WO2018068177A1

WO2018068177A1 - Method and device for eliminating noise by comb filtering and frequency domain adaptive equalization device

Info

Publication number: WO2018068177A1
Application number: PCT/CN2016/101651
Authority: WO
Inventors: 马文凯; 查如冰迈克尔; 魏帅; 赵曼; 陈俊; 李良川
Original assignee: 华为技术有限公司
Priority date: 2016-10-10
Filing date: 2016-10-10
Publication date: 2018-04-19
Also published as: CN109716664B; CN109716664A

Abstract

Embodiments of the invention provide a method and device for eliminating noise by comb filtering and a frequency domain adaptive equalization device. The method for eliminating noise by comb filtering comprises: acquiring first frequency domain data, wherein the number of frequency points of the first frequency domain data is N; transforming the first frequency domain data into first time domain data by means of N-point IFFT; determining a data position of aliasing noise generated in the first time domain data according to a data position correspondence relationship between the first time domain data and preset reference data, and acquiring second time domain data by resetting data at the data position; and performing comb filtering on the second time domain data to obtain second frequency domain data after the aliasing noise is eliminated, wherein the number of frequency points of the second frequency domain data is M. The method for eliminating noise by comb filtering provided by the embodiments of the invention can eliminate aliasing noise by means of comb filtering and improve the performance of frequency domain comb filtering.

Description

梳状滤波噪声消除方法、装置及频域自适应均衡装置Comb filtering noise elimination method, device and frequency domain adaptive equalization device

技术领域Technical field

本发明涉及数据处理技术领域，尤其涉及一种梳状滤波噪声消除方法、装置及频域自适应均衡装置。The present invention relates to the field of data processing technologies, and in particular, to a comb filtering noise canceling method and apparatus, and a frequency domain adaptive equalizing apparatus.

背景技术Background technique

近些年，具备长的冲击响应的滤波器在自适应数字数据处理中被广泛应用，例如在频域自适应均衡装置中，传输链路中的色度色散(Chromatic dispersion，简称CD)，偏振模色散(Polarization Mode Dispersion，简称PMD)和载波恢复(Carrier Recovery，简称CR)都需要通过滤波器进行均衡补偿。一种有效的降低大抽头滤波器复杂度的方法是频域算法，将时域的卷积运算转换为频域的相乘运算。In recent years, filters with long impulse response have been widely used in adaptive digital data processing, such as in the frequency domain adaptive equalization device, Chromatic dispersion (CD) in the transmission link, polarization Both Polarization Mode Dispersion (PMD) and Carrier Recovery (CR) require equalization compensation through filters. An effective method for reducing the complexity of the large tap filter is the frequency domain algorithm, which converts the time domain convolution operation into the frequency domain multiplication operation.

目前，在频域自适应均衡装置中，经常使用高点数的快速傅立叶变换(Fast Fourier Transformation，简称FFT)/快速傅立叶逆变换(Inverse Fast Fourier Transformation，简称IFFT)，高点数的FFT/IFFT使得计算复杂度高、装置成本代价大。为了降低计算复杂度以及降低成本，通常在频域自适应均衡装置中增加频域梳状滤波单元和时域梳状滤波单元，使得数据分辨率降低D倍，从而将FFT/IFFT的点数降低D倍，降低资源功耗。At present, in the frequency domain adaptive equalization device, Fast Fourier Transformation (FFT)/Inverse Fast Fourier Transformation (IFFT) is often used, and high-point FFT/IFFT makes calculation The complexity is high and the cost of the device is high. In order to reduce the computational complexity and reduce the cost, the frequency domain comb filter unit and the time domain comb filter unit are usually added in the frequency domain adaptive equalization device, so that the data resolution is reduced by D times, thereby reducing the number of points of the FFT/IFFT. Times, reducing resource power consumption.

但是，频域梳状滤波单元会产生混叠噪声，导致梳状滤波性能较差。However, the frequency domain comb filter unit generates aliasing noise, resulting in poor comb filtering performance.

发明内容Summary of the invention

本发明实施例提供一种梳状滤波噪声消除方法、装置及频域自适应均衡装置，可以在实现梳状滤波的基础上消除混叠噪声，提升了频域梳状滤波的性能。Embodiments of the present invention provide a comb filtering noise canceling method and apparatus, and a frequency domain adaptive equalizing apparatus, which can eliminate aliasing noise on the basis of implementing comb filtering, and improve performance of frequency domain comb filtering.

第一方面，本发明实施例提供一种梳状滤波噪声消除方法，该方法包括：获取第一频域数据，第一频域数据的频点数为N，将第一频域数据通过N点IFFT转换为第一时域数据，根据第一时域数据与预设参考数据之间的数据位置对应关系，确定第一时域数据中产生混叠噪声的数据位置，并在数据位置处将数据置零获取第二时域数据，对第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据，第二频域数据的频点数为M。In a first aspect, an embodiment of the present invention provides a method for canceling comb noise, the method comprising: acquiring first frequency domain data, the frequency of the first frequency domain data is N, and the first frequency domain data is passed through an N point IFFT. Converting to the first time domain data, determining a data location of the aliasing noise in the first time domain data according to a data position correspondence relationship between the first time domain data and the preset reference data, and at the data location The data is zeroed to obtain the second time domain data, and the second time domain data is comb-filtered to obtain the second frequency domain data after the aliasing noise is eliminated, and the frequency number of the second frequency domain data is M.

通过第一方面提供的梳状滤波噪声消除方法，实现了对频域数据的梳状滤波，在进行梳状滤波之前，先通过预设参考数据确定可能产生混叠噪声的数据位置，将相应数据位置处的数据置0，这样就可以避免产生混叠噪声。相比现有技术中的频域梳状滤波装置，可以在实现梳状滤波的基础上消除混叠噪声，提升了频域梳状滤波的性能。Through the comb filtering noise elimination method provided by the first aspect, the comb filtering of the frequency domain data is realized, and before the comb filtering, the data position of the aliasing noise may be determined by using the preset reference data, and the corresponding data is The data at the location is set to 0 to avoid aliasing noise. Compared with the frequency domain comb filter device in the prior art, the aliasing noise can be eliminated on the basis of the comb filter, and the performance of the frequency domain comb filter is improved.

可选的，在第一方面的一种可能的实施方式中，对第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据，包括：对第二时域数据进行时域梳状滤波，获得M点时域数据，将M点时域数据通过M点FFT转换为第二频率数据。Optionally, in a possible implementation manner of the first aspect, comb filtering the second time domain data to obtain the second frequency domain data after the anti-aliasing noise is obtained, including: performing the second time domain data The time domain comb filtering obtains the M point time domain data, and converts the M point time domain data into the second frequency data through the M point FFT.

通过该可能的实施方式提供的梳状滤波噪声消除方法，通过先对第二时域数据进行时域梳状滤波然后进行数据的时频转换，这样，在将时域数据转换为频域数据时，仅需要对M点时域数据进行M点FFT处理，降低了FFT的点数，进一步降低了计算复杂度以及成本，提升了频域梳状滤波的性能。The comb filter noise cancellation method provided by the possible implementation method performs time domain comb filtering on the second time domain data and then performs time-frequency conversion of the data, so that when the time domain data is converted into the frequency domain data, Only M point FFT processing is needed for the M point time domain data, which reduces the number of FFT points, further reduces the computational complexity and cost, and improves the performance of the frequency domain comb filtering.

可选的，在第一方面的一种可能的实施方式中，对第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据，包括：将第二时域数据通过N点FFT转换为N点频域数据，对N点频域数据进行频域梳状滤波，获得第二频域数据。Optionally, in a possible implementation manner of the first aspect, the second time domain data is comb filtered to obtain the second frequency domain data after the aliasing noise is removed, including: passing the second time domain data The N-point FFT is converted into N-point frequency domain data, and frequency-domain comb filtering is performed on the N-point frequency domain data to obtain second frequency domain data.

通过该可能的实施方式提供的梳状滤波噪声消除方法，通过先对第二时域数据进行时频转换然后进行频域梳状滤波，由于频域梳状滤波的实现复杂度低于时域梳状滤波的复杂度，因此进一步降低了计算复杂度以及成本，提升了频域梳状滤波的性能。The comb filter noise cancellation method provided by the possible implementation manner is characterized in that the complexity of the frequency domain comb filter is lower than that of the time domain comb by first performing time-frequency conversion on the second time domain data and then performing frequency domain comb filtering. The complexity of the filtering, thus further reducing the computational complexity and cost, and improving the performance of the frequency domain comb filtering.

第二方面，本发明实施例提供一种梳状滤波噪声消除装置，该装置可以包括：获取模块、数据转换模块、设置模块和混叠噪声消除模块，其中，获取模块用于获取第一频域数据，第一频域数据的频点数为N，数据转换模块用于将第一频域数据通过N点IFFT转换为第一时域数据，设置模块用于根据第一时域数据与预设参考数据之间的数据位置对应关系，确定第一时域数据中产生混叠噪声的数据位置，并在数据位置处将数据置零获取第二时域数据，混叠噪声消除模块用于对第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据，第二频域数据的频点数为M。In a second aspect, an embodiment of the present invention provides a comb filter noise canceling apparatus, where the apparatus may include: an acquisition module, a data conversion module, a setting module, and an aliasing noise cancellation module, where the acquiring module is configured to acquire the first frequency domain. Data, the number of frequency points of the first frequency domain data is N, the data conversion module is configured to convert the first frequency domain data into the first time domain data through the N point IFFT, and the setting module is configured to use the first time domain data and the preset reference The data position correspondence between the data, determining the data position of the first time domain data to generate aliasing noise, and zeroing the data at the data position to obtain the second time domain data, and the aliasing noise cancellation module is used for the second Time domain data is comb filtered to obtain anti-aliasing noise The second frequency domain data after the sound, the frequency number of the second frequency domain data is M.

