阅读说明：本技术 一种光纤通信系统中非线性损伤预补偿方法 (Nonlinear damage precompensation method in optical fiber communication system ) 是由 忻向军 张琦 蒋锦坤 王曦朔 姚海鹏 刘炳春 田清华 田凤 王拥军 杨雷静 于 2020-11-23 设计创作，主要内容包括：本发明提供了一种光纤通信系统中非线性损伤预补偿方法及装置,涉及光纤通信系统领域。其中,上述方法包括：根据符号之间的时间相关性和相互作用,在发射端进行高阶非线性建模,预测当前时刻输出；根据瞬时平方误差最小化准则,计算使预测值和接收误差最小的各阶预补偿抽头系数；根据时延组合和符号组合在各个阶次下建立二维查找表存储预补偿抽头系数；根据累加平均原则,更新查找表中元素；根据训练得到的查找表,对与模型匹配的滤波器进行配置,从而得到预补偿的信号。应用本发明实施例提供的方案进行非线性预补偿,可以提高系统的非线性容忍度,充分利用发射端计算资源和功耗潜能,降低接收端复杂度,平衡发射端与接收端的功耗成本。本发明同时还提供了一种预均衡与预补偿的联合方法,根据级联支路合并原则,将预均衡抽头系数乘入到含有相同时延项的与补偿抽头系数中,实现滤波器的共享。应用本发明实例提供的联合方法,可以提高发射端硬件资源利用率。(The invention provides a method and a device for pre-compensating nonlinear damage in an optical fiber communication system, and relates to the field of optical fiber communication systems. Wherein, the method comprises the following steps: according to the time correlation and interaction among the symbols, performing high-order nonlinear modeling at a transmitting end, and predicting the output at the current moment; calculating each order pre-compensation tap coefficient which minimizes the predicted value and the receiving error according to the instantaneous square error minimization criterion; establishing a two-dimensional lookup table at each order according to the time delay combination and the symbol combination to store a pre-compensation tap coefficient; updating elements in the lookup table according to an accumulation average principle; and configuring the filter matched with the model according to the lookup table obtained by training so as to obtain a precompensated signal. By applying the scheme provided by the embodiment of the invention to carry out nonlinear precompensation, the nonlinear tolerance of the system can be improved, the computing resources and the power consumption potential of the transmitting terminal are fully utilized, the complexity of the receiving terminal is reduced, and the power consumption cost of the transmitting terminal and the receiving terminal is balanced. The invention also provides a combined method of pre-equalization and pre-compensation, which multiplies the pre-equalization tap coefficient into the compensation tap coefficient containing the same time delay term according to the merging principle of the cascade branches to realize the sharing of the filter. The application of the combination method provided by the embodiment of the invention can improve the utilization rate of the hardware resources of the transmitting terminal.)

1. A method for pre-compensating nonlinear damage in an optical fiber communication system is characterized by comprising the following steps:

according to the time correlation and interaction among the symbols, performing high-order nonlinear modeling at a transmitting end, and predicting the output at the current moment;

calculating each order pre-compensation tap coefficient which minimizes the predicted value and the receiving error according to the instantaneous square error minimization criterion;

establishing a two-dimensional lookup table at each order to store a precompensation tap coefficient according to the time delay combination and the symbol combination of the mode under the time delay combination;

updating elements in the lookup table according to an accumulation average principle;

and configuring the filter matched with the model according to the lookup table obtained by training so as to obtain a precompensated signal.

2. The higher-order nonlinear precompensation model according to claim 1, characterized in that the predicted symbols of the precompensation algorithm are expressed at the receiving end as polynomials composed of symbols in a fixed-length transmitted symbol sequence (pattern) according to the time correlation and interaction between symbols, and have the following time-dependent, higher-order nonlinear form:

where y (k) is the predicted symbol at the current time, x (k- τ)_j) For a time delay of τ_jFrom pattern X (k-M: k: k + M), the length of the pattern being 2M +1,for pre-compensation of the tap coefficients, D is the polynomial order under consideration for the pre-compensation.

3. The non-linear pre-compensation algorithm of claim 1 and claim 2, comprising:

establishing a two-dimensional lookup table at each order to store a precompensation tap coefficient according to the time delay combination and the symbol combination of the mode under the time delay combination;

and during pre-compensation, pre-compensation tap coefficients corresponding to all orders are searched in the table according to the sending mode to configure the filter.

