CN101335551B - SINR estimation method based on multi-antenna diversity scheme of DFT-S-GMC system - Google Patents

Abstract

The invention relates to an SINR (Signal Interference-to-Noise Ratio) estimation method based on DFT-S-GMC (Discrete Fourier Transform Spread Generalized Multi-Carrier) system multiple antenna diversity schemes. The effective signal interference-to-noise ratio for the DFT-S-GMC multiple antenna diversity scheme can be calculated accurately in a way that: a mathematical model for signal input and output in the DFT-S-GMC system is constructed firstly, then according to channel frequency response, channel noise variance and equilibrium approaches, an equilibrium factor of a frequency-domain equalization sub-carrier in the DFT-S-GMC system is set up; average power of the wanted signal of the DFT-S-GMC system receiving end, average power and noise variance of inter-signal interference are calculated respectively based on the mathematical model and the frequency-domain equilibrium factor; then the effective signal interference-to-noise ratio is calculated. The invention can be used for both mapping interfaces from link-level simulation to system-level simulation for the DFT-S-GMC multiple antenna diversity transmission scheme, and the link adaptation technology and wireless resource scheduling technique, etc. providing support for Adaptive Coding Modulation based on the transmission scheme.

Description

SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme

Technical field

The present invention relates to a kind of SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme.

Background technology

Single-carrier frequency division multiple access (SC-FDMA) is that put forward in the world in recent years a kind of both possessed single carrier communication peak-to-average force ratio characteristic; Possess multi-carrier communication again and realize the novel fdma communication systems of simple and scheduling of resource flexible nature, be mainly used in the up link solution of wide-band mobile communication.At present, SC-FDMA has two kinds of implementations, a kind of SC-FDMA [5-8] that is based on OFDM (A) technology, a kind of SC-FDMA that is based on the bank of filters conversion.SC-FDMA for based on OFDMA technology has two kinds of ways of realization again, and a kind of is the SC-FDMA that handles through frequency domain, just based on the orthogonal frequency division multiplexing multiple access (DFT-S-OFDMA) of DFT spread-spectrum; Another kind is the SC-FDMA that handles through time domain.For the DFT-S-OFDMA system, the data symbol after each subscriber-coded modulation through a DFT conversion than small point (identical with the number of sub carrier wave that distributes), transmits the data map after the conversion earlier then to the subcarrier that distributes.Transmit because DFT-S-OFDMA spreads to each data symbol on the subcarrier of all distribution, make its transmission signals have the characteristic of single-carrier signal.Therefore, compare with the OFDMA system, this system can obviously reduce the transmission signals peak-to-average force ratio.Yet,, therefore also have the multiple access that synchronous error is caused and disturb responsive defective because DFT-S-OFDM also is based on the OFDM transmission.Two kinds of implementation methods of SC-FDMA that time domain is handled.A kind of is that the modulated symbols data block is directly added Cyclic Prefix, through behind the shaping filter, again through the specific frequency spectrum shift of user, realizes frequency division multiple access, and its transmission signals has continuous frequency spectrum; Another kind is that the modulated symbols data block is repeated cascade earlier, adds Cyclic Prefix then.Then through behind the shaping filter, again through the specific frequency spectrum shift of user, realize frequency division multiple access, its transmission signals has discrete spectrum.In fact, adopt the system of this implementation method to be also referred to as the frequency division multiplexing multiple access that interweaves (IFDMA) system.The SC-FDMA that time domain is handled has lower peak-to-average force ratio than DFT-S-OFDMA, but with respect to the DFT-S-OFDMA based on the OFDM technology, its availability of frequency spectrum obviously reduces.In addition, IFDMA disturbs equally very responsive for the multiple access that synchronous error is caused.Based on the SC-FDMA of bank of filters conversion, promptly, similar with DFT-S-OFDM based on generalized multi-carrier (GMC) frequency division multiple access scheme of DFT (DFT), adopt DFT to carry out frequency domain spread spectrum, to reduce the transmission signals peak-to-average force ratio.But different with DFT-S-OFDM is; DFT-S-GMC adopts inverse filterbank conversion (IFBT) to realize frequency division multiplexing and frequency division multiple access; Because the bandwidth of each subband of DFT-S-GMC is bigger with respect to carrier wave frequency deviation and Doppler frequency shift; Have simultaneously certain frequency domain protection between each subband at interval, the frequency spectrum of each subband has precipitous attenuation outside a channel in addition, disturbs between the multi-user that these characteristics make this scheme cause carrier wave frequency deviation and timing error to have stronger robustness.

