CN112039815A - Interference elimination method applied to filter bank multi-carrier system and application thereof - Google Patents

Abstract

The invention discloses an interference elimination method applied to a filter bank multi-carrier system and its application. The steps are as follows: a transmitting end sends a group of N sequences of sub-carrier sequence groups, and performs real part and imaginary part processing respectively on them. Two frequency-domain real-valued sequences are obtained; the two frequency-domain real-valued sequences are respectively subjected to prototype filtering, IFFT, signal subtraction processing (two groups of adjacent signals are subtracted and then subtracted) and signal rearrangement processing (using odd numbers first and then even numbers) Rearrange the sequence in the order), merge the two, and send them to the receiving end; the receiving end divides the received signal sequence group into two branches, and then performs prototype filtering, FFT, signal restoration processing and real part operation respectively, namely must receive the sequence signal. Under the premise of not changing the QAM modulation mode and not reducing the data transmission efficiency, the present invention not only weakens the orthogonality condition between the filters and more effectively eliminates the inherent interference of the imaginary part, but also further suppresses the inter-subcarrier and inter-symbol interference. residual interference.

Description

A kind of interference cancellation method applied to filter bank multi-carrier system and its application

technical field

The invention belongs to the technical field of filter bank multi-carrier modulation, relates to an interference elimination method applied to a filter bank multi-carrier system and its application, in particular to an interference elimination method of FBMC/QAM in a high-speed mobile 5G communication channel environment Design.

Background technique

With the wide application of broadband wireless communication technology and the continuous development of the Internet of Things, the new multi-carrier modulation technology has become an important key technology in 5G mobile communication. Spectrum utilization, the ability to adapt to the needs of 5G communications such as multiple services and emerging applications. Filter Bank Multicarrier (FBMC) technology replaces the traditional Orthogonal Frequency Division Multiplexing (OFDM) technology by introducing a prototype filter with good time frequency localization (TFL) characteristics. The rectangular window function filters the subcarriers, the out-of-band emission (OOBE) performance of the system is greatly improved, and the requirement for orthogonality between subcarriers is also greatly reduced, making it applicable In ultra-high-speed mobile scenarios; at the same time, the system no longer needs to insert cyclic prefix (CP) and guard bands between sub-carriers, which will further improve the spectrum utilization of the system; in addition, the prototype filter The function can also flexibly control the bandwidth and overlapping degree of each sub-carrier, so that the system can flexibly allocate the available time-frequency resources and support asynchronous transmission between adjacent sub-carriers, which can well meet the requirements of the Internet of Things (IoT). ), machine to machine (M2M) and other diverse communication scenarios. Therefore, FBMC has been widely regarded at home and abroad as the technology most likely to become the modulation scheme of the 5G communication physical layer.

However, the traditional FBMC adopts the modulation method of offset quadrature amplitude modulation (OQAM). The introduction of OQAM realizes the full-rate transmission of FBMC data signals, but also makes some key technologies in OFDM cannot be directly used. For FBMC/OQAM; in addition, since FBMC/OQAM only satisfies the real number domain orthogonality between subcarriers, each pilot symbol demodulated will be affected by the inherent interference of surrounding adjacent symbols in the imaginary number domain, which will seriously affect Therefore, it is necessary to try to eliminate or suppress it.

Since the inherent interference of the imaginary part is caused by the introduction of OQAM, some experts and scholars have proposed a system design scheme of FBMC/QAM in recent years [Document 1: Hyungju Nam, Moonchang Choi, and SeongbaeHan.A New Filter-Bank Multicarrier System with Two Prototype Filters for QAMSymbols Transmission and Reception[J].IEEE Transactions on Communications,2016,15(9):5998-6009.], [Document 2: Taehyun Lee,Dongkyu Sim,Bongsung Seo,etal.Channel Estimation Scheme in Oversampled Frequency Domain for FBMC-QAMSystems Based on Prototype Filter Set[J].IEEE Transactions on VehicularTechnology,2019,68(1):728-739.], it firstly separates the original signal from the transmitter by parity, and then uses two Different prototype filters filter the odd-numbered signal and the even-numbered signal respectively. When the filter satisfies a certain orthogonality condition, the inherent interference of the imaginary part can be completely eliminated, and the loss of the data transmission rate will not be caused at the same time. Since FBMC/QAM can further overcome the shortcomings of serious inherent interference of its imaginary part and poor compatibility with OFDM and MIMO on the basis of inheriting a series of advantages of traditional FBMC/OQAM, it has more advantages than FBMC/OQAM in 5G communication. broader application prospects.