可选的，在第二方面的一种可能的实施方式中，混叠噪声消除模块具体用于：对第二时域数据进行时域梳状滤波，获得M点时域数据，将M点时域数据通过M点FFT转换为第二频率数据。Optionally, in a possible implementation manner of the second aspect, the aliasing noise cancellation module is specifically configured to perform time domain comb filtering on the second time domain data to obtain M point time domain data, and when the M point is The domain data is converted to the second frequency data by the M-point FFT.

可选的，在第二方面的一种可能的实施方式中，混叠噪声消除模块具体用于：将第二时域数据通过N点FFT转换为N点频域数据，对N点频域数据进行频域梳状滤波，获得第二频域数据。Optionally, in a possible implementation manner of the second aspect, the aliasing noise cancellation module is specifically configured to: convert the second time domain data into N-point frequency domain data by N-point FFT, and perform N-point frequency domain data on the N-point frequency domain data. Frequency domain comb filtering is performed to obtain second frequency domain data.

第三方面，本发明实施例提供一种频域自适应均衡装置，该装置可以包括：时域到频域变换单元、色散补偿单元、频域自适应均衡单元、频域到时域变换单元、载波恢复单元、硬判决单元、误差计算单元、时域误差梳状滤波单元、频域自适应均衡系数更新单元、插值上采样单元，以及梳状滤波噪声消除装置。In a third aspect, an embodiment of the present invention provides a frequency domain adaptive equalization apparatus, where the apparatus may include: a time domain to frequency domain transform unit, a dispersion compensation unit, a frequency domain adaptive equalization unit, a frequency domain to time domain transform unit, A carrier recovery unit, a hard decision unit, an error calculation unit, a time domain error comb filter unit, a frequency domain adaptive equalization coefficient update unit, an interpolation upsampling unit, and a comb filter noise canceling device.

其中，时域到频域变换单元，用于将输入的时域数据进行数据交叠，将数据交叠后的数据从时域变换到频域并输出至色散补偿单元。色散补偿单元，用于对时域到频域变换单元输出的数据在频域进行色散补偿，将色散补偿后的数据分别输出至梳状滤波噪声消除装置和频域自适应均衡单元。频域自适应均衡单元，用于根据差值上采样单元输出的抽头系数对色散补偿单元输出的数据进行频域均衡，将频域均衡后的数据输出至频域到时域变换单元。频域到时域变换单元，用于将频域自适应均衡单元输出的数据从频域变换到时域后进行数据解交叠，将数据解交叠后的数据输出至载波恢复单元。载波恢复单元，用于对频域到时域变换单元输出的数据进行载波恢复，将载波恢复后的数据分别输出至误差计算单元和硬判决单元。硬判决单元，用于对载波恢复单元输出的数据进行硬判决，将硬判决后的数据输出至误差计算单元。误差计算单元，用于根据载波恢复单元输出的数据和硬判决单元输出的数据计算数据误差，将数据误差输出至时域误差梳状滤波单元。时域误差梳状滤波单元，用于对误差计算单元输出的数据误差进行时域梳状滤波，将时域梳状滤波后的数据输出至频域自适应均衡系数更新单元。频域自适应均衡系数更新单元，用于根据时域误差梳状滤波单元输出的数据和梳状滤波噪声消除装置输出的数据进行频域自适应抽头系数更新，并将抽头系数更新结果输出至插值上采样单元。插值上采样单元，用于将频域自适应均衡系数更新单元输出的抽头系数更新结果进行插值上采样，将差值上采样后的抽头系数输出至频域自适应均衡单元。The time domain to frequency domain transform unit is configured to overlap the input time domain data, convert the data overlapped data from the time domain to the frequency domain, and output the data to the dispersion compensation unit. The dispersion compensation unit is configured to perform dispersion compensation on the data outputted by the time domain to the frequency domain transform unit in the frequency domain, and output the dispersion compensated data to the comb filter noise canceling device and the frequency domain adaptive equalization unit respectively. The frequency domain adaptive equalization unit is configured to perform frequency domain equalization on the data output by the dispersion compensation unit according to the tap coefficients output by the difference upsampling unit, and output the frequency domain equalized data to the frequency domain to the time domain transform unit. The frequency domain to time domain transform unit is configured to transform the data outputted by the frequency domain adaptive equalization unit from the frequency domain to the time domain, and perform data de-interlacing, and output the data after the data de-interleaving to the carrier recovery unit. The carrier recovery unit is configured to perform carrier recovery on the data output from the frequency domain to the time domain transform unit, and output the data after the carrier recovery to the error calculation unit and the hard decision unit respectively. The hard decision unit is configured to perform hard decision on the data output by the carrier recovery unit, and output the hard-decised data to the error calculation unit. The error calculation unit is configured to calculate a data error according to the data output by the carrier recovery unit and the data output by the hard decision unit, and output the data error to the time domain error comb filter unit. The time domain error comb filtering unit is configured to perform time domain comb filtering on the data error output by the error calculating unit, and output the time domain comb filtered data to the frequency domain adaptive equalization coefficient updating unit. a frequency domain adaptive equalization coefficient updating unit, configured to perform frequency domain adaptive tap coefficient update according to data output by the time domain error comb filtering unit and data output by the comb filtering noise canceling device, and output the tap coefficient update result to the interpolation Upsampling unit. Interpolation upsampling unit for more adaptive frequency domain adaptive equalization coefficients The tap coefficient update result output by the new unit is subjected to interpolation upsampling, and the tap coefficients after the difference upsampling are output to the frequency domain adaptive equalization unit.

其中，梳状滤波噪声消除装置包括：获取模块、数据转换模块、设置模块、混叠噪声消除模块和发送模块。The comb filter noise canceling device comprises: an acquisition module, a data conversion module, a setting module, an aliasing noise canceling module and a sending module.

获取模块用于获取色散补偿单元输出的数据，色散补偿单元输出的数据的频点数为N，数据转换模块用于将色散补偿单元输出的数据通过N点IFFT转换为第一时域数据，设置模块用于根据第一时域数据与时域误差梳状滤波单元输出的数据之间的数据位置对应关系，确定第一时域数据中产生混叠噪声的数据位置，并在数据位置处将数据置零获取第二时域数据，混叠噪声消除模块用于对第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据，第二频域数据的频点数为M，发送模块用于将第二频域数据输出至频域自适应均衡系数更新单元。The obtaining module is configured to obtain data output by the dispersion compensation unit, and the frequency of the data output by the dispersion compensation unit is N, and the data conversion module is configured to convert the data output by the dispersion compensation unit into the first time domain data through the N-point IFFT, and set the module. And determining, according to a data position correspondence relationship between the first time domain data and the data output by the time domain error comb filtering unit, determining a data position of the aliasing noise generated in the first time domain data, and setting the data at the data location The second time domain data is obtained by zero, and the aliasing noise cancellation module is configured to comb filter the second time domain data to obtain second frequency domain data after the aliasing noise is eliminated, and the frequency number of the second frequency domain data is M, The sending module is configured to output the second frequency domain data to the frequency domain adaptive equalization coefficient updating unit.

可选的，在第三方面的一种可能的实施方式中，该装置还可以包括：样值截取单元和硬判决截取单元。其中，样值截取单元用于在载波恢复单元输出的数据中截取一部分输出至误差计算单元。硬判决截取单元用于在硬判决单元输出的数据中截取一部分输出至误差计算单元。其中，样值截取单元截取的数据在载波恢复单元输出的数据中的位置与硬判决截取单元截取的数据在硬判决单元输出的数据中的位置相同。Optionally, in a possible implementation manner of the third aspect, the apparatus may further include: a sample intercepting unit and a hard decision intercepting unit. The sample value intercepting unit is configured to intercept a part of the output to the error calculating unit in the data output by the carrier recovery unit. The hard decision intercepting unit is configured to intercept a part of the output to the error calculating unit in the data output by the hard decision unit. The position of the data intercepted by the sample intercepting unit in the data output by the carrier recovery unit is the same as the position of the data intercepted by the hard decision intercepting unit in the data output by the hard decision unit.

通过该可能的实施方式提供的频域自适应均衡装置，通过样值截取单元在载波恢复单元输出的数据中截取一部分数据，通过硬判决截取单元在硬判决单元输出的数据中截取一部分数据，由于样值截取单元截取的数据在载波恢复单元输出的数据中的位置与硬判决截取单元截取的数据在硬判决单元输出的数据中的位置相同，因此，误差计算单元根据截取的数据计算数据误差，可以大幅降低数据处理时延，进而减小抽头系数更新环路时延，提升了频域自适应滤波装置的***性能。The frequency domain adaptive equalization apparatus provided by the possible implementation section intercepts a part of data in the data output by the carrier recovery unit by the sample intercepting unit, and intercepts a part of the data in the data output by the hard decision unit by the hard decision intercepting unit, The position of the data intercepted by the sample intercepting unit in the data output by the carrier recovery unit is the same as the position of the data intercepted by the hard decision intercepting unit in the data output by the hard decision unit. Therefore, the error calculating unit calculates the data error according to the intercepted data. The data processing delay can be greatly reduced, thereby reducing the tap coefficient update loop delay, and improving the system performance of the frequency domain adaptive filtering device.

可选的，在第三方面的一种可能的实施方式中，混叠噪声消除模块具体用于：对第二时域数据进行时域梳状滤波，获得M点时域数据，将M点时域数据通过M点FFT转换为第二频率数据。Optionally, in a possible implementation manner of the third aspect, the aliasing noise cancellation module is specifically configured to perform time domain comb filtering on the second time domain data to obtain M point time domain data, where the M point is The domain data is converted to the second frequency data by the M-point FFT.

