阅读说明：本技术 并行变步长cma均衡算法、装置、电子设备及存储介质 (Parallel variable-step-size CMA (China Mobile alliance) equalization algorithm, device, electronic equipment and storage medium ) 是由 李一航 王锐 李倩 王晓东 王忠丽 于 2020-06-16 设计创作，主要内容包括：本申请公开了一种并行变步长CMA均衡算法、装置、电子设备及存储介质。包括：各个支路将输入信号与滤波器抽头矩阵的转置相乘得出当前时刻的输出值；根据所述输出值,得到每个支路当前时刻的误差值；根据所述误差值,计算汇总结果；通过汇总结果r,计算步长μ；各个支路通过所述步长、所述误差值以及输入信号计算用于下一时刻均衡的滤波器抽头,同时读入下一组输入信号。本申请方法在变步长计算中只设计四则运算,算法结构简单,占用资源较少,便于实现并行结构。针对高速数字通信中常用的并行处理结构进行了专门的设计,更方便用于高速无线通信当中；步长的改变与时间无关,可以在时变信道中正常工作。(The application discloses a parallel variable-step CMA (constant amplitude adaptive equalizer) equalization algorithm, a device, electronic equipment and a storage medium. The method comprises the following steps: each branch circuit multiplies the input signal by the transpose of the filter tap matrix to obtain the output value at the current moment; obtaining the error value of each branch at the current moment according to the output value; calculating a summary result according to the error value; calculating the step length mu through the summary result r; each branch calculates a filter tap for equalization at the next time through the step size, the error value and the input signal, and simultaneously reads in the next group of input signals. The method only designs four arithmetic operations in the variable step length calculation, has simple algorithm structure and less occupied resources, and is convenient for realizing a parallel structure. The special design is carried out aiming at the common parallel processing structure in the high-speed digital communication, so that the method is more convenient to be used in the high-speed wireless communication; the step size changes are time-independent and can work properly in time-varying channels.)

1. A parallel variable-step CMA equalization algorithm is characterized by comprising the following steps:

step S1, each branch multiplies the input signal by the transpose of the filter tap matrix to obtain the output value of the current moment;

step S2, obtaining an error value of each branch circuit at the current moment according to the output value;

step S3, calculating a summary result according to the error value;

step S4, calculating step length through the summary result;

step S5, each branch calculates a filter tap for equalization at the next time through the step size, the error value and the input signal, and reads in the next set of input signals at the same time, and then goes to step S1.

2. The parallel variable-step CMA equalization algorithm of claim 1, wherein the step S2 comprises:

s2.1, calculating a modulus of the output value, and subtracting the squared modulus from the square of the given modulus stored in the register to obtain a modulus difference;

and S2.2, multiplying the modulus difference by the output value to obtain the error value of each branch at the current moment.

3. The parallel variable-step CMA equalization algorithm of claim 1, wherein the step S3 comprises:

s3.1, calculating the square r of the error value modulus value by each branch_i；

Step S3.2, passing_iAnd carrying out summary calculation to obtain a summary result r.

4. The parallel variable-step CMA equalization algorithm of claim 3 wherein the summary result:

R＝[r₁，r₂，...，r_m]^T，S＝[s₁，s₂，...，s_m]

wherein j is the number of 0 in S, M is the number of branches, M is the number of error values, R is an error module value square matrix calculated by each branch, and S is a coefficient matrix.

5. The parallel variable-step CMA equalization algorithm of claim 1, wherein the compute step method:

wherein, mu₀For the step factor, a is a real number greater than 1 for limiting the maximum value of μ, and b is a real number slightly greater than 0 for limiting the minimum value of μ and preventing an error in the r-0 calculation.

6. The parallel variable-step-size CMA equalization algorithm of claim 3, wherein S takes a value of:

when a certain value R in the matrix R_iAnd the average value

7. A parallel variable-step CMA equalization device is realized by adopting a parallel variable-step CMA equalization algorithm of any one of claims 1 to 5, and is characterized by comprising the following steps: the device comprises an output value calculation module, an error value calculation module, a summary module and a step length calculation module;

the output value calculation module, the error value calculation module, the summary module and the step length calculation module are sequentially linked; the output value calculation module obtains an output value and inputs the output value to the error calculation module, the step length calculation module obtains a step length calculation result, and the step length calculation result is fed back to the output value calculation module;

the output value calculating module is used for calculating an output value, and each branch multiplies the input signal by the transpose of a filter tap matrix to obtain an output value at the current moment;

the error value calculation module is used for calculating an error value and obtaining the error value of each branch at the current moment according to the output value;

the summarizing module is used for calculating a summarizing value and calculating a summarizing result according to the error value;

and the step length calculating module is used for calculating the step length and calculating the step length through the summary result.