However, FBMC/QAM itself is not perfect, and it still has the following problems in interference suppression:

(1) The reason why the FBMC/QAM proposed by Hyungju Nam et al. can completely eliminate the inherent interference of the imaginary part is that it relies on the orthogonality condition satisfied between its two different prototype filters. It is difficult to know: ① The orthogonality conditions that need to be satisfied between the two prototype filters are 4 (ie (27a) to (27d) in Reference 1), which are relatively complicated; ② The four orthogonality conditions are ideal without However, there are often various interference and noise in the actual 5G communication channel, which will undoubtedly have a certain impact on the orthogonality conditions between the filters, resulting in the imaginary part Deterioration of inherent interference rejection;

(2) It is pointed out in Reference 2 that due to the influence of multipath fading in the wireless channel, the orthogonality between the two filters in FBMC/QAM is easily destroyed, which will make the difference between different subcarriers and symbols. Therefore, how to effectively suppress the residual interference has become a crucial issue whether FBMC/QAM can be used in 5G communication applications; at the same time, Reference 2 also attributes the suppression of residual interference to Suppression of inter-carrier interference (ICI) and inter-symbol interference (ISI) in the system.

Therefore, it is of great practical significance to develop an interference suppression method of FBMC/QAM with less residual interference.

SUMMARY OF THE INVENTION

The purpose of the present invention is to overcome the defect that the residual interference of the prior art FBMC/QAM technology is too large and thus affect the system performance, and to provide a FBMC/QAM interference suppression method with small residual interference.

To achieve the above object, the present invention provides the following technical solutions:

An interference elimination method applied to a filter bank multi-carrier system, the steps of which are as follows;

(1) The complex-valued sequence generator at the transmitting end sends a group of sub-carrier sequence groups of N sequences, and takes the real part and the imaginary part respectively to obtain two frequency-domain real-valued sequences;

(2) The two frequency-domain real-valued sequences obtained in step (1) are respectively subjected to prototype filtering, N-point inverse fast Fourier transform (IFFT), signal subtraction processing and signal rearrangement processing in turn, and then merge the two, and combine them. It is sent to the receiving end, and the signal subtraction processing is to subtract two groups of adjacent signals in the time domain sequence obtained after N-point inverse fast Fourier transform and then subtract them. The signal rearrangement processing refers to using Rearrange the order of odd numbers first and then even numbers;

(3) After the receiving end divides the received signal sequence group into two branches, the received sequence signal is obtained by performing prototype filtering, N-point fast Fourier transform (FFT), signal restoration processing and real part operation respectively in turn. The signal restoration processing refers to restoration of the signal rearrangement processing.

The interference elimination method applied to the filter bank multi-carrier system of the present invention has good advantages in system design, system analysis, imaginary part inherent interference elimination and residual interference suppression, and the details are as follows:

In terms of system design, compared with document 1, the frequency domain signal is subjected to parity sampling and then processed separately. The present invention separates the real part and the imaginary part of the frequency domain signal and then processes it separately. The processing method of the present invention is simpler. efficient;

In terms of system analysis, all the analyses in document 1 are based on the premise of an ideal channel and no noise, and the analysis in document 2 only reflects the influence of multipath fading and noise in wireless channels, and does not involve The effect of the frequency offset factor term e ^j2πεi/N on the signal caused by fast time-varying fading in wireless channels, and this scheme considers the effects of fast time-varying fading, multipath fading and noise in actual wireless channels, so it is more in line with The actual situation of 5G communication;

和

显然本方案能够对滤波器之间的正交性条件做一定程度上的减弱，因而更加有利于虚部固有干扰的消除和系统的实现；In terms of the inherent interference cancellation of the imaginary part, the orthogonality conditions that need to be satisfied between the two prototype filters in Document 1 are 4 (ie (27a) to (27d) in Document 1), and this scheme will eliminate the imaginary part. The scheme of inherent interference is changed to take the real part of R"(m') and T"(m') directly, so the orthogonality conditions are reduced to 2, namely:

and

Obviously, this scheme can weaken the orthogonality conditions between filters to a certain extent, so it is more conducive to the elimination of the inherent interference of the imaginary part and the realization of the system;