可选的，在第三方面的一种可能的实施方式中，混叠噪声消除模块具体用于：将第二时域数据通过N点FFT转换为N点频域数据，对N点频域数据进行频域梳状滤波，获得第二频域数据。Optionally, in a possible implementation manner of the third aspect, the aliasing noise cancellation module is specifically configured to: convert the second time domain data into N-point frequency domain data by N-point FFT, and compare the frequency of the N-point frequency domain. According to the frequency domain comb filtering, the second frequency domain data is obtained.

结合上述第一方面以及第一方面的各可能的实施方式、第二方面以及第二方面的各可能的实施方式、第三方面以及第三方面的各可能的实施方式，N为大于1的整数，N为M的整数倍，M为大于1的整数。In combination with the first aspect described above and the possible embodiments of the first aspect, the second aspect, and the possible implementations of the second aspect, the third aspect, and the possible implementations of the third aspect, N is an integer greater than one. , N is an integer multiple of M, and M is an integer greater than 1.

本发明实施例提供了一种梳状滤波噪声消除方法、装置及频域自适应均衡装置，其中，梳状滤波噪声消除方法包括：获取第一频域数据，第一频域数据的频点数为N，将第一频域数据通过N点IFFT转换为第一时域数据，根据第一时域数据与预设参考数据之间的数据位置对应关系，确定第一时域数据中产生混叠噪声的数据位置，并在数据位置处将数据置零获取第二时域数据，对第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据，第二频域数据的频点数为M。本发明实施例提供的梳状滤波噪声消除方法，通过预设参考数据确定第一频域数据中可能产生混叠噪声的数据位置，将相应数据位置处的数据置0，可以在实现梳状滤波的基础上消除混叠噪声，提升频域梳状滤波的性能。The embodiment of the present invention provides a comb filtering noise canceling method and device, and a frequency domain adaptive equalizing apparatus, wherein the comb filtering noise canceling method includes: acquiring first frequency domain data, and the frequency number of the first frequency domain data is N, converting the first frequency domain data into the first time domain data by using the N point IFFT, and determining the aliasing noise in the first time domain data according to the data position correspondence between the first time domain data and the preset reference data Data location, and zeroing the data at the data location to obtain second time domain data, comb filtering the second time domain data, obtaining second frequency domain data after eliminating aliasing noise, and second frequency domain data The frequency point is M. The comb filtering noise canceling method provided by the embodiment of the present invention determines the data position of the first frequency domain data that may generate aliasing noise by using the preset reference data, and sets the data at the corresponding data position to 0, and can implement comb filtering. Eliminate aliasing noise and improve the performance of frequency domain comb filtering.

附图说明DRAWINGS

为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作一简单地介绍，显而易见地，下面描述中的附图是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, a brief description of the drawings used in the embodiments or the prior art description will be briefly described below. Obviously, the drawings in the following description It is a certain embodiment of the present invention, and other drawings can be obtained from those skilled in the art without any inventive labor.

图1为现有的频域自适应均衡装置的结构示意图；1 is a schematic structural diagram of a conventional frequency domain adaptive equalization apparatus;

图2A为现有的频域梳状滤波单元输出的数据格式示意图；2A is a schematic diagram of a data format output by a conventional frequency domain comb filter unit;

图2B为现有的时域误差梳状滤波单元输出的数据格式示意图；2B is a schematic diagram of a data format of an output of a conventional time domain error comb filter unit;

图3为本发明实施例一提供的梳状滤波噪声消除方法的流程图；3 is a flowchart of a comb filter noise canceling method according to Embodiment 1 of the present invention;

图4为本发明实施例一提供的梳状滤波噪声消除装置的结构示意图；4 is a schematic structural diagram of a comb filter noise canceling apparatus according to Embodiment 1 of the present invention;

图5为本发明实施例一提供的频域自适应均衡装置的结构示意图；FIG. 5 is a schematic structural diagram of a frequency domain adaptive equalization apparatus according to Embodiment 1 of the present invention;

图6为本发明实施例一提供的样值截取单元输出的数据格式示意图；6 is a schematic diagram of a data format output by a sample intercepting unit according to Embodiment 1 of the present invention;

图7为本发明实施例二提供的频域自适应均衡装置的结构示意图。FIG. 7 is a schematic structural diagram of a frequency domain adaptive equalization apparatus according to Embodiment 2 of the present invention.

具体实施方式 detailed description

为使本发明实施例的目的、技术方案和优点更加清楚，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。The technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. It is a partial embodiment of the invention, and not all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

本发明的说明书和权利要求书及上述附图中的术语“第一”、“第二”、“第三”、“第四”等(如果存在)是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便这里描述的本发明的实施例例如能够以除了在这里图示或描述的那些以外的顺序实施。此外，术语“包括”和“具有”以及他们的任何变形，意图在于覆盖不排他的包含，例如，包含了一系列步骤或单元的过程、方法、***、产品或设备不必限于清楚地列出的那些步骤或单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它步骤或单元。The terms "first", "second", "third", "fourth", etc. (if present) in the specification and claims of the present invention and the above figures are used to distinguish similar objects without having to use To describe a specific order or order. It is to be understood that the data so used may be interchanged as appropriate, such that the embodiments of the invention described herein can be implemented, for example, in a sequence other than those illustrated or described herein. In addition, the terms "comprises" and "comprises" and "the" and "the" are intended to cover a non-exclusive inclusion, for example, a process, method, system, product, or device that comprises a series of steps or units is not necessarily limited to Those steps or units may include other steps or units not explicitly listed or inherent to such processes, methods, products or devices.

本发明实施例提供的梳状滤波噪声消除方法和装置，主要应用于频域自适应均衡装置中，当然也可以应用于其他需要消除梳状滤波混叠噪声的场景中，本发明对应用场景并不做特别限制。The comb filter noise cancellation method and device provided by the embodiments of the present invention are mainly applied to a frequency domain adaptive equalization device, and can also be applied to other scenarios that need to eliminate comb filter aliasing noise. No special restrictions are imposed.

图1为现有的频域自适应均衡装置的结构示意图。如图1所示，现有的频域自适应均衡装置包括：时域到频域变换单元11、色散补偿单元12、频域自适应均衡单元13、频域到时域变换单元14、载波恢复单元15、硬判决单元16、误差计算单元17、时域误差梳状滤波单元18、频域自适应均衡系数更新单元19和频域梳状滤波单元20。FIG. 1 is a schematic structural diagram of a conventional frequency domain adaptive equalization apparatus. As shown in FIG. 1, the existing frequency domain adaptive equalization apparatus includes: a time domain to frequency domain transform unit 11, a dispersion compensation unit 12, a frequency domain adaptive equalization unit 13, a frequency domain to time domain transform unit 14, and carrier recovery. The unit 15, the hard decision unit 16, the error calculation unit 17, the time domain error comb filter unit 18, the frequency domain adaptive equalization coefficient update unit 19, and the frequency domain comb filter unit 20.

其中，在数据路径上，时域到频域变换单元11将接收到的信号x(t)进行数据交叠后通过FFT变换为频域信号X(f)，色散补偿单元12进行色散补偿，接着，频域自适应均衡单元13在频域进行自适应均衡以补偿信道的PMD和SOP等损伤，然后，频域到时域变换单元14通过IFFT以及数据解交叠后输出时域信号Tx(t)，载波恢复单元15进行载波恢复以补偿收端和发端之间的频差以及相位噪声，输出信号Rx(t)，最后，硬判决单元16根据信号Rx(t)进行硬判决得到信号Dx(t)。Wherein, on the data path, the time domain to frequency domain transform unit 11 performs data overlap on the received signal x(t) and then transforms it into a frequency domain signal X(f) by FFT, and the dispersion compensation unit 12 performs dispersion compensation, and then The frequency domain adaptive equalization unit 13 performs adaptive equalization in the frequency domain to compensate for impairments such as PMD and SOP of the channel, and then the frequency domain to time domain transform unit 14 outputs the time domain signal Tx by using IFFT and data de-interlacing. The carrier recovery unit 15 performs carrier recovery to compensate for the frequency difference and phase noise between the terminating and transmitting terminals, and outputs a signal Rx(t). Finally, the hard decision unit 16 performs a hard decision based on the signal Rx(t) to obtain a signal Dx ( t).

在系数更新路径上，频域梳状滤波单元20使得色散补偿单元12输出的频域数据的频域抽样分辨率降低D倍，误差计算单元17用于计算载波恢复单元15输出的数据Rx(t)和硬判决单元16输出的数据Dx(t)之间的误差数据，该误差数据为时域数据，时域误差梳状滤波单元18使得误差计算单元17输出的时域数据的分辨率降低D倍，频域自适应均衡系数更新单元19根据频域梳状滤波单元20输出的数据和时域误差梳状滤波单元18输出的数据进行抽头系数的迭代更新，然后送给频域自适应均衡单元13。其中，频域自适应均衡系数更新单元19进行的FFT/IFFT的点数降低了D倍，因此降低了计算复杂度以及资源功耗。On the coefficient update path, the frequency domain comb filtering unit 20 causes the output of the dispersion compensation unit 12 The frequency domain sampling resolution of the frequency domain data is reduced by D times, and the error calculating unit 17 is configured to calculate error data between the data Rx(t) output by the carrier recovery unit 15 and the data Dx(t) output by the hard decision unit 16, which The error data is time domain data, and the time domain error comb filtering unit 18 reduces the resolution of the time domain data output by the error calculating unit 17 by D times, and the frequency domain adaptive equalization coefficient updating unit 19 outputs according to the frequency domain comb filtering unit 20. The data output by the data and time domain error comb filtering unit 18 is iteratively updated by the tap coefficients and then sent to the frequency domain adaptive equalization unit 13. The number of points of the FFT/IFFT performed by the frequency domain adaptive equalization coefficient updating unit 19 is reduced by D times, thereby reducing computational complexity and resource power consumption.