Except the multiple access jamming performance with robust, the DFT-S-GMC transmission plan also can be supported link adaptation techniques such as frequency domain dispatching and adaptive coding and modulating flexibly.Yet, realize that these technological keys are must be at effective Signal to Interference plus Noise Ratio of the accurate estimating received signal of receiving terminal (for up link, being the base station).And in system of broadband wireless communication, because channel is the fading channel of temporal dispersion, in order to obtain high data transfer rate; Adopt multi-antenna technology, utilize space diversity can obtain better transmission property, and to the DFT-S-GMC system; What use is the space-time block coding scheme and single carrier least mean-square error frequency domain equalization (SC MMSE-FDE) technology of similar Alamouti structure; See also Fig. 1 to Fig. 3, wherein, Fig. 1 is a DFT-S-GMC system transmitter structure sketch map; Fig. 2 is a DFT-S-GMC system receiver structural representation; Fig. 3 supposes that k the modulated symbols that n IFBT conversion imported constantly is ak (n), 0≤k≤K-1 for the emitting structural sketch map that the DFT-S-GMC system divides the collection emission; 0≤n≤D-1, the number of sub-bands that K takies for the active user, D is illustrated in IFBT number of symbols multiplexing in each data block transmitted.Through the conversion of K point discrete Fourier, the output signal does

$A_{k\&prime;}\left(n\right)=\frac{1}{\sqrt{K}}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{a}_k\left(n\right)\exp (-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;kk}\&prime;/K),0\&le;k\&prime;\&le;K-1;0\&le;n\&le;D-1---\left(1\right)$

The subband mapping is transmitted each element map in the DFT spread spectrum output signal sequence to corresponding subband.Mapping mode can be for concentrating mapping and disperseing the mapping dual mode.

For disperseing mapping, mapping is output as

$b_m\left(n\right)=\left\{\begin{array}{c}{A}_{k\&prime;}\left(n\right),m=C+k\&prime;\&times;R;0\&le;m\&le;M-1;0\&le;k\&prime;\&le;K-1;0\&le;n\&le;D-1\\ 0,\mathrm{otherwise}\end{array}\right.---(2-a)$

For concentrating mapping, mapping is output as

$b_m\left(n\right)=\left\{\begin{array}{c}{A}_{k\&prime;}\left(n\right),m=C+k\&prime;;0\&le;m\&le;M-1;0\&le;k\&prime;\&le;K-1;0\&le;n\&le;D-1\\ 0,\mathrm{otherwise}\end{array}\right.---(2-b)$

Wherein, C is specific user's a subband side-play amount, and M is the sub-band sum of system, and R is subband mapping interval.

Through inverse filterbank conversion (IFBT), L centrifugal pump of n IFBT symbol of transmission does

$g_t\left(n\right)=\underset{m=0}{\overset{M-1}{\mathrm{\&Sigma;}}}{b}_m\left(n\right)f_p\left(t\right)\exp (\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;mt}/M),0\&le;t\&le;L-\mathrm{1,0}\&le;n\&le;D-1---\left(3\right)$

F wherein _p(t) be the impulse response of bank of filters prototype filter, this prototype filter satisfies the shift-orthogonal condition

$\underset{t=0}{\overset{L-1}{\mathrm{\&Sigma;}}}{f}_p\left(t\right)f_p^{*}(t-\mathrm{kN})=\left\{\begin{array}{c}1,k=0\\ 0,k\&NotEqual;0\end{array}\right.---\left(4\right)$

Wherein, N is the shift-orthogonal interval of prototype filter, subscript " * " expression conjugation.The inverse filterbank conversion is divided into some subband transmission signals with broad-band channel, and is quasiorthogonal between each subband.Disturb for reducing each intersubband, prototype filter satisfies the quasiorthogonal condition of frequency domain

$\underset{t=0}{\overset{L-1}{\mathrm{\&Sigma;}}}{f}_p\left(t\right)f_p^{*}\left(t\right)\exp [\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;}(m-m\&prime;)t/M]\left\{\begin{array}{c}=1,m=m\&prime;\\ <\mathrm{\&xi;},m\&NotEqual;m\&prime;\end{array}\right.---\left(5\right)$

Wherein ξ is than 1 much little constant, representes the maximum interference between each subband.If the shift-orthogonal of prototype filter N at interval can make to have the certain protection frequency band between each subband, to reduce the interference between the adjacent sub-bands greater than system's sub-band sum M.Prototype filter can adopt root raised cosine filter, and constituting length through the afterbody zero padding is the filter of L, and design L is the integral multiple of system's sub-band sum M, and then IFBT can use the fast algorithm implementation based on FFT.