其中2I-I₁的作用是将时域信号的输出变为“两组相邻信号相减后再相减”的形式，因此不仅可以实现与文献3(Yuping Zhao,and S.-G.Haggman.Intercarrier interference self-cancellationscheme for OFDM mobile communication systems[J].IEEE Transactions onCommunications,2001,49(7):1185-1191.)中同样的ICI抑制效果，而且还不会像文献3中那样导致发送数据的传输效率下降至原来的一半；I₂和

的作用是将发射端经过2I-I₁处理之后的数据按照“先奇数后偶数”的顺序进行重新排列，并在接收端再进行还原，这样可以使得原来两个相邻发送信号之间的间隔扩大至N/2，在无线信道传输过程中能够进一步分散多径衰落中的突发性错误并增强接收端的纠错能力，最终更有效地抑制信号中的ISI。In terms of residual interference suppression, Reference 2 attributes the suppression of residual interference to the suppression of ICI and ISI. In this scheme, the time domain signal is multiplied by I ₂ ·(2I-I ₁ ) and

Among them, the function of 2I- _I1 is to change the output of the time domain signal into the form of "subtracting two sets of adjacent signals and then subtracting them". .Intercarrier interference self-cancellation scheme for OFDM mobile communication systems[J].IEEE Transactions onCommunications,2001,49(7):1185-1191.) has the same ICI suppression effect, but also does not cause the transmission of data as in Reference 3 The transmission efficiency is reduced to half of the original; I ₂ and

The function is to rearrange the data after the 2I-I ₁ processing at the transmitting end in the order of "odd first, then even", and restore it at the receiving end, so as to make the interval between the original two adjacent transmitted signals Expanding to N/2, it can further disperse burst errors in multipath fading and enhance the error correction capability of the receiving end during wireless channel transmission, and ultimately suppress ISI in the signal more effectively.

The above advantages ensure that the present invention can not only weaken the orthogonality condition between the filters and eliminate the inherent interference of the imaginary part more effectively, but also can further suppress the sub-interference Residual interference between carriers and symbols.

As the preferred technical solution:

A kind of interference elimination method applied to filter bank multi-carrier system as above, the concrete operation of described signal subtraction processing is to multiply the time domain sequence obtained after N-point inverse fast Fourier transform by (2I-I ₁ );

The specific operation of the signal rearrangement processing is to multiply the time domain sequence obtained after the signal subtraction processing by I ₂ ;

The I is an N-order unit matrix, and I ₁ and I ₂ are as follows:

A kind of interference elimination method applied to filter bank multi-carrier system as above, in step (2) and step (3) for the prototype filter corresponding to the prototype filter of the same sequence of the same type and for different sequences of prototypes Filtering corresponds to different types of prototype filters.

Here, the present invention provides a specific scheme of the interference cancellation method applied to the filter bank multi-carrier system:

(1) The transmitter sends a set of N-point frequency-domain complex-valued sequences X _m,n :

X _m,n =[X _0,n ,X _1,n ,...,X _N-1,n ] ^T ,

Among them, 0≤m≤N-1 is the number of frequency-domain subcarriers, n∈Z={0,1,...,N-1} is the number of time-domain sequences;

(2) Take the real part and the imaginary part of X _m,n respectively to obtain two frequency domain real-valued sequences _Am,n and B _m,n :

(3) Multiply A _m,n and B _m,n by the prototype filters f(i-nN) and g(i-nN) respectively (f(i-nN) and g(i-nN) have different types ), and then perform Inverse Fast Fourier Transform (IFFT) respectively to obtain two time-domain sequences a _m,n and b _m,n :

Among them, i is the time index (time index) of two prototype filters f(i-nN) and g(i-nN);

(4) Multiply each of a _m,n and b _m,n by two elementary matrices 2I-I ₁ and I ₂ to obtain a' _m,n and b' _m,n :

Among them, I is an N-order unit matrix, and I ₁ and I ₂ are two N-order elementary matrices, which can be expressed as:

(5) Combine the two time-domain sequences a' _m,n and b' _m,n to obtain the final transmitted signal s'(i):

(6) Send s'(i) to the receiving end. When passing through the wireless channel, due to the influence of fast time-varying fading, multipath fading and noise in the channel, the time domain signal obtained by the receiving end is r'( i):

Among them, ε is the frequency offset factor, h(i) and L are the impulse response function and the number of paths of the multipath fading channel, respectively, and w(i) is the additive white Gaussian noise encountered when the transmitted signal propagates in the channel ;

(7) The receiving end divides r'(i) into two branches and processes them respectively. The two branches are first multiplied by the prototype filters f ^* (i-n'N) and g ^* (i-n'N) prototype filters (here f ^* (i-n'N) and f(i -nN) is the same type, g ^* (i-n'N) is the same type as g(i-nN)), and then perform N-point Fast Fourier Transform (FFT) to get the receiving end at the first Two frequency domain signals R'(m') and T'(m') in m' subcarriers:

Among them, 0≤m'≤N-1, 0≤n'≤N-1, ^* represents the conjugate operation, H(m'), W ₁ (m') and W ₂ (m') are respectively h(i ), w(i)·f ^* (i-n'N) and N-point FFT transform of w(i)·g ^* (i-n'N);

得到：(8) Multiply R'(m') and T'(m') by an N-order elementary matrix respectively

get:

根据离散傅里叶变换(DiscreteFourier Transform,DFT)的性质FFT{x_m,n±l}＝FFT{x_m,n}e^±j2πim'l/N，因此有：in,

According to the properties of Discrete Fourier Transform (DFT), FFT{x _m,n±l }=FFT{x _m,n }e ^±j2πim'l/N , so there are:

Therefore, it can be further obtained:

where H" ₁ and H" ₂ are the desired channel coefficients of R"(m') and T"(m') at (m', n'), respectively; and I" ₁ and I" ₂ are the interference terms of R"(m') and T"(m') at point (m', n') respectively (including the inherent interference and residual interference of the imaginary part), which is an imaginary number (due to H(m) '), I" ₁ and I" ₂ are not necessarily pure imaginary numbers), which can be expressed as:

Among them, S _m'-m is the ICI coefficient between the mth subcarrier and the m'th subcarrier, which can be expressed as:

The ICI coefficients between adjacent subcarriers in OFDM are very close, and FBMC/QAM has good compatibility with OFDM, so using this method can suppress the ICI coefficients between adjacent subcarriers to a large extent. It can greatly reduce I ₁ ” and I ₂ ”;

(8) Take the real part operation of R"(m') and T"(m') respectively, and obtain the frequency domain signal values R"_m',n' and T" at the point (m', n') _{m', n'} :

Among them, 0≤m'≤N-1, 0≤n'≤N-1, H" _1R (and H" _2R ), I" _1R (and I" _2R ), W" _1R (m') (and W ” _2R (m’)) is the real part of H” ₁ (and H” ₂ ), I” ₁ (and I” ₂ ), W” ₁ (m’) (and W” ₂ (m')), respectively results after.

The present invention also provides a device for applying the above-mentioned interference elimination method applied to a filter bank multi-carrier system, including a central processing unit, a memory, a program, a transmitter and a receiver;

The program is stored in the memory, and when executed by the central processing unit, causes the apparatus to perform the interference cancellation method applied to a filter bank multi-carrier system as described above.

Beneficial effects:

(1) An interference elimination method applied to a filter bank multi-carrier system of the present invention, without changing the QAM modulation mode and without reducing the data transmission efficiency, can not only weaken the orthogonality condition between the filters but also It can more effectively eliminate the inherent interference of the imaginary part, and at the same time, it can further suppress the residual interference generated between subcarriers and symbols, which has great application prospects;

(2) The device of the present invention has a simple structure, good interference suppression effect, small inherent interference and residual interference of the imaginary part, and good application prospect.

Description of drawings

1 is a structural diagram of a subcarrier sequence X _m,n in Embodiment 1;

Fig. 2 is the structure diagram of the time domain sequence a' _{m, n} in the embodiment 1;

FIG. 3 is a working flow chart of the interference elimination method applied to the filter bank multi-carrier system according to the present invention.

