CN105049386A - Active interference elimination method in UFMC system - Google Patents

Abstract

The invention discloses an active interference elimination method in a UFMC system. The method comprises the steps that a transmission end adds a certain number of interference cancellation subcarriers at two sides of sub-bands symmetrically through a blank space between sub-bands of the UFMC system, enables raw data information and modulation data of the interference cancellation subcarriers to serve as new data information, and carries out modulation, filtering and transmission through the UFMC system, wherein the first sub-band and the last sub-band only operate at one side; a receiving end carries out time domain windowing preprocessing and serial-to-parallel conversion at first, then carries out FFT transformation through a filter matched with the transmission side, and finally obtains a new estimation value of data information through a zero-forcing equalizer, and recovers the raw data information after removing the modulated data of the interference cancellation subcarriers employed by all sub-bands. The method maintains the good properties of the UFMC system and enables the effect of interference inhibition between the sub-bands of the UFMC system through the combination of the UFMC with active interference cancellation, thereby improving the error bit ratio of the system, and improving the communication quality.

Description

Active interference clearance method in a kind of UFMC system

Technical field

The present invention relates to wireless communication field, be specifically related to the active interference clearance method in a kind of UFMC system.

Background technology

As everyone knows, LTE is with OFDM as the baseband signal form carrying data with its current evolution form and WiFi, and OFDM is current most important multi-carrier modulation technology.OFDM becomes N number of subchannel to carry out parallel transmission information whole Channel division by N number of subcarrier.It mainly contains following advantage: the first, the high availability of frequency spectrum.Not only there is no guard band between each subchannel of OFDM, and between adjacent channel, the secondary lobe of signal spectrum is overlapped.In addition, each subchannel of OFDM can adopt different multi-system modulation schemes, which further improves the spectrum efficiency of system.The second, implementation procedure is fairly simple.The mode based on FFT can be adopted to realize modulation and demodulation.3rd, anti-multipath fading ability is strong.By being many narrow channels by whole channel distribution, the decline on each subchannel can regard near flat as, and often every sub-channels only needs the equalizer of single tap.Though be like this, OFDM has very large shortcoming: the first, higher sidelobe level.Because each subcarrier has carried out rectangle windowing in time domain, corresponding frequency spectrum is sinc shape, and this can cause the interference of intercarrier (ICI) and intersymbol interference (ISI) usually; The second, ofdm system peak-to-average force ratio is higher, and this can improve the requirement of the amplifier in radio frequency circuit, thus increases cost; Three, OFDM is very responsive to the location calibration on time-frequency, and the little carrier wave frequency deviation in any point all can destroy the orthogonality between subcarrier, causes ICI, entire system performance is declined.

Along with developing rapidly of science and technology, following mobile radio system has two large main trend: Internet of Things and arrive the transformation of customer-centric centered by community, if adopt strict synchronization mechanism again, very large signaling consumption and system energy consumption will be brought, be applied to actual conditions, can very do not calculate, therefore this trend decides the main development direction that the synchronization mechanism of non-critical is future wireless system.Consider ofdm system recited above only strict synchronous under just can show self good characteristic, so the demand faced the future, OFDM is no longer dominant technology, and designing novel multi-carrier modulation technology is a problem demanding prompt solution.Typical technical scheme has following two kinds at present:

(1) FBMC is (based on the multi-carrier modulation of filter bank, FilterBankbasedonMulti-Carrier): transmitting terminal adopts the mapping mode of OQAM, spectrum division is become multiple orthogonal sub-band by it, then carries out filtering operation to each subcarrier.There is following characteristics: a) have lower sidelobe level, inter-sub-carrier interference is less; B) can realize time-frequency efficiency is 1, but prerequisite is filter length is tending towards infinite, this can bring new problem: the rising and falling time of filter can make the efficiency of short burst communication very low, and short burst is very important for the MTC (MachineTypeCommunication) in future; C) the OQAM mechanism adopted, can not be directly compatible with all types of mimo system; D) in theory, FBMC has good characteristic, but when configuration, actual infeasible.

(2) UFMC (general filtering multi-carrier modulation, UniversalFilteredMulti-Carrier): spectrum division is become a series of sub-band comprising several subcarriers by it, then carries out filtering operation to each sub-band.There is following characteristics: a) support fragmented spectrum communication; B) compared with OFDM, spectral sidelobes level has lacked tens dB, very low; C) interference for time and frequency-offset and intercarrier has higher robustness; D) owing to adopting undemanding time-frequency to calibrate, so reduce signaling consumption, this is also for access introduces much new selection; E) key is the design of filter, has certain complexity.

Consider the advantage of UFMC technology, the thought of (AIC) is eliminated again: it is based on spectrum pool technology in conjunction with existing active interference, be applicable to a kind of interference mitigation technology under cognitive OFDM environment, spectrum pool is divided into the subcarrier of several numbers by cognitive user, frequency spectrum two side position of subcarrier is occupied in next-door neighbour primary user frequency range correspondence, place the subcarrier that some modulate special weighted factor, be referred to as interference cancellation subcarrier, by modulating these subcarriers with weighted factor, thus make the data subcarrier of cognitive user, the i.e. initial subcarrier sending data message modulation, to disturb outward at the band that primary user's frequency range produces with interference cancellation subcarrier and cancel out each other, by AF panel to enough little, thus reach the object of protection primary user communication, we propose the active interference clearance method in a kind of UFMC system, by each sub-band of UFMC as the subcarrier one by one in cognitive OFDM, it is protected, thus make interference between sub-band reach lower.

Summary of the invention

The object of the invention is to propose a kind of innovation scheme, make UFMC entire system performance better.The method can not only retain the advantage of UFMC system itself, and can improve communication quality, reduces the design complexities of UFMC median filter to a certain extent.