4. A lookup table as claimed in claim 1 and claim 3 comprising:

establishing two dimensions according to a time delay combination and a symbol combination of a mode under the time delay combination;

calculating a precompensation tap coefficient according to an instantaneous square error minimization criterion;

updating the pre-compensation tap coefficient according to the principle of accumulation average;

the d table is a look-up table of precompensated tap coefficients of a d-th order polynomial, the contents of which are combined according to time delay tau₁,τ₂,...,τ_dSymbol combination x (k-tau) of sum mode under the time delay combination₁),x(k-τ₂),...,x(k-τ_d) A pre-compensation tap coefficient look-up table stored for the index.

5. A joint method of pre-equalization and pre-compensation, wherein a tap coefficient of pre-equalization is multiplied to a tap coefficient of pre-compensation having the same delay term, so as to realize the joint of pre-equalization and pre-compensation, and share a first-order filter, the method comprising:

common tap processing: for the taps existing in both pre-equalization and pre-compensation, multiplying the pre-equalization tap coefficient by the pre-compensation tap coefficient cascaded with the pre-equalization tap coefficient;

and (3) judging the residual tap: judging whether unprocessed taps exist or not, and if not, finishing the combination of pre-equalization and pre-compensation; if yes, the residual equalization tap processing is carried out.

And (3) residual balanced tap processing: for the remaining taps that are only present in the pre-equalization, the retention is done and the tap coefficients are not processed.

6. An apparatus for pre-compensating for nonlinear impairments in an optical fiber communication system, applied to a transmitting end, the apparatus comprising:

the tap coefficient calculation module is used for calculating the pre-compensation tap coefficient, the accumulated value of the pre-compensation tap coefficient, the accumulated number of the tap coefficients, the accumulated average value of the pre-compensation tap coefficient and the joint tap coefficient;

the table look-up storage module is used for storing the data according to the structure of a look-up table;

and the filter configuration module is used for filtering according to the data of the lookup table to realize precompensation.

7. An electronic device comprising a processor, a memory, a communication interface, and a communication bus, wherein:

a processor for executing a program stored in the memory to perform the method steps of any of claims 1-5;

a memory for storing computer programs and data;

a communication interface for communication between the electronic device and the outside;

communication bus: for communication between the processor, the memory and the communication interface.

Technical Field

The invention relates to the field of optical communication, in particular to a nonlinear optical fiber communication system sensitive to system cost, especially receiving end cost, and a nonlinear damage precompensation method and device.

Background

With the development of the optical fiber communication system to ultra-high speed and ultra-long distance, the number of wavelength division multiplexing channels, the single channel rate and the signal power are continuously increased, the nonlinear effect becomes an important factor restricting the system performance, and the nonlinear pre-compensation and post-compensation effects directly influence the transmission performance of the system, so that the method has important practical significance for the research of nonlinear pre-compensation and post-compensation schemes. In order to ensure the reliability of transmission and the usability of the device, a solution with high precision and low cost needs to be provided. At present, a pre-compensation scheme widely used at a transmitting end adopts first-order nonlinear modeling, so that the precision of channel estimation and pre-compensation is difficult to guarantee; the post-compensation scheme based on series expansion, which is used in large quantity at the receiving end, causes the complexity of the receiving end to be improved due to the complex implementation structure, further increases the power consumption cost of the receiving end, and is not beneficial to hardware implementation and popularization in networks. In order to avoid further aggravating the mismatching of the power consumption cost of the transmitting terminal and the receiving terminal, the potential of the transmitting terminal is fully utilized, and the research of the nonlinear damage precompensation method based on the transmitting terminal is significant.

The invention mainly designs a novel nonlinear precompensation scheme suitable for an optical fiber communication system, and provides a precompensation algorithm based on high-order nonlinear modeling and a tap coefficient lookup table. The transmitting end considers the high-order terms which have correlation in time and interact among a plurality of symbols to establish a nonlinear model and uses a lookup table to complete the configuration of a nonlinear filter, thereby achieving the aim of precompensation. The invention also provides a pre-equalization and pre-compensation combined method through tap coefficient operation, which realizes the sharing of the first-order filter tap and improves the utilization rate of the hardware resources of the transmitting terminal.

