WO2016107236A1 - 一种光纤通信中自适应均衡滤波装置、方法和存储介质 - Google Patents

一种光纤通信中自适应均衡滤波装置、方法和存储介质 Download PDF

Info

Publication number
WO2016107236A1
WO2016107236A1 PCT/CN2015/091139 CN2015091139W WO2016107236A1 WO 2016107236 A1 WO2016107236 A1 WO 2016107236A1 CN 2015091139 W CN2015091139 W CN 2015091139W WO 2016107236 A1 WO2016107236 A1 WO 2016107236A1
Authority
WO
WIPO (PCT)
Prior art keywords
group
odd
time domain
frequency
domain signal
Prior art date
Application number
PCT/CN2015/091139
Other languages
English (en)
French (fr)
Inventor
姚扬中
蔡轶
黄源良
曾文琪
李运鹏
顾国华
Original Assignee
深圳市中兴微电子技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳市中兴微电子技术有限公司 filed Critical 深圳市中兴微电子技术有限公司
Publication of WO2016107236A1 publication Critical patent/WO2016107236A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
    • H04B10/25—Arrangements specific to fibre transmission

Definitions

  • the present invention relates to the field of optical communications, and in particular, to an adaptive equalization filtering apparatus, method, and storage medium for optical fiber communication.
  • FIG. 1 is a schematic diagram of a typical digital coherent receiver in polarization multiplexing optical fiber communication.
  • a digital coherent receiver receives two polarized optical signals, and each polarized optical signal is divided into a mutual polarization beam splitter (PBS). Two orthogonal polarization signals.
  • the PBS output polarized light signal is mixed with the local oscillator optical signal through a 90° optical mixer (90° hybrid).
  • the mixed optical signal is converted into a baseband electrical signal by a balanced photodetector (PD).
  • PD balanced photodetector
  • the photoelectrically converted electrical signal has two signals for each polarization state, but the four signals do not correspond to the original four signals, because after the transmission channel, there is crosstalk between the two polarization states, and the polarization state also rotates.
  • the photoelectrically converted electrical signal is converted by an analog-to-digital converter (ADC) For digital signals.
  • ADC analog-to-digital converter
  • the universal equalization filtering technique can be used to process the ADC into a digital signal.
  • CD chromatic dispersion
  • PMD polarization mode dispersion
  • Residual chromatic dispersion and polarization mode dispersion compensation is achieved by an adaptive equalizer that uses an adaptive algorithm to update the coefficients to track the polarization mode dispersion that compensates dynamically over time.
  • the existing number of taps in the time domain adaptive equalization filtering device is large, the number of filter banks is large, and the required digital logic circuit is large in scale and large in power consumption. If the input signal is transformed from the time domain to the frequency domain and then equalized and filtered, it is expected to simplify the scale of the equalization filter circuit; however, it is necessary to add a set of conversion circuits and a set of signals that convert the input signal from the time domain to the frequency domain. The output signal is converted from the frequency domain to the time domain conversion circuit, which brings about an increase in circuit scale.
  • the existing time domain adaptive equalization filtering device and the frequency domain adaptive equalization filtering device all have the problems of large circuit scale and large power consumption. There is an urgent need for an optimized apparatus and method for implementing adaptive equalization filtering.
  • Embodiments of the present invention are directed to providing an adaptive equalization filtering apparatus, method, and storage medium, which can reduce circuit scale and power consumption of implementing an adaptive equalization filtering apparatus.
  • An embodiment of the present invention provides an apparatus for adaptive equalization filtering in optical fiber communication, where the apparatus includes: a data grouping unit, a time-frequency converting unit, an equalization filtering unit, an inverse Fourier transform unit, and a data combining unit;
  • the data packet unit is configured to group the received time domain signals according to a sample index to obtain a packet time domain signal
  • the time-frequency converting unit is configured to convert the packet time domain signal obtained by the data packet unit into a packet frequency domain signal
  • the equalization filtering unit is configured to perform equalization filtering on the packet frequency domain signal obtained by the time-frequency conversion unit according to the frequency domain equalization filter coefficient;
  • the inverse Fourier transform unit is configured to perform inverse Fourier transform on the packet frequency domain signal that is equalized and filtered by the equalization filtering unit to obtain an equalized filtered packet time domain signal;
  • the data merging unit is configured to combine the equalized and filtered packet time domain signals of the inverse Fourier transform unit to obtain a complete equalized filtered time domain signal.
  • the time-frequency conversion unit includes:
  • An odd-time time-frequency conversion module configured to convert an odd-group time domain signal into an odd-group frequency domain signal
  • the even group time-frequency conversion module is configured to convert the even-group time domain signal into the even-group frequency domain signal
  • the equalization filtering unit comprises:
  • the odd-group equalization filtering module is configured to perform equalization filtering on the odd-group frequency domain signals obtained by the odd-group time-frequency conversion module according to the odd-group frequency-domain equalization filter coefficients;
  • the even group equalization filtering module is configured to perform equalization filtering on the even group frequency domain signals obtained by the even group time-frequency conversion module according to the even group frequency domain equalization filter coefficients;
  • the inverse Fourier transform unit comprises:
  • the odd-group inverse Fourier transform module is configured to perform inverse Fourier transform on the odd-group frequency domain signals that are equalized and filtered by the odd-group equalization filtering module, to obtain an equal-filtered odd-group time domain signal;
  • the even-group inverse Fourier transform module is configured to perform inverse Fourier transform on the even-group frequency domain signals that are equalized and filtered by the even-group equalization filtering module, to obtain an equal-filtered odd-group time domain signal.
  • the data grouping unit further includes:
  • the data segmentation unit is configured to divide the grouped time domain signal into preset N segments, and the segmentation time domain signals overlap, and the overlap length is not less than the filter length minus one;
  • the time-frequency conversion unit further includes N segment time-frequency conversion modules; wherein each segment time-frequency conversion module is configured to convert the segment time domain signal obtained by the corresponding data segmentation unit into points Segment frequency domain signal;
  • the equalization filtering unit further includes N segment equalization filtering modules; wherein each segment equalization filtering module is configured to obtain a frequency domain obtained by the corresponding segment time-frequency conversion module according to the frequency domain equalization filter coefficient The signal is equalized and filtered;
  • the inverse Fourier transform unit further includes N segment inverse Fourier transform modules; wherein each segment inverse Fourier transform module is configured to perform equalization and filtering of the frequency domain signal by the corresponding segment equalization filtering module. Inverse Fourier transform to obtain an equalized filtered time domain signal;
  • the data merging unit further includes a segment data merging unit configured to remove the equalized filtered time domain signals belonging to the same group, and perform overlapping to obtain a complete time domain signal;
  • the N is an integer greater than one.
  • the device further includes:
  • a coefficient update unit configured to calculate a time domain equalization filter coefficient according to a preset equalization filter coefficient update algorithm
  • the coefficient time-frequency converting unit is configured to perform a Fourier transform on the time-domain equalization filter coefficients calculated by the coefficient updating unit to obtain a frequency-domain equalization filter coefficient.
  • the coefficient update unit when the data packet unit divides the received time domain signals into odd groups and even groups according to the sample index, the coefficient update unit includes:
  • the odd group coefficient updating module is configured to calculate an odd group time domain equalization filter coefficient according to a preset equalization filter coefficient update algorithm
  • the even group coefficient updating module is configured to calculate an even group time domain equalization filter coefficient according to a preset equalization filter coefficient update algorithm
  • the coefficient time-frequency conversion unit comprises:
  • the odd-group coefficient time-frequency converting unit is configured to perform Fourier transform on the odd-group time-domain equalizing filter coefficients calculated by the odd-group coefficient updating module to obtain an odd-group frequency-domain equalizing filter coefficient;
  • the even-group coefficient time-frequency conversion unit is configured to perform Fourier transform on the even-group time-domain equalization filter coefficients calculated by the even-group coefficient update module to obtain an even-group frequency-domain equalization filter coefficient.
  • An embodiment of the present invention further provides an adaptive equalization filtering method in optical fiber communication, where the method includes:
  • the received time domain signals are grouped according to the sample index to obtain a packet time domain signal
  • the grouping the received time domain signals according to the sample index includes:
  • the received time domain signals are divided into odd groups and even groups according to sample index
  • the converting the packet time domain signal into a packet frequency domain signal comprises:
  • the equalizing and filtering the packet frequency domain signal according to the frequency domain equalization filter coefficient comprises:
  • Equalizing and filtering the odd frequency domain signal according to the odd frequency domain equalization filter coefficient performing equalization filtering on the even frequency domain signal according to the even frequency domain equalization filter coefficient
  • performing the inverse Fourier transform on the equalized and filtered packet frequency domain signal to obtain the equalized and filtered packet time domain signal including:
  • the odd-group time domain signal is subjected to inverse Fourier transform on the even-filtered even-group frequency domain signal to obtain an equal-filtered odd-group time domain signal.
  • the method further includes:
  • the grouped time domain signal is divided into preset N segments, and the segmented time domain signals overlap, and the overlap length is not less than the filter length minus one;