Active interference clearance method concrete steps in UFMC system are as follows:

1) suppose that the interference caused i-th sub-band comes from the i-th-1 sub-band and the i-th+1 sub-band, by i-th-1, an i and i+1 sub-band processes as an entirety, and this subcarrier index occupied by three sub-bands is from left to right followed successively by: [v3:v4], [v1:v2] and [v5:v6];

2) carry out multi-system modulation to original bit stream to map

Respectively the original bit stream sent is uploaded to three sub-bands and carry out multi-system modulation mapping, obtain treating that the coded bit stream through UFMC modulation is respectively X _lI, X _iand X _rI, length is respectively v4-v3+1, v2-v1+1 and v6-v5+1;

3) the solving of interference cancellation sub-carrier modulation data

3.1) after IFFT conversion being carried out to the frequency domain symbol sent, obtain corresponding time-domain signal, then Cyclic Prefix is added to this time-domain signal, obtain new time-domain signal, and carry out μ and doubly rise sampling, obtain the spectral model S of signal, this spectral model is expressed as the frequency domain symbol of transmission and the form of matrix multiple, is spectral coefficient matrix Q by this defined matrix _g;

3.2) the i-th-1 sub-band places interference cancellation subcarrier i-th sub-band both sides

The i-th-1 sub-band respectively places c i-th sub-band both sides ₁individual interference cancellation subcarrier, solves the data X of the interference cancellation subcarrier-modulated that the i-th-1 sub-band uses by least mean-square estimate _lC, make the data subcarrier of the i-th-1 sub-band and interference cancellation subcarrier minimum in total interference at i-th sub-band place, the interference spectral coefficient matrix Q that the data subcarrier of the i-th-1 sub-band causes _gsubmatrix and X _lIproduct representation, the line label at this submatrix place is [u*v1:u*v2], and row label is [v3:v4], the interference spectral coefficient matrix Q that interference cancellation subcarrier causes _gsubmatrix and X _lCproduct representation, the line label at this submatrix place is [u*v1:u*v2], and row label is [v1-c ₁: v1-1, v2+1:v2+c ₁], by X _lIand X _lCas data message X new on the i-th-1 sub-band _i-1, wherein X _i-1=[0 ..., 0, X _lI, 0 ..., 0, X _lC, 0 ..., 0] _{1 × N}, N is that IFFT counts;

3.3) the i-th+1 sub-band places interference cancellation subcarrier i-th sub-band both sides

Consistent with the solution procedure of the i-th-1 sub-band, by least mean-square estimate, solve the data X of the interference cancellation subcarrier-modulated that the i-th+1 sub-band uses _rC, by X _rIand X _rCas data message X new on the i-th+1 sub-band _i+1, wherein X _i+1=[0 ..., 0, X _rC, 0 ..., 0, X _rI, 0 ..., 0] _{1 × N};

4) modulation of UFMC sends:

By step 2) and step 3) obtain X _i-1, X _iand X _i+1after, carry out modulation through IFFT conversion and filtering two processes successively respectively and send;

5) process of receiving terminal:

5.1) first receiving terminal carries out time-domain windowed preliminary treatment and serial to parallel conversion to the received signal, obtains parallel signal;

5.2) parallel signal obtained is carried out following operation respectively on corresponding sub-band: the filter first by matching with transmitting terminal, then FFT conversion is carried out, zero forcing equalizer is finally utilized to eliminate ISI, obtain the estimated value of new data message, after the data that the interference cancellation subcarrier removing the use of each sub-band is modulated, original data message can be recovered.

Described step 3) in 3.1), 3.2) and 3.3) concrete steps be:

3.1) define represent to length to be d ₂vector carry out d ₁point Fourier transform, the element in matrix is: $F_{n,m}=e^{\frac{-j2\mathrm{\πnm}}{d_1}},0\≤m\≤d_2-\mathrm{1,0}\≤n\≤d_1-1$

Wherein m, n are element subscript;

To the frequency domain symbol vectors X=[X sent ₀, X ₁... X _n-1], do IFFT conversion, obtaining the time domain x expression formula after converting is: $x=\frac{1}{N}{F^{*}}_{N\×N}{X}^T$

Wherein, N is that IFFT counts, () ^*represent complex conjugate operation, for Fourier transform matrix;

Add Cyclic Prefix to each UFMC symbol, definition C matrix is as follows:

$C={\left[\begin{array}{cc}{0}_{\mathrm{NG}\×(N-\mathrm{NG})}& I_{\mathrm{NG}}\\ I_{N-\mathrm{NG}}& 0_{(N-\mathrm{NG})\×\mathrm{NG}}\\ 0_{\mathrm{NG}\×(N-\mathrm{NG})}& I_{\mathrm{NG}}\end{array}\right]}_{(N+\mathrm{NG})\×N}$

Wherein, 0 is 0 matrix, and I is unit matrix, and NG is the length of Cyclic Prefix;

After the time-domain signal x of transmission is added Cyclic Prefix, obtain new time-domain signal x', the computing formula of x' is x'=Cx;

To new time-domain signal x', carry out μ and doubly rise sampling, obtain the spectral model S of signal, the computing formula of S is:

$S=F_{\mathrm{\μN}\×(N+\mathrm{NG})}{x}^{\′}=F_{\mathrm{\μN}\×(N+\mathrm{NG})}\mathrm{Cx}=\frac{1}{N}{F}_{\mathrm{\μN}\×(N+\mathrm{NG})}{{\mathrm{CF}}^{*}}_{N\×N}{X}^T$

Wherein, μ for rise decimation factor, F _{μ N × (N+NG)}for step 3.1) middle definition d ₁for uN, d ₂for N+NG, () ^trepresent transpose operation;

Definition dimension is the spectral coefficient matrix of uN × N

3.2) the i-th-1 sub-band places interference cancellation subcarrier i-th sub-band both sides

The i-th-1 sub-band respectively places c in the both sides of i-th sub-band ([v1:v2]) ₁individual interference cancellation subcarrier, the position of interference cancellation subcarrier is [v1-c ₁: v1-1, v1:v2, v2+1:v2+c ₁], the subcarrier index of the i-th-1 sub-band modulation transmission data is [v3:v4];

Data subcarrier X _lIthe interference produced at i-th sub-band is Q _lIx _lI ^t, the data X of interference cancellation subcarrier-modulated _lCthe interference produced at i-th sub-band is Q _lCx _lC ^t, calculate two sub-carriers produce and interference, computing formula is:

I _i-1＝||Q _lCX _lC ^T+Q _lIX _lI ^T|| ²

Wherein, X _lI=[X _i-1(v3), X _i-1(v3+1) ... X _i-1(v4)],

X _lC＝[X _i-1(v1-c ₁),...X _i-1(v1-1),X _i-1(v2+1),...X _i-1(v2+c ₁)]，

$Q_{\mathrm{lI}}=Q_{G[v3:v4]}^{[u*v1:v*v2]},$ $Q_{\mathrm{lC}}=Q_{G[v1-c_1:v1-1,v2+1:v2+c_1]}^{[u*v1:u*v2]},$ Wherein with subscript all represent step 3.1) in matrix Q _grow, the equal representing matrix Q of subscript _grow, || || ²represent the quadratic sum of vector element; X in formula _i-1after introducing interference cancellation subcarrier, the i-th-1 son frequently