Detailed ways

The specific embodiments of the present invention will be further described below with reference to the accompanying drawings.

Example 1

An interference elimination method applied to a filter bank multi-carrier system, the steps are as follows (work flow chart is shown in Figure 3):

Step S1: The complex-valued sequence generator at the transmitting end sends a group of N sequences of subcarrier sequence groups (the structure of which is shown in Figure 1), which is denoted as:

X _m,n =[X _0,n ,X _1,n ,...,X _N-1,n ] ^T ,0≤m≤N-1,n∈Z={0,1,...N-1},

Next, divide X _{m and n} into two paths for processing respectively;

Step S2: Extract the real part of X _m,n and perform a series of processing to obtain the final time domain sequence group a' _m,n of the first uplink branch, including:

Step S2.1: Take the real part of X _m,n to obtain:

A _m,n =Re{X _m,n }=[A _0,n ,A _1,n ,...,A _N-1,n ] ^T ;

Step S2.2: Multiply Am _,n by the prototype filter f(i-nN), and then pass through an N-point IFFT operator to obtain:

Among them, i is the time index (time index) of the prototype filter f(i-nN);

Step S2.3: multiply a _{m, n} by two N-order elementary matrices 2I-I ₁ and I ₂ (here I is an N-order unit matrix) to obtain a' _m,n (its structure is shown in Figure 2 Show):

in,

Step S3: Extract the imaginary part of X _m,n and perform a series of processing to obtain the final time-domain sequence group b' _m,n of the second uplink branch, including:

Step S3.1: take the imaginary part of X _m,n to obtain:

B _m,n =Im{X _m,n }=[B _0,n ,B _1,n ,...,B _N-1,n ] ^T ;

Step S3.2: Multiply B _m,n by the prototype filter g(i-nN), and then pass through an N-point IFFT operator to obtain:

where i is the time index of the prototype filter g(i-nN);

Step S3.3: Multiply b _m,n by two N-order elementary matrices 2I-I ₁ and I ₂ successively to obtain:

Step S4: Combine the time-domain sequence a' _{m,n of} the first uplink branch and the time-domain sequence b'm,n of the second uplink branch bit-wise to obtain:

Step S5: Send s'(i) to the receiving end, which will be affected by fast time-varying fading, multipath fading and noise in the wireless channel during the transmission process, and obtain at the receiving end:

Among them, ε is the frequency offset factor, h(i) and L are the impulse response function and the number of paths of the multipath fading channel, respectively, and w(i) is the additive white Gaussian noise encountered when the transmitted signal propagates in the channel . Next, r'(i) is divided into two channels to be processed separately;

Step S6: Perform a series of processing on r'(i) on the first downlink branch to obtain the frequency domain signal value R'm _' ,n' at the point (m',n'), including:

Step S6.1: First pass r'(i) through the prototype filter f ^* (i-n'N), and then pass through an N-point FFT operator to obtain the frequency domain signal of the receiving end in the m'th subcarrier :

时，有：Among them, W ₁ (m') is the result of w(i)·f ^* (i-n'N) after the N-point FFT transformation of the first downlink branch; H(m') is the FFT transformation of h(i) . When R'(m') satisfies

, there are:

得到：Step S6.2: Multiply R'(m') by an elementary matrix of order N

get:

同时有：in,

Also have:

Therefore, it can be further obtained:

Among them, H" ₁ and I" ₁ are the desired channel coefficient and interference term of R"(m') at the point (m', n'), respectively. I" ₁ can be expressed as:

Among them, S _m'-m is the ICI coefficient between the mth subcarrier and the m'th subcarrier, which can be expressed as:

Step S6.3: Perform real part operation on R"(m') to obtain the frequency domain signal value at the point (m', n'):

R” _{m', n'} ＝H” _1R +I” _1R +W” _1R (m') 0≤m'≤1,0≤n'≤1

Among them, H" _1R , I" _1R and W" _1R (m') are the results of taking the real part of H" ₁ , I" ₁ and W" ₁ (m'), respectively.