但是，现有的频域梳状滤波单元20会产生混叠噪声。具体的，图2A为现有的频域梳状滤波单元输出的数据格式示意图，图2B为现有的时域误差梳状滤波单元输出的数据格式示意图，其中，为了直观地说明如何引入混叠噪声，图2A中色散补偿单元输出的数据为时域数据，频域梳状滤波的原理按照时域梳状滤波的原理进行说明。如图2A所示，色散补偿单元12输出的数据包括data1～data4，频域梳状滤波单元20对所有的数据(data1～data4)进行抽样处理完成梳状滤波。如图2B所示，误差计算单元17输出的数据包括zero、zero、newdata3和newdata3，其中，zero表示数据为0，这是由于频域到时域变换单元14进行的数据解交叠处理而导致的，时域误差梳状滤波单元18仅对newdata3和newdata3进行了梳状滤波。可见，经过频域梳状滤波单元20处理后的数据与经过时域误差梳状滤波单元18处理后的数据不对应，因此，频域梳状滤波单元20产生了混叠噪声。However, the existing frequency domain comb filter unit 20 generates aliasing noise. Specifically, FIG. 2A is a schematic diagram of a data format outputted by a conventional frequency domain comb filter unit, and FIG. 2B is a schematic diagram of a data format output by a conventional time domain error comb filter unit, wherein, in order to visually explain how to introduce aliasing Noise, the data output by the dispersion compensation unit in Fig. 2A is time domain data, and the principle of frequency domain comb filtering is described in accordance with the principle of time domain comb filtering. As shown in FIG. 2A, the data output from the dispersion compensating unit 12 includes data1 to data4, and the frequency domain comb filtering unit 20 performs sampling processing on all data (data1 to data4) to complete comb filtering. As shown in FIG. 2B, the data output by the error calculating unit 17 includes zero, zero, newdata3, and newdata3, where zero indicates that the data is 0, which is caused by the data de-overlap processing by the frequency domain to time domain transform unit 14. The time domain error comb filtering unit 18 only comb filters newdata3 and newdata3. It can be seen that the data processed by the frequency domain comb filtering unit 20 does not correspond to the data processed by the time domain error comb filtering unit 18, and therefore, the frequency domain comb filtering unit 20 generates aliasing noise.

本发明实施例提供的梳状滤波噪声消除方法、装置及频域自适应均衡装置，用于消除梳状滤波过程中产生的混叠噪声。The comb filtering noise canceling method and device and the frequency domain adaptive equalizing device provided by the embodiments of the present invention are used for eliminating aliasing noise generated in the comb filtering process.

图3为本发明实施例一提供的梳状滤波噪声消除方法的流程图，本发明实施例提供的梳状滤波噪声消除方法，执行主体可以为梳状滤波噪声消除装置。如图3所示，本发明实施例提供的梳状滤波噪声消除方法，可以包括：FIG. 3 is a flowchart of a comb filter noise cancellation method according to Embodiment 1 of the present invention. The comb filter noise cancellation method provided by the embodiment of the present invention may be a comb filter noise cancellation device. As shown in FIG. 3, the comb filter noise cancellation method provided by the embodiment of the present invention may include:

步骤101、获取第一频域数据。Step 101: Acquire first frequency domain data.

其中，第一频域数据的频点数为N，N为大于1的整数。The frequency of the first frequency domain data is N, and N is an integer greater than 1.

步骤102、将第一频域数据通过N点IFFT转换为第一时域数据。Step 102: Convert the first frequency domain data into the first time domain data by N-point IFFT.

步骤103、根据第一时域数据与预设参考数据之间的数据位置对应关系，确定第一时域数据中产生混叠噪声的数据位置，并在数据位置处将数据置零获取第二时域数据。Step 103: Corresponding to a data location between the first time domain data and the preset reference data And determining a data location in the first time domain data that generates aliasing noise, and zeroing the data at the data location to obtain second time domain data.

步骤104、对第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据。Step 104: Perform comb filtering on the second time domain data to obtain second frequency domain data after eliminating aliasing noise.

其中，第二频域数据的频点数为M，N为M的整数倍，M为大于1的整数。The frequency of the second frequency domain data is M, N is an integer multiple of M, and M is an integer greater than 1.

具体的，获取的第一频域数据的频点数为N，通过N点IFFT将第一频域数据转换为第一时域数据。预设参考数据为时域数据，且与第一时域数据具有位置上的对应关系，可以理解，若第一时域数据与预设参考数据相对应，则对第一时域数据进行梳状滤波之后将不会产生混叠噪声，若第一时域数据与预设参考数据不对应，则对第一时域数据进行梳状滤波之后将产生混叠噪声。其中，第一时域数据与预设参考数据是否对应，是指第一时域数据与预设参考数据对应位置的数据是否均为0或者均为非0，若均为0或者均为非0则认为第一时域数据与预设参考数据相对应。因此，通过第一时域数据与预设参考数据之间的数据位置对应关系，可以确定在第一时域数据中哪些数据位置将会产生混叠噪声。Specifically, the obtained frequency domain of the first frequency domain data is N, and the first frequency domain data is converted into the first time domain data by the N point IFFT. The preset reference data is time domain data and has a positional correspondence with the first time domain data. It can be understood that if the first time domain data corresponds to the preset reference data, the first time domain data is combed. After filtering, no aliasing noise will be generated. If the first time domain data does not correspond to the preset reference data, aliasing noise will be generated after comb filtering the first time domain data. Whether the data of the first time domain data and the preset reference data corresponds to whether the data of the first time domain data and the preset reference data is 0 or both are non-zero, if all are 0 or are non-zero The first time domain data is considered to correspond to the preset reference data. Therefore, by the data position correspondence between the first time domain data and the preset reference data, it can be determined which data locations in the first time domain data will generate aliasing noise.

在确定了第一时域数据中产生混叠噪声的数据位置后，将该数据位置处的数据置0获得第二时域数据，第二时域数据与预设参考数据是相对应的，对第二时域数据进行梳状滤波，就可以避免产生混叠噪声。而且，第二时域数据进行梳状滤波后得到的第二频域数据的频点数降低为M，这样，如果需要进一步对第二频域数据进行FFT或者IFFT处理时，就可以降低FFT/IFFT的点数，降低了计算复杂度以及成本。After determining the data location of the aliasing noise in the first time domain data, setting the data at the data location to 0 to obtain the second time domain data, the second time domain data corresponding to the preset reference data, The second time domain data is comb filtered to avoid aliasing noise. Moreover, the frequency of the second frequency domain data obtained by comb filtering in the second time domain data is reduced to M, so that if FFT or IFFT processing is further performed on the second frequency domain data, the FFT/IFFT can be reduced. The number of points reduces computational complexity and cost.

其中，第一时域数据中产生混叠噪声的数据位置通常为连续的一段数据位置。Wherein, the data position in the first time domain data to generate aliasing noise is usually a continuous data position.

可见，本发明实施例提供的梳状滤波噪声消除方法，实现了对频域数据的梳状滤波，在进行梳状滤波之前，先通过预设参考数据确定可能产生混叠噪声的数据位置，将相应数据位置处的数据置0，这样就可以避免产生混叠噪声。相比现有技术中的频域梳状滤波装置，可以在实现梳状滤波的基础上消除混叠噪声，提升了频域梳状滤波的性能。It can be seen that the comb filtering noise cancellation method provided by the embodiment of the present invention implements comb filtering on frequency domain data, and before performing comb filtering, first determines data positions that may generate aliasing noise through preset reference data, and The data at the corresponding data location is set to 0 to avoid aliasing noise. Compared with the frequency domain comb filter device in the prior art, the aliasing noise can be eliminated on the basis of the comb filter, and the performance of the frequency domain comb filter is improved.

需要说明的是，N和M的值根据实际需要进行设置，本发明实施例对此不做特别限制。It should be noted that the values of N and M are set according to actual needs, and this embodiment of the present invention No special restrictions are imposed.

需要说明的是，本发明实施例对于预设参考数据的设置不做特别限制，根据需要进行设置。It should be noted that, in the embodiment of the present invention, the setting of the preset reference data is not particularly limited, and is set as needed.

可选的，作为一种具体的实现方式，步骤104，对第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据，可以包括：Optionally, as a specific implementation, in step 104, performing comb filtering on the second time domain data to obtain the second frequency domain data after the aliasing noise is removed, which may include:

对第二时域数据进行时域梳状滤波，获得M点时域数据。Time domain comb filtering is performed on the second time domain data to obtain M point time domain data.

将M点时域数据通过M点FFT转换为第二频率数据。The M-point time domain data is converted to the second frequency data by the M-point FFT.

具体的，通过先对第二时域数据进行时域梳状滤波然后进行数据的时频转换，这样，在将时域数据转换为频域数据时，仅需要对M点时域数据进行M点FFT处理，降低了FFT的点数，进一步降低了计算复杂度以及成本，提升了频域梳状滤波的性能。Specifically, by performing time domain comb filtering on the second time domain data and then performing time-frequency conversion of the data, when converting the time domain data into the frequency domain data, only the M point time domain data needs to be M point. The FFT processing reduces the number of points of the FFT, further reduces the computational complexity and cost, and improves the performance of the frequency domain comb filter.