Disclosure of Invention

An object of an embodiment of the present invention is to provide a method and an apparatus for pre-compensating for nonlinear impairments in an optical fiber communication system, so as to solve the problems encountered by the current methods for pre-compensating and post-compensating for nonlinear impairments, including:

the nonlinear precompensation scheme based on first-order nonlinear modeling has the problems of insufficient modeling precision and poor precompensation performance;

the problem that the power consumption cost of a transmitting end is not matched with that of a receiving end is aggravated due to the fact that the complexity of the receiving end is increased and the cost is increased caused by a post-compensation scheme based on a series expansion method;

the problem of hardware resource waste caused by the fact that pre-equalization and pre-compensation cannot be effectively combined.

In order to solve the above problem, the embodiment of the present invention is implemented as follows:

in a first aspect, an embodiment of the present invention provides a nonlinear precompensation method, including:

high-order nonlinear modeling: according to the time correlation and interaction among the symbols, a high-order polynomial item model is established at a transmitting end, and the current time output is predicted;

pre-compensation tap coefficient calculation: calculating each order pre-compensation tap coefficient which minimizes the predicted value and the receiving error according to the instantaneous square error minimization criterion;

pre-compensation tap coefficient storage: establishing a two-dimensional lookup table at each order to store a precompensation tap coefficient according to the time delay combination and the symbol combination of the mode under the time delay combination;

updating a lookup table: updating elements in the lookup table according to an accumulation average principle;

filter configuration: and configuring the filter matched with the model according to the lookup table obtained by training so as to obtain a precompensated signal.

In the above-mentioned high-order nonlinear modeling, the following steps are included:

setting the highest order D of the intersymbol interaction considered by pre-compensation and the time correlation length 2M +1, and modeling and calculating the predicted symbol according to the following expression:

where the symbol time correlation length is understood to be a time window, only considering that the symbols within the window have a time correlation will have an effect on the current output, the window moving according to the movement of the current time instant. The pattern X (k-M: k: k + M) is a symbol within a window centered on the current time, X (k- τ)_j) For intra-mode delay of τ_jThe transmission symbol of (2) is transmitted,to pre-compensate the tap coefficients, y (k) is the predicted sign at the current time.

In the above pre-compensation tap coefficient calculation, the following steps are included:

assuming that the reception at the receiving end at the current time is r (k), the instantaneous square error is calculated according to the following expression:

wherein X_nIs an n-th order termConstituent column vectors, W_nThe corresponding pre-compensated tap coefficient column vectors.

The pre-compensation tap coefficient vector is calculated according to:

when the input mode is determined, the pre-compensation tap coefficient vector in the mode is determined accordingly, and the result is recorded asIn precompensating tap coefficient vectorFor precompensating tap coefficients in a modelA column vector of components.

In the above pre-compensation tap coefficient storage, the following steps are included:

and establishing 3D lookup tables which are respectively a pre-compensation tap coefficient accumulated value lookup table, an accumulated tap coefficient number lookup table and a pre-compensation tap coefficient accumulated average value lookup table according to the set interaction highest order D, wherein an element in the initialization table is 0. The d-th table is a related element lookup table with the order of d, and the contents of the related element lookup table are respectively as follows:

combined by time delay (tau)₁,τ₂,...,τ_d) For vertical indexing, the symbol combination (x (k-tau)) of the pattern under the time delay combination₁),x(k-τ₂),...,x(k-τ_d) For horizontal indexing, the pre-compensation tap coefficient accumulation value stored at the position specified by two indexesNumber ofAnd the accumulated mean value

In the above lookup table update, the following steps are included:

according to the principle of accumulation average, the elements in the lookup table are respectively updated as follows:

in a second aspect, an embodiment of the present invention provides a joint method of pre-equalization and pre-compensation, including:

common tap processing: for the taps existing in both pre-equalization and pre-compensation, multiplying the pre-equalization tap coefficient by the pre-compensation tap coefficient cascaded with the pre-equalization tap coefficient;

and (3) judging the residual tap: judging whether unprocessed taps exist or not, and if not, finishing the combination of pre-equalization and pre-compensation; if yes, the residual equalization tap processing is carried out.