Bring new data message, X _i-1expression formula be:

X _i-1＝[0,...,0,X _lI,0,...,0,X _lC,0,...,0] _1×N

＝[0,...0,X _i-1(v3),X _i-1(v3+1),...X _i-1(v4),0,...0,X _i-1(v1-c ₁),...X _i-1(v1-1),X _i-1(v2+1),...X _i-1(v2+c ₁),0,...0] _1×N

Solve X _lC, make these two kinds interference be reduced to minimum at frequency range [v1:v2] place, namely meet following formula:

$X_{\mathrm{lC}}=\underset{X_{\mathrm{lC}}}{\arg \min }{\left|\right|Q_{\mathrm{lC}}{X_{\mathrm{lC}}}^T+Q_{\mathrm{lI}}{X_{\mathrm{lI}}}^T\left|\right|}^2$

Above formula is undertaken solving by least mean-square estimate:

X _lC＝-(Q _lC ^HQ _lC) ^-1Q _lC ^HQ _lIX _lI ^T

3.3) the i-th+1 sub-band places interference cancellation subcarrier i-th sub-band both sides

With step 3.2) in solution procedure consistent, the i-th+1 sub-band respectively places c in the both sides of i-th sub-band ([v1:v2]) ₂individual interference cancellation subcarrier, the position of interference cancellation subcarrier is [v1-c ₂: v1-1, v1:v2, v2+1:v2+c ₂], the subcarrier index of the i-th+1 sub-band modulation transmission data is [v5:v6];

Data subcarrier X _rIthe interference produced at i-th sub-band is Q _rIx _rI ^t, interference cancellation subcarrier X _rCthe interference produced at i-th sub-band is Q _rCx _rC ^t, what this two sub-carrier produced with interference is:

I _i+1＝||Q _rCX _rC ^T+Q _rIX _rI ^T|| ²

Wherein, X _rI=[X _i+1(v5), X _i+1(v5+1) ... X _i+1(v6)],

X _rC＝[X _i+1(v1-c ₂),...X _i+1(v1-1),X _i+1(v2+1),...X _i+1(v2+c ₂)]，

$Q_{\mathrm{rI}}=Q_{G[v5:v6]}^{[u*v1:v*v2]}$ With $Q_{\mathrm{rC}}=Q_{G[v1-c_2:v1-1,v2+1:v2+c_2]}^{[u*v1:u*v2]},$ Wherein with subscript all represent step 3.1) in matrix Q _grow, the equal representing matrix Q of subscript _grow, the X in formula _i+1after introducing interference cancellation subcarrier, data message new on the i-th+1 sub-band, X _i+1expression formula be:

X _i+1＝[0,...,0,X _rC,0,...,0,X _rI,0,...,0] _1×N

=[0 ... 0, X _i+1(v1-c ₂) ... X _i+1(v1-1), X _i+1(v2+1) ... X _i+1(v2+c ₂), 0 ... 0, X _i+1(v5), X _i+1(v5+1) ... X _i+1(v6), 0 ... 0] _{1 × N}undertaken solving by least mean-square estimate:

X _rC＝-(Q _rC ^HQ _rC) ^-1Q _rC ^HQ _rIX _rI ^T

Required X _rCbe the data of the interference cancellation subcarrier-modulated that i-th sub-band uses.

Described step 4) in 4.1) and 4.2) concrete steps be:

4.1) IFFT conversion

Realize IFFT conversion by Fourier's matrix, this Fourier's defined matrix is represent to length to be d ₂vector carry out d ₁point IFFT conversion, the element in the middle of Fourier's matrix is as follows:

$V_{n,m}=\frac{1}{\sqrt{d_1}}{e}^{\frac{j2\mathrm{\πnm}}{d_1}},0\≤m\≤d_2-\mathrm{1,0}\≤n\≤d_1-1$

4.2) filtering

For filtering, complete this linear convolution operation with toeplitz matrix, the centre centre frequency of filter being moved to sub-band is needed for different sub-bands;

Arranging UFMC system mesarcs filter length is L, and the expression formula of impulse response h is: h={h [0], h [1] ..., h [L-1] }, corresponding frequency domain response is H;

The subcarrier index calculating the i-th-1 sub-mid-band frequency place is: by the frequency domain response H translation of prototype filter individual unit, obtains the frequency domain response H of the i-th-1 sub-band _i-1,

The impulse response h of the i-th-1 sub-band _i-1for: h _i-1={ h _i-1[0], h _i-1[1] ..., h _i-1[L-1] }, wherein, $h_{i-1}\left[n\right]=h\left[n\right]e^{\left(\frac{j2\mathrm{\π}{k}_{i-1}n}{N}\right)},0\≤n\≤L-1;$

Definition toeplitz matrix T _i-1for:

Get its front N row to be designated as with realize the filtering of the i-th-1 sub-band;

Consistent with the solution procedure of i-1 sub-band, for i-th sub-band, translation on frequency domain individual unit, obtains frequency domain response H _iwith impulse response h _i, h _iin element be: $h_i\left[n\right]=h\left[n\right]e^{\left(\frac{j2\mathrm{\π}{k}_{i}n}{N}\right)},0\≤n\≤L-1;$ For the i-th+1 sub-band, translation on frequency domain $k_{i+1}=\frac{v6-v5}{2}+v5$ Individual unit, obtains frequency domain response H _i+1with impulse response h _i+1, h _i+1in element be: by h _i-1[n] uses h respectively _i[n] and h _i+1[n] replaces and obtains the toeplitz matrix of i-th sub-band and the i-th+1 sub-band, then gets its front N row respectively, obtains the toeplitz matrix that i-th sub-band and the i-th+1 sub-band realize filtering and is respectively with

Through above-mentioned steps 1), step 2), step 3) and step 4), transmitting terminal send time-domain signal x _totalfor:

x _total＝T·V·X

Wherein, $T=[T_{i-1}^N,T_i^N,T_{i+1}^N],$

$X={[X_{i-1}^T,X_i^T,X_{i+1}^T]}^T,$

V=diag{V _i-1, V _i, V _i+1represent with matrix V _i-1, V _iand V _i+1for the diagonal matrix that diagonal element is formed, V, V _i-1, V _iand V _i+1represent respectively and three sub-bands realize step 4.1) in Fourier's matrix of IFFT conversion and d ₁and d ₂value factor band and different.