可选的，作为另一种具体的实现方式，步骤104，对第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据，可以包括：Optionally, as another specific implementation, in step 104, performing comb filtering on the second time domain data to obtain the second frequency domain data after the aliasing noise is removed, which may include:

将第二时域数据通过N点FFT转换为N点频域数据；Converting the second time domain data into N-point frequency domain data by N-point FFT;

对N点频域数据进行频域梳状滤波，获得第二频域数据。Frequency domain comb filtering is performed on the N-point frequency domain data to obtain second frequency domain data.

具体的，通过先对第二时域数据进行时频转换然后进行频域梳状滤波，由于频域梳状滤波的实现复杂度低于时域梳状滤波的复杂度，因此进一步降低了计算复杂度以及成本，提升了频域梳状滤波的性能。Specifically, by performing time-frequency conversion on the second time domain data and then performing frequency domain comb filtering, since the implementation complexity of the frequency domain comb filtering is lower than the complexity of the time domain comb filtering, the computation complexity is further reduced. Degree and cost increase the performance of frequency domain comb filtering.

本发明实施例提供了一种梳状滤波噪声消除方法，包括：获取第一频域数据，将第一频域数据通过N点IFFT转换为第一时域数据，根据第一时域数据与预设参考数据之间的数据位置对应关系，确定第一时域数据中产生混叠噪声的数据位置，并在数据位置处将数据置零获取第二时域数据，对第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据。本发明实施例提供的梳状滤波噪声消除方法，可以在实现梳状滤波的基础上消除混叠噪声，提升了频域梳状滤波的性能。An embodiment of the present invention provides a comb filtering noise cancellation method, including: acquiring first frequency domain data, converting first frequency domain data into first time domain data by using N point IFFT, according to first time domain data and pre The data position correspondence relationship between the reference data is determined, the data position of the aliasing noise generated in the first time domain data is determined, and the data is zeroed at the data position to obtain the second time domain data, and the second time domain data is combed. Filtering, obtaining the second frequency domain data after the aliasing noise is eliminated. The comb filter noise elimination method provided by the embodiment of the invention can eliminate the aliasing noise on the basis of the comb filter and improve the performance of the frequency domain comb filter.

图4为本发明实施例一提供的梳状滤波噪声消除装置的结构示意图，本发明实施例提供的梳状滤波噪声消除装置，用于执行图3所示实施例提供的梳状滤波噪声消除方法。如图4所示，本发明实施例提供的梳状滤波噪声消除装置，可以包括： 4 is a schematic structural diagram of a comb filter noise canceling apparatus according to Embodiment 1 of the present invention, and a comb filter noise canceling apparatus according to an embodiment of the present invention is configured to perform a comb filter noise canceling method provided by the embodiment shown in FIG. . As shown in FIG. 4, the comb filter noise canceling apparatus provided by the embodiment of the present invention may include:

获取模块31，用于获取第一频域数据。第一频域数据的频点数为N，N为大于1的整数。The obtaining module 31 is configured to acquire first frequency domain data. The frequency of the first frequency domain data is N, and N is an integer greater than 1.

数据转换模块32，用于将第一频域数据通过N点IFFT转换为第一时域数据。The data conversion module 32 is configured to convert the first frequency domain data into the first time domain data by N-point IFFT.

设置模块33，用于根据第一时域数据与预设参考数据之间的数据位置对应关系，确定第一时域数据中产生混叠噪声的数据位置，并在数据位置处将数据置零获取第二时域数据。The setting module 33 is configured to determine, according to a data location correspondence relationship between the first time domain data and the preset reference data, a data location of the aliasing noise generated in the first time domain data, and set the data to zero at the data location. Second time domain data.

混叠噪声消除模块34，用于对第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据。第二频域数据的频点数为M，N为M的整数倍，M为大于1的整数。The aliasing noise cancellation module 34 is configured to perform comb filtering on the second time domain data to obtain second frequency domain data after the aliasing noise is eliminated. The frequency of the second frequency domain data is M, N is an integer multiple of M, and M is an integer greater than 1.

可选的，混叠噪声消除模块34具体用于：Optionally, the aliasing noise cancellation module 34 is specifically configured to:

将第二时域数据通过N点FFT转换为N点频域数据。The second time domain data is converted to N-point frequency domain data by an N-point FFT.

本发明实施例提供的梳状滤波噪声消除装置，用于执行图3所示方法实施例提供的梳状滤波噪声消除方法，其技术原理和技术效果类似，此处不再赘述。The comb filter noise canceling apparatus provided in the embodiment of the present invention is used to perform the comb filter noise canceling method provided by the method embodiment shown in FIG. 3, and the technical principle and technical effects thereof are similar, and details are not described herein again.

图5为本发明实施例一提供的频域自适应均衡装置的结构示意图。如图5所示，本发明实施例提供的频域自适应均衡装置，可以包括：FIG. 5 is a schematic structural diagram of a frequency domain adaptive equalization apparatus according to Embodiment 1 of the present invention. As shown in FIG. 5, the frequency domain adaptive equalization apparatus provided by the embodiment of the present invention may include:

时域到频域变换单元41、色散补偿单元42、频域自适应均衡单元43、频域到时域变换单元44、载波恢复单元45、硬判决单元46、误差计算单元47、时域误差梳状滤波单元48、频域自适应均衡系数更新单元49、插值上采样单元50，以及梳状滤波噪声消除装置51。Time domain to frequency domain transform unit 41, dispersion compensation unit 42, frequency domain adaptive equalization unit 43, frequency domain to time domain transform unit 44, carrier recovery unit 45, hard decision unit 46, error calculation unit 47, time domain error comb The filter unit 48, the frequency domain adaptive equalization coefficient update unit 49, the interpolation upsampling unit 50, and the comb filter noise canceling means 51.

时域到频域变换单元41，用于将输入的时域数据进行数据交叠，将数据交叠后的数据从时域变换到频域并输出至色散补偿单元42。The time domain to frequency domain transform unit 41 is configured to perform data overlap on the input time domain data, transform the data overlapped data from the time domain to the frequency domain, and output the data to the dispersion compensation unit 42.

色散补偿单元42，用于对时域到频域变换单元41输出的数据在频域进行色散补偿，将色散补偿后的数据分别输出至梳状滤波噪声消除装置52和频域自适应均衡单元43。The dispersion compensation unit 42 is configured to perform dispersion compensation on the data output from the time domain to frequency domain transform unit 41 in the frequency domain, and output the dispersion compensated data to the comb filter noise canceling device 52 and Frequency domain adaptive equalization unit 43.

频域自适应均衡单元43，用于根据差值上采样单元输出的抽头系数对色散补偿单元42输出的数据进行频域均衡，将频域均衡后的数据输出至频域到时域变换单元44。The frequency domain adaptive equalization unit 43 is configured to perform frequency domain equalization on the data output by the dispersion compensation unit 42 according to the tap coefficients output by the difference upsampling unit, and output the frequency domain equalized data to the frequency domain to time domain transform unit 44. .

频域到时域变换单元44，用于将频域自适应均衡单元43输出的数据从频域变换到时域后进行数据解交叠，将数据解交叠后的数据输出至载波恢复单元45。The frequency domain to time domain transforming unit 44 is configured to perform data de-overlap after transforming the data output by the frequency domain adaptive equalization unit 43 from the frequency domain to the time domain, and output the data after the data de-overlapped to the carrier recovery unit 45. .

载波恢复单元45，用于对频域到时域变换单元44输出的数据进行载波恢复，将载波恢复后的数据分别输出至误差计算单元47和硬判决单元46。The carrier recovery unit 45 is configured to perform carrier recovery on the data output from the frequency domain to time domain transform unit 44, and output the carrier restored data to the error calculation unit 47 and the hard decision unit 46, respectively.

硬判决单元46，用于对载波恢复单元45输出的数据进行硬判决，将硬判决后的数据输出至误差计算单元47。The hard decision unit 46 is configured to perform hard decision on the data output by the carrier recovery unit 45, and output the hard-decised data to the error calculation unit 47.

误差计算单元47，用于根据载波恢复单元45输出的数据和硬判决单元46输出的数据计算数据误差，将数据误差输出至时域误差梳状滤波单元48。The error calculating unit 47 is configured to calculate a data error according to the data output by the carrier recovery unit 45 and the data output by the hard decision unit 46, and output the data error to the time domain error comb filtering unit 48.

时域误差梳状滤波单元48，用于对误差计算单元47输出的数据误差进行时域梳状滤波，将时域梳状滤波后的数据输出至频域自适应均衡系数更新单元49。The time domain error comb filtering unit 48 is configured to perform time domain comb filtering on the data error output by the error calculating unit 47, and output the time domain comb filtered data to the frequency domain adaptive equalization coefficient updating unit 49.

频域自适应均衡系数更新单元49，用于根据时域误差梳状滤波单元48输出的数据和梳状滤波噪声消除装置51输出的数据进行频域自适应抽头系数更新，并将抽头系数更新结果输出至插值上采样单元50。The frequency domain adaptive equalization coefficient updating unit 49 is configured to perform frequency domain adaptive tap coefficient update according to the data output by the time domain error comb filtering unit 48 and the data output by the comb filtering noise removing device 51, and update the tap coefficient. Output to the interpolation upsampling unit 50.

插值上采样单元50，用于将频域自适应均衡系数更新单元49输出的抽头系数更新结果进行插值上采样，将差值上采样后的抽头系数输出至频域自适应均衡单元43。The interpolation upsampling unit 50 is configured to perform interpolation upsampling on the tap coefficient update result output by the frequency domain adaptive equalization coefficient updating unit 49, and output the difference upsampled tap coefficients to the frequency domain adaptive equalization unit 43.