And (3) residual balanced tap processing: for taps that are only present in the pre-equalization, the retention is done and the tap coefficients are not processed.

In the common tap processing described above, the following steps are included:

the number of taps used for pre-equalization is set to be 2N +1, for taps of a pre-equalization filter with time delay of-M: M, the same taps exist in a pre-compensated first-order filter, the pre-equalization taps are removed, and tap coefficients are directly multiplied by pre-compensated tap coefficients of each order containing the same time delay item:

whereinFor a time delay of τ_iPre-equalizing the tap coefficients.

In the above residual tap judgment, the method includes the following steps:

and judging whether N is larger than M or not for the number of taps 2N +1 of the pre-equalization first-order filter and the number of taps 2M +1 of the pre-compensation first-order filter. If not, the pre-equalization first-order tap is less than the pre-compensation first-order tap, and the combination of pre-equalization and pre-compensation is completed; and if so, performing residual balanced tap processing.

In the above-described residual equalization tap processing, the following steps are included:

when N is more than M, the pre-balanced first-order tap is more than the pre-compensated first-order tap, the tap which does not exist in the pre-compensation is reserved, and the tap coefficient is not processed:

in a third aspect, an embodiment of the present invention provides a nonlinear precompensation apparatus, including:

the tap coefficient calculation module is used for calculating the pre-compensation tap coefficient, the accumulated value of the pre-compensation tap coefficient, the accumulated number of the tap coefficients, the accumulated average value of the pre-compensation tap coefficient and the joint tap coefficient;

the table look-up storage module is used for storing the data according to the structure of a look-up table;

and the filter configuration module is used for filtering according to the data of the lookup table to realize precompensation.

In the above tap coefficient calculation module, the tap coefficient calculation module includes:

the tap coefficient updating submodule is used for calculating the pre-compensation tap coefficient, the accumulated value of the pre-compensation tap coefficient and the accumulated number of the tap coefficients;

the tap coefficient average value calculation submodule is used for calculating the accumulated average value of the precompensation tap coefficient;

and the tap coefficient joint submodule is used for calculating the joint tap coefficient.

In a fourth aspect, an embodiment of the present invention provides an electronic device, which is characterized by including a processor, a memory, a communication interface, and a communication bus, where:

a processor for executing a program stored in the memory to perform the method steps of any of claims 1-5;

a memory for storing computer programs and data;

a communication interface for communication between the electronic device and the outside;

communication bus: for communication between the processor, the memory and the communication interface.

The embodiment of the invention has the following beneficial effects:

in the scheme provided by the embodiment of the invention, the output is predicted by adopting the high-order nonlinear model, the configuration of the precompensation tap is realized by using a lookup table mode, the advantages of high response speed of the precompensation scheme and accurate high-order nonlinear modeling compensation effect are integrated, the nonlinear effect resistance of the system is improved, the complexity and the network construction cost of the receiving end are reduced, the computing resource and the power consumption potential of the transmitting end are fully utilized, and the power consumption cost of the transmitting end and the receiving end is balanced. Because the same tap exists in the pre-equalization filter and the pre-compensation filter at the transmitting end, the pre-equalized tap coefficient is multiplied by the pre-compensation tap coefficient of the cascade branch according to the merging principle of the cascade branch, thereby realizing the sharing of the filter and improving the utilization rate of hardware resources of the transmitting end adopting the combined pre-equalization and pre-compensation method.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention, are incorporated in and constitute a part of this specification. The exemplary embodiments of the present invention and their description are intended to illustrate the present invention and should not be construed as limiting the present invention. In the drawings:

fig. 1 is a schematic flowchart of a nonlinear impairment compensation method in an optical fiber communication system according to an embodiment of the present invention.

Fig. 2 is a structural diagram of a precompensation filter corresponding to a high-order nonlinear model according to an embodiment of the present invention.

Fig. 3 is a schematic diagram illustrating storage of elements of a lookup table of each stage according to an index according to an embodiment of the present invention.

Fig. 4 is a schematic diagram illustrating an updating process of each level of lookup table elements according to an embodiment of the present invention.

Fig. 5 is a schematic diagram of a joint method of pre-equalization and pre-compensation according to an embodiment of the present invention.