The present invention is applicable to access on a large scale the wireless communication system under environment, and supports fragmented spectrum communication, and system is without the need for strict net synchronization capability, and this is that access introduces much new selection.In addition, compared to FBMC, due to the filter using length shorter, this can make system response time shorten, and the short burst communication therefore for future has good applicability.

The present invention combines by being eliminated with active interference by UFMC, not only can retain the superperformance of UFMC system, and the interference that can reduce to a great extent between sub-band, thus improve the error bit ability of system, improve the reliability of transmission, in addition, annoyance level one timing when between UFMC sub-band, compared to simple UFMC system, adopt these association schemes can reduce the design complexities of UFMC median filter to a certain extent.

Accompanying drawing explanation

Fig. 1 is that the active interference in cognitive OFDM eliminates schematic diagram;

Fig. 2 is the schematic diagram that UFMC and AIC of the present invention combines;

Fig. 3 is active interference clearance method structured flowchart in UFMC system of the present invention;

Fig. 4 is when Cyclic Prefix (CP) length is 16, respectively places the fluting that 2 interference cancellation subcarriers are formed in protection frequency range [15:25] both sides;

Fig. 5 is when Cyclic Prefix (CP) length is 32, respectively places the fluting that 2 interference cancellation subcarriers are formed in protection frequency range [15:25] both sides;

Fig. 6 is that UFMC and AIC of the present invention combines the normalized power spectral density of left and right sub-band;

Fig. 7 is that UFMC and AIC of the present invention combines the normalized power spectral density of left and right sub-band associating;

Fig. 8 is the normalized power spectral density of about UFMC sub-band of the present invention;

Fig. 9 is the normalized power spectral density of about UFMC sub-band of the present invention associating;

Figure 10 be OFDM of the present invention UFMC the error bit ability curve that combines of UFMC and AIC.

Embodiment

Below in conjunction with the drawings and specific embodiments, the present invention is described in further detail.

As shown in Figure 1, as shown in Figure 2, the transceiver schematic diagram of the inventive method as shown in Figure 3 for the schematic diagram that UFMC and AIC combines for the introducing background of AIC.

Active interference clearance method in a kind of UFMC system of the present invention, concrete implementation step is as follows:

1) setting comes from the 3rd sub-band and the 5th sub-band to the 4th interference that sub-band causes, processed as an arrangement by 3rd, 4 and 5 sub-band, this subcarrier index occupied by three sub-bands is from left to right followed successively by: [7:18], [25:36] and [43:54].

2) original bit stream carries out multi-system modulation mapping:

2.1) to the original bit stream [q that the 3rd sub-band transmits ₁, q ₂, q ₃..., q ₂₄] carry out QPSK modulation mapping, obtain treating the coded bit stream X through UFMC modulation _lI=[b ₁, b ₂, b ₃..., b ₁₂];

2.2) to the original bit stream [w that the 4th sub-band transmits ₁, w ₂, w ₃..., w ₂₄] carry out QPSK modulation mapping, obtain treating the coded bit stream X through UFMC modulation ₄=[a ₁, a ₂, a ₃..., a ₁₂];

2.3) to the original bit stream [r that the 5th sub-band transmits ₁, r ₂, r ₃..., r ₂₄] carry out QPSK modulation mapping, obtain treating the coded bit stream X through UFMC modulation _rI=[s ₁, s ₂, s ₃..., s ₁₂];

3) the solving of interference cancellation subcarrier-modulated weighted factor:

First the basis in order to provide UFMC symbol to add Cyclic Prefix, invention has been simulation analysis, and result is as shown in accompanying drawing 3 and accompanying drawing 4: in accompanying drawing 4, and as CP=16, the groove depth of formation is about 35db, and the energy cost of consumption is 13.48%; In accompanying drawing 5, the groove depth formed as CP=32 is about 35db, and the energy cost of consumption only has 5.98%.Both of these case, the energy cost consumed is all far below the energy cost consuming 30.66% when not using Cyclic Prefix.Therefore adopt CP can play the effect reducing band outward leakage, thus reduce the energy that interference cancellation subcarrier consumes, below derivation can consider to use Cyclic Prefix.

3.1) setting FFT points N is 128, and circulating prefix-length NG is 32, and up-sampling multiple μ is 2, spectral coefficient matrix Q _gexpression formula is:

$Q_G=\frac{1}{N}{F}_{\mathrm{\μN}\×(N+\mathrm{NG})}{{\mathrm{CF}}^{*}}_{N\×N}=\frac{1}{128}{F}_{256\×(128+32)}{{\mathrm{CF}}^{*}}_{128\×128}$

Wherein, $C={\left[\begin{array}{cc}{0}_{32\×(128-32)}& I_{32}\\ I_{128-32}& 0_{(128-32)\×32}\\ 0_{32\×(128-32)}& I_{32}\end{array}\right]}_{(128+32)\×128}$ 0 represents 0 matrix, I representation unit matrix,

Element in F is: $F_{n,m}=\exp \left(\frac{-j2\mathrm{\πnm}}{256}\right),0\≤n\≤\mathrm{255,0}\≤m\≤159,$

F ^*in element be: ${F^{*}}_{h,g}=\exp \left(\frac{j2\mathrm{\πhg}}{128}\right),0\≤h\≤\mathrm{127,0}\≤g\≤127;$

3.2) the 3rd sub-band places interference cancellation subcarrier the 4th sub-band both sides

3rd sub-band respectively places 2 interference cancellation subcarriers the 4th sub-band ([25:36]) both sides, the position of interference cancellation subcarrier is [23:24,25:36,37:38], the subcarrier index of the 3rd sub-band modulation transmission data is [7:18];

Data subcarrier X _lIthe interference produced at the 4th sub-band is Q _lIx _lI ^t, the data X of interference cancellation subcarrier-modulated _lCthe interference produced at the 4th sub-band is Q _lCx _lC ^t, what this two sub-carrier produced with interference is:

I ₃＝||Q _lCX _lC ^T+Q _lIX _lI ^T|| ²

Wherein, X _lI=[b ₁, b ₂..., b ₁₂]=[X ₃(7), X ₃(8) ... X ₃(18)],

X _lC＝[X ₃(23),X ₃(24),X ₃(25),...X ₃(38)]，

$Q_{\mathrm{lI}}=Q_{G[7:18]}^{[2*25:2*36]}$ With $Q_{\mathrm{lC}}=Q_{G[23:\mathrm{24,25}:\mathrm{36,37}:38]}^{[2*25:2*36]}$ The equal representing matrix Q of subscript _grow, the equal representing matrix Q of subscript _grow, || || ²represent the quadratic sum of vector element;