梳状滤波噪声消除装置51可以包括：获取模块、数据转换模块、设置模块、混叠噪声消除模块和发送模块。The comb filter noise canceling device 51 may include an acquisition module, a data conversion module, a setting module, an aliasing noise canceling module, and a transmitting module.

获取模块，用于获取色散补偿单元42输出的数据。色散补偿单元42输出的数据的频点数为N，N为大于1的整数。The acquisition module is configured to acquire data output by the dispersion compensation unit 42. The number of frequency points of the data output by the dispersion compensation unit 42 is N, and N is an integer greater than one.

数据转换模块，用于将色散补偿单元42输出的数据通过N点IFFT转换为第一时域数据。The data conversion module is configured to convert the data output by the dispersion compensation unit 42 into the first time domain data by N-point IFFT.

设置模块，用于根据第一时域数据与时域误差梳状滤波单元48输出的数据之间的数据位置对应关系，确定第一时域数据中产生混叠噪声的数据位置，并在数据位置处将数据置零获取第二时域数据。a setting module, configured to output according to the first time domain data and the time domain error comb filtering unit 48 The data position correspondence between the data determines the data position in the first time domain data where aliasing noise is generated, and zeros the data at the data position to obtain the second time domain data.

混叠噪声消除模块，用于对第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据。第二频域数据的频点数为M，N为M的整数倍，M为大于1的整数。The aliasing noise cancellation module is configured to comb filter the second time domain data to obtain second frequency domain data after eliminating aliasing noise. The frequency of the second frequency domain data is M, N is an integer multiple of M, and M is an integer greater than 1.

发送模块，用于将第二频域数据输出至频域自适应均衡系数更新单元49。And a sending module, configured to output the second frequency domain data to the frequency domain adaptive equalization coefficient updating unit 49.

本发明实施例提供的频域自适应均衡装置，包括的梳状滤波噪声消除装置51与图4所示实施例提供的梳状滤波噪声消除装置原理相似，此处不再赘述。The combo filter noise canceling apparatus 51 of the embodiment of the present invention is similar to the comb filter noise canceling apparatus provided in the embodiment shown in FIG. 4, and details are not described herein again.

需要说明的是，本发明实施例对于频域自适应均衡装置中各个单元模块的具体实现方式不做特别限定，只要实现相应功能即可，例如：各个单元模块可以通过电子元器件组成的电路实现，也可以通过芯片电路实现。It should be noted that, in the embodiment of the present invention, the specific implementation manner of each unit module in the frequency domain adaptive equalization device is not particularly limited, as long as the corresponding function is implemented, for example, each unit module can be implemented by a circuit composed of electronic components. It can also be realized by a chip circuit.

可选的，混叠噪声消除模块具体用于：Optionally, the aliasing noise cancellation module is specifically used to:

可选的，本发明实施例提供的频域自适应均衡装置，还可以包括：样值截取单元52和硬判决截取单元53。Optionally, the frequency domain adaptive equalization apparatus provided by the embodiment of the present invention may further include: a sample intercepting unit 52 and a hard decision intercepting unit 53.

样值截取单元52，用于在载波恢复单元45输出的数据中截取一部分输出至误差计算单元47。The sample value intercepting unit 52 is configured to intercept a part of the data outputted from the carrier recovery unit 45 to the error calculating unit 47.

硬判决截取单元53，用于在硬判决单元46输出的数据中截取一部分输出至误差计算单元47。The hard decision intercepting unit 53 is configured to intercept a part of the data outputted from the hard decision unit 46 to the error calculating unit 47.

其中，样值截取单元52截取的数据在载波恢复单元45输出的数据中的位置与硬判决截取单元53截取的数据在硬判决单元46输出的数据中的位置相同。The position of the data intercepted by the sample value extracting unit 52 in the data output by the carrier recovery unit 45 is the same as the position of the data intercepted by the hard decision intercepting unit 53 in the data output by the hard decision unit 46.

具体的，在频域自适应均衡装置中，误差计算单元47需要根据载波恢复单元45和硬判决单元46输出的数据计算数据误差，以实现与色散补偿单元42输出的数据相对应，但是，载波恢复单元45通常采用并行处理结构，如果等待并行的全部数据或者并行的任意一组数据全部处理完毕将会大幅增大频域自适应均衡系数更新单元49进行抽头系数更新的环路延时，进而影响***性能，例如，载波恢复单元45中并行的一组数据的数据长度为512，如果每个时钟节拍处理8个，那么需要512/8＝64个时钟节拍才能处理完该组数据，这个延时是不能接受的。Specifically, in the frequency domain adaptive equalization apparatus, the error calculation unit 47 needs to calculate a data error according to the data output by the carrier recovery unit 45 and the hard decision unit 46 to implement the dispersion compensation sheet. The data outputted by the element 42 corresponds to each other. However, the carrier recovery unit 45 generally adopts a parallel processing structure. If all data in parallel or any set of data in parallel is processed, the frequency domain adaptive equalization coefficient updating unit 49 is greatly increased. The loop delay of the tap coefficient update is performed, thereby affecting system performance. For example, the data length of a parallel set of data in the carrier recovery unit 45 is 512. If each clock beat is processed 8, then 512/8=64 is required. The clock tick can process the set of data, and this delay is unacceptable.

因此，在本发明实施例中，通过样值截取单元52在载波恢复单元45输出的数据中截取一部分数据，通过硬判决截取单元53在硬判决单元46输出的数据中截取一部分数据，由于样值截取单元52截取的数据在载波恢复单元45输出的数据中的位置与硬判决截取单元53截取的数据在硬判决单元46输出的数据中的位置相同，因此，误差计算单元47根据截取的数据计算数据误差，可以大幅降低数据处理时延，进而减小抽头系数更新环路时延，提升了频域自适应滤波装置的***性能。Therefore, in the embodiment of the present invention, a part of data is intercepted by the sample value extracting unit 52 in the data outputted by the carrier recovery unit 45, and a part of the data is intercepted by the hard decision intercepting unit 53 in the data output by the hard decision unit 46, due to the sample value. The position of the data intercepted by the intercepting unit 52 in the data output from the carrier recovery unit 45 is the same as the position of the data intercepted by the hard decision intercepting unit 53 in the data output by the hard decision unit 46, and therefore, the error calculating unit 47 calculates based on the intercepted data. The data error can greatly reduce the data processing delay, thereby reducing the tap coefficient update loop delay and improving the system performance of the frequency domain adaptive filtering device.

下面通过具体示例进行详细说明。The following is a detailed description by way of a specific example.

图6为本发明实施例一提供的样值截取单元输出的数据格式示意图。如图6所示，在载波恢复单元中，一共并行处理5组数据，分别为s1～s5，其中，s1段数据对应的载波恢复之后输出的数据包括new1和new2，通过样值截取单元截取一部分数据，样值截取单元最终输出的数据为new1。FIG. 6 is a schematic diagram of a data format output by a sample intercepting unit according to Embodiment 1 of the present invention. As shown in FIG. 6, in the carrier recovery unit, a total of five sets of data are processed in parallel, which are respectively s1 to s5, wherein the data output after the carrier corresponding to the s1 segment data includes new1 and new2, and the sample intercepting unit intercepts a part. Data, the final output of the sample interception unit is new1.

本发明实施例提供了一种频域自适应均衡装置，包括梳状滤波噪声消除装置，梳状滤波噪声消除装置将时域误差梳状滤波单元输出的数据作为预设参考数据，根据预设参考数据对色散补偿单元输出的数据进行梳状滤波以及噪声消除，由于在梳状滤波时没有引入混叠噪声，因此提升了频域自适应均衡装置的***性能。The embodiment of the invention provides a frequency domain adaptive equalization device, comprising a comb filter noise canceling device, and the comb filter noise canceling device uses the data output by the time domain error comb filter unit as preset reference data, according to the preset reference. The data combs and eliminates the data output by the dispersion compensation unit. Since no aliasing noise is introduced during the comb filtering, the system performance of the frequency domain adaptive equalization device is improved.

图7为本发明实施例二提供的频域自适应均衡装置的结构示意图，本发明实施例在图7所示实施例的基础上，提供了频域自适应均衡装置的一种具体实现结构。如图7所示，本发明实施例提供的频域自适应均衡装置，可以包括：FIG. 7 is a schematic structural diagram of a frequency domain adaptive equalization apparatus according to Embodiment 2 of the present invention. The embodiment of the present invention provides a specific implementation structure of a frequency domain adaptive equalization apparatus based on the embodiment shown in FIG. 7. As shown in FIG. 7, the frequency domain adaptive equalization apparatus provided by the embodiment of the present invention may include:

时域到频域变换单元、色散补偿单元、频域自适应均衡单元、频域到时域变换单元、载波恢复单元、硬判决单元、样值截取单元、硬判截取单元、误差计算单元、时域误差梳状滤波单元、频域自适应均衡系数更新单元、插值上采样单元，以及梳状滤波噪声消除装置。Time domain to frequency domain transform unit, dispersion compensation unit, frequency domain adaptive equalization unit, frequency domain to time domain transform unit, carrier recovery unit, hard decision unit, sample interception unit, hard interpretation intercept unit, An error calculation unit, a time domain error comb filter unit, a frequency domain adaptive equalization coefficient update unit, an interpolation upsampling unit, and a comb filter noise canceling device.

