of the mixed light to the current signal, and the ADC module completes the conversion of the current signal to the digital signal. The current signal is converted into a digital signal by an analog-to-digital converter (ADC) for processing by a digital signal processor (DSP).

[0133] refer to Figure 5 The present application provides a two-stage frequency deviation estimation device for a receiver based on FFT+CIC-mSDFT, comprising:

[0134] A signal acquisition module 51 is configured to acquire received digital signals;

[0135] The preprocessing module 52 is configured to preprocess the digital signal to eliminate the symbol phase and phase noise to obtain a preprocessed signal;

[0136] The frequency offset estimation module 53 is configured to use the FFT algorithm to perform a rough frequency offset estimation on the preprocessed signal to obtain a first frequency corresponding to the maximum value of the amplitude-spectrum frequency;

[0137] The frequency offset interpolation module 54 is configured to use the CIC-mSDFT algorithm to perform an accurate frequency offset estimation on the first frequency to obtain a second frequency corresponding to the maximum value of the amplitude spectrum;

[0138] The signal recovery module 55 is configured to correct the pre-processed signal using the second frequency that meets the calibration condition to eliminate the influence of the frequency deviation and obtain a recovered signal.

[0139] The pre-processing module 52 is specifically configured as follows:

[0140] Performing quadratic processing on the digital signal to eliminate the symbol phase information therein;

[0141] The digital signal with symbol phase information eliminated is subjected to sliding average processing to remove phase noise therein to obtain a preprocessed signal.

[0142] The digital signal is expressed as:

[0143] x(n)=exp{j[θ s (n)+ΔωnT O+θ L (n)+θ n (n)]};

[0144] In the formula, θ s (n) is the symbol phase information of the nth sampling signal, θ s (n) represents the modulation information; Δω is the angular frequency information of the frequency deviation to be estimated, T0 = 1 / f s is the sampling period, ΔωnT0=2πΔfnT0 is the phase error introduced by the frequency deviation, θ L (n) is the phase noise caused by the laser line width, θ n (n) is the phase noise caused by the spontaneous emission of the optical amplifier;

[0145] The digital signal with symbol phase information eliminated is expressed as:

[0146] x 4 (n) = exp{j[4ΔωnT O +4θ L (n)+4θ n (n)]}.

[0147] The frequency offset estimation module 53 is specifically configured as follows:

[0148] Calculating a first spectrum of the preprocessed signal using a predetermined number of FFT points;

[0149] In the first spectrum, the peak search is performed to find the first frequency f0 and the first index k corresponding to the maximum value of the spectrum.

[0150] The frequency offset interpolation module 54 is configured as follows:

[0151] Perform a times frequency interpolation in the upper and lower spectrum lines of the first frequency f0, and use the CIC-mSDFT algorithm to calculate the second spectrum of the interpolation frequency point;

[0152] In the second spectrum, the second frequency Δf corresponding to the maximum value of the spectrum is searched by peak value. est and the corresponding second index k final .

[0153] The two-stage frequency deviation estimation device for a receiver based on FFT+CIC-mSDFT also includes a continuous monitoring module configured as follows:

[0154] Performing a recursive operation on the second frequency to confirm whether the second frequency meets the accuracy requirement;

[0155] If the second frequency does not meet the accuracy requirement, an early warning is issued and the second frequency is re-estimated;

[0156] If the second frequency meets the accuracy requirement, it is confirmed that the second frequency meets the calibration condition.

[0157] The continuous monitoring module is specifically configured as follows:

[0158] Two estimation boundaries are set at the upper and lower rth spectrum line positions of the second frequency at the current moment;

[0159] Using the recursive property of the CIC-mSDFT algorithm, calculate the third spectrum of the second frequency at the current moment and the fourth spectrum of the upper and lower spectral lines;

[0160] If the amplitude of the fourth spectrum at the current moment is higher than the corresponding estimated boundary, and the first amplitude is not less than θ times the second amplitude, then it is confirmed that the second frequency at the current moment meets the accuracy requirement, otherwise, it does not meet the requirement; wherein the first amplitude is the amplitude of the third spectrum at the current moment, the second amplitude is the amplitude of the third spectrum at the previous moment, and θ is a value between (0,1].