After introducing interference cancellation subcarrier, data message X new on the 3rd sub-band ₃dimension is 128, and expression formula is:

X ₃＝[0,...,0,X _lI,0,...,0,X _lC,0,...,0] _1×128

＝[0,...0,X ₃(7),X ₃(8),...X ₃(18),0,...0,X ₃(23),X ₃(24),X ₃(25),...X ₃(38),0,...0] _1×128

According to the basic thought of interference cancellation, solve X _lCmake these two kinds interference be reduced to minimum at frequency range [25:36] place, namely meet following formula:

$X_{\mathrm{lC}}=\underset{X_{\mathrm{lC}}}{\arg \min }{\left|\right|Q_{\mathrm{lC}}{X_{\mathrm{lC}}}^T+Q_{\mathrm{lI}}{X_{\mathrm{lI}}}^T\left|\right|}^2$

Above formula can be undertaken solving by least mean-square estimate:

X _lC＝-(Q _lC ^HQ _lC) ^-1Q _lC ^HQ _lIX _lI ^T

3.3) the 5th sub-band places interference cancellation subcarrier the 4th sub-band both sides

5th sub-band respectively places 2 interference cancellation subcarriers the 4th sub-band ([25:36]) both sides, the position of interference cancellation subcarrier is [23:24,25:36,37:38], the subcarrier index of the 5th sub-band modulation transmission data is [43:54];

Data subcarrier X _rIthe interference produced at the 4th sub-band is Q _rIx _rI ^t, the data X of interference cancellation subcarrier-modulated _rCthe interference produced at the 4th sub-band is Q _rCx _rC ^t, what this two sub-carrier produced with interference is:

I ₅＝||Q _rCX _rC ^T+Q _rIX _rI ^T|| ²

Wherein, X _rI=[s ₁, s ₂..., s ₁₂]=[X ₅(43), X ₅(44) ... X ₅(54)],

X _rC＝[X ₅(23),X ₅(24),X ₅(25),...X ₅(38)]，

$Q_{\mathrm{rI}}=Q_{G[43:54]}^{[2*25:2*36]},$ $Q_{\mathrm{rC}}=Q_{G[23:\mathrm{24,25}:\mathrm{36,37}:38]}^{[2*25:2*36]}$ Subscript representing matrix Q _grow, subscript representing matrix Q _grow, || || ²represent the quadratic sum of vector element;

After introducing interference cancellation subcarrier, data message X new on the 5th sub-band ₅dimension is 128, and expression formula is:

X ₅＝[0,...,0,X _rC,0,...,0,X _rI,0,...,0] _1×128

=[0 ... 0, X ₅(23), X ₅(24), X ₅(25) ... X ₅(38), 0 ... 0, X ₅(43), X ₅(44) ... X ₅(54), 0 ... 0] _{1 × 128}similarly, undertaken solving can obtain by least mean-square estimate:

X _rC＝-(Q _rC ^HQ _rC) ^-1Q _rC ^HQ _rIX _rI ^T

4) modulation of UFMC sends:

By step 1), step 2) and step 3), obtain X ₃, X ₄and X ₅, then carry out modulation through IFFT conversion and filtering two processes successively respectively and send;

4.1) IFFT conversion

We are to the data message X of each sub-band ₃, X ₄and X ₅, all carry out 128 IFFT conversion, Fourier's matrix V that three sub-bands use ₃, V ₄and V ₅be: V _{128 × 128}, the element in matrix is:

$V_{n,m}=\frac{1}{\sqrt{d_1}}{e}^{\frac{j2\mathrm{\πnm}}{d_1}},0\≤m\≤127,0\≤n\≤127.$

4.2) filtering

For filtering, this linear convolution operation is completed with toeplitz matrix, the centre centre frequency of filter being moved to sub-band is needed for different sub-bands, if the prototype filter that UFMC system uses is length is 16, side lobe attenuation is the Chebyshev filter of 40dB, and its impulse response coefficient is: h={h [0], h [1],, h [15] }, corresponding frequency domain response is H;

The subcarrier index calculating the 3rd sub-mid-band frequency place is: 12.5, by the frequency domain response H translation of prototype filter individual unit, obtains the frequency domain response H of the 3rd sub-band ₃,

Impulse response h ₃for: h ₃={ h ₃[0], h ₃[1] ..., h ₃[15] }, wherein, toeplitz matrix is:

Get its front 128 row to be designated as with realize the filtering of the 3rd sub-band;

For the 4th sub-band, translation on frequency domain individual unit, obtains frequency domain response H ₄with impulse response h ₄, h ₄in element be: for the 5th sub-band, translation on frequency domain individual unit, obtains frequency domain response H ₅with impulse response h ₅, h ₅in element be: by h ₃[n] uses h respectively ₄[n] and h ₅[n] replaces and obtains the toeplitz matrix of the 4th sub-band and the 5th sub-band, then get its front 128, obtain the 4th sub-band and the 5th sub-band and realize filtering with

Through above-mentioned steps 1), step 2), step 3) and step 4), transmitting terminal send time-domain signal be:

x _total＝T·V·X

Wherein, $T=[T_3^{128},T_4^{128},T_5^{128}],X={[X_3^T,X_4^T,X_5^T]}^T,$

V=diag{V ₃, V ₄, V ₅represent with matrix V ₃, V ₄and V ₅for the diagonal matrix V that diagonal element is formed, V ₃, V ₄and V ₅represent respectively and three sub-bands realize step 4) in Fourier's matrix V of IFFT conversion _{128 × 128};

3rd sub-band and the 4th sub-band are after AIC and UFMC process, and corresponding signal spectrum is: $s_3=Q_G{X}_3^T{H}_3,S_r=Q_G{X}_r^T{H}_r$

Convenient in order to describe below, the 3rd sub-band, the 4th sub-band and the 5th sub-band are called left sub-band, dynatron frequency band and right sub-band.By drawing out spectrogram corresponding to above-mentioned equation, just can find out after adopting AIC very intuitively, the interference suppressioning effect that between two the sub-band centerings in left and right, sub-band brings, as illustrated shown in accompanying drawing 6 and accompanying drawing 7, it respectively places 2 interference cancellation subcarriers in dynatron frequency band [25:36] both sides, the interference suppressioning effect that two, adjacent left and right sub-band obtains at the sub-band place of centre, can obviously find out: left sub-band is on average roughly at the groove depth at dynatron frequency band place: 85dB, be roughly in the decay of right sub-band: about 70dB, also right sub-band is protected while protection dynatron frequency band, two sub-bands in left and right are on average roughly at the associating groove depth at dynatron frequency band place: about 170dB.The effect of right sub-band is consistent with left sub-band.