其中，在数据路径上，时域到频域变换单元接收两个偏振符号序列path x和path y，通过数据交叠和N点FFT分别将path x和path y转换为频域信号X(f)和Y(f)，之后，两个色散补偿单元分别对X(f)和Y(f)进行色散补偿，输出色散补偿后的信号Rx(f)和Ry(f)。接着，频域自适应均衡单元在频域根据差值上采样单元输出的抽头系数Wxx(f)、Wyx(f)、Wxy(f)和Wyy(f)，对Rx(f)和Ry(f)进行频域自适应均衡以补偿信道的PMD和SOP等损伤，输出两个频域信号Tx(f)和Ty(f)。然后，频域到时域变换单元分别对Tx(f)和Ty(f)进行N点IFFT和数据解交叠，将频域信号Tx(f)和Ty(f)分别转换为时域信号Tx(t)和Ty(t)。两个载波恢复单元分别对Tx(t)和Ty(t)进行载波恢复以补偿收端和发端之间的频差以及相位噪声，分别输出时域信号Rx(t)和Ry(t)，最后，硬判决单元分别根据信号Rx(t)和Ry(t)进行硬判决得到信号Dx(t)和Dy(t)。Wherein, in the data path, the time domain to frequency domain transform unit receives two polarization symbol sequences path x and path y, and converts path x and path y into frequency domain signals X(f) by data overlap and N point FFT, respectively. And Y(f), after which the two dispersion compensation units respectively perform dispersion compensation on X(f) and Y(f), and output the dispersion-compensated signals Rx(f) and Ry(f). Next, the frequency domain adaptive equalization unit compares the tap coefficients Wxx(f), Wyx(f), Wxy(f), and Wyy(f) output by the difference upsampling unit in the frequency domain, to Rx(f) and Ry(f). Perform frequency domain adaptive equalization to compensate for impairments such as PMD and SOP of the channel, and output two frequency domain signals Tx(f) and Ty(f). Then, the frequency domain to time domain transform unit respectively performs N-point IFFT and data de-overlap on Tx(f) and Ty(f), and converts frequency domain signals Tx(f) and Ty(f) into time domain signals Tx, respectively. (t) and Ty(t). The two carrier recovery units perform carrier recovery on Tx(t) and Ty(t) respectively to compensate the frequency difference and phase noise between the receiving end and the transmitting end, respectively outputting the time domain signals Rx(t) and Ry(t), and finally The hard decision unit performs hard decisions based on the signals Rx(t) and Ry(t) to obtain signals Dx(t) and Dy(t), respectively.

在系数更新路径上，两个梳状滤波噪声消除装置分别对两个色散补偿单元输出的频域信号Rx(f)和Ry(f)进行频域滤波，将频域抽样分辨率降低D(D＝N/M)倍，分别输出信号Hx(f)和Hy(f)。样值截取单元截取Rx(t)的一部分，硬判截取单元截取Dx(t)的一部分，误差计算单元根据样值截取单元和硬判截取单元的输出计算数据误差，该误差数据为时域数据，时域误差梳状滤波单元将该误差数据的分辨率降低D倍后输出至频域自适应均衡系数更新单元。频域自适应均衡系数更新单元根据时域误差梳状滤波单元的输出数据以及Hx(f)和Hy(f)，通过M点FFT和M点IFFT运算输出抽头系数更新结果，插值上采样单元将抽头系数更新结果进行插值上采样获得抽头系数Wxx(f)、Wyx(f)、Wxy(f)和Wyy(f)，并将抽头系数Wxx(f)、Wyx(f)、Wxy(f)和Wyy(f)输出至频域自适应均衡单元。On the coefficient update path, the two comb filter noise canceling devices respectively perform frequency domain filtering on the frequency domain signals Rx(f) and Ry(f) output by the two dispersion compensation units, and reduce the frequency domain sampling resolution by D (D). =N/M) times, and outputs signals Hx(f) and Hy(f), respectively. The sample intercepting unit intercepts a part of Rx(t), and the hard interpretation intercepting unit intercepts a part of Dx(t), and the error calculating unit calculates a data error according to the output of the sample value intercepting unit and the hard judgment intercepting unit, and the error data is time domain data. The time domain error comb filter unit reduces the resolution of the error data by D times and outputs the signal to the frequency domain adaptive equalization coefficient update unit. The frequency domain adaptive equalization coefficient updating unit outputs the tap coefficient update result by the M point FFT and the M point IFFT operation according to the output data of the time domain error comb filter unit and Hx(f) and Hy(f), and the interpolation upsampling unit will The tap coefficient update result is subjected to interpolation upsampling to obtain the tap coefficients Wxx(f), Wyx(f), Wxy(f), and Wyy(f), and the tap coefficients Wxx(f), Wyx(f), Wxy(f), and Wyy(f) is output to the frequency domain adaptive equalization unit.

可见，通过梳状滤波噪声消除装置以及时域误差梳状滤波单元，在频域自适应均衡系数更新单元中进行的FFT/IFFT的点数为M点，降低了FFT/IFFT的点数。而且，通过梳状滤波噪声消除装置，在实现频域梳状滤波时没有引入混叠噪声，因此提升了频域自适应均衡装置的***性能。It can be seen that, by the comb filter noise canceling device and the time domain error comb filter unit, the number of points of the FFT/IFFT performed in the frequency domain adaptive equalization coefficient updating unit is M points, and the number of points of the FFT/IFFT is lowered. Moreover, the comb filter noise canceling device does not introduce aliasing noise when implementing frequency domain comb filtering, thereby improving the system performance of the frequency domain adaptive equalization device.

本领域普通技术人员可以理解：实现上述各方法实施例的全部或部分步骤可以通过程序指令相关的硬件来完成。前述的程序可以存储于一计算机可读取存储介质中。该程序在执行时，执行包括上述各方法实施例的步骤；而前述的存储介质包括：ROM、RAM、磁碟或者光盘等各种可以存储程序代码的介质。One of ordinary skill in the art will appreciate that all or part of the steps to implement the various method embodiments described above may be accomplished by hardware associated with the program instructions. The aforementioned program can be stored in a computer Read in the storage medium. The program, when executed, performs the steps including the foregoing method embodiments; and the foregoing storage medium includes various media that can store program codes, such as a ROM, a RAM, a magnetic disk, or an optical disk.

最后应说明的是：以上各实施例仅用以说明本发明的技术方案，而非对其限制；尽管参照前述各实施例对本发明进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分或者全部技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本发明各实施例技术方案的范围。 Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, and are not intended to be limiting; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art will understand that The technical solutions described in the foregoing embodiments may be modified, or some or all of the technical features may be equivalently replaced; and the modifications or substitutions do not deviate from the technical solutions of the embodiments of the present invention. range.

Claims

一种梳状滤波噪声消除方法，其特征在于，包括：A comb filter noise cancellation method, comprising:

获取第一频域数据；所述第一频域数据的频点数为N，N为大于1的整数；Obtaining first frequency domain data; the number of frequency points of the first frequency domain data is N, and N is an integer greater than 1;

将所述第一频域数据通过N点快速傅立叶逆变换IFFT转换为第一时域数据；Converting the first frequency domain data into first time domain data by an N-point fast Fourier inverse transform (IFFT);

根据所述第一时域数据与预设参考数据之间的数据位置对应关系，确定所述第一时域数据中产生混叠噪声的数据位置，并在所述数据位置处将数据置零获取第二时域数据；Determining, according to the data location correspondence between the first time domain data and the preset reference data, a data location in the first time domain data that generates aliasing noise, and setting the data to zero at the data location Second time domain data;

对所述第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据；所述第二频域数据的频点数为M，N为M的整数倍，M为大于1的整数。Performing comb filtering on the second time domain data to obtain second frequency domain data after eliminating aliasing noise; the frequency number of the second frequency domain data is M, N is an integer multiple of M, and M is greater than 1 The integer.
根据权利要求1所述的方法，其特征在于，所述对所述第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据，包括：The method according to claim 1, wherein the combing the second time domain data to obtain the second frequency domain data after the aliasing noise is removed, comprising:

对所述第二时域数据进行时域梳状滤波，获得M点时域数据；Performing time domain comb filtering on the second time domain data to obtain M point time domain data;

将所述M点时域数据通过M点快速傅立叶变换FFT转换为所述第二频率数据。The M-point time domain data is converted into the second frequency data by an M-point fast Fourier transform FFT.
根据权利要求1所述的方法，其特征在于，所述对所述第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据，包括：The method according to claim 1, wherein the combing the second time domain data to obtain the second frequency domain data after the aliasing noise is removed, comprising:

将所述第二时域数据通过N点FFT转换为N点频域数据；Converting the second time domain data to N-point frequency domain data by N-point FFT;

对所述N点频域数据进行频域梳状滤波，获得所述第二频域数据。Performing frequency domain comb filtering on the N-point frequency domain data to obtain the second frequency domain data.
一种梳状滤波噪声消除装置，其特征在于，包括：A comb filter noise canceling device, comprising:

获取模块，用于获取第一频域数据；所述第一频域数据的频点数为N，N为大于1的整数；An acquiring module, configured to acquire first frequency domain data; the number of frequency points of the first frequency domain data is N, and N is an integer greater than one;

数据转换模块，用于将所述第一频域数据通过N点快速傅立叶逆变换IFFT转换为第一时域数据；a data conversion module, configured to convert the first frequency domain data into first time domain data by an N-point fast Fourier inverse transform (IFFT);

设置模块，用于根据所述第一时域数据与预设参考数据之间的数据位置对应关系，确定所述第一时域数据中产生混叠噪声的数据位置，并在所述数据位置处将数据置零获取第二时域数据；a setting module, configured to determine, according to a data location correspondence relationship between the first time domain data and the preset reference data, a data location in the first time domain data that generates aliasing noise, and at the data location Zeroing the data to obtain second time domain data;