  • the equalizing and filtering the packet time domain signals are combined to obtain a complete equalized filtered time domain signal, including: after the equalizing and filtering of the N segment time domain signals, removing the overlap and combining a complete set of time domain signals;
  • the N is an integer greater than one.
  • the equalizing and filtering the packet frequency domain signal according to the frequency domain equalization filter coefficient satisfies the following expression:
  • n is a segmentation index
  • a xh,even , A xv,even , A yh,even , A yv,even are even-group frequency domain equalization filter coefficients
  • a xh, odd , A xv, odd , A yh, odd , A yv, odd are odd-frequency frequency-domain equalization filter coefficients
  • H even, n , V even, n are the n-th even-group frequency domain data
  • H odd, n , V odd, n are the n-th odd-frequency frequency domain Data
  • X n (k) is the n-th phase equalized filtered x-frequency domain signal
  • Y n (k) is the n-th stage equalized y-frequency domain signal.
  • the method for calculating the frequency domain equalization filter coefficient is:
  • the time domain equalization filter coefficients are converted to obtain frequency domain equalization filter coefficients based on a Fourier transform technique.
  • the embodiment of the invention further provides a storage medium, wherein the storage medium stores computer executable instructions, and the computer executable instructions are configured to perform the adaptive equalization filtering method in the optical fiber communication according to the embodiment of the invention.
  • the adaptive equalization filtering device, the method and the storage medium in the optical fiber communication provided by the embodiments of the present invention, the adaptive equalization filtering device provided by the embodiment of the present invention by grouping and segmenting the input time domain signals according to the sample index, The signal is divided into a plurality of small segments, and the filter coefficients are also divided into two parts according to the index parity. Then, the frequency domain filtering method is used, the required Fourier transform has fewer points, and the Fourier transform with smaller points is simple to implement and the circuit scale is small. Low power consumption and low latency.
  • FIG. 1 is a schematic structural diagram of a typical digital coherent receiver in polarization multiplexing optical fiber communication
  • FIG. 2 is a schematic structural diagram of a structure of an adaptive equalization filtering device in optical fiber communication according to an embodiment of the present invention
  • FIG. 3 is a schematic structural diagram of an apparatus for adaptive equalization filtering in an optical fiber communication according to an application scenario
  • FIG. 4 is a schematic diagram of a data packet segmentation unit according to an embodiment of the present invention.
  • FIG. 5 is a flowchart of a method for adaptive equalization filtering in optical fiber communication according to an embodiment of the present invention.
  • FIG. 2 is a schematic structural diagram of an apparatus for adaptive equalization filtering in optical fiber communication according to an embodiment of the present invention.
  • the apparatus includes: a data grouping unit 201, a time-frequency converting unit 202, an equalization filtering unit 203, and an inverse Fourier. Transform unit 204 and data merge unit 205; wherein
  • the data grouping unit 201 is configured to group the received time domain signals according to a sample index. Obtaining a packet time domain signal
  • the time-frequency converting unit 202 is configured to convert the packet time domain signal obtained by the data packet unit 201 into a packet frequency domain signal;
  • the equalization filtering unit 203 is configured to perform equalization filtering on the packet frequency domain signal obtained by the time-frequency converting unit 202 according to the frequency domain equalization filter coefficient;
  • the inverse Fourier transform unit 204 is configured to perform inverse Fourier transform on the packet frequency domain signal that is equalized and filtered by the equalization filtering unit 203, to obtain an equalized and filtered packet time domain signal;
  • the data combining unit 205 is configured to combine the equalized and filtered packet time domain signals of the inverse Fourier transform unit 204 to obtain a complete equalized filtered time domain signal.
  • the data packet unit 201 divides the received time domain signals into odd groups and even groups according to the sample index, that is, the original time domain signals are divided into odd groups according to whether the time domain signal sample index is odd or even.
  • the time domain signal and the even group time domain signal; here, the original time domain signal refers to a time domain signal that is not equalized filtered.
  • the time-frequency converting unit 202 includes:
  • An odd-time time-frequency conversion module configured to convert an odd-group time domain signal into an odd-group frequency domain signal
  • the even group time-frequency conversion module is configured to convert the even-group time domain signals into even-group frequency domain signals.
  • the equalization filtering unit 203 includes:
  • the odd-group equalization filtering module is configured to perform equalization filtering on the odd-group frequency domain signals obtained by the odd-group time-frequency conversion module according to the odd-group frequency-domain equalization filter coefficients;
  • the even group equalization filtering module is configured to perform equalization filtering on the even group frequency domain signals obtained by the even group time-frequency conversion module according to the even group frequency domain equalization filter coefficients.
  • the inverse Fourier transform unit 204 includes:
  • the odd-group inverse Fourier transform module is configured to perform inverse Fourier transform on the odd-group frequency domain signals that are equalized and filtered by the odd-group equalization filtering module, to obtain an equal-filtered odd-group time domain signal;
  • the even-group inverse Fourier transform module is configured to perform inverse Fourier transform on the even-group frequency domain signals that are equalized and filtered by the even-group equalization filtering module, to obtain an equal-filtered odd-group time domain signal.
  • the grouped time domain signal may be further segmented, that is, in the adaptive equalization filtering device, the data grouping unit 201 further includes:
  • the data segmentation unit is configured to divide the grouped time domain signal into preset N segments, and the segmentation time domain signals overlap, and the overlap length is not less than the filter length minus one;
  • the time-frequency conversion unit 202 further includes N segmentation time-frequency conversion modules; wherein each segment time-frequency conversion module is configured to convert the segmentation time domain signal obtained by the corresponding data segmentation unit into Segmented frequency domain signal;
  • the equalization filtering unit 203 further includes N segment equalization filtering modules; wherein each segment equalization filtering module is configured to obtain a frequency of the corresponding segment time-frequency conversion module according to the frequency domain equalization filter coefficient.
  • the domain signal is equalized and filtered;
  • the inverse Fourier transform unit 204 further includes N segment inverse Fourier transform modules; wherein each segment inverse Fourier transform module is configured to perform equalization and filtering of the frequency domain signal by the corresponding segment equalization filtering module. Performing an inverse Fourier transform to obtain an equalized filtered time domain signal;
  • the data merging unit 205 further includes a segment data merging unit configured to remove the equalized and filtered time domain signals belonging to the same group, and perform overlapping to obtain a complete time domain signal;
  • N is an integer greater than 1.
  • the adaptive equalization filter control apparatus further includes: a coefficient update unit and a coefficient time-frequency conversion unit; wherein
  • a coefficient update unit configured to calculate a time domain equalization filter coefficient according to a preset equalization filter coefficient update algorithm
  • the coefficient time-frequency converting unit is configured to perform a Fourier transform on the time-domain equalization filter coefficients calculated by the coefficient updating unit to obtain a frequency-domain equalization filter coefficient.
  • the coefficient update unit when the data packet unit divides the received time domain signals into odd groups and even groups according to the sample index, the coefficient update unit includes:
  • the odd group coefficient updating module is configured to calculate an odd group time domain equalization filter coefficient according to a preset equalization filter coefficient update algorithm
  • the even group coefficient updating module is configured to calculate an even group time domain equalization filter coefficient according to a preset equalization filter coefficient update algorithm
  • the coefficient time-frequency conversion unit comprises:
  • the odd-group coefficient time-frequency converting unit is configured to perform Fourier transform on the odd-group time-domain equalizing filter coefficients calculated by the odd-group coefficient updating module to obtain an odd-group frequency-domain equalizing filter coefficient;
  • the even-group coefficient time-frequency conversion unit is configured to perform Fourier transform on the even-group time-domain equalization filter coefficients calculated by the even-group coefficient update module to obtain an even-group frequency-domain equalization filter coefficient.
  • Each of the above units and modules may be implemented by a central processing unit (CPU), a microprocessor (MPU), a digital signal processor (DSP), or a field programmable gate array (FPGA) located in a digital coherent receiver. )achieve.
  • CPU central processing unit
  • MPU microprocessor
  • DSP digital signal processor
  • FPGA field programmable gate array
  • FIG. 3 is a schematic structural diagram of an apparatus for adaptive equalization filtering in an optical fiber communication according to an embodiment of the present invention.
  • the adaptive equalization filtering apparatus receives two time domain signals h and v. In each clock cycle (ie, one beat), 128 data samples are collected on each time domain signal.
  • the device includes: a data grouping unit 401, a time-frequency converting unit 402, and an equalization. Filtering unit 403, inverse Fourier transform unit 404, data combining unit 405, coefficient updating unit 406 and coefficient time-frequency converting unit 407;
  • the obtained packet segmentation result is the same;
  • FIG. 4 is a schematic diagram of a data packet segmentation unit according to an embodiment of the present invention, as shown in FIG. 4.
  • the polarization signal v also performs the above-described packet segmentation operation. For the 136 samples of h and v, they can be recorded as:
  • the sample of 136 is divided into 5 segments, each segment is overlapped by 6 samples, each segment has a length of 32, and then divided into two groups according to the sample index. Then the data after segmentation is satisfied by formula (1) and formula (2):
  • the time-frequency converting unit 402 performs a 16-point Fourier transform on the packet segmentation signals output by the data packet unit 401, and obtains a frequency domain signal and outputs the signal to the equalization filtering unit 403.
  • the Fourier transform of h odd,n is denoted as H odd,n ,h even
  • the Fourier transform of n is denoted as H even,n ,v odd
  • the Fourier transform of n is denoted as the Fourier transform of V odd,n ,v even,n Is V even,n .
  • the circuit scale of the time-frequency conversion unit can be reduced.
  • the equalization filtering unit 403 performs equalization filtering on the frequency domain signal output by the time-frequency converting unit 402 according to the frequency domain equalization filter coefficient output by the coefficient time-frequency converting unit 406, and equalizes the frequency domain signal of the x frequency domain.
  • a y frequency domain signal is output to the inverse Fourier transform unit 404; In the middle, the calculation method of the equalization filter satisfies the formula (3):
  • n is a segmentation index
  • a xh,even , A xv,even , A yh,even , A yv,even are even-group frequency domain equalization filter coefficients
  • a xh, odd , A xv, odd , A yh, odd , A yv, odd are odd-frequency frequency-domain equalization filter coefficients
  • H even, n , V even, n are the n-th even-group frequency domain data
  • H odd, n , V odd, n are the n-th odd-frequency frequency domain Data
  • X n (k) is the n-th phase equalized filtered x-frequency domain signal
  • Y n (k) is the n-th stage equalized y-frequency domain signal.
  • the inverse Fourier transform unit 404 is configured to perform a 16-point inverse Fourier transform on the frequency domain signal output by the equalization filtering unit 403, to obtain an equalized filtered x time domain signal and a y time domain signal, and output the data to the data combination.
  • the data combining unit 405 is configured as an x or y time domain signal (ie, an equalized filtered time domain signal) of the packet segmentation outputted by the inverse Fourier transform unit 404, and the overlapping samples are removed for the odd group data, and merged.
  • An odd group time domain signal is formed; likewise, the overlapping samples are removed for the even group data, and merged into an even group time domain signal; finally, the odd group and the even group data are merged into a complete time domain signal xout and aout output.
  • the coefficient updating unit 406 performs update calculation on the time domain equalization filter coefficients, and the coefficient update methods that can be adopted include: Constant Modulus Algorithm (CMA) and Decision-Assisted Minimum Mean Square Error Method (DDLMS, Decision-Directed Least). -Mean-Square), or a multi-mode blind equalization algorithm, and other adaptive coefficient update algorithms for the time domain are also applicable to the embodiments of the present invention.
  • the coefficient time-frequency converting unit 407 divides the time domain coefficients a xh outputted by the coefficient updating unit 406 into two groups according to whether the index is odd or even, and transforms it into the frequency domain by using Fourier transform technology to obtain A xh. Odd and A xh,even are the same. For a xv , a yh , and a yv , the two are divided into two groups according to whether the index is odd or even , and are transformed into the frequency domain by Fourier transform technique. The number of Fourier transform points used is 16 . .
  • the existing adaptive multiplexing filter device for polarization multiplexed optical communication receiver is a butterfly filter structure with a total of four sets of filter coefficients.
  • the filter length is long and the coefficients are complex.
  • the circuit scale is large and the power consumption is large.
  • the existing frequency domain equalization filtering method although the resources consumed by the equalization filtering are saved, the Fourier transform and the inverse Fourier transform circuit with larger points and the circuit scale of the Fourier transform and the inverse Fourier transform with larger points are additionally added. Large, high power consumption, and the delay caused by the transformation is large.
  • the adaptive equalization filtering device divides the signal into a plurality of segments according to the sample index and segmentation according to the sample index, and the filter coefficients are also divided into two parts according to the index parity, and then Using the frequency domain filtering method, the Fourier transform of the required Fourier transform is small, and the Fourier transform with a small number of points is simple to implement, the circuit scale is small, the power consumption is small, and the delay is small.
  • the implementation of the adaptive equalization filtering time domain implementation of 7 taps (ie, 128-point Fourier transform) and the embodiment of the present invention (ie, 16-point Fourier transform) are implemented in a fixed-point manner, and the equalization filtering device of the embodiment of the present invention is more balanced than the existing one.
  • the implementation of the filtering device should save the circuit scale to 50%.
  • FIG. 5 is a flowchart of a method for adaptive equalization filtering in optical fiber communication according to an embodiment of the present invention. As shown in FIG. 5, the method includes:
  • Step 601 Group the received time domain signals according to the sample index to obtain a packet time domain signal.
  • Step 602 Convert the packet time domain signal into a packet frequency domain signal.
  • Step 603 Perform equalization filtering on the packet frequency domain signal according to the frequency domain equalization filter coefficient.
  • Step 604 Perform inverse Fourier transform on the equalized and filtered packet frequency domain signal to obtain an equalized and filtered packet time domain signal.
  • Step 605 Combine the equalized and filtered packet time domain signals to obtain a complete equalized filtered time domain signal.
  • the grouping the received time domain signals according to the sample index in the step 601 includes: dividing the received time domain signals into odd groups and even groups according to the sample index, that is, indexing according to the time domain signal samples. Whether the odd or even number divides the original time domain signal into an odd group time domain signal and an even group time domain signal; here, the original time domain signal refers to a time domain signal that is not equalized filtered;
  • converting the packet time domain signal into a packet frequency domain signal in step 602 comprising: converting an odd group time domain signal into an odd group frequency domain signal, and converting the even group time domain signal into an even group frequency domain. signal;
  • the equalizing and filtering the packet frequency domain signal according to the frequency domain equalization filter coefficient in the step 603, comprising: performing equalization filtering on the odd frequency domain signal according to the odd frequency domain equalization filter coefficient;
  • the domain equalization filter coefficient performs equalization filtering on the dual frequency domain signals;
  • performing the inverse Fourier transform on the equalized and filtered packet frequency domain signal in step 604 to obtain the equalized and filtered packet time domain signal including:
  • the grouped time domain signal may be further segmented, and the method further includes:
  • the grouped time domain signal is divided into preset N segments, and the segmented time domain signals overlap, and the overlap length is not less than the filter length minus one;
  • step 605 the equalizing and filtering the packet time domain signals are combined to obtain a complete equalized filtered time domain signal, including: equalizing and filtering the same group of the N segments.
  • the overlap is removed and combined to obtain a complete set of time domain signals; the N is an integer greater than one.
  • step 603 the equalizing and filtering the packet frequency domain signal according to the frequency domain equalization filter coefficient satisfies the formula (4):
  • n is a segmentation index
  • a xh,even , A xv,even , A yh,even , A yv,even are even-group frequency domain equalization filter coefficients
  • a xh, odd , A xv, odd , A yh, odd , A yv, odd are odd-frequency frequency-domain equalization filter coefficients
  • H even, n , V even, n are the n-th even-group frequency domain data
  • H odd, n , V odd, n are the n-th odd-frequency frequency domain Data
  • X n (k) is the n-th phase equalized filtered x-frequency domain signal
  • Y n (k) is the n-th stage equalized y-frequency domain signal.
  • the method for calculating the frequency domain equalization filter coefficients is:
  • the time domain equalization filter coefficients are converted to obtain frequency domain equalization filter coefficients based on a Fourier transform technique.
  • the calculating the time domain equalization filter coefficients according to the preset equalization filter coefficient update algorithm including: updating according to the preset equalization filter coefficients.
  • the algorithm calculates an odd-group time-domain equalization filter coefficient; and calculates an even-group time-domain equalization filter coefficient according to a preset equalization filter coefficient update algorithm;
  • the transforming the time domain equalization filter coefficients to obtain frequency domain equalization filter coefficients based on a Fourier transform technique including:
  • the embodiment of the invention further provides a storage medium, wherein the storage medium stores computer executable instructions, and the computer executable instructions are configured to perform the adaptive equalization filtering method in the optical fiber communication according to the embodiment of the invention.
  • embodiments of the present invention can be provided as a method, system, or computer program product. Accordingly, the present invention can take the form of a hardware embodiment, a software embodiment, or a combination of software and hardware. Moreover, the invention can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) including computer usable program code.
  • the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
  • the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
  • the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.
  • the input time domain signal is grouped and segmented according to the sample index, and the signal is divided into a plurality of small segments, and the filter coefficients are also divided into two parts according to the index parity, and then the frequency domain filtering method is used, and the required
  • the Fourier transform has fewer points and a smaller Fourier transform.
  • the implementation is simple, the circuit scale is small, the power consumption is small, and the delay is small.

