WO2017097077A1

WO2017097077A1 - Data processing method and apparatus

Info

Publication number: WO2017097077A1
Application number: PCT/CN2016/104752
Authority: WO
Inventors: 辛雨
Original assignee: 中兴通讯股份有限公司
Priority date: 2015-12-11
Filing date: 2016-11-04
Publication date: 2017-06-15
Also published as: CN106878219A

Abstract

The present invention provides a data processing method and apparatus. The method comprises: receiving a data signal, and then linearly processing the data signal through a specified processing algorithm; and performing fast Fourier transformation on the processed data signal. The present invention can transform a large matrix into a plurality of low-dimensional matrices for operation, thereby reducing the operation processing complexity of a receiving end.

Description

Method and device for data processing

Technical field

The present invention relates to the field of communications, and in particular, to a method and apparatus for data processing at a receiving end.

Background technique

Long Term Evolution (LTE) is a wireless cellular communication technology of 4G (4th Generation Mobile Communication System). LTE adopts OFDM (Orthogonal Frequency Division Multiplexing) technology, and time-frequency resources composed of subcarriers and OFDM symbols form a radio physical time-frequency resource of the LTE system. At present, OFDM technology has been widely used in wireless communication. Due to the Cyclic Prefix (CP), the CP-OFDM system can solve the multipath delay problem well and divide the frequency selective channel into a set of parallel flat channels, which simplifies the channel. Estimate the method and have a higher channel estimation accuracy. However, the performance of the CP-OFDM system is sensitive to the frequency offset and time offset between adjacent sub-bands. This is mainly because the spectrum leakage of the system is relatively large, so it is easy to cause inter-subband interference. Moreover, the CP also takes up time resources and reduces spectrum efficiency.

At present, major companies are beginning to study the 5G (Fifth Generation Mobile Communication System) technology, and GFDM (Generalized Frequency Division Multiplexing) is likely to be adopted in 5G. The GFDM system performs coding and modulation processing at the transmitting end and decoding and demodulation processing at the receiving end in units of one subframe (or data block). Generally, one subframe includes a plurality of subcarriers and a plurality of symbols. Assuming that the number of subcarriers is K and the number of symbols is L, the data of one subframe is N=K×L. Since the sub-carriers and inter-symbols of the GFDM system are non-orthogonal, the data in one sub-frame interacts and interferes with each other. At the receiving end, the received data signal of one subframe is the data signal after the interaction of the N data, and the data signals need to be interfered to separate the N data. How to perform better interference processing at the receiving end is still a problem that needs to be solved.

Recently, some literatures have proposed some processing methods at the receiving end of the GFDM system. One method is to calculate the baseband processing matrix equivalent to the transmitting end, and then use the ZF (Zero Forcing) algorithm or the MMSE (Minimum Mean Squared Error). The mean square error estimation algorithm inverts the matrix to separate the N data. Since the dimension of the matrix is relatively high and increases with the increase of one sub-frame data, the calculation amount of the receiving end is relatively large and the complexity is relatively high. . Another method is to separate the N data by serial interference cancellation, which is more complicated. Therefore, it is an important problem that the current technology needs to solve in the receiving end of the GFDM system to propose a processing method with low complexity and good performance.

In other new multi-carrier systems, a processing method with low complexity and good performance is also needed. Therefore, we hope to propose a good processing method at the receiving end, which is as suitable as possible in a variety of systems based on time-frequency physical resources.

Summary of the invention

The technical problem to be solved by the present invention is to provide a data processing method and apparatus for overcoming the defects of large operation volume and high complexity existing in the existing data processing technology of the receiving end in the multi-carrier system.

In order to solve the above technical problem, the present invention provides a data processing method, which is applied to a multi-carrier system, and the method includes:

After receiving the data signal, linearly processing the data signal by specifying a processing algorithm; and

A fast Fourier transform operation is performed on the processed data signal.

The method further has the following features: after the receiving the data signal, the method further includes:

The data signal is digital-to-analog converted into a discrete data signal, and the discrete data signal is linearly processed by a specified processing algorithm.

The above method also has the following feature: the linear processing of the discrete data signal by a specified processing algorithm, including:

Grouping the discrete data signals;

A plurality of matrices are constructed using the specified processing algorithm, and the discrete data signals of the respective groups are linearly operated, respectively.