混叠噪声消除模块，用于对所述第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据；所述第二频域数据的频点数为M，N为M的整数倍，M为大于1的整数。An aliasing noise cancellation module, configured to comb filter the second time domain data to obtain cancellation Second frequency domain data after aliasing noise; the frequency of the second frequency domain data is M, N is an integer multiple of M, and M is an integer greater than 1.
根据权利要求4所述的装置，其特征在于，所述混叠噪声消除模块具体用于：The apparatus according to claim 4, wherein the aliasing noise cancellation module is specifically configured to:

对所述第二时域数据进行时域梳状滤波，获得M点时域数据；Performing time domain comb filtering on the second time domain data to obtain M point time domain data;

将所述M点时域数据通过M点快速傅立叶变换FFT转换为所述第二频率数据。The M-point time domain data is converted into the second frequency data by an M-point fast Fourier transform FFT.
根据权利要求4所述的装置，其特征在于，所述混叠噪声消除模块具体用于：The apparatus according to claim 4, wherein the aliasing noise cancellation module is specifically configured to:

将所述第二时域数据通过N点FFT转换为N点频域数据；Converting the second time domain data to N-point frequency domain data by N-point FFT;

对所述N点频域数据进行频域梳状滤波，获得所述第二频域数据。Performing frequency domain comb filtering on the N-point frequency domain data to obtain the second frequency domain data.
一种频域自适应均衡装置，其特征在于，包括：时域到频域变换单元、色散补偿单元、频域自适应均衡单元、频域到时域变换单元、载波恢复单元、硬判决单元、误差计算单元、时域误差梳状滤波单元、频域自适应均衡系数更新单元、插值上采样单元，以及梳状滤波噪声消除装置；A frequency domain adaptive equalization apparatus, comprising: a time domain to frequency domain transform unit, a dispersion compensation unit, a frequency domain adaptive equalization unit, a frequency domain to time domain transform unit, a carrier recovery unit, a hard decision unit, An error calculation unit, a time domain error comb filter unit, a frequency domain adaptive equalization coefficient update unit, an interpolation upsampling unit, and a comb filter noise canceling device;

所述时域到频域变换单元，用于将输入的时域数据进行数据交叠，将数据交叠后的数据从时域变换到频域并输出至所述色散补偿单元；The time domain to frequency domain transform unit is configured to perform data overlap on the input time domain data, transform the data overlapped data from the time domain to the frequency domain, and output the data to the dispersion compensation unit;

所述色散补偿单元，用于对所述时域到频域变换单元输出的数据在频域进行色散补偿，将色散补偿后的数据分别输出至所述梳状滤波噪声消除装置和所述频域自适应均衡单元；The dispersion compensation unit is configured to perform dispersion compensation on the data outputted by the time domain to frequency domain transform unit in a frequency domain, and output the dispersion compensated data to the comb filter noise canceling apparatus and the frequency domain, respectively. Adaptive equalization unit

所述频域自适应均衡单元，用于根据所述差值上采样单元输出的抽头系数对所述色散补偿单元输出的数据进行频域均衡，将频域均衡后的数据输出至所述频域到时域变换单元；The frequency domain adaptive equalization unit is configured to perform frequency domain equalization on the data output by the dispersion compensation unit according to the tap coefficients output by the difference upsampling unit, and output the frequency domain equalized data to the frequency domain. To the time domain transform unit;

所述频域到时域变换单元，用于将所述频域自适应均衡单元输出的数据从频域变换到时域后进行数据解交叠，将数据解交叠后的数据输出至所述载波恢复单元；The frequency domain to time domain transform unit is configured to perform data de-overlap after transforming data output by the frequency domain adaptive equalization unit from a frequency domain to a time domain, and output data that is overlapped by the data to the Carrier recovery unit;

所述载波恢复单元，用于对所述频域到时域变换单元输出的数据进行载波恢复，将载波恢复后的数据分别输出至所述误差计算单元和所述硬判决单元；The carrier recovery unit is configured to perform carrier recovery on the data output by the frequency domain to the time domain transform unit, and output the data after the carrier recovery to the error calculation unit and the hard decision unit respectively;

所述硬判决单元，用于对所述载波恢复单元输出的数据进行硬判决，将硬判决后的数据输出至所述误差计算单元；The hard decision unit is configured to perform hard decision on the data output by the carrier recovery unit, and The data after the hard decision is output to the error calculation unit;

所述误差计算单元，用于根据所述载波恢复单元输出的数据和所述硬判决单元输出的数据计算数据误差，将所述数据误差输出至所述时域误差梳状滤波单元；The error calculation unit is configured to calculate a data error according to the data output by the carrier recovery unit and the data output by the hard decision unit, and output the data error to the time domain error comb filter unit;

所述时域误差梳状滤波单元，用于对所述误差计算单元输出的所述数据误差进行时域梳状滤波，将时域梳状滤波后的数据输出至所述频域自适应均衡系数更新单元；The time domain error comb filtering unit is configured to perform time domain comb filtering on the data error output by the error calculating unit, and output time domain comb filtered data to the frequency domain adaptive equalization coefficient. Update unit

所述频域自适应均衡系数更新单元，用于根据所述时域误差梳状滤波单元输出的数据和所述梳状滤波噪声消除装置输出的数据进行频域自适应抽头系数更新，并将抽头系数更新结果输出至所述插值上采样单元；The frequency domain adaptive equalization coefficient updating unit is configured to perform frequency domain adaptive tap coefficient update according to the data output by the time domain error comb filtering unit and the data output by the comb filtering noise canceling device, and the tap The coefficient update result is output to the interpolation upsampling unit;

所述插值上采样单元，用于将所述频域自适应均衡系数更新单元输出的所述抽头系数更新结果进行插值上采样，将差值上采样后的抽头系数输出至所述频域自适应均衡单元；The interpolation upsampling unit is configured to perform interpolation upsampling on the tap coefficient update result output by the frequency domain adaptive equalization coefficient updating unit, and output the difference upsampled tap coefficients to the frequency domain adaptive Equalization unit

所述梳状滤波噪声消除装置包括：获取模块、数据转换模块、设置模块、混叠噪声消除模块和发送模块；The comb filter noise cancellation device includes: an acquisition module, a data conversion module, a setting module, an aliasing noise cancellation module, and a sending module;

所述获取模块，用于获取所述色散补偿单元输出的数据；所述色散补偿单元输出的数据的频点数为N，N为大于1的整数；The acquiring module is configured to acquire data output by the dispersion compensation unit; the frequency of the data output by the dispersion compensation unit is N, and N is an integer greater than 1.

所述数据转换模块，用于将所述色散补偿单元输出的数据通过N点快速傅立叶逆变换IFFT转换为第一时域数据；The data conversion module is configured to convert the data output by the dispersion compensation unit into the first time domain data by an N-point fast Fourier inverse transform (IFFT);

设置模块，用于根据所述第一时域数据与所述时域误差梳状滤波单元输出的数据之间的数据位置对应关系，确定所述第一时域数据中产生混叠噪声的数据位置，并在所述数据位置处将数据置零获取第二时域数据；a setting module, configured to determine, according to a data position correspondence relationship between the first time domain data and data output by the time domain error comb filtering unit, a data position of the aliasing noise generated in the first time domain data And zeroing the data at the data location to obtain second time domain data;

混叠噪声消除模块，用于对所述第二时域数据进行梳状滤波，获得消除混叠噪声后的第二频域数据；所述第二频域数据的频点数为M，N为M的整数倍，M为大于1的整数；An aliasing noise cancellation module, configured to perform comb filtering on the second time domain data to obtain second frequency domain data after eliminating aliasing noise; the frequency number of the second frequency domain data is M, and N is M Integer multiple, M is an integer greater than one;

所述发送模块，用于将所述第二频域数据输出至所述频域自适应均衡系数更新单元。The sending module is configured to output the second frequency domain data to the frequency domain adaptive equalization coefficient updating unit.
根据权利要求7所述的装置，其特征在于，所述装置还包括：样值截取单元和硬判决截取单元；The device according to claim 7, wherein the device further comprises: a sample intercepting unit and a hard decision intercepting unit;

所述样值截取单元，用于在所述载波恢复单元输出的数据中截取一部分输出至所述所述误差计算单元；The sample intercepting unit is configured to intercept a part of data output by the carrier recovery unit Output to the error calculation unit;

所述硬判决截取单元，用于在所述硬判决单元输出的数据中截取一部分输出至所述误差计算单元；The hard decision intercepting unit is configured to intercept a part of the data outputted by the hard decision unit to the error calculating unit;

其中，所述样值截取单元截取的数据在载波恢复单元输出的数据中的位置与所述硬判决截取单元截取的数据在硬判决单元输出的数据中的位置相同。The position of the data intercepted by the sample intercepting unit in the data output by the carrier recovery unit is the same as the position of the data intercepted by the hard decision intercepting unit in the data output by the hard decision unit.
根据权利要求7或8所述的装置，其特征在于，所述混叠噪声消除模块具体用于：The apparatus according to claim 7 or 8, wherein the aliasing noise cancellation module is specifically configured to:

对所述第二时域数据进行时域梳状滤波，获得M点时域数据；Performing time domain comb filtering on the second time domain data to obtain M point time domain data;

将所述M点时域数据通过M点快速傅立叶变换FFT转换为所述第二频率数据。The M-point time domain data is converted into the second frequency data by an M-point fast Fourier transform FFT.
根据权利要求7或8所述的装置，其特征在于，所述混叠噪声消除模块具体用于：The apparatus according to claim 7 or 8, wherein the aliasing noise cancellation module is specifically configured to:

将所述第二时域数据通过N点FFT转换为N点频域数据；Converting the second time domain data to N-point frequency domain data by N-point FFT;

对所述N点频域数据进行频域梳状滤波，获得所述第二频域数据。 Performing frequency domain comb filtering on the N-point frequency domain data to obtain the second frequency domain data.