[0161] The two-stage frequency deviation estimation device of the receiver based on FFT+CIC-mSDFT also includes:

[0162] A receiving end receives an optical signal sent from a transmitting end;

[0163] An optical filter, filtering the received optical signal to obtain a filtered signal;

[0164] An optical mixer mixes the filtered signal with the generated reference light to obtain a mixed light;

[0165] Photodetector, which converts the mixed light into a current signal;

[0166] ADC module converts the current signal into a digital signal.

[0167] The two-stage frequency deviation estimation device of the receiver based on FFT+CIC-mSDFT is applied to the receiving end, and the receiving end and the transmitting end transmit digital signals through the transmission path; the transmitting end maps and modulates the data source through 16 / 64QAM to obtain an I / Q modulated signal, and then generates a modulated optical signal through a laser; and transmits the modulated optical signal to the receiving end through the transmission path.

[0168] The embodiment of the present application discloses a two-stage frequency deviation estimation device for a receiver based on FFT+CIC-mSDFT, which collects received digital signals; preprocesses the digital signals to eliminate symbol phase and phase noise to obtain preprocessed signals; uses FFT algorithm to roughly estimate the frequency deviation of the preprocessed signals to obtain the first frequency corresponding to the maximum value of the amplitude spectrum; uses CIC-mSDFT algorithm to accurately estimate the frequency deviation of the first frequency to obtain the second frequency corresponding to the maximum value of the amplitude spectrum; uses the second frequency that meets the calibration conditions to correct the preprocessed signal to eliminate the influence of the frequency deviation to obtain a restored signal. The CIC-mSDFT algorithm proposed in the present application has excellent recursive properties. For the spectrum estimation of each new input moment, it only takes one complex multiplication operation to update the spectrum information of a single spectral line. Since the complexity of the recursive operation is extremely low, it can realize real-time frequency deviation estimation and compensation on DSP, which is suitable for high-order M-QAM modulated signals and has early warning characteristics when the frequency deviation exceeds the monitoring range.

[0169] Those skilled in the art can understand that the accompanying drawings are only schematic diagrams of one embodiment, and the modules or processes in the accompanying drawings are not necessarily required to implement the present application.

[0170] Those skilled in the art can understand that the modules in the device in the embodiment can be distributed in the device in the embodiment according to the description of the embodiment, or can be changed accordingly and located in one or more devices different from the embodiment. The modules in the above embodiment can be combined into one module, or can be further divided into multiple sub-modules.

[0171] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A two-stage receiver frequency offset estimation method based on FFT+CIC-mSDFT, characterized in that: include: Collecting received digital signals; Preprocessing the digital signal to eliminate symbol phase and phase noise to obtain a preprocessed signal; Using an FFT algorithm to roughly estimate the offset frequency of the preprocessed signal to obtain a first frequency corresponding to a maximum value of the amplitude-frequency spectrum; Using the CIC-mSDFT algorithm to perform accurate frequency estimation on the first frequency to obtain a second frequency corresponding to the maximum value of the amplitude-spectrum ratio; The preprocessed signal is corrected using a second frequency that meets the calibration condition to eliminate the influence of the frequency deviation to obtain a restored signal.

2. The two-stage receiver frequency offset estimation method based on FFT+CIC-mSDFT according to claim 1, characterized in that: The preprocessing of the digital signal to eliminate the symbol phase and phase noise to obtain a preprocessed signal includes: Performing quadratic processing on the digital signal to eliminate symbol phase information therein; The digital signal with symbol phase information eliminated is subjected to sliding average processing to remove phase noise therein to obtain a preprocessed signal.

3. The two-stage receiver frequency offset estimation method based on FFT+CIC-mSDFT according to claim 2, characterized in that: The digital signal is expressed by the formula: x(n)=exp{j[θ s (n)+DωnT O +θ L (n)+θ n (n)]}; In the formula, θ s (n) is the symbol phase information of the nth sampling signal, θ s (n) represents the modulation information; Δω is the angular frequency information of the frequency deviation to be estimated, T0 = 1 / f s is the sampling period, ΔωnT0=2πΔfnT0 is the phase error introduced by the frequency deviation, θ L (n) is the phase noise caused by the laser line width, θ n (n) is the phase noise caused by the spontaneous emission of the optical amplifier; The digital signal eliminating the symbol phase information is expressed by the formula: x 4 (n)=exp{j[4ΔωnT O +4θ L (n)+4θ n (n)]}。 4. The two-stage receiver frequency offset estimation method based on FFT+CIC-mSDFT according to claim 1, characterized in that: The method of using the FFT algorithm to roughly estimate the offset frequency of the preprocessed signal to obtain the first frequency corresponding to the maximum value of the amplitude-spectrum ratio comprises: Calculating a first spectrum of the preprocessed signal using a predetermined number of FFT points; In the first spectrum, a peak search is performed to find a first frequency f0 and a first index k corresponding to the maximum value of the spectrum.