The above method also has the following feature: the grouping the discrete data signals includes:

The discrete data signals are sampled and grouped.

The method further has the following feature: after the linear processing of the data signal by the specified processing algorithm, the method further includes:

Interpolation and shift superposition operations are performed on the data signals of each group, and time domain data of a plurality of symbols is output.

The above method further has the following feature: performing a fast Fourier transform operation on the processed data signal, the method comprising:

A fast Fourier transform operation is performed on each time domain data of each symbol, and frequency domain data of a plurality of symbols is output.

The method further has the following feature: after performing the fast Fourier transform operation on the processed data signal, the method further includes:

Detecting frequency domain data after fast Fourier transform operation by specifying a detection algorithm;

The frequency domain data is demodulated.

The above method also has the following features: the specified detection algorithm includes: a minimum mean square error estimation algorithm or a zero forcing algorithm.

The above method also has the following characteristics:

The discrete data signal is a discrete time domain data signal comprising a plurality of symbolized data signals.

The above method also has the following characteristics:

The specified processing algorithm includes a minimum mean square error estimation algorithm or a zero forcing algorithm.

In order to solve the above problems, the present invention also provides an apparatus for data processing, including:

a first processing module, configured to perform linear processing on the data signal by a specified processing algorithm after receiving the data signal;

The second processing module is configured to perform a fast Fourier transform operation on the processed data signal.

The above device also has the following features:

The first processing module, after receiving the data signal, is further configured to: perform digital-to-analog conversion of the data signal into a discrete data signal, and perform linear processing on the discrete data signal by using a specified processing algorithm.

The above device also has the following features:

The first processing module, performing linear processing on the discrete data signal by specifying a processing algorithm includes: grouping the discrete data signals; constructing a plurality of matrices using the specified processing algorithm, respectively, for each group Discrete data signals are linearly operated.

The above device also has the following features:

The first processing module, the grouping the discrete data signals includes: sampling the discrete data signals, wherein the discrete data signals are discrete time domain data signals, and data signals including multiple symbols.

The above device also has the following features:

After the linear processing of the data signal by the processing algorithm, the first processing module is further configured to: perform interpolation and shift superposition operations on the data signals of each group, and output time domain data of the plurality of symbols.

The above device also has the following features:

The second processing module performs a fast Fourier transform operation on the processed data signal, including: performing a fast Fourier transform operation on each time domain data signal of each symbol, and outputting frequency domain data of the plurality of symbols.

The above device also has the following features:

After the fast Fourier transform operation is performed on the processed data signal, the second processing module is further configured to: detect the frequency domain data after the fast Fourier transform operation by using a specified detection algorithm; and perform the frequency domain data on the frequency domain data Performing demodulation, the specified detection algorithm includes: a minimum mean square error estimation algorithm or a zero forcing algorithm.

In summary, the present invention provides a data processing method and apparatus, which can transform a large matrix operation into a plurality of small-dimensional matrices for operation, thereby reducing computational processing complexity at the receiving end; and the method of the present invention enables processing at the receiving end. Can be compatible with other more launch solutions.

DRAWINGS

The drawings are intended to provide a further understanding of the invention, and are intended to be a In the drawing:

1 is a flowchart of a method for data processing according to an embodiment of the present invention;

2 is a schematic diagram of a linear operation of a discrete data signal by a receiving end according to an embodiment of the present invention;

3 is a schematic diagram of a matrix linear operation of a discrete data signal by a receiving end according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of an apparatus for data processing according to an embodiment of the present invention.

Detailed ways

As mentioned in the background, the receiving end of the GFDM system performs a matrix linear operation on the entire sub-frame or data block, and then obtains the frequency domain data of the entire sub-frame. After obtaining the frequency domain data of the entire subframe, the receiving end uses the detection operation of the MMSE or ZF algorithm, and then outputs the data of the transmitting end after passing through the demodulation module. The matrix used in this kind of operation has a very large dimension, so the amount of computation is very large and the complexity is high.

The embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be noted that, in the case of no conflict, the features in the embodiments and the embodiments in the present application may be arbitrarily combined with each other.

FIG. 1 is a flowchart of a method for data processing according to an embodiment of the present invention. As shown in FIG. 1 , the method in this embodiment includes:

S11. After receiving the data signal, linearly processing the data signal by using a specified processing algorithm.