Subsequently, the shift-orthogonal of pressing prototype filter is N at interval, the displacement IFBT symbol that D length is L that adds up, and it is output as

$s\left(t\right)=\underset{n=0}{\overset{D-1}{\mathrm{\&Sigma;}}}{g}_{t-\mathrm{nN}}\left(n\right),0\&le;t\&le;(D-1)N+L-1---\left(6\right)$

For reducing the interference of intersubband, the transition band of the frequency response of subband should be precipitous as far as possible.At this moment, the corresponding prototype filter coefficient of many Methods of Subband Filter Banks will be very long, thereby the add up signal of output of causing being shifted has very long hangover.If this signal is directly sent, with the availability of frequency spectrum that greatly reduces system.For improving spectrum efficiency, redispatch away after the brachymemma of process waveform earlier through the signal of too much sub-band filter.If directly will clip through the hangover in the signal of too much sub-band filter, then on the one hand can cause distorted signals, cause the spectrum leakage that transmits on the other hand, cause the band of signal to disturb outward.For overcoming above-mentioned defective, the DFT-S-GMC system adopts loop-around data to become block method, promptly earlier the add up length of output of displacement is respectively T for the data sequence of (D-1) N+L is divided into length ₁=(L-N)/2, T ₂=D * N and T ₃=(L-N)/2 three segment data pieces; Then the first segment data piece is added to the afterbody of the second segment data piece, the 3rd segment data piece is added to the stem of the second segment data piece, the data block of acquisition, promptly the live part of S-GMC symbol is the continuous loop-around data pieces of head and the tail.Circulation adds up and is output as

$x\left(t\right)=s(t+T_1)R_{T_2}\left(t\right)+s(t+{\mathrm{<figure-callout\; id="1"\; label="T"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">T</figure-callout>}}_1+T_2)R_{T_3}\left(t\right)+s\left(t\right)R_{T_1}\left(t\right),0\&le;t\&le;N\&times;D-1---\left(7\right)$

Wherein

$R_T\left(t\right)=\left\{\begin{array}{c}1\mathrm{,}0\&le;t\&le;T-1\\ 0,\mathrm{otherwise}\end{array}\right.---\left(8\right)$

At last; The complete S-GMC symbol that the loop-around data piece that generates is constituted; Through shaping filter; The transmission order of diversity that uses two antennas to send similar Alamouti structure in the time at two pieces the signal behind the shaping filter is 2 sets of signals, and before every, adds Cyclic Prefix, concrete coding and to add cyclic prefix scheme following:

If n symbolic representation of last k the transmission block of antenna i (i=1,2) is x _i ^(k)(n).At moment k=0,2,4 ..., reading a pair of length the signal that behind shaping filter, obtains is the piece x of N ₁ ^k(n) and x ₂ ^k(n) (0≤n≤N-1).With reference to the STBC structure of Alamouti, adopted following emission diversity scheme (like Fig. 3):

${x_1}^{(k+1)}\left(n\right)=-{{\stackrel{\&OverBar;}{x}}_2}^{\left(k\right)}\left({(-n)}_N\right))$ With ${x_2}^{(k+1)}\left(n\right)={{\stackrel{\&OverBar;}{x}}_1}^{\left(k\right)}\left({(-n)}_N\right)),---\left(9\right)$

n＝0，1，...，N-1；k＝0，2，4...

Wherein,

(.) _NRepresent complex conjugate and mould N operation respectively.Promptly to launch length in the time at k piece be the data block x of N to first antenna ₁ ^k(n) (0≤n≤N-1), and second antenna to launch length in the time at k piece be the data block x of N ₂ ^k(n) (0≤n≤N-1) is the x of N and in the time block of the k+1 of first antenna, send length ₁ ^(k+1)(n) (0≤n≤N-1), in the time block of the k+1 of second antenna, sending length is the x of N ₂ ^(k+1)(n) (0≤n≤N-1).