Step S7: Perform a series of processing on r'(i) on the second downlink branch to obtain the frequency domain signal value T"_m',n' at the point (m', n'), including:

Step S7.1: First pass r'(i) through the prototype filter g ^* (i-n'N), and then pass through an N-point FFT operator to obtain the frequency domain signal of the receiving end in the m'th subcarrier :

时，有：Among them, W ₂ (m') is the result of w(i)·g ^* (i-n'N) after the N-point FFT transformation of the second downstream branch; H(m') is the FFT transformation of h(i) . When T'(m') satisfies

, there are:

得到：Step S7.2: Multiply T'(m') by an elementary matrix of order N

get:

同时有：in,

Also have:

Therefore, it can be further obtained:

Among them, H” ₂ and I” ₂ are the expected channel coefficients and interference terms of T”(m’) at the point (m’, n’), respectively. I” ₂ can be expressed as:

Among them, S _m'-m is the ICI coefficient between the mth subcarrier and the m'th subcarrier, which can be expressed as:

Step S7.3: Take the real part operation on T"(m') to obtain the frequency domain signal value at the point (m', n'):

T” _{m', n'} ＝H” _2R +I” _2R +W” _2R (m') 0≤m'≤1,0≤n'≤1

Among them, H" _2R , I" _2R and W" _2R (m') are the results of taking the real part of H" ₂ , I" ₂ and W" ₂ (m'), respectively.

It has been verified that the interference elimination method applied to the filter bank multi-carrier system of the present invention can not only weaken the orthogonality condition between the filters but also be more effective on the premise of not changing the QAM modulation mode and not reducing the data transmission efficiency. It can effectively eliminate the inherent interference of the imaginary part, and at the same time, it can further suppress the residual interference generated between sub-carriers and between symbols, which has great application prospects.

Example 2

A device for applying an interference elimination method, comprising a central processing unit, a memory, a program, a transmitter and a receiver;

The program is stored in the memory, and when the program is executed by the central processing unit, causes the apparatus to perform the same interference cancellation method applied to the filter bank multi-carrier system as in Embodiment 1.

It has been verified that the device of the present invention has a simple structure, good interference suppression effect, small inherent interference and residual interference of the imaginary part, and good application prospect.

Although the specific embodiments of the present invention have been described above, those skilled in the art should understand that these are only examples, and various changes may be made to these embodiments without departing from the principle and essence of the present invention. Revise.

Claims (4)

1. an interference cancellation method applied to a filter bank multi-carrier system, is characterized in that, its steps are as follows; (1) The complex-valued sequence generator at the transmitting end sends a group of sub-carrier sequence groups of N sequences, and takes the real part and the imaginary part respectively to obtain two frequency-domain real-valued sequences; (2) The two frequency-domain real-valued sequences obtained in step (1) are respectively subjected to prototype filtering, N-point inverse fast Fourier transform, signal subtraction processing and signal rearrangement processing, and then merge the two, and send them to the receiver. On the other hand, the signal subtraction processing is to subtract two groups of adjacent signals in the time-domain sequence obtained after N-point inverse fast Fourier transform, and then subtract them. The order of even numbers is rearranged; (3) After the receiving end divides the received signal sequence group into two branches, it sequentially performs prototype filtering, N-point fast Fourier transform, signal restoration processing and real part operation to obtain the received sequence signal. The restoration process refers to restoration of the signal rearrangement process. 2. a kind of interference elimination method applied to filter bank multi-carrier system according to claim 1, is characterized in that, the concrete operation of described signal subtraction processing is to obtain after N-point inverse fast Fourier transform The time domain sequence is multiplied by (2I-I ₁ ); The specific operation of the signal rearrangement processing is to multiply the time domain sequence obtained after the signal subtraction processing by I ₂ ; The I is an N-order unit matrix, and I ₁ and I ₂ are as follows:

3. a kind of interference elimination method that is applied to filter bank multi-carrier system according to claim 1 is characterized in that, the prototype filter corresponding to the prototype filtering of same sequence in step (2) and step (3) The types of the prototype filters are the same and the prototype filters corresponding to different sequences are of different types. 4. the device of applying a kind of interference elimination method applied to filter bank multi-carrier system as described in any one of claim 1～3, is characterized in that, comprises central processing unit, memory, program, transmitting end and receiving end ; The program is stored in the memory, and when the program is executed by the central processing unit, causes the apparatus to execute the filter bank multi-carrier system according to any one of claims 1 to 3. Interference cancellation method.

Images