S12. Perform a fast Fourier transform operation on the processed data signal.

The specified processing algorithm may include an MMSE (Minimum Mean Squared Error) algorithm or a ZF (Zero Forcing) algorithm.

Regarding the matrix linear operation of the entire sub-frame or the data block, the embodiment of the present invention proposes a step-by-step method, which is divided into two types of operations with different characteristics, namely, the linear operation of the MMSE algorithm (or the ZF algorithm) and the FFT (Fast Fourier Transform, Fast Fourier Transform) operation. After receiving the data signal, the receiving end performs linear operation of the MMSE algorithm or the ZF algorithm, and then performs an FFT operation.

The benefits of splitting into two different types of features are:

(1) By sampling grouping, multiple small-dimensional matrices can be used to operate, reducing computational complexity;

(2) After dividing into two different types of operations, other types of operations can be added between the two types of operations, which facilitates the use of a new scheme for transmitting data at the transmitting end of the GFDM system (for example, each of the GFDM system data signals) The symbols are not added to the CP. After the scheme of this embodiment, each symbol of the data signal at the transmitting end can increase the CP to improve the anti-multipath channel capability. At the receiving end, it can be proposed in this embodiment. Add a CP-to-CP operation between the two types of operations). That is to say, after dividing into two different types of operations, the processing at the receiving end can be compatible with other more transmission schemes.

(3) Using multiple small-dimension matrices to operate, it is also possible to reduce the transmission of interference errors between data, thereby improving data demodulation performance at the receiving end.

The MMSE algorithm is a minimum mean square error algorithm, and the MMSE is also called LMMSE (Linear Minimum Mean Square Error). The MMSE algorithm in the embodiment of the present invention includes MMSE, MMSE-IRC, EMMSE- IRC algorithm. MMSE-IRC (Linear Minimum Mean Square Error-Interference Rejection Combining) refers to the MMSE algorithm for interference suppression combining. When there is interference between data, some methods are used to obtain interference-related information, which is used in the linear operation of the MMSE algorithm. These interference related information can play a role in suppressing interference. EMMSE-IRC (Enhanced Linear Minimum Mean Square Error-Interference Rejection Combining) is Refers to the enhanced MMSE algorithm for interference suppression, which is that the interference-related information obtained by the better method is more accurate, so that it can better suppress the interference.

The data signal in step S11 is a discrete data signal after the receiver performs digital-to-analog conversion on the received signal.

The linear operation of the MMSE algorithm or the ZF algorithm is to linearly operate the data signal using a matrix, specifically: first grouping the discrete data signals, and then performing linear operations using the plurality of matrices respectively.

The plurality of matrices are constructed by an MMSE algorithm or a ZF algorithm.

In a preferred embodiment, the discrete data signals are grouped, specifically: sampling and grouping the discrete data signals.

In a preferred embodiment, the discrete data signal is a discrete time domain data signal containing data for a plurality of symbols.

In a preferred embodiment, the data signal is outputted by the MMSE algorithm or the linear operation of the ZF algorithm, and the time domain data of the plurality of symbols is output.

In a preferred embodiment, after the data signal passes the linear operation of the MMSE algorithm or the ZF algorithm, the data of each group is further subjected to interpolation and shift superposition operations, and time domain data of a plurality of symbols is output.

In a preferred embodiment, the FFT operation is to perform an FFT operation on the time domain data of each symbol to output frequency domain data of a plurality of symbols.

In a preferred embodiment, the frequency domain data after the FFT operation is subjected to a detection operation by a specified detection algorithm (for example, an MMSE algorithm or a ZF algorithm), and then demodulated to output data of the transmitting end.

The following is a specific example.

Embodiment 1

It is assumed that one subframe (or one data block) of the GFDM system contains K subcarriers and L symbols, and the data series transmitted by the transmitting end in one subframe is d, d is a matrix of N rows and 1 column or is called a vector (including There are vectors of N elements), and there is K multiplied by L equal to N, that is, K × L = N. After the data series is processed by the baseband of the GFDM system, it becomes A*d, "*" is a matrix multiplication operation, A is the equivalent matrix of the baseband processing of the GFDM system of M rows and N columns, and both M and N are positive integers (if baseband When processing, there is sampling, then M>N; if there is no sampling, then M=N. In this example, it is assumed that there is no sampling, so M=N=K×L). It is assumed that the wireless channel is an AWGN (Additive White Gaussian Noise) channel. After the transmitted data passes through the AWGN channel, the data series r received by the receiving end is:

r=A*d+n

Where r is a matrix of M rows and 1 column, which is a data signal received by the receiving end, and the data signal is a discrete data signal after the receiving end performs digital-to-analog conversion on the received signal. n is a matrix of M rows and 1 column, which is a noise vector.