8. The parallel variable-step CMA equalization apparatus of claim 6 wherein the output value calculation module is implemented using an FIR filter.

9. An electronic device, comprising a memory for storing a computer program and a processor for executing the computer program to cause the computer device to perform a parallel variable-step CMA equalization algorithm according to any of claims 1-5.

10. A readable storage medium having stored thereon computer program instructions which, when read and executed by a processor, perform a parallel variable-step CMA equalization algorithm as claimed in any one of claims 1-5.

Technical Field

The present application relates to the field of channel equalization and wireless communication technologies, and in particular, to a parallel variable-step CMA equalization algorithm, apparatus, electronic device, and storage medium.

Background

At present, in high-speed wireless communication, the communication quality is seriously affected by intersymbol interference caused by multipath effects, deviation generated by timing sampling and in-band amplitude inconsistency of a radio frequency device and an ADC (analog to digital converter). The effect of intersymbol interference becomes increasingly severe as the symbol rate increases. To solve this problem, an equalization module needs to be added.

High-speed signals are often processed in parallel in an FPGA, and the equalization algorithm is discussed based on processing of the high-speed parallel signals. The system mainly comprises an equalizing filter of each branch, a branch error calculation module and a summary calculation module.

The traditional CMA equalization algorithm is a fixed step size and therefore slower in convergence speed than the variable step size equalization algorithm.

The variable-step CMA equalization principle is that an error factor e is calculated by the square difference between the modulus of the received signal and a prescribed modulus. And updating the step size mu of the algorithm according to e. And calculating the step length calculated by each branch together to obtain a uniform step length. And updates the tap value w of the filter in accordance therewith.

At present, relevant patents are mainly applied to research of a variable-step-size CMA (China Union, China and telecommunications) balance algorithm by Nanjing university, Wuhan-Pod research institute limited company, Nanjing information technology university and southwest technology university, but the relevant patents often have the defects of complex calculation, more resource occupation and the like, and meanwhile, the problem of how to adjust the algorithm in a parallel state is not considered, so that the method is difficult to be used in high-speed wireless digital communication.

Wherein, patent one is a patent application proposed by Nanjing university, and the patent application number is as follows: CN201710082645.8, entitled "a satellite channel multimode blind equalization algorithm", proposes an equalization algorithm that realizes switching between two equalization modes by error decision. The method relates to a part of a variable-step CMA (constant-amplitude adaptive equalization) algorithm, which relates to a large amount of natural exponent operation and power operation, is difficult to realize on an FPGA (field programmable gate array), needs a large amount of operation resources, greatly increases the cost under the condition of high-speed parallel, and is difficult to realize.

Patent two is the patent application that wuhan post academy of science research ltd provided, patent application number: CN201811362088.6, entitled "CMA-based signal equalization method and system", proposes a CMA equalization algorithm for optical communication based on bit error rate determination and adjustment step size factor. The method relies on error code judgment in step size adjustment, which complicates the system structure in wireless communication and has large hysteresis in step size adjustment.

Patent three is the patent application that Nanjing information science and technology university proposed, patent application number: CN200910028459.1, entitled "dual-mode variable-step blind equalization method based on orthogonal wavelet packet transformation", proposes a dual-mode variable-step equalization algorithm based on orthogonal wavelet packet transformation for underwater acoustic communication. However, the implementation of the orthogonal wavelet transform on the FPGA occupies a large amount of resources, and the problem is aggravated in a parallel state, so that the implementation cost of the algorithm in high-speed communication is too high.

Patent four is the patent application that southwest science and technology university provided, patent application number: CN201810578167.4 entitled "CMA blind equalization variable step length optimization method for MPSK signal", proposes a variable step length CMA equalization algorithm based on inverse proportion function. The operation is simpler than other algorithms, but the step size approaches 0 indefinitely over time, which makes the algorithm unable to work in time-varying channels.

Aiming at the problems of complex operation, no consideration of a parallel mode and incapability of coping with a time-varying channel in the related art, an effective solution is not provided at present.

Disclosure of Invention

The present application mainly aims to provide a parallel variable-step CMA equalization algorithm, an apparatus, an electronic device, and a storage medium, which occupy less system resources, are simple to implement, and work exclusively in a variable-step CMA equalization algorithm in a parallel mode, so as to solve the problems of complex operation, no consideration of the parallel mode, and incapability of coping with a time-varying channel in the related art.