Abstract

本发明实施例公开了一种光纤通信中自适应均衡滤波装置、方法和存储介质，所述装置包括：数据分组单元，配置为将接收的时域信号按照样点索引分组，得到分组时域信号；时频转换单元，配置为将所述数据分组单元得到的分组时域信号转换为分组频域信号；均衡滤波单元，配置为根据频域均衡滤波系数，对所述时频转换单元得到的分组频域信号进行均衡滤波；逆傅立叶变换单元，配置为对所述均衡滤波单元进行均衡滤波后的分组频域信号进行逆傅里叶变换，得到均衡滤波后的分组时域信号；数据合并单元，配置为将所述逆傅立叶变换单元得到均衡滤波后的分组时域信号进行合并，得到完整的均衡滤波后时域信号。

Description

一种光纤通信中自适应均衡滤波装置、方法和存储介质 技术领域
本发明涉及光通信领域，尤其涉及一种光纤通信中自适应均衡滤波装置、方法和存储介质。
背景技术
随着互联网流量的增加，在干线系统的光通信系统中需要更大的容量。当每波长比特率增加时，在传输路径上的色度色散，偏振模色散以及各种非线性效应的波形失真而导致信息质量的退化变得很严重。数字相干技术与和非相干技术相比具有明显优势，在高速光通信系统中应用广泛。光相干接收机中，通过将信号光和本振光进行混频，信号光的幅度和相位信息被搬移到基带信号中，因而光相干检测保留了光场的所有信息，可以发挥数字信号处理技术的功能和性能上的优势。采用电均衡技术能够近乎完全地补偿光信号的线性失真，如补偿色度色散(CD)，偏振模色散(PMD)等。
图1为偏振复用光纤通信中典型的数字相干接收机结构示意图，如图1所示，数字相干接收机接收两路偏振光信号，每一路偏振光信号被一个偏振分光器(PBS)分成相互正交的两个偏振态信号。PBS输出偏振光信号通过90°光混频器(90°hybrid)与本振光信号进行混频。混频后的光信号通过平衡光电检测器(PD，Photo electronic Detector)转换为基带电信号。光电转换后的电信号为每个偏振态有两路信号，但这四路信号与原始的四路信号并不对应，因为经过传输信道后，两个偏振态之间有串扰、偏振态也有旋转，所以接收端此处的两个偏振态、每个偏振态有两路正交信号与发射信号没有对应关系。光电转换后的电信号通过模数变换器(ADC)转换 为数字信号。可以通用均衡滤波技术对ADC转换为数字信号进行处理。
色度色散(CD)的值一般较大，色度色散和偏振模色散(PMD)的均衡一般分两部分完成，首先，补偿静态色散，这里的均衡器通常不能使用标准的自适应算法进行系数更新，如果要补偿40000ps/nm色度色散，则滤波器抽头数要达到大几百甚至上千，通常利用快速傅立叶变换技术进行频域快速卷积。
残余的色度色散和偏振模色散的补偿通过自适应均衡器来实现，自适应均衡器采用自适应算法对系数进行更新，以跟踪补偿随时间动态变化的偏振模色散。
由于现有的在时域自适应均衡滤波装置中抽头数较多，滤波器组数较多导致需要的数字逻辑电路规模大、功耗大。若将输入信号由时域变换到频域后再进行均衡滤波，则有望简化均衡滤波电路的规模；但如此又必须增加一组将输入信号由时域转为频域的变换电路和一组将输出信号由频域转为时域的变换电路，又带来了电路规模的增加。现有的时域自适应均衡滤波装置和频域自适应均衡滤波装置都存在电路规模大、功耗大的问题。迫切需要一种优化的实现自适应均衡滤波的装置和方法。
发明内容
本发明实施例期望提供一种自适应均衡滤波装置、方法和存储介质，能减小实现自适应均衡滤波装置的电路规模和功耗。
本发明实施例的技术方案是这样实现的：
本发明实施例提供一种光纤通信中自适应均衡滤波装置，所述装置包括：数据分组单元、时频转换单元、均衡滤波单元、逆傅立叶变换单元和数据合并单元；其中，
所述数据分组单元，配置为将接收的时域信号按照样点索引分组，得到分组时域信号；
所述时频转换单元，配置为将所述数据分组单元得到的分组时域信号转换为分组频域信号；
所述均衡滤波单元，配置为根据频域均衡滤波系数，对所述时频转换单元得到的分组频域信号进行均衡滤波；
所述逆傅立叶变换单元，配置为对所述均衡滤波单元进行均衡滤波后的分组频域信号进行逆傅里叶变换，得到均衡滤波后的分组时域信号；
所述数据合并单元，配置为将所述逆傅立叶变换单元得到均衡滤波后的分组时域信号进行合并，得到完整的均衡滤波后时域信号。
作为一种实施方式，当所述数据分组单元将接收的时域信号按照样点索引分为奇组和偶组时，所述时频转换单元包括：
奇组时频转换模块，配置为将奇组时域信号转换为奇组频域信号；
偶组时频转换模块，配置为将偶组时域信号转换为偶组频域信号；
相应的，所述均衡滤波单元包括：
奇组均衡滤波模块，配置为根据奇组频域均衡滤波系数对所述奇组时频转换模块得到的奇组频域信号进行均衡滤波；
偶组均衡滤波模块，配置为根据偶组频域均衡滤波系数对所述偶组时频转换模块得到的偶组频域信号进行均衡滤波；
相应的，所述逆傅立叶变换单元包括：
奇组逆傅立叶变换模块，配置为对所述奇组均衡滤波模块进行均衡滤波后的奇组频域信号进行逆傅里叶变换，得到均衡滤波后的奇组时域信号；
偶组逆傅立叶变换模块，配置为对所述偶组均衡滤波模块进行均衡滤波后的偶组频域信号进行逆傅里叶变换，得到均衡滤波后的奇组时域信号。
作为一种实施方式，所述数据分组单元还包括：
数据分段单元，配置为将分组后的时域信号分为预设的N段，分段时域信号之间有重叠，重叠长度不小于滤波器长度减1；
相应的，所述时频转换单元还包括N个分段时频转换模块；其中，每个分段时频转换模块，配置为将对应的数据分段单元得到的分段时域信号转换为分段频域信号；
相应的，所述均衡滤波单元还包括N个分段均衡滤波模块；其中，每个分段均衡滤波模块，配置为根据频域均衡滤波系数，将对应的分段时频转换模块得到的频域信号进行均衡滤波；
相应的，所述逆傅立叶变换单元还包括N个分段逆傅立叶变换模块；其中，每个分段逆傅立叶变换模块，配置为将对应的分段均衡滤波模块进行均衡滤波后的频域信号进行逆傅里叶变换，得到均衡滤波后的时域信号；
相应的，所述数据合并单元还包括分段数据合并单元，配置为将同属于一组的均衡滤波后的时域信号，去掉重叠，进行合并得到一组完整的时域信号；
所述N为大于1的整数。
作为一种实施方式，所述装置还包括：
系数更新单元，配置为按照预设的均衡滤波系数更新算法计算时域均衡滤波系数；
系数时频转换单元，配置为将所述系数更新单元计算出的时域均衡滤波系数进行傅立叶变换，得到频域均衡滤波系数。
作为一种实施方式，当所述数据分组单元将接收的时域信号按照样点索引分为奇组和偶组时，所述系数更新单元包括：
奇组系数更新模块，配置为按照预设的均衡滤波系数更新算法计算奇组时域均衡滤波系数；
偶组系数更新模块，配置为按照预设的均衡滤波系数更新算法计算偶组时域均衡滤波系数；
相应的，所述系数时频转换单元包括：
奇组系数时频转换单元，配置为将所述奇组系数更新模块计算出的奇组时域均衡滤波系数进行傅立叶变换，得到奇组频域均衡滤波系数；
偶组系数时频转换单元，配置为将所述偶组系数更新模块计算出的偶组时域均衡滤波系数进行傅立叶变换，得到偶组频域均衡滤波系数。
本发明实施例还提供一种光纤通信中自适应均衡滤波方法，所述方法包括：
将接收的时域信号按照样点索引分组，得到分组时域信号；
将所述分组时域信号转换为分组频域信号；
根据频域均衡滤波系数，对所述分组频域信号进行均衡滤波；
对均衡滤波后的分组频域信号进行逆傅里叶变换，得到均衡滤波后的分组时域信号；
将所述均衡滤波后的分组时域信号进行合并，得到完整的均衡滤波后时域信号。
作为一种实施方式，所述将接收的时域信号按照样点索引分组包括：
将接收的时域信号按照样点索引分为奇组和偶组；
相应的，所述将所述分组时域信号转换为分组频域信号，包括：
将奇组时域信号转换为奇组频域信号，将偶组时域信号转换为偶组频域信号；
相应的，所述根据频域均衡滤波系数，对所述分组频域信号进行均衡滤波，包括：
根据奇组频域均衡滤波系数对奇组频域信号进行均衡滤波；根据偶组频域均衡滤波系数对偶组频域信号进行均衡滤波；
相应的，所述对均衡滤波后的分组频域信号进行逆傅里叶变换，得到均衡滤波后的分组时域信号，包括：
对均衡滤波后的奇组频域信号进行逆傅里叶变换，得到均衡滤波后的 奇组时域信号；对均衡滤波后的偶组频域信号进行逆傅里叶变换，得到均衡滤波后的奇组时域信号。
作为一种实施方式，所述方法还包括：
将分组后的时域信号分为预设的N段，分段时域信号之间有重叠，重叠长度不小于滤波器长度减1；