The data series r(i) are first sampled and grouped, and the sampling multiple is S=M/L. The method of sampling by stepwise displacement is shown in Fig. 2. In Figure 2, the data series r(i) is first sampled using S times to obtain the first 1 set of data series E1(j), then shift the data series r(i) by one bit, then use S times to sample, obtain the second set of data series E2(j), and then shift the data series of one bit. (i) After shifting one more bit, use S times to sample, obtain the third group data series E3(j), and so on, and obtain the total S group data series: E1(j), E2(j),... , ES (j).

From the equivalent matrix A of the baseband processing of the GFDM system, the S small matrices a1, a2, ..., aS are decomposed. The dimensions of the S matrices are L, and the S matrices obtained by the MMSE algorithm are respectively:

MMSE-a1=a1 ^H *(a1*a1 ^H +(1/SNR)*I) ^-1 ,

MMSE-a2=a2 ^H *(a2*a2 ^H +(1/SNR)*I) ^-1 ,

...

MMSE-aS=aS ^H *(aS*aS ^H +(1/SNR)*I) ^-1 .

Among them, MMSE-a1 represents the linear operation matrix obtained by using MMSE algorithm of a1. Similarly, MMSE-a2 and MMSE-aS respectively represent the linear operation matrix obtained by using MMSE algorithm with a2 and aS. The superscript H indicates a conjugate transpose operation on the matrix. The SNR represents the signal to noise ratio of the data signal received by the receiving end. I denotes an identity matrix. The superscript -1 indicates the inversion of the matrix.

Then, the obtained S group data series F1(j), F2(j), ..., FS(j) are respectively interpolated using the S rate, and then shifted and added, and finally the data series O(i) is output.

The data series O(i) has a length of M, which is time domain data including L symbols in sequence, and then sequentially operates on the time domain data segments of each symbol using an FFT algorithm, and finally obtains frequency domain data of each symbol.

The frequency domain data after the FFT operation is subjected to frequency domain equalization, and then subjected to a detection operation by the MMSE (or ZF) algorithm, and then demodulated to output data of the transmitting end. Usually the frequency domain equalization is combined with the detection operation of the MMSE (or ZF) algorithm.

Embodiment 2

It is assumed that one subframe (or one data block) of the GFDM system contains K subcarriers and L symbols, and the data series transmitted by the transmitting end in one subframe is d, d is a matrix of N rows and 1 column or is called a vector (including There are vectors of N elements), and there is K multiplied by L equal to N, that is, K × L = N. After the data series is processed by the baseband of the GFDM system, it becomes A*d, "*" is a matrix multiplication operation, A is the equivalent matrix of the baseband processing of the GFDM system of M rows and N columns, and both M and N are positive integers (if baseband When processing, there is sampling, then M>N; if there is no sampling, then M=N. In this example, it is assumed that there is no sampling, so M=N=K×L). It is assumed that the wireless channel is an AWGN channel. After the transmitted data passes through the AWGN channel, the data series r received by the receiving end is:

r=A*d+n

Where r is a matrix of M rows and 1 column, which is a data signal received by the receiving end, and the data signal is a discrete data signal after the receiving end performs digital-to-analog conversion on the received signal. n is a matrix of M rows and 1 column, which is noise Sound vector.

The data series r(i) are first sampled and grouped, and the sampling multiple is S=M/L. The method of sampling by stepwise displacement is shown in Fig. 2. In Fig. 2, the data series r(i) is first sampled using S times to obtain the first group data series E1(j), and then the data series r(i) is shifted by one bit, and then S times is used for sampling. The second group data series E2(j), then shift one bit of the data series r(i) one bit later, then use S times to sample, obtain the third group data series E3(j), and so on. , a total of S group data series can be obtained: E1 (j), E2 (j), ..., ES (j).