In order to achieve the above object, the present application provides a parallel variable step size CMA equalization algorithm, which includes the following steps:

step S1, each branch multiplies the input signal by the transpose of the filter tap matrix to obtain the output value at the current time.

This step may be implemented by a FIR filter.

Step S2: and obtaining the error value of each branch circuit at the current moment according to the output value.

The method comprises the following specific steps:

and S2.1, calculating the modulus of the output value, and subtracting the squared modulus from the square of the given modulus stored in the register to obtain a modulus difference.

And S2.2, multiplying the modulus difference by the output value to obtain the error value of each branch at the current moment.

Steps S1 through S2 are consistent with the conventional CMA equalization algorithm, and the difference between the two algorithms is that the conventional CMA equalization algorithm does not need the following steps, and proceeds directly to step S7.

Step S3, calculating a summary result according to the error value;

the method comprises the following specific steps:

s3.1, calculating the square r of the error value modulus value by each branch_i。

Step S3.2, passing_iAnd carrying out summary calculation to obtain a summary result r.

Wherein j is the number of 0 in S, M is the number of branches, M is the number of error values, R is an error module value square matrix calculated by each branch, and S is a coefficient matrix.

Step S4, calculating the step size mu through the summary result r, wherein the expression is as follows:

wherein, mu₀A is a real number greater than 1 for limiting the maximum value of μ, and b is a real number slightly greater than 0 for limiting the maximum value of μSmall value and prevent errors in the r-0 calculation.

The calculation method only involves four arithmetic operations, and is easier to realize on a hardware platform compared with other variable step size CMA algorithms.

Step S5, each branch calculates the filter tap for the next equalization through the step size, the error value and the input signal, and reads in the next set of input signals at the same time, and goes to step S1.

To this end, from step S1 to step S5, a parallel variable step CMA equalization algorithm is completed.

The value of S is when_iTaking 1/M, where r is r_iAverage value of (a). However, in a severe channel environment, the influence of noise on the system may make R extremely susceptible to some maximum or minimum values deviating from the normal range. In the parallel computing mode of the algorithm, the average values can be deviated if the influence of the average values is not removed. S can thus be modified to:

when a certain value R in the matrix R_iAnd the average valueThe mean ratio is greater than or less than the threshold r_up，r_downIts corresponding coefficient in the coefficient matrix is modified to.

Due to the CMA equalization

Substantially equal at two adjacent times, so that the square r' of the error modulus at the previous time is substituted

To reduce the computation time.

A parallel variable-step CMA equalization device is realized by adopting a parallel variable-step CMA equalization algorithm, and comprises the following steps: the device comprises an output value calculation module, an error value calculation module, a summary module and a step length calculation module;

the output value calculation module, the error value calculation module, the summary module and the step length calculation module are sequentially linked; the output value calculation module obtains an output value and inputs the output value to the error calculation module, the step length calculation module obtains a step length calculation result, and the step length calculation result is fed back to the output value calculation module;

the output value calculating module is used for calculating an output value, and each branch multiplies the input signal by the transpose of a filter tap matrix to obtain an output value at the current moment;

the error value calculation module is used for calculating an error value and obtaining the error value of each branch at the current moment according to the output value;

the summarizing module is used for calculating a summarizing value and calculating a summarizing result according to the error value;

and the step length calculating module is used for calculating the step length and calculating the step length through the summary result.

An electronic device comprising a memory for storing a computer program and a processor for executing the computer program to cause the computer device to execute a parallel variable-step CMA equalization algorithm according to said one.

A readable storage medium having stored thereon computer program instructions which, when read and executed by a processor, perform a parallel variable-step CMA equalization algorithm.

The beneficial technical effects are as follows:

compared with other variable-step CMA equalization algorithms, the parallel variable-step CMA equalization algorithm, the device, the electronic equipment and the storage medium provided by the invention have the following beneficial effects:

1. the algorithm only designs four arithmetic operations in the variable step length calculation, and the algorithm has a simple structure.

2. Less occupied resources and convenient realization of a parallel structure.

3. The parallel processing structure is specially designed for the common parallel processing structure in high-speed digital communication, and is more convenient to be used in high-speed wireless communication.