相应的，所述将所述均衡滤波后的分组时域信号进行合并，得到完整的均衡滤波后时域信号，包括：将所述N段时域信号经过均衡滤波后，去掉重叠，进行合并得到一组完整的时域信号；
所述N为大于1的整数。
作为一种实施方式，所述根据频域均衡滤波系数，对所述分组频域信号进行均衡滤波满足以下表达式：
Figure PCTCN2015091139-appb-000001
其中，n为分段索引，Axh,even，Axv,even，Ayh,even，Ayv,even为偶组频域均衡滤波系数，Axh,odd，Axv,odd，Ayh,odd，Ayv,odd为奇组频域均衡滤波系数；Heven,n，Veven,n为第n段偶组频域数据；Hodd,n，Vodd,n为第n段奇组频域数据；Xn(k)为第n段均衡滤波后x频域信号，Yn(k)为第n段均衡后y频域信号。
作为一种实施方式，所述频域均衡滤波系数的计算方法为：
按照预设的均衡滤波系数更新算法计算时域均衡滤波系数；
基于傅立叶变换技术，将所述时域均衡滤波系数转换得到频域均衡滤波系数。
本发明实施例还提供了一种存储介质，所述存储介质中存储有计算机可执行指令，所述计算机可执行指令配置为执行本发明实施例所述的光纤通信中自适应均衡滤波方法。
本发明实施例所提供的光纤通信中自适应均衡滤波装置、方法和存储介质，本发明实施例提供的自适应均衡滤波装置，通过将输入的时域信号按照样点索引分组、分段，将信号分成多个小段，而滤波器系数亦按索引奇偶分成两部分，之后再使用频域滤波方式，所需的傅立叶变换的点数少，点数较小的傅立叶变换，实现简单，电路规模较小，功耗小，延迟也小。
附图说明
图1为偏振复用光纤通信中典型的数字相干接收机结构示意图；
图2为本发明实施例提供的光纤通信中自适应均衡滤波装置的组成结构示意图；
图3为本发明提供的一种光纤通信中自适应均衡滤波装置实施例在一种应用场景下的结构示意图；
图4为本发明实施例提供的数据分组分段单元示意图；
图5为本发明实施例提供的一种光纤通信中自适应均衡滤波方法的流程图。
具体实施方式
为了更清楚地说明本发明实施例和技术方案，下面将结合附图及实施例对本发明的技术方案进行更详细的说明，显然，所描述的实施例是本发明的一部分实施例，而不是全部实施例。基于本发明的实施例，本领域普通技术人员在不付出创造性劳动的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。
图2为本发明实施例提供的光纤通信中自适应均衡滤波装置的结构示意图，如图2所示，所述装置包括：数据分组单元201、时频转换单元202、均衡滤波单元203、逆傅立叶变换单元204和数据合并单元205；其中，
所述数据分组单元201，配置为将接收的时域信号按照样点索引分组， 得到分组时域信号；
所述时频转换单元202，配置为将所述数据分组单元201得到的分组时域信号转换为分组频域信号；
所述均衡滤波单元203，配置为根据频域均衡滤波系数，对所述时频转换单元202得到的分组频域信号进行均衡滤波；
所述逆傅立叶变换单元204，配置为对所述均衡滤波单元203进行均衡滤波后的分组频域信号进行逆傅里叶变换，得到均衡滤波后的分组时域信号；
所述数据合并单元205，配置为将所述逆傅立叶变换单元204得到均衡滤波后的分组时域信号进行合并，得到完整的均衡滤波后时域信号。
在一个实施例中，上述数据分组单元201将接收的时域信号按照样点索引分为奇组和偶组，即按照时域信号样点索引是奇数还是偶数将原始时域信号分为奇组时域信号和偶组时域信号；这里，所述原始时域信号是指未经均衡滤波的时域信号。此时，上述自适应均衡滤波装置中，所述时频转换单元202包括：
奇组时频转换模块，配置为将奇组时域信号转换为奇组频域信号；
偶组时频转换模块，配置为将偶组时域信号转换为偶组频域信号。
相应的，所述均衡滤波单元203包括：
奇组均衡滤波模块，配置为根据奇组频域均衡滤波系数对所述奇组时频转换模块得到的奇组频域信号进行均衡滤波；
偶组均衡滤波模块，配置为根据偶组频域均衡滤波系数对所述偶组时频转换模块得到的偶组频域信号进行均衡滤波。
相应的，所述逆傅立叶变换单元204包括：
奇组逆傅立叶变换模块，配置为对所述奇组均衡滤波模块进行均衡滤波后的奇组频域信号进行逆傅里叶变换，得到均衡滤波后的奇组时域信号；
偶组逆傅立叶变换模块，配置为对所述偶组均衡滤波模块进行均衡滤波后的偶组频域信号进行逆傅里叶变换，得到均衡滤波后的奇组时域信号。
在一个实施例中，为进一步减小实现自适应均衡滤波装置的电路规模，可将分组后的时域信号进一步分段，即上述自适应均衡滤波装置中，所述数据分组单元201还包括：
数据分段单元，配置为将分组后的时域信号分为预设的N段，分段时域信号之间有重叠，重叠长度不小于滤波器长度减1；
相应的，所述时频转换单元202还包括N个分段时频转换模块；其中，每个分段时频转换模块，配置为将对应的数据分段单元得到的分段时域信号转换为分段频域信号；
相应的，所述均衡滤波单元203还包括N个分段均衡滤波模块；其中，每个分段均衡滤波模块，配置为根据频域均衡滤波系数，将对应的分段时频转换模块得到的频域信号进行均衡滤波；
相应的，所述逆傅立叶变换单元204还包括N个分段逆傅立叶变换模块；其中，每个分段逆傅立叶变换模块，配置为将对应的分段均衡滤波模块进行均衡滤波后的频域信号进行逆傅里叶变换，得到均衡滤波后的时域信号；
相应的，所述数据合并单元205还包括分段数据合并单元，配置为将同属于一组的均衡滤波后的时域信号，去掉重叠，进行合并得到一组完整的时域信号；
这里，N为大于1的整数。
在一个实施例中，所述自适应均衡滤波控制装置中还包括：系数更新单元和系数时频转换单元；其中，
系数更新单元，配置为按照预设的均衡滤波系数更新算法计算时域均衡滤波系数；
系数时频转换单元，配置为将所述系数更新单元计算出的时域均衡滤波系数进行傅立叶变换，得到频域均衡滤波系数。
具体的，当所述数据分组单元将接收的时域信号按照样点索引分为奇组和偶组时，所述系数更新单元包括：
奇组系数更新模块，配置为按照预设的均衡滤波系数更新算法计算奇组时域均衡滤波系数；
偶组系数更新模块，配置为按照预设的均衡滤波系数更新算法计算偶组时域均衡滤波系数；
相应的，所述系数时频转换单元包括：
奇组系数时频转换单元，配置为将所述奇组系数更新模块计算出的奇组时域均衡滤波系数进行傅立叶变换，得到奇组频域均衡滤波系数；
偶组系数时频转换单元，配置为将所述偶组系数更新模块计算出的偶组时域均衡滤波系数进行傅立叶变换，得到偶组频域均衡滤波系数。
上述各个单元及各个模块在实际应用中，均可由位于数字相干接收机中的中央处理器(CPU)、微处理器(MPU)、数字信号处理器(DSP)、或现场可编程门阵列(FPGA)实现。
图3为本发明提供的一种光纤通信中自适应均衡滤波装置实施例在一种应用场景下的结构示意图，在所述应用场景下，自适应均衡滤波装置接收两路时域信号h和v；在每一个时钟周期(即一拍)内，在每路时域信号上采集128个数据样点，如图3所示，所述装置包括：数据分组单元401、时频转换单元402、均衡滤波单元403、逆傅立叶变换单元404、数据合并单元405、系数更新单元406和系数时频转换单元407；
具体的，所述数据分组单元401对128点并行输入的h偏振信号，与上一拍输入的128点数据重叠8个样点，得到136个样点。把136个样点按照样点索引分为奇组时域信号和偶组时域信号，再将每组时域信号分成 N＝5个数据段，每段间重叠3个样点；显然，也可以先将136个样点分成5个数据段，每段间重叠6个样点，每小段数据再按索引的奇偶分组，得到的分组分段结果是相同的；图4为本发明实施例提供的数据分组分段单元示意图，如图4所示。同理，偏振信号v也进行上述分组分段操作。对于h和v的所述136个样点，可分别记为：
h(i)，v(i)，i＝1，…，136；
将136的样点分成5小段，每段重叠6个样点，每小段长度为32，再按样点索引分成奇偶两组。则分组分段后的数据为满足公式(1)和公式(2)：
Figure PCTCN2015091139-appb-000002
Figure PCTCN2015091139-appb-000003
其中，n＝1,…,5，i＝1,…,16；hodd,n表示奇组h时域信号的第n段数据，heven,n表示偶组h时域信号的第n段数据，vodd,n表示奇组v时域信号的第n段数据；veven,n表示偶组v时域信号的第n段数据。