From the equivalent matrix A of the baseband processing of the GFDM system, the S small matrices a1, a2, ..., aS are decomposed. The dimensions of the S matrices are L, and the S matrices obtained by the ZF algorithm are respectively:

ZF-a1=a1 ^-1 ,

ZF-a2=a2 ^-1 ,

...

ZF-aS=aS ^-1 ,

or

ZF-a1=(a1 ^H *a1) ^-1 *a1 ^H ,

ZF-a2=(a2 ^H *a2) ^-1 *a2 ^H ,

...

ZF-aS=(aS ^H *aS) ^-1 *aS ^H ,

Then, the S group data series: E1(j), E2(j), ..., ES(j) are linearly operated using the matrices MMSE-a1, MMSE-a2, ..., MMSE-aS, respectively, to obtain the S group data series F1 ( j), F2(j), ..., FS(j), ie:

F1(j)=MMSE-a1*E1(j),

F2(j)=MMSE-a2*E2(j),

...

FS(j)=MMSE-aS*ES(j).

The frequency domain data after the FFT operation is subjected to frequency domain equalization, and then subjected to a detection operation by the MMSE (or ZF) algorithm, and then demodulated to output data of the transmitting end.

Embodiment 3

The difference between this embodiment and the first embodiment is that the operations of shifting, sampling grouping, interpolation, and shift addition in FIG. 2 are equivalent to using some operations of the matrix in a specific implementation, as shown in FIG. 3.

For the data series r(i), the columns of the K(K=M/L) data composition matrix are sequentially intercepted, so that the matrix R(K, L) of the K rows and L columns is constructed. This operation is similar to the operation of the reshape function in Matlab. Each row of data of this matrix R(K, L) is equivalent to each set of data series sampled in Fig. 2, for example, the first row is the data series E1(i). The number of rows K of the matrix is equal to the number of groups S (S = K = M / L). Then, each row of the matrix R(K, L) is sequentially linearly operated using the MMSE-a1, MMSE-a2, ..., MMSE-aS matrix to obtain each new row of the matrix, thus obtaining a new matrix R' (K) , L), and then each column of the matrix R' (K, L) data is sequentially connected into the data series O (i), this operation is similar to the operation of the reshape function in Matlab. The sequential arrangement of each column of data of the matrix R'(K, L) into the data series O(i) corresponds to the interpolation and shift addition operations of FIG. The other is the same as the implementation one.

The receiving end processing scheme of this embodiment can be used in other multi-carrier systems in addition to the GFDM system.

The receiving end processing method of the present invention can transform the operation of one large matrix into a plurality of matrixes of small dimensions for operation, thereby reducing the processing complexity of the receiving end; and the method of the present invention makes the processing of the receiving end compatible with other more transmitting schemes. .

FIG. 4 is a schematic diagram of a device for data processing according to an embodiment of the present invention. As shown in FIG. 4, the device in this embodiment includes:

In a preferred embodiment, the first processing module, after receiving the data signal, is further configured to: digitally convert the data signal into a discrete data signal, and then linearize the discrete data signal by using a specified processing algorithm. deal with.

In a preferred embodiment, the first processing module linearly processes the discrete data signal by specifying a processing algorithm, including: grouping the discrete data signals; constructing a plurality of matrices using the specified processing algorithm, The discrete data signals of the respective groups are linearly operated, respectively.

In a preferred embodiment, the first processing module, the grouping the discrete data signals comprises: sampling the discrete data signals, where the discrete data signals are discrete time domain data signals, including multiple Symbolic data signal.

In a preferred embodiment, after the linear processing of the data signal by the specified processing algorithm, the first processing module is further configured to: perform interpolation and shift superposition operations on each group of data signals, and output multiple symbols. Time domain data.

In a preferred embodiment, the second processing module performs fast Fourier transform on the processed data signal. The changing operation includes: performing a fast Fourier transform operation on each time domain data signal of each symbol, and outputting frequency domain data of the plurality of symbols.

In a preferred embodiment, after the fast Fourier transform operation is performed on the processed data signal, the second processing module is further configured to: detect the frequency domain data after the fast Fourier transform operation by using a specified detection algorithm. Demodulating the frequency domain data, the specified detection algorithm comprising: a minimum mean square error estimation algorithm or a zero forcing algorithm.