4. The step size changes are time-independent and can work properly in time-varying channels.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this application, serve to provide a further understanding of the application and to enable other features, objects, and advantages of the application to be more apparent. The drawings and their description illustrate the embodiments of the invention and do not limit it. In the drawings:

FIG. 1 is a flow chart of a parallel variable step CMA equalization algorithm provided according to an embodiment of the present application;

fig. 2 is a structural diagram of a parallel variable-step CMA equalization algorithm provided in an embodiment of the present application;

FIG. 3 is a graph showing the variation of the step size μ with the square r of the error modulus provided by an embodiment of the present application;

FIG. 4 is a diagram of a change in a main tap value of a filter using a fixed-step algorithm and an equalization module of the present algorithm according to an embodiment of the present application;

fig. 5 is a constellation diagram of output signals in four cases provided according to the embodiment of the present application, (a) is an output signal without equalization, (b) is an output signal subjected to fixed-step CMA equalization, (c) is an output signal subjected to variable-step CMA equalization for calculating step sizes respectively, and (d) is an output signal of the algorithm of the present application. .

Detailed Description

In order to make the technical solutions better understood by those skilled in the art, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only partial embodiments of the present application, but not all embodiments. 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 application.

It should be noted that the terms "first," "second," and the like in the description and claims of this application and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It should be understood that the data so used may be interchanged under appropriate circumstances such that embodiments of the application described herein may be used. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

In this application, the terms "upper", "lower", "left", "right", "front", "rear", "top", "bottom", "inner", "outer", "middle", "vertical", "horizontal", "lateral", "longitudinal", and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings. These terms are used primarily to better describe the present application and its embodiments, and are not used to limit the indicated devices, elements or components to a particular orientation or to be constructed and operated in a particular orientation.

Moreover, some of the above terms may be used to indicate other meanings besides the orientation or positional relationship, for example, the term "on" may also be used to indicate some kind of attachment or connection relationship in some cases. The specific meaning of these terms in this application will be understood by those of ordinary skill in the art as appropriate.

In addition, the term "plurality" shall mean two as well as more than two.

It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.

A parallel variable-step CMA equalization algorithm, an apparatus, an electronic device, and a storage medium are described in detail below with reference to fig. 1, where the parallel variable-step CMA equalization algorithm includes the following steps:

step S1, each branch converts the input signal X to [ X ═ X₁，x₂，...，x_m]^TAnd filter tap W ═ W₁，w₂，...，w_m]^TThe output value y of the current moment is obtained by transposition multiplication_i. This step may be implemented by a FIR filter.

In the process of processing high-speed digital signals, input signals enter an equalization module in a parallel mode, each branch comprises an FIR filter, and signals in the filter at the current moment can be represented as X ═ X₁，x₂，...，x_m]^TThe filter tap value is expressed as W ═ W₁，w₂，...，w_m]^TAs shown in fig. 2, the output value calculation module in fig. 2 is an FIR filter.

The two are transposed and multiplied to obtain the output value y of the current time_i。

Step S2, calculating y_iModulus value of_iI, squared with the square d of a given modulus value stored in a register²And subtracting to obtain the module value difference between the two.

Step S3, sum the modulus difference with y_iMultiplying to obtain the error value e of each branch circuit at the current moment_i。

e_i＝y_i*(d²-|y_i|²)

Steps S1 through S3 are consistent with the conventional CMA equalization algorithm, and the difference between the two algorithms is that the conventional CMA equalization algorithm does not need the following steps, and proceeds directly to step S7.

In high-rate wireless communications, the symbol rate tends to be greater than the clock rate of the system, which causes all modules of the system to operate in parallel. In the equalization process, M parallel paths will generate M error values at the same time. If a step size change algorithm is to be introduced in the parallel structure, it is necessary to consider how to calculate the step size using multiple error values.

The simplest method is to use the error value e of each sub-filter_i(n) calculating the corresponding step size mu_i(n)。However, there is a problem that the separate calculation generates non-uniform step sizes, which causes the convergence time of each sub-filter to be different from the steady-state error, thereby affecting the final equalization effect.

The algorithm adopts a method that after error values are calculated for each branch, the error values are processed, and then a unified mu is calculated by using results.

Step S4, branch calculation e_iSquare of modulus r_i＝|e_i|²And the calculation is used for the subsequent steps.

Calculating the modulus squared r of the error for each path_iAnd then the error is sent to a summary calculation module to calculate the square value r of the total error modulus.

Step S5, the summary calculation module passes through r_iAnd calculating r.

Compared with the method for respectively calculating the step length of each path, the method saves calculation resources, and the algorithm for respectively calculating generates non-uniform step lengths, which can cause the difference between the convergence time and the steady-state error of each sub-filter, thereby influencing the final equalization effect.