所述时频转换单元402将所述数据分组单元401输出的分组分段信号分别进行16点傅立叶变换，得到频域信号并输出给所述均衡滤波单元403。hodd,n的傅立叶变换记为Hodd,n，heven,n的傅立叶变换记为Heven,n，vodd,n的傅立叶变换记为Vodd,n，veven,n的傅立叶变换记为Veven,n。在实践中，10(＝N*2)个实现16点傅立叶变换所需的电路规模小于1个实现128点傅立叶变换所需的电路规模，因此，将时域信号样点进行分组分段处理，可降低时频转换单元的电路规模。
所述均衡滤波单元403根据所述系数时频转换单元406输出的频域均衡滤波系数，对所述时频转换单元402输出的频域信号进行均衡滤波，并将均衡后的x频域信号和y频域信号输出给所述逆傅立叶变换单元404；其 中，均衡滤波的计算方法满足公式(3)：
Figure PCTCN2015091139-appb-000004
这里，n为分段索引，Axh,even，Axv,even，Ayh,even，Ayv,even为偶组频域均衡滤波系数，Axh,odd，Axv,odd，Ayh,odd，Ayv,odd为奇组频域均衡滤波系数；Heven,n，Veven,n为第n段偶组频域数据；Hodd,n，Vodd,n为第n段奇组频域数据；Xn(k)为第n段均衡滤波后x频域信号，Yn(k)为第n段均衡后y频域信号。
所述逆傅立叶变换单元404配置为对所述均衡滤波单元403输出的频域信号进行16点逆傅立叶变换，得到均衡滤波后的x时域信号和y时域信号，并输出给所述数据合并单元405。在实践中，10(＝N*2)个实现16点逆傅立叶变换所需的电路规模小于1个实现128点逆傅立叶变换所需的电路规模，同样，所述逆傅立叶变换单元404的电路规模也得到了降低。
所述数据合并单元405配置为对所述逆傅立叶变换单元404输出的分组分段的x或y时域信号(即经过均衡滤波后的时域信号)，对于奇组数据去掉重叠样点，合并成一段奇组时域信号；同样，对于偶组数据去掉重叠样点，合并成一段偶组时域信号；最终将奇组和偶组数据再合并为完整的时域信号xout和yout输出。
所述系数更新单元406针对时域均衡滤波系数进行更新计算，可以采用的系数更新方法包括：恒模算法(CMA，Constant Modulus Algorithm)、判决辅助最小均方误差法算法(DDLMS，Decision-Directed Least-Mean-Square)、或者多模盲均衡算法，其他针对时域的自适应系数更新算法，对于本发明实施例也是适用的。偏振复用光通信系统的自适应均衡滤波装置中有四组系数：axh(i)、axv(i)、ayh(i)、ayv(i)，i＝0,…,M-1。本实施例中，滤波器抽头个数M＝7。
所述系数时频转换单元407对所述系数更新单元406输出的时域系数axh，按照索引是奇数还是偶数分成两组，并利用傅立叶变换技术，将其变换到频域，得到Axh,odd和Axh,even同理，对于axv、ayh、ayv分别按照索引是奇数还是偶数分成两组，并利用傅立叶变换技术，将其变换到频域，所使用的傅立叶变换点数为16。
现有的偏振复用光通讯接收机自适应均衡滤波装置为蝶形滤波器结构，共4组滤波器系数，为了有效补偿信道失真，滤波器长度较长，且系数为复数。使用现有的时域均衡滤波方式时，电路规模较大，功耗大。使用现有的频域均衡滤波方式时，虽节省了均衡滤波所消耗的资源，但额外增加了点数较大的傅立叶变换和逆傅立叶变换电路，点数较大的傅立叶变换和逆傅立叶变换的电路规模大、功耗大，同时实现变换带来的延迟大。采用本发明实施例提供的自适应均衡滤波装置，与现有技术相比，按照样点索引分组、分段，将信号分成多个小段，而滤波器系数亦按索引奇偶分成两部分，之后再使用频域滤波方式，所需的傅立叶变换的点数少，点数较小的傅立叶变换，实现简单，电路规模较小，功耗小，延迟也小。定点化评估了7个抽头(即128点傅立叶变换)的自适应均衡滤波时域实现方式与本发明实施例(即16点傅立叶变换)实现方式，本发明实施例均衡滤波装置比现有的均衡滤波装置实现方式要节约电路规模达到50％。
图5是本发明实施例提供的一种光纤通信中自适应均衡滤波方法的流程图，如图5所示，所述方法包括：
步骤601，将接收的时域信号按照样点索引分组，得到分组时域信号。
步骤602，将所述分组时域信号转换为分组频域信号。
步骤603，根据频域均衡滤波系数，对所述分组频域信号进行均衡滤波。
步骤604，对均衡滤波后的分组频域信号进行逆傅里叶变换，得到均衡滤波后的分组时域信号。
步骤605，将所述均衡滤波后的分组时域信号进行合并，得到完整的均衡滤波后时域信号。
在一个实施例中，步骤601中所述将接收的时域信号按照样点索引分组包括：将接收的时域信号按照样点索引分为奇组和偶组，即按照时域信号样点索引是奇数还是偶数将原始时域信号分为奇组时域信号和偶组时域信号；这里，所述原始时域信号是指未经均衡滤波的时域信号；
相应的，步骤602中所述将所述分组时域信号转换为分组频域信号，包括：将奇组时域信号转换为奇组频域信号，将偶组时域信号转换为偶组频域信号；
相应的，步骤603中所述根据频域均衡滤波系数，对所述分组频域信号进行均衡滤波，包括：根据奇组频域均衡滤波系数对奇组频域信号进行均衡滤波；根据偶组频域均衡滤波系数对偶组频域信号进行均衡滤波；
相应的，步骤604中所述对均衡滤波后的分组频域信号进行逆傅里叶变换，得到均衡滤波后的分组时域信号，包括：
对均衡滤波后的奇组频域信号进行逆傅里叶变换，得到均衡滤波后的奇组时域信号；对均衡滤波后的偶组频域信号进行逆傅里叶变换，得到均衡滤波后的奇组时域信号。
在一个实施例中，为进一步减小实现自适应均衡滤波装置的电路规模，可将分组后的时域信号进一步分段，所述方法还包括：
将分组后的时域信号分为预设的N段，分段时域信号之间有重叠，重叠长度不小于滤波器长度减1；
相应的，步骤605中，所述将所述均衡滤波后的分组时域信号进行合并，得到完整的均衡滤波后时域信号，包括：将所述N段中同属于一组的均衡滤波后的时域信号经过均衡滤波后，去掉重叠，进行合并得到一组完整的时域信号；所述N为大于1的整数。
本实施例中，步骤603中，所述根据频域均衡滤波系数，对所述分组频域信号进行均衡滤波满足公式(4)：
Figure PCTCN2015091139-appb-000005
其中，n为分段索引，Axh,even，Axv,even，Ayh,even，Ayv,even为偶组频域均衡滤波系数，Axh,odd，Axv,odd，Ayh,odd，Ayv,odd为奇组频域均衡滤波系数；Heven,n，Veven,n为第n段偶组频域数据；Hodd,n，Vodd,n为第n段奇组频域数据；Xn(k)为第n段均衡滤波后x频域信号，Yn(k)为第n段均衡后y频域信号。
在一个实施例中，所述频域均衡滤波系数的计算方法为：
按照预设的均衡滤波系数更新算法计算时域均衡滤波系数；
基于傅立叶变换技术，将所述时域均衡滤波系数转换得到频域均衡滤波系数。
其中，当将接收的时域信号按照样点索引分为奇组和偶组时，所述按照预设的均衡滤波系数更新算法计算时域均衡滤波系数，包括：按照预设的均衡滤波系数更新算法计算奇组时域均衡滤波系数；按照预设的均衡滤波系数更新算法计算偶组时域均衡滤波系数；
相应的，所述基于傅立叶变换技术，将所述时域均衡滤波系数转换得到频域均衡滤波系数，包括：
将所述奇组时域均衡滤波系数进行傅立叶变换，得到奇组频域均衡滤波系数；将所述偶组时域均衡滤波系数进行傅立叶变换，得到偶组频域均衡滤波系数。
本发明实施例还提供了一种存储介质，所述存储介质中存储有计算机可执行指令，所述计算机可执行指令配置为执行本发明实施例所述的光纤通信中自适应均衡滤波方法。
本领域内的技术人员应明白，本发明的实施例可提供为方法、系统、或计算机程序产品。因此，本发明可采用硬件实施例、软件实施例、或结合软件和硬件方面的实施例的形式。而且，本发明可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器和光学存储器等)上实施的计算机程序产品的形式。
本发明是参照根据本发明实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
以上所述，仅为本发明的较佳实施例而已，并非用于限定本发明的保护范围。
工业实用性
本发明实施例通过将输入的时域信号按照样点索引分组、分段，将信号分成多个小段，而滤波器系数亦按索引奇偶分成两部分，之后再使用频域滤波方式，所需的傅立叶变换的点数少，点数较小的傅立叶变换，实现简单，电路规模较小，功耗小，延迟也小。