One of ordinary skill in the art will appreciate that all or a portion of the steps described above can be accomplished by a program that instructs the associated hardware, such as a read-only memory, a magnetic or optical disk, and the like. Alternatively, all or part of the steps of the above embodiments may also be implemented using one or more integrated circuits. Correspondingly, each module/unit in the foregoing embodiment may be implemented in the form of hardware or in the form of a software function module. The invention is not limited to any specific form of combination of hardware and software.

The above is only a preferred embodiment of the present invention, and of course, the present invention may be embodied in various other embodiments without departing from the spirit and scope of the invention. Corresponding changes and modifications are intended to be included within the scope of the appended claims.

Industrial applicability

The invention is applicable to the field of communication, and is used to realize the operation of transforming the operation of one large matrix into a plurality of small-dimension matrices, thereby reducing the processing complexity of the receiving end, and the method of the invention makes the processing of the receiving end compatible with other more. Launch plan.

Claims

A method of data processing, applied to a multi-carrier system, the method comprising:

After receiving the data signal, linearly processing the data signal by specifying a processing algorithm; and

A fast Fourier transform operation is performed on the processed data signal.
The method of claim 1, wherein after the receiving the data signal, the method further comprises:

The data signal is digital-to-analog converted into a discrete data signal, and the discrete data signal is linearly processed by a specified processing algorithm.
The method of claim 2 wherein said linearly processing said discrete data signal by a specified processing algorithm comprises:

Grouping the discrete data signals;

A plurality of matrices are constructed using the specified processing algorithm, and the discrete data signals of the respective groups are linearly operated, respectively.
The method of claim 3 wherein said grouping said discrete data signals comprises:

The discrete data signals are sampled and grouped.
The method of claim 3 or 4, wherein after the linear processing of the data signal by a specified processing algorithm, the method further comprises:

Interpolation and shift superposition operations are performed on the data signals of each group, and time domain data of a plurality of symbols is output.
The method of claim 5 wherein said performing a fast Fourier transform operation on said processed data signal comprises:

A fast Fourier transform operation is performed on each time domain data of each symbol, and frequency domain data of a plurality of symbols is output.
The method of claim 6 wherein after the performing a fast Fourier transform operation on the processed data signal, the method further comprises:

Detecting frequency domain data after fast Fourier transform operation by specifying a detection algorithm;

The frequency domain data is demodulated.
The method of claim 7 wherein

The specified detection algorithm includes: a minimum mean square error estimation algorithm or a zero forcing algorithm.
The method of claim 2, wherein

The discrete data signal is a discrete time domain data signal comprising a plurality of symbolized data signals.
The method of any of claims 1-4, 6-9, wherein:

The specified processing algorithm includes a minimum mean square error estimation algorithm or a zero forcing algorithm.
A device for data processing, comprising:

a first processing module configured to linearly process the data signal by specifying a processing algorithm after receiving the data signal; and

The second processing module is configured to perform a fast Fourier transform operation on the processed data signal.
The device of claim 11 wherein

The first processing module, after receiving the data signal, is further configured to: perform digital-to-analog conversion of the data signal into a discrete data signal, and perform linear processing on the discrete data signal by using a specified processing algorithm.
The device of claim 12, wherein

The first processing module, performing linear processing on the discrete data signal by specifying a processing algorithm includes: grouping the discrete data signals; constructing a plurality of matrices using the specified processing algorithm, respectively, for each group Discrete data signals are linearly operated.
The device of claim 13 wherein

The first processing module, the grouping the discrete data signals includes: sampling the discrete data signals, wherein the discrete data signals are discrete time domain data signals, and data signals including multiple symbols.
The apparatus according to claim 13 or 14, wherein

After the linear processing of the data signal by the processing algorithm, the first processing module is further configured to: perform interpolation and shift superposition operations on the data signals of each group, and output time domain data of the plurality of symbols.
The device of claim 15 wherein

The second processing module performs a fast Fourier transform operation on the processed data signal, including: performing a fast Fourier transform operation on each time domain data signal of each symbol, and outputting frequency domain data of the plurality of symbols.
The device of claim 16 wherein

After the fast Fourier transform operation is performed on the processed data signal, the second processing module is further configured to: detect the frequency domain data after the fast Fourier transform operation by using a specified detection algorithm; and perform the frequency domain data on the frequency domain data Performing demodulation, the specified detection algorithm includes: a minimum mean square error estimation algorithm or a zero forcing algorithm.