R of each branch received by the collecting calculation module_iThe method of calculating the square r of the error modulus is as follows.

r＝SR，R＝[r₁，r₂，...，r_m]^T，S＝[s₁，s₂，...，s_m]

In the general case, s_iCan be taken as 1/M, in which case r is r_iAverage value of (a). However, in a severe channel environment, the influence of noise on the system may make R extremely susceptible to some maximum or minimum values deviating from the normal range. In the parallel computing mode of the algorithm, the average values can be deviated if the influence of the average values is not removed. S can thus be modified to:

when a certain value R in the matrix R_iAnd the average value

Is greater than or less than a threshold value r_up，r_downIts corresponding coefficient in the coefficient matrix is modified to. For the sake of calculation, in the present algorithm, 0 is taken. At the same time, the calculation method of r should be modified to:

j is the number of 0S in S

Due to the CMA equalization

Substantially equal at two adjacent times, so the error modulus value squared r' at the previous time is substitutedTo reduce the computation time.

Step S6, calculating a step size μ by r, which is expressed as:

wherein mu₀For the step factor, a is a real number greater than 1 for limiting the maximum value of μ, and b is a real number slightly greater than 0 for limiting the minimum value of μ and preventing an error in the r-0 calculation.

When the equalization filter does not realize equalization, the system error is larger, r is also larger, the value of a in the formula is dominant, and the value of mu is larger at the moment, so that the system convergence speed is accelerated.

When the equalizing filter basically realizes equalization, the system error is smaller, r is smaller, and in the formula

According to the leading position, the value of mu is close to mu₀The system convergence accuracy increases. Figure 3 pictorially shows the transformation of mu with the change of r.

The calculation method only involves four arithmetic operations, and is easier to realize on a hardware platform compared with other variable step size CMA algorithms.

Step S7, each branch passes through step size mu and error e_iAnd calculating the filter tap W' for the next equalization of the input signal X while reading in the next set of input signals, and repeating step S1.

W`＝W+μe_iX

To this end, from step S1 to step S7, a parallel variable step CMA equalization algorithm is completed.

A parallel variable-step CMA equalization apparatus, as shown in fig. 2, implemented by using a parallel variable-step CMA equalization algorithm, comprising: the device comprises an output value calculation module, an error value calculation module, a summary module and a step length calculation module;

the output value calculation module, the error value calculation module, the summary module and the step length calculation module are sequentially linked; the output value calculation module obtains an output value and inputs the output value to the error calculation module, the step length calculation module obtains a step length calculation result, and the step length calculation result is fed back to the output value calculation module;

the output value calculating module is used for calculating an output value, and each branch multiplies the input signal by the transpose of a filter tap matrix to obtain an output value at the current moment;

the error value calculation module is used for calculating an error value and obtaining the error value of each branch at the current moment according to the output value;

the summarizing module is used for calculating a summarizing value and calculating a summarizing result according to the error value;

and the step length calculating module is used for calculating the step length and calculating the step length through the summary result.

An electronic device comprising a memory for storing a computer program and a processor for executing the computer program to cause the computer device to execute a parallel variable-step CMA equalization algorithm according to said one.

A readable storage medium having stored thereon computer program instructions which, when read and executed by a processor, perform a parallel variable-step CMA equalization algorithm.

In order to verify the performance of the algorithm, a related multipath channel environment is set for simulation, and the difference between the multipath channel environment and a fixed-step-size CMA equalization algorithm and a variable-step-size CMA equalization algorithm for respectively calculating step sizes is compared.

The input signal is set to be a QPSK signal with a symbol rate of 2.5 Gsps. The adopted channel is an exponential fading channel. The relevant parameters involved in the variable step algorithm are: mu.s₀＝0.005，a＝20，b＝0.01。

The input signal is equalized by using a fixed-step CMA equalization algorithm and the present algorithm, and the change situation of the tap value of the main tap of the filter is obtained as shown in fig. 4. It can be seen that the convergence rate of the algorithm is significantly faster than that of the fixed-step CMA equalization algorithm.

Fig. 5 shows the output results of various algorithms. a is an output signal without equalization, b is an output signal after constant-step CMA equalization, c is an output signal of variable-step CMA equalization for calculating step length respectively, and d is an output signal of the algorithm of the application. It can be seen that the output results of b and d are substantially consistent after the system converges, and c generates little distortion.

Table 1 is the error rate results for the 4 cases, which numerically demonstrates that the unified calculation step size algorithm is more accurate than the individual calculation step size results.

TABLE 1 error Rate results for the four cases

The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.