In order to give prominence to the interference suppressioning effect adopting AIC to bring better, we have also been made simulation analysis to the spectrogram of the UFMC system not adopting AIC, as illustrated in Figure 8 and 9 reference.Can find out: in UFMC system, left sub-band is on average roughly at the groove depth at dynatron frequency band place: 70dB, is roughly: about 75dB in the decay of right sub-band, also protects right sub-band while protection dynatron frequency band; Two sub-bands in left and right are on average roughly at the associating groove depth at dynatron frequency band place: about 140dB.The effect of right sub-band is consistent with left sub-band.

Contrast accompanying drawing 6, accompanying drawing 7 and accompanying drawing 8, accompanying drawing 9; can obtain: relative to not adopting for AIC scheme; the UFMC system that have employed AIC achieves good interference suppressioning effect at dynatron frequency band place; moreover; while protection dynatron frequency band, protection can also be formed to another sub-band.

5) process of receiving terminal:

Consider additive Gaussian noise channels, the frequency response H=I of channel, I representation dimension is the unit vector of N+L-1=143, and now receiving terminal does not need to do zero forcing equalization, and random generation length is that the multiple gaussian variable of N+L-1=143 is as Gaussian noise.

5.1) first receiving terminal carries out time-domain windowed preliminary treatment and serial to parallel conversion to the received signal;

5.2) parallel signal obtained is carried out following operation respectively on corresponding sub-band: the filter first by matching with transmitting terminal, then FFT conversion is carried out, obtain the estimated value of new data message, the data that the interference cancellation subcarrier finally removing the use of each sub-band is modulated, can recover original data message.

Concrete operations are as follows:

The time-domain signal that step 1 receives: $\begin{array}{c}y=x_{\mathrm{total}}+z\\ =T\·V\·X+z\end{array};$

Step 2 receiving terminal matched filtering: T ^-1y=VX+T ^-1z;

Step 3FFT converts: $\begin{array}{c}{V}^{-1}{T}^{-1}y=X+V^{-1}{T}^{-1}z\\ =X+V^{-1}{T}^{-1}z\end{array};$

Step 4, by above-mentioned three steps, estimates the estimated value of the new data message of transmitting terminal then the data message that the interference cancellation subcarrier removing the use of each sub-band is modulated, obtains the estimated value of original data message again through QPSK demodulation, recover three sub-bands successively and upload the original bit stream sent

By using the above-mentioned whole process comprising modulation and demodulation, can the error bit ability of quantitative analysis the inventive method, and can, by contrasting with OFDM, UFMC scheme, make the advantage of the inventive method more outstanding, strengthen confidence level.

Accompanying drawing 10 is the simulation result of the active interference clearance method example in above-mentioned UFMC system, and contains the error bit ability curve of OFDM scheme and UFMC scheme, and the simulation parameter of three kinds of schemes arranges as shown in table 1.

Table 1OFDM UFMC UFMC and AIC combine simulation parameter setting

Observe simulation result, can find out significantly: OFDM error bit ability is relatively the poorest, and UFMC takes second place, the inventive method is optimum, and the UFMC that have employed AIC has good error bit ability compared to UFMC, the two is under same SNR, and bit error rate gap is very large.

The inventive method is respectively from normalized power spectral density and bit error rate angle, the method and UFMC scheme have been carried out contrast simulation, obviously can find out: the inventive method is all better than the latter in both cases, absolutely prove that the inventive method is reasonable.

Claims (3)

1. the active interference clearance method in UFMC system, is characterized in that, the method concrete steps are as follows:

1) suppose that the interference caused i-th sub-band comes from the i-th-1 sub-band and the i-th+1 sub-band, by i-th-1, an i and i+1 sub-band processes as an entirety, and this subcarrier index occupied by three sub-bands is from left to right followed successively by: [v3:v4], [v1:v2] and [v5:v6];

2) carry out multi-system modulation to original bit stream to map

Respectively the original bit stream sent is uploaded to three sub-bands and carry out multi-system modulation mapping, obtain treating that the coded bit stream through UFMC modulation is respectively X _lI, X _iand X _rI, length is respectively v4-v3+1, v2-v1+1 and v6-v5+1;

3) the solving of interference cancellation sub-carrier modulation data

3.1) after IFFT conversion being carried out to the frequency domain symbol sent, obtain corresponding time-domain signal, then Cyclic Prefix is added to this time-domain signal, obtain new time-domain signal, and carry out μ and doubly rise sampling, obtain the spectral model S of signal, this spectral model is expressed as the frequency domain symbol of transmission and the form of matrix multiple, is spectral coefficient matrix Q by this defined matrix _g;

3.2) the i-th-1 sub-band places interference cancellation subcarrier i-th sub-band both sides

The i-th-1 sub-band respectively places c i-th sub-band both sides ₁individual interference cancellation subcarrier, solves the data X of the interference cancellation subcarrier-modulated that the i-th-1 sub-band uses by least mean-square estimate _lC, make the data subcarrier of the i-th-1 sub-band and interference cancellation subcarrier minimum in total interference at i-th sub-band place, the interference spectral coefficient matrix Q that the data subcarrier of the i-th-1 sub-band causes _gsubmatrix and X _lIproduct representation, the line label at this submatrix place is [u*v1:u*v2], and row label is [v3:v4], the interference spectral coefficient matrix Q that interference cancellation subcarrier causes _gsubmatrix and X _lCproduct representation, the line label at this submatrix place is [u*v1:u*v2], and row label is [v1-c ₁: v1-1, v2+1:v2+c ₁], by X _lIand X _lCas data message X new on the i-th-1 sub-band _i-1, wherein X _i-1=[0 ..., 0, X _lI, 0 ..., 0, X _lC, 0 ..., 0] _{1 × N}, N is that IFFT counts;

3.3) the i-th+1 sub-band places interference cancellation subcarrier i-th sub-band both sides

Consistent with the solution procedure of the i-th-1 sub-band, by least mean-square estimate, solve the data X of the interference cancellation subcarrier-modulated that the i-th+1 sub-band uses _rC, by X _rIand X _rCas data message X new on the i-th+1 sub-band _i+1, wherein X _i+1=[0 ..., 0, X _rC, 0 ..., 0, X _rI, 0 ..., 0] _{1 × N};

4) modulation of UFMC sends

By step 2) and step 3) obtain X _i-1, X _iand X _i+1after, carry out modulation through IFFT conversion and filtering two processes successively respectively and send;

5) process of receiving terminal

5.1) first receiving terminal carries out time-domain windowed preliminary treatment and serial to parallel conversion to the received signal, obtains parallel signal;

5.2) parallel signal obtained is carried out following operation respectively on corresponding sub-band: the filter first by matching with transmitting terminal, then FFT conversion is carried out, zero forcing equalizer is finally utilized to eliminate ISI, obtain the estimated value of new data message, after the data that the interference cancellation subcarrier removing the use of each sub-band is modulated, original data message can be recovered.