In addition, each sends piece front interpolation length is the Cyclic Prefix of v, is used for eliminating IBI (inter-block-interference), and when the time domain channel impulse response remained unchanged within a certain period of time, corresponding frequency domain channel matrix was what circulate.At last, the transmitted power of every antenna is half under the single-antenna case, fixes thereby satisfy total transmitting power.

At last, through digital-to-analogue conversion, baseband signal is upconverted to radio frequency.Through radio frequency sending module and transmitting antenna, transmitter output radiofrequency signal.

Receiving terminal receives the signal process Time and Frequency Synchronization that receives through radio frequency, removes Cyclic Prefix, after the channel estimating operation, carries out frequency domain equalization, wherein mainly comprises:

Sampled signal is to received signal carried out Fourier Tranform;

The Fourier transform that receives signal is carried out linear transformation;

Signal after utilizing the interior estimated channel fading coefficients of two time blocks to the sampled signal Fourier Tranform carries out the single-point equilibrium;

Signal after the equilibrium is carried out contrary Fourier Tranform;

Subsequent to after the equilibrium and what transformed to that the signal of time domain accomplishes is the inverse operation opposite with transmitting terminal, just do not giving unnecessary details in detail here.

From the above mentioned; Multi-antenna diversity transmission plan for the employing of DFT-S-GMC system; In order to support link adaptation techniques; Need be at effective Signal to Interference plus Noise Ratio of the accurate estimating received signal of receiving terminal, therefore accurately estimating the effective Signal to Interference plus Noise Ratio that receives signal has become the technical task that those skilled in the art need to be resolved hurrily.

Summary of the invention

The object of the present invention is to provide a kind of SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme, with realize accurately calculating DFT-S-GMC system multi-antenna diversity scheme Signal to Interference plus Noise Ratio, satisfy the needs of link circuit self-adapting.

In order to achieve the above object, the SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme provided by the invention comprises step: 1) set up the Mathematical Modeling between the signal input and output of said DFT-S-GMC system; 2), set the equalizing coefficient of the frequency domain equalization subcarrier of said DFT-S-GMC system according to channel frequency response, interchannel noise variance and the equalization methods of said DFT-S-GMC system; 3) according to the average power of said Mathematical Modeling and the said DFT-S-GMC system receiving terminal of said frequency domain equalization coefficient calculations useful signal; 4) calculate the average power of disturbing between said receiving end signal according to the average power of said receiving terminal useful signal; 5) according to the noise variance of said Mathematical Modeling and the corresponding noise of the said receiving terminal of said frequency domain equalization coefficient calculations; 6) calculate said SINR according to the average power of disturbing between the average power of said receiving terminal useful signal, said receiving end signal, said receiving terminal noise variance.

Wherein, said DFT-S-GMC system is the system that comprises 2 a plurality of reception antennas of transmitting antenna, and said Mathematical Modeling is:

Wherein, subscript " T " expression transposition, subscript " H " expression conjugate transpose;

Y receives signal frequency-domain to represent,

$Y=\left[\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="Y"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">Y</figure-callout>}}_1\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ Y_{\mathrm{Nr}}\end{array}\right]=\left[\begin{array}{c}{Y_1}^{\left(k\right)}\\ {{\stackrel{\&OverBar;}{Y}}_1}^{(k+1)}\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ {Y_{\mathrm{Nr}}}^{\left(k\right)}\\ {{\stackrel{\&OverBar;}{Y}}_{\mathrm{Nr}}}^{(k+1)}\end{array}\right]=\left[\begin{array}{cc}{\mathrm{\&Lambda;}}_{11}& {\mathrm{\&Lambda;}}_{21}\\ {{\mathrm{\&Lambda;}}_{21}}^{*}& -{{\mathrm{\&Lambda;}}_{11}}^{*}\\ \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;\\ {\mathrm{\&Lambda;}}_{1\mathrm{Nr}}& {\mathrm{\&Lambda;}}_{2\mathrm{Nr}}\\ {{\mathrm{\&Lambda;}}_{2\mathrm{Nr}}}^{*}& -{{\mathrm{\&Lambda;}}_{1\mathrm{Nr}}}^{*}\end{array}\right]\left[\begin{array}{c}{X_1}^{\left(k\right)}\\ {X_2}^{\left(k\right)}\end{array}\right]+\left[\begin{array}{c}{Z_1}^{\left(k\right)}\\ {{\stackrel{\&OverBar;}{Z}}_1}^{(k+1)}\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ {Z_{\mathrm{Nr}}}^{\left(k\right)}\\ {{\stackrel{\&OverBar;}{Z}}_{\mathrm{Nr}}}^{(k+1)}\end{array}\right]$