Claims (11)

  1. 一种光纤通信中自适应均衡滤波装置，所述装置包括：数据分组单元、时频转换单元、均衡滤波单元、逆傅立叶变换单元和数据合并单元；其中，
    所述数据分组单元，配置为将接收的时域信号按照样点索引分组，得到分组时域信号；
    所述时频转换单元，配置为将所述数据分组单元得到的分组时域信号转换为分组频域信号；
    所述均衡滤波单元，配置为根据频域均衡滤波系数，对所述时频转换单元得到的分组频域信号进行均衡滤波；
    所述逆傅立叶变换单元，配置为对所述均衡滤波单元进行均衡滤波后的分组频域信号进行逆傅里叶变换，得到均衡滤波后的分组时域信号；
    所述数据合并单元，配置为将所述逆傅立叶变换单元得到的分组时域信号进行合并，得到完整的均衡滤波后时域信号。
  2. 根据权利要求1所述的自适应均衡滤波装置，其中，当所述数据分组单元将接收的时域信号按照样点索引分为奇组和偶组时，所述时频转换单元包括：
    奇组时频转换模块，配置为将奇组时域信号转换为奇组频域信号；
    偶组时频转换模块，配置为将偶组时域信号转换为偶组频域信号；
    相应的，所述均衡滤波单元包括：
    奇组均衡滤波模块，配置为根据奇组频域均衡滤波系数对所述奇组时频转换模块得到的奇组频域信号进行均衡滤波；
    偶组均衡滤波模块，配置为根据偶组频域均衡滤波系数对所述偶组时频转换模块得到的偶组频域信号进行均衡滤波；
    相应的，所述逆傅立叶变换单元包括：
    奇组逆傅立叶变换模块，配置为对所述奇组均衡滤波模块进行均衡滤波后的奇组频域信号进行逆傅里叶变换，得到均衡滤波后的奇组时域信号；
    偶组逆傅立叶变换模块，配置为对所述偶组均衡滤波模块进行均衡滤波后的偶组频域信号进行逆傅里叶变换，得到均衡滤波后的奇组时域信号。
  3. 根据权利要求1或2所述的自适应均衡滤波装置，其中，所述数据分组单元还包括：
    数据分段单元，配置为将分组后的时域信号分为预设的N段，分段时域信号之间有重叠，重叠长度不小于滤波器长度减1；
    相应的，所述时频转换单元还包括N个分段时频转换模块；其中，每个分段时频转换模块，配置为将对应的数据分段单元得到的分段时域信号转换为分段频域信号；
    相应的，所述均衡滤波单元还包括N个分段均衡滤波模块；其中，每个分段均衡滤波模块，配置为根据频域均衡滤波系数，将对应的分段时频转换模块得到的频域信号进行均衡滤波；
    相应的，所述逆傅立叶变换单元还包括N个分段逆傅立叶变换模块；其中，每个分段逆傅立叶变换模块，配置为将对应的分段均衡滤波模块进行均衡滤波后的频域信号进行逆傅里叶变换，得到均衡滤波后的时域信号；
    相应的，所述数据合并单元还包括分段数据合并单元，配置为将同属于一组的均衡滤波后的时域信号，去掉重叠，进行合并得到一组完整的时域信号；
    所述N为大于1的整数。
  4. 根据权利要求1所述的自适应均衡滤波装置，其中，所述装置还包括：
    系数更新单元，配置为按照预设的均衡滤波系数更新算法计算时域均衡滤波系数；
    系数时频转换单元，配置为将所述系数更新单元计算出的时域均衡滤波系数进行傅立叶变换，得到频域均衡滤波系数。
  5. 根据权利要求4所述的自适应均衡滤波装置，其中，当所述数据分组单元将接收的时域信号按照样点索引分为奇组和偶组时，所述系数更新单元包括：
    奇组系数更新模块，配置为按照预设的均衡滤波系数更新算法计算奇组时域均衡滤波系数；
    偶组系数更新模块，配置为按照预设的均衡滤波系数更新算法计算偶组时域均衡滤波系数；
    相应的，所述系数时频转换单元包括：
    奇组系数时频转换单元，配置为将所述奇组系数更新模块计算出的奇组时域均衡滤波系数进行傅立叶变换，得到奇组频域均衡滤波系数；
    偶组系数时频转换单元，配置为将所述偶组系数更新模块计算出的偶组时域均衡滤波系数进行傅立叶变换，得到偶组频域均衡滤波系数。
  6. 一种光纤通信中自适应均衡滤波方法，所述方法包括：
    将接收的时域信号按照样点索引分组，得到分组时域信号；
    将所述分组时域信号转换为分组频域信号；
    根据频域均衡滤波系数，对所述分组频域信号进行均衡滤波；
    对均衡滤波后的分组频域信号进行逆傅里叶变换，得到均衡滤波后的分组时域信号；
    将所述均衡滤波后的分组时域信号进行合并，得到完整的均衡滤波后时域信号。
  7. 根据权利要求6所述的自适应均衡滤波方法，其中，所述将接收的时域信号按照样点索引分组包括：
    将接收的时域信号按照样点索引分为奇组和偶组；
    相应的，所述将所述分组时域信号转换为分组频域信号，包括：
    将奇组时域信号转换为奇组频域信号，将偶组时域信号转换为偶组频域信号；
    相应的，所述根据频域均衡滤波系数，对所述分组频域信号进行均衡滤波，包括：
    根据奇组频域均衡滤波系数对奇组频域信号进行均衡滤波；根据偶组频域均衡滤波系数对偶组频域信号进行均衡滤波；
    相应的，所述对均衡滤波后的分组频域信号进行逆傅里叶变换，得到均衡滤波后的分组时域信号，包括：
    对均衡滤波后的奇组频域信号进行逆傅里叶变换，得到均衡滤波后的奇组时域信号；对均衡滤波后的偶组频域信号进行逆傅里叶变换，得到均衡滤波后的奇组时域信号。
  8. 根据权利要求6或7所述的自适应均衡滤波方法，其中，所述方法还包括：
    将分组后的时域信号分为预设的N段，分段时域信号之间有重叠，重叠长度不小于滤波器长度减1；
    相应的，所述将所述均衡滤波后的分组时域信号进行合并，得到完整的均衡滤波后时域信号，包括：将所述N段中同属于一组的均衡滤波后的时域信号经过均衡滤波后，去掉重叠，进行合并得到一组完整的时域信号；
    所述N为大于1的整数。
  9. 根据权利要求8所述的自适应均衡滤波方法，其中，所述根据频域均衡滤波系数，对所述分组频域信号进行均衡滤波满足以下表达式：
    Figure PCTCN2015091139-appb-100001
    其中，n为分段索引，Axh,even，Axv,even，Ayh,even，Ayv,even为偶组频域均衡滤波系数，Axh,odd，Axv,odd，Ayh,odd，Ayv,odd为奇组频域均衡滤波系数；Heven,n，Veven,n为第n段偶组频域数据；Hodd,n，Vodd,n为第n段奇组频域数据；Xn(k)为第n段均衡滤波后x频域信号，Yn(k)为第n段均衡后y频域信号。
  10. 根据权利要求6至9任一项所述的自适应均衡滤波方法，其中，所述频域均衡滤波系数的计算方法为：
    按照预设的均衡滤波系数更新算法计算时域均衡滤波系数；
    基于傅立叶变换技术，将所述时域均衡滤波系数转换得到频域均衡滤波系数。
  11. 一种存储介质，所述存储介质中存储有计算机可执行指令，所述计算机可执行指令配置为执行权利要求6至10任一项所述的光纤通信中自适应均衡滤波方法。
PCT/CN2015/091139 2014-12-30 2015-09-29 一种光纤通信中自适应均衡滤波装置、方法和存储介质 WO2016107236A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201410840453.5A CN105812060B (zh) 2014-12-30 2014-12-30 一种光纤通信中自适应均衡滤波装置和方法
CN201410840453.5 2014-12-30