Fig. 6 is a schematic diagram of a nonlinear pre-compensation apparatus according to an embodiment of the present invention.

Detailed Description

The technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. These embodiments are merely exemplary and are not intended to limit the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Example 1:

at present, video-based multimedia and broadband mobile services are increasingly demanding for large bandwidth and high data rate transmission. For such application scenarios, a pulse amplitude modulation (PAM-4) system based on intensity modulation direct detection (IM/DD) is highly regarded for its simple structure and low power consumption. However, the performance of the PAM-4 system is severely limited by the nonlinear damage generated in the transmission channel, and therefore, the method has great significance for the research of the nonlinear precompensation algorithm of the PAM-4 system.

In the PAM-4 system, the transmitting end adopts the nonlinear pre-compensation method as shown in fig. 1, including the following steps S101 to S110.

S101, setting the highest order D of the inter-symbol interaction considered by pre-compensation, the symbol time correlation length 2M +1 and the number of pre-equalization taps 2N +1, wherein the structure of the pre-compensation filter is shown in figure 2. Initializing, using the time delay combination as a longitudinal index and the symbol combination under the time delay combination as a transverse index through a table look-up storage module 602, establishing a 3D two-dimensional look-up table, wherein the element is 0, and the content in the table is respectively a pre-compensation tap coefficient accumulated value, the number of accumulated tap coefficients and a pre-compensation tap coefficient accumulated average value. The process of storing the elements in the lookup tables of each order according to the designated positions of the indexes is shown in fig. 3, and the process of reading is reverse to that in the figure.

S102, pre-equalizing the training sequence to obtain a tap coefficient omega_τ1.

And S103, calculating a pre-compensation tap coefficient by using the pre-equalized data through the tap coefficient updating submodule 601A. For PAM-4 signals, each symbol x in the sequence belongs to { + -1, + -3 }, a time window with the length of 2M +1 is established by taking the current time k as the center, and the sequence in the window is the pattern x (k-M: k: k + M). And calculating a pre-compensation tap coefficient according to the time delay combination and the symbol combination of the mode under the time delay combination, and then shifting the current time k to the right, wherein the current time k is k + 1.

S104: the tap coefficient updating submodule 601A is used for calculating the accumulated value of the pre-compensation tap coefficients and the number of the accumulated tap coefficients, and storing the accumulated value and the number of the accumulated tap coefficients in a corresponding lookup table through a lookup table storage module. With the right shift of the current time, the contents of the two lookup tables are continuously updated.

S105, judging whether the time window center is outside the training sequence, if so, finishing the calculation of the two variables, and executing the step S106; if not, go to step S103.

S106: the tap coefficient average value calculation submodule 601B divides the accumulated value of the pre-compensation tap coefficient by the number of accumulated tap coefficients at the corresponding position to obtain the accumulated average value of the pre-compensation tap coefficient of each filter order, and stores the accumulated average value in the corresponding lookup table through the lookup table storage module 602. The updating process of the elements in the lookup table is shown in fig. 4, where the solid black line represents storage and the dashed black line represents reading.

S107: common tap processing, for taps existing in the pre-compensation, directly multiplying the pre-equalized tap coefficients by the pre-equalized tap coefficients cascaded with the pre-equalized tap coefficients through the tap coefficient combination submodule 601C, and removing the pre-equalized tap through the table look-up storage module 602, so as to realize sharing of a first-order filter, improve the utilization rate of hardware resources of the transmitting terminal adopting the combined pre-equalization and pre-compensation method, and the process is shown in fig. 5.

S108: and residual tap judgment, namely judging whether N is greater than M, and judging whether residual balanced tap processing is required. If yes, go to step S109; if not, go to step S110.

S109: and (4) residual equalization tap processing, reserving the excessive part of taps, and not processing tap coefficients.

S110: and after the establishment of the tap coefficient lookup table of each order of filter matched with the current channel is completed, signal transmission can be carried out. During transmission, a sequence in a time window is selected as a sending mode in a signal frame according to the current moment, and a corresponding filter tap coefficient is found from each order lookup table by using a time delay combination and a symbol combination corresponding to the time delay combination in the mode as indexes through a filter configuration module 603. And the filter is configured according to the found tap coefficient, so that the compensation of the current signal can be completed.