5. The two-stage receiver frequency offset estimation method based on FFT+CIC-mSDFT according to claim 1, characterized in that: The using of the CIC-mSDFT algorithm to accurately estimate the offset frequency of the first frequency to obtain the second frequency corresponding to the maximum value of the amplitude-spectrum ratio comprises: Perform a times frequency interpolation in the upper and lower spectrum lines of the first frequency f0, and calculate the second spectrum of the interpolation frequency point using the CIC-mSDFT algorithm; In the second spectrum, the second frequency Δf corresponding to the maximum value of the spectrum is searched by peak value. est and the corresponding second index k final .

6. The two-stage receiver frequency offset estimation method based on FFT+CIC-mSDFT according to claim 1, characterized in that: Before correcting the preprocessed signal by using the second frequency to eliminate the influence of the frequency deviation to obtain a restored signal, the estimation method further includes: Performing a recursive operation on the second frequency to confirm whether the second frequency meets the accuracy requirement; If the second frequency does not meet the accuracy requirement, an early warning is issued and the second frequency is re-estimated; If the second frequency meets the accuracy requirement, it is confirmed that the second frequency meets the calibration condition.

7. The two-stage receiver frequency offset estimation method based on FFT+CIC-mSDFT according to claim 6, characterized in that: The performing a recursive operation on the second frequency to confirm whether the second frequency meets the accuracy requirement includes: Two estimation boundaries are set at the upper and lower rth spectrum line positions of the second frequency at the current moment; Using the recursive property of the CIC-mSDFT algorithm, calculate the third spectrum of the second frequency at the current moment and the fourth spectrum of the upper and lower spectral lines; If the amplitude of the fourth spectrum at the current moment is higher than the corresponding estimated boundary, and the first amplitude is not less than θ times the second amplitude, then it is confirmed that the second frequency at the current moment meets the accuracy requirement, otherwise, it does not meet the requirement; wherein the first amplitude is the amplitude of the third spectrum at the current moment, the second amplitude is the amplitude of the third spectrum at the previous moment, and θ is a value between (0,1].

8. The two-stage receiver frequency offset estimation method based on FFT+CIC-mSDFT according to claim 1, characterized in that: Before collecting the received digital signal, the estimation method further includes: receiving an optical signal sent from a transmitting end; filtering the received optical signal to obtain a filtered signal; Mixing the filtered signal with the generated reference light to obtain mixed light; Converting the mixed light into a current signal; The current signal is converted into a digital signal.

9. The two-stage receiver frequency offset estimation method based on FFT+CIC-mSDFT according to any one of claim 8, characterized in that: The two-stage frequency deviation estimation method of the receiver based on FFT+CIC-mSDFT is applied to the receiving end, and the receiving end and the transmitting end transmit the digital signal through a transmission path; the transmitting end performs 16 / 64QAM mapping and signal modulation on the data source to obtain an I / Q modulated signal, and then generates a modulated optical signal through a laser; The modulated optical signal is transmitted to the receiving end through the transmission path.

10. A two-stage frequency deviation estimation device for a receiver based on FFT+CIC-mSDFT, characterized in that: include: A signal acquisition module is configured to acquire received digital signals; A preprocessing module is configured to preprocess the digital signal to eliminate symbol phase and phase noise to obtain a preprocessed signal; A frequency offset estimation module is configured to perform a rough frequency offset estimation on the preprocessed signal using an FFT algorithm to obtain a first frequency corresponding to a maximum value of the amplitude-frequency spectrum; A frequency offset interpolation module is configured to use a CIC-mSDFT algorithm to perform an accurate frequency offset estimation on the first frequency to obtain a second frequency corresponding to a maximum value of the amplitude-spectrum ratio; The signal recovery module is configured to correct the preprocessed signal using a second frequency that meets the calibration condition to eliminate the influence of the frequency deviation and obtain a recovered signal.

Images