Publications (1)

Publication Number Publication Date
WO2016107236A1 true WO2016107236A1 (zh) 2016-07-07

Family

ID=56284147

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/091139 WO2016107236A1 (zh) 2014-12-30 2015-09-29 一种光纤通信中自适应均衡滤波装置、方法和存储介质

Country Status (2)

Country Link
CN (1) CN105812060B (zh)
WO (1) WO2016107236A1 (zh)

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018068177A1 (zh) * 2016-10-10 2018-04-19 华为技术有限公司 梳状滤波噪声消除方法、装置及频域自适应均衡装置
CN109038577B (zh) * 2018-09-11 2021-07-09 广东石油化工学院 负荷分解中功率信号自适应重构方法
CN108918932B (zh) * 2018-09-11 2021-01-15 广东石油化工学院 负荷分解中功率信号自适应滤波方法
CN109257069B (zh) * 2018-09-11 2021-09-21 广东石油化工学院 电力线通信信号自适应滤波方法
CN109951174B (zh) * 2019-04-16 2021-07-13 上海交通大学 一种快速傅里叶变换滤波方法
CN110247703A (zh) * 2019-04-25 2019-09-17 清华大学 一种高速可见光通信系统中的分段频域前均衡方法
CN110492946B (zh) * 2019-04-29 2021-04-27 华中科技大学 一种采用单探测器的光场重建与色散补偿系统与方法
CN110768728B (zh) * 2019-04-29 2021-06-15 华中科技大学 一种偏振无关光场重建与码间干扰补偿系统与方法
CN112713941A (zh) 2019-10-24 2021-04-27 富士通株式会社 静态均衡器系数的确定装置及方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101478510A (zh) * 2009-02-17 2009-07-08 上海高清数字科技产业有限公司 一种自适应均衡器及使用该均衡器的接收机系统
CN101692628A (zh) * 2009-09-10 2010-04-07 复旦大学 基于单边带调制的单载波频域均衡技术的光纤通信系统
CN102868651A (zh) * 2012-10-08 2013-01-09 中国电子科技集团公司第五十四研究所 一种基于频域均衡的新型解调方法
CN103763226A (zh) * 2014-01-06 2014-04-30 南京复实通讯科技有限公司 混合时频域均衡的方法
CN103959693A (zh) * 2011-12-01 2014-07-30 三菱电机株式会社 接收装置和方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101478510A (zh) * 2009-02-17 2009-07-08 上海高清数字科技产业有限公司 一种自适应均衡器及使用该均衡器的接收机系统
CN101692628A (zh) * 2009-09-10 2010-04-07 复旦大学 基于单边带调制的单载波频域均衡技术的光纤通信系统
CN103959693A (zh) * 2011-12-01 2014-07-30 三菱电机株式会社 接收装置和方法
CN102868651A (zh) * 2012-10-08 2013-01-09 中国电子科技集团公司第五十四研究所 一种基于频域均衡的新型解调方法
CN103763226A (zh) * 2014-01-06 2014-04-30 南京复实通讯科技有限公司 混合时频域均衡的方法

Also Published As

Publication number Publication date
CN105812060B (zh) 2018-08-03
CN105812060A (zh) 2016-07-27

Similar Documents

Publication Publication Date Title
WO2016107236A1 (zh) 一种光纤通信中自适应均衡滤波装置、方法和存储介质
US10135543B2 (en) Clock recovery method, device and system and computer storage medium
JP2016521084A (ja) マルチブランチフィルタバンク構造を用いた波長分散補償モジュールを有する光受信機
US8861981B2 (en) Optical signal compensation device
JP6476716B2 (ja) チャネル間の非線形劣化の補償装置及び補償方法
JP7248072B2 (ja) ディジタル信号処理装置
JP6301478B2 (ja) 信号を送信及び受信するための方法、並びに対応する装置及びシステム
US10171177B2 (en) Digital signal processor, digital optical receiver using the same, and digital signal processing method
US9496965B2 (en) Chromatic dispersion compensator with integrated anti-aliasing filter and resampler
JP6376211B2 (ja) 波長分散補償装置、波長分散補償方法及び通信装置
WO2019130002A1 (fr) Système de démodulation ou de recherche en aveugle des caractéristiques de signaux de télécommunication numérique
WO2016101541A1 (zh) 时钟恢复均衡装置与方法、计算机存储介质
WO2017077802A1 (ja) 波長分散推定回路、光受信装置及び波長分散量推定方法
KR20210090676A (ko) 신호 처리 방법, 장치, 설비 및 컴퓨터 저장 매체
US9768883B2 (en) Low power equalizer for an optical receiver
CN106911409B (zh) 群时延信道系统的优化方法及其装置
CN110061780B (zh) 一种色散补偿方法及装置
JP6233318B2 (ja) 周波数オフセット補償装置および周波数オフセット補償方法
US11108464B2 (en) Apparatus and method for determining coefficients of fixed equalizer
CN115398827B (zh) 用于补偿光传输系统中的群速度色散的数字滤波器装置
WO2018068177A1 (zh) 梳状滤波噪声消除方法、装置及频域自适应均衡装置
JP2016208417A (ja) 光受信装置
WO2019129759A1 (fr) Rétro-propagation d'erreur pour une chaine de démodulation en aveugle d'un signal de télécommunication numérique
Radic Coherent Signal Processing: Fundamentals and Applications

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15874921

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15874921

Country of ref document: EP

Kind code of ref document: A1