2. the active interference clearance method in UFMC system as claimed in claim 1, is characterized in that: described step 3) in step 3.1), step 3.2) and step 3.3) concrete steps be:

3.1) define represent to length to be d ₂vector carry out d ₁point Fourier transform, the element in matrix is: $F_{n,m}=e^{\frac{-j2\mathrm{\πnm}}{d_1}},0\≤m\≤d_2-\mathrm{1,0}\≤0\≤n\≤d_1-1$

Wherein m, n are element subscript;

To the frequency domain symbol vectors X=[X sent ₀, X ₁... X _n-1], do IFFT conversion, obtaining the time domain x expression formula after converting is: $x=\frac{1}{N}{F^{*}}_{N\×N}{X}^T$

Wherein, N is that IFFT counts, () ^*represent complex conjugate operation, for Fourier transform matrix;

Add Cyclic Prefix to each UFMC symbol, definition C matrix is as follows:

$C={\left[\begin{array}{cc}{0}_{\mathrm{NG}\×(N-\mathrm{NG})}& I_{\mathrm{NG}}\\ I_{N-\mathrm{NG}}& 0_{(N-\mathrm{NG})\×\mathrm{NG}}\\ 0_{\mathrm{NG}\×(N-\mathrm{NG})}& I_{\mathrm{NG}}\end{array}\right]}_{(N+\mathrm{NG})\×N}$

Wherein, 0 is 0 matrix, and I is unit matrix, and NG is the length of Cyclic Prefix;

After the time-domain signal x of transmission is added Cyclic Prefix, obtain new time-domain signal x', the computing formula of x' is x'=Cx;

To new time-domain signal x', carry out μ and doubly rise sampling, obtain the spectral model S of signal, the computing formula of S is:

$S=F_{\mathrm{\μN}\×(N+\mathrm{NG})}{x}^{\′}=F_{\mathrm{\μN}\×(N+\mathrm{NG})}\mathrm{Cx}=\frac{1}{N}{F}_{\mathrm{\μN}\×(N+\mathrm{NG})}{{\mathrm{CF}}^{*}}_{N\×N}{X}^T$

Wherein, μ for rise decimation factor, F _{μ N × (N+NG)}for step 3.1) middle definition d ₁for uN, d ₂for N+NG, () ^trepresent transpose operation;

Definition dimension is the spectral coefficient matrix of uN × N

3.2) the i-th-1 sub-band places interference cancellation subcarrier i-th sub-band both sides

The i-th-1 sub-band respectively places c in the both sides of i-th sub-band ([v1:v2]) ₁individual interference cancellation subcarrier, the position of interference cancellation subcarrier is [v1-c ₁: v1-1, v1:v2, v2+1:v2+c ₁], the subcarrier index of the i-th-1 sub-band modulation transmission data is [v3:v4];

Data subcarrier X _lIthe interference produced at i-th sub-band is Q _lIx _lI ^t, the data X of interference cancellation subcarrier-modulated _lCthe interference produced at i-th sub-band is Q _lCx _lC ^t, calculate two sub-carriers produce and interference, computing formula is:

I _i-1＝||Q _lCX _lC ^T+Q _lIX _lI ^T|| ²

Wherein, X _lI=[X _i-1(v3), X _i-1(v3+1) ... X _i-1(v4)],

X _lC＝[X _i-1(v1-c ₁),…X _i-1(v1-1),X _i-1(v2+1),…X _i-1(v2+c ₁)]，

$Q_{\mathrm{lI}}=Q_{G[v3:v4]}^{\text{[v*v1:u*v2]}},Q_{\mathrm{lC}}{=Q}_{G[v1-c_1:v1-1,v2+1:v2+c_1]}^{[u*v1:u*v2]},$ Wherein with subscript all represent step 3.1) in matrix Q _grow, the equal representing matrix Q of subscript _grow, || || ²represent the quadratic sum of vector element; X in formula _i-1after introducing interference cancellation subcarrier, data message new on the i-th-1 sub-band, X _i-1expression formula be:

X _i-1＝[0,…,0,X _lI,0,…,0,X _lC,0,…,0] _1×N

＝[0,…0,X _i-1(v3),X _i-1(v3+1),…X _i-1(v4),0,…0,X _i-1(v1-c ₁),…X _i-1(v1-1),X _i-1(v2+1),…X _i-1(v2+c ₁),0,…0] _1×N

Solve X _lC, make these two kinds interference be reduced to minimum at frequency range [v1:v2] place, namely meet following formula:

$X_{\mathrm{lC}}=\underset{X_{\mathrm{lC}}}{\arg \mathrm{}\min }{\left|\right|Q_{\mathrm{lC}}{X_{\mathrm{lC}}}^T+Q_{\mathrm{lI}}{X_{\mathrm{lI}}}^T\left|\right|}^2$

Above formula is undertaken solving by least mean-square estimate:

X _lC＝-(Q _lC ^HQ _lC) ^-1Q _lC ^HQ _lIX _lI ^T

3.3) the i-th+1 sub-band places interference cancellation subcarrier i-th sub-band both sides

With step 3.2) in solution procedure consistent, the i-th+1 sub-band respectively places c in the both sides of i-th sub-band ([v1:v2]) ₂individual interference cancellation subcarrier, the position of interference cancellation subcarrier is [v1-c ₂: v1-1, v1:v2, v2+1:v2+c ₂], the subcarrier index of the i-th+1 sub-band modulation transmission data is [v5:v6];

Data subcarrier X _rIthe interference produced at i-th sub-band is Q _rIx _rI ^t, interference cancellation subcarrier X _rCthe interference produced at i-th sub-band is Q _rCx _rC ^t, what this two sub-carrier produced with interference is:

I _i+1＝||Q _rCX _rC ^T+Q _rIX _rI ^T|| ²

Wherein, X _rI=[X _i+1(v5), X _i+1(v5+1) ... X _i+1(v6)],

X _rC＝[X _i+1(v1-c ₂),…X _i+1(v1-1),X _i+1(v2+1),…X _i+1(v2+c ₂)]，

$Q_{\mathrm{rI}}=Q_{G[v5:v6]}^{\text{[u*v1:u*v2]}}$ With $Q_{\mathrm{rC}}{=Q}_{G[v1-c_2:v1-1,v2+1:v2+c_2]}^{[u*v1:u*v2]},$ Wherein with subscript all represent step 3.1) in matrix Q _grow, the equal representing matrix Q of subscript _grow, the X in formula _i+1after introducing interference cancellation subcarrier, data message new on the i-th+1 sub-band, X _i+1expression formula be:

X _i+1＝[0,…,0,X _rC,0,…,0,X _rI,0,…,0] _1×N

＝[0,…0,X _i+1(v1-c ₂),…X _i+1(v1-1),X _i+1(v2+1),…X _i+1(v2+c ₂),0,…0,X _i+1(v5),X _i+1(v5+1),…X _i+1(v6),0,…0] _1×N

Undertaken solving by least mean-square estimate:

X _rC＝-(Q _rC ^HQ _rC) ^-1Q _rC ^HQ _rIX _rI ^T

Required X _rCbe the data of the interference cancellation subcarrier-modulated that i-th sub-band uses.

3. the active interference clearance method in UFMC system as claimed in claim 1, is characterized in that: described step 4) in step 4.1) and step 4.2) concrete steps be:

4.1) IFFT conversion

Realize IFFT conversion by Fourier's matrix, this Fourier's defined matrix is represent to length to be d ₂vector carry out d ₁point IFFT conversion, the element in the middle of Fourier's matrix is as follows:

$V_{n,m}=\frac{1}{\sqrt{d_1}}{e}^{\frac{j2\mathrm{\πnm}}{d_1}},0\≤m\≤d_2-\mathrm{1,0}\≤n\≤d_1-1$

4.2) filtering

For filtering, complete this linear convolution operation with toeplitz matrix, the centre centre frequency of filter being moved to sub-band is needed for different sub-bands;

Arranging UFMC system mesarcs filter length is L, and the expression formula of impulse response h is: h={h [0], h [1] ..., h [L-1] }, corresponding frequency domain response is H;

The subcarrier index calculating the i-th-1 sub-mid-band frequency place is: by the frequency domain response H translation of prototype filter individual unit, obtains the frequency domain response H of the i-th-1 sub-band _i-1,

The impulse response h of the i-th-1 sub-band _i-1for: h _i-1={ h _i-1[0], h _i-1[1] ..., h _i-1[L-1] }, wherein, $h_{i-1}\left[n\right]=h\left[n\right]e^{\left(\frac{j2\mathrm{\π}{k}_{i-1}n}{N}\right)},0\≤n\≤L-1;$

Definition toeplitz matrix T _i-1for:

$T_{i-1}={\left[\begin{array}{cccccccc}{h}_{i-1}\left[0\right]& 0& 0& \·\·\·& 0& 0& \·\·\·& 0\\ h_{i-1}\left[1\right]& h_{i-1}\left[0\right]& 0& \·\·\·& 0& 0& \·\·\·& 0\\ h_{i-1}\left[2\right]& h_{i-1}\left[1\right]& h_{i-1}\left[0\right]& \·\·\·& 0& 0& \·\·\·& 0\\ \·& \·& \·& & \·& \·& & \·\\ \·& \·& \·& & \·& \·& & \·\\ \·& \·& \·& & \·& \·& & \·\\ h_{i-1}[L-1]& h_{i-1}[L-2]& h_{i-1}[L-3]& \·\·\·& h_{i-1}\left[0\right]& 0& \·\·\·& 0\\ 0& h_{i-1}[L-1]& h_{i-1}[L-2]& \·\·\·& h_{i-1}\left[1\right]& h_{i-1}\left[0\right]& \·\·\·& 0\\ 0& 0& h_{i-1}[L-1]& \·\·\·& h_{i-1}\left[2\right]& h_{i-1}\left[1\right]& \·\·\·& 0\\ 0& 0& 0& & \·& \·& & \·\\ \·& \·& \·& \·\·\·& \·& \·& & \·\\ \·& \·& \·& & \·& \·& & \·\\ \·& \·& \·& \·\·\·& h_{i-1}[L-2]& h_{i-1}[L-3]& \·\·\·& 0\\ 0& 0& 0& \·\·\·& h_{i-1}[L-1]& h_{i-1}[L-2]& \·\·\·& h_{i-1}\left[0\right]\end{array}\right]}_{(N+L-1)\×(N+L-1)}$

Get its front N row to be designated as with realize the filtering of the i-th-1 sub-band;

Consistent with the solution procedure of i-1 sub-band, for i-th sub-band, translation on frequency domain individual unit, obtains frequency domain response H _iwith impulse response h _i, h _iin element be: $h_i\left[n\right]=h\left[n\right]e^{\left(\frac{j2\mathrm{\π}{k}_{i}n}{N}\right)},0\≤n\≤L-1;$ For the i-th+1 sub-band, translation on frequency domain $k_{i+1}=\frac{v6-v5}{2}+v5$ Individual unit, obtains frequency domain response H _i+1with impulse response h _i+1, h _i+1in element be: by h _i-1[n] uses h respectively _i[n] and h _i+1[n] replaces and obtains the toeplitz matrix of i-th sub-band and the i-th+1 sub-band, then gets its front N row respectively, obtains the toeplitz matrix that i-th sub-band and the i-th+1 sub-band realize filtering and is respectively with

Through above-mentioned steps 1), step 2), step 3) and step 4), transmitting terminal send time-domain signal x _totalfor:

x _total＝T·V·X

Wherein, $T=[T_{i-1}^N,T_i^N,T_{i+1}^N],X={[X_{i-1}^T,X_i^T,X_{i+1}^T]}^T,$ V=diag{V _i-1, V _i, V _i+1represent with matrix V _i-1, V _iand V _i+1for the diagonal matrix that diagonal element is formed, V, V _i-1, V _iand V _i+1represent respectively and three sub-bands realize step 4.1) in Fourier's matrix of IFFT conversion and d ₁and d ₂value factor band and different.