Z is that variance is σ ²The N point DFT conversion output vector of time domain AWGN noise vector;

D _KFor the length of the generalized multi-carrier transmission system transmitting terminal of said DFT spread spectrum transmission is the modulation symbol vector of K, the number of sub-bands that K also takies for transmitting terminal;

F _KBe K * K DFT spread spectrum matrix, and $F_K{F}_K^H=I_K,$ I _KBe K * K unit matrix;

T _{M, K}Be M * K subband mapping matrix, have only K element in its M * K element for " 1 ", all the other be " 0 ", when hope will pass through k element map that K point FFT conversion exports upload to m subband defeated, then with T _{M, K}The element of the capable k of m row be changed to " 1 ";

Be M point inverse filterbank conversion (IFBT) matrix, wherein, F _MBe M point FFT conversion unitary matrice, and $F_M{F}_M^H=I_M,$ Γ _{L, M}Be the cascade extended matrix of L * M, and Γ _{L, M}=[I _M, I _M..., I _M] ^T, I _MBe the unit matrix of M * M, L is the integral multiple of M,

_LFor L * L is a diagonal matrix, its diagonal element is many Methods of Subband Filter Banks prototype filter L dot factor f _p(t), t=0,1 ..., L-1;

${\mathrm{\&Omega;}}_{N,L}=\left[\begin{array}{c}{I}_L\\ 0_{(N-L)\&times;L}\end{array}\right],$ I _LBe the unit matrix of L * L, 0 _{(N-L) * L}Be (N-L) * L null matrix;

Be N * N diagonal matrix, its diagonal element vector ${[{\widetilde{\mathrm{\&Lambda;}}}_0{\widetilde{\mathrm{\&Lambda;}}}_1\&CenterDot;\&CenterDot;\&CenterDot;{\widetilde{\mathrm{\&Lambda;}}}_m\&CenterDot;\&CenterDot;\&CenterDot;{\widetilde{\mathrm{\&Lambda;}}}_{N-1}]}^T,$ Wherein ${\widetilde{\mathrm{\&Lambda;}}}_m=\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_m^{i,r}\right|}^2,$ And

Represent the channel frequency response of i transmit antennas to m frequency domain equalization subcarrier between the r root reception antenna; W is N * N diagonal matrix, its diagonal element vector [ω ₀ω ₁ω _mω _N-1] ^TBe the frequency domain equalization coefficient;

Be M point bank of filters conversion (FBT) matrix;

Be K * M subband solutions mapping matrix;

Be K * K _IDFTThe despreading matrix.

When said DFT-S-GMC system adopts zero forcing equalization, said step 2) in set the equalizing coefficient ω of m frequency domain equalization subcarrier correspondence _mFor ${\mathrm{\&omega;}}_m=\frac{1}{{\left|{\widetilde{\mathrm{\&Lambda;}}}_m\right|}^2}=\frac{1}{{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_m^{i,r}\right|}^2\right|}^2},$ Wherein,

Be the channel frequency response of i transmit antennas to m frequency domain equalization subcarrier between the r root reception antenna; When said DFT-S-GMC system adopts least mean-square error balanced, then said step 2) sets the equalizing coefficient ω of m frequency domain equalization subcarrier correspondence in _mFor ${\mathrm{\&omega;}}_m=\frac{1}{{\left|{\widetilde{\mathrm{\&Lambda;}}}_m\right|}^2+\frac{1}{\mathrm{SNR}}}=\frac{1}{{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_m^{i,r}\right|}^2\right|}^2+{\mathrm{\&sigma;}}^2},$ Wherein,

Be the channel frequency response of i transmit antennas to m frequency domain equalization subcarrier between the r root reception antenna, σ ²Noise variance for the frequency domain equalization subcarrier.

Said step 3) comprises step:

(1) according to the channel frequency response of i transmit antennas to the balanced subcarrier of multipath channel frequency domain between the r root reception antenna ${[{\mathrm{\&Lambda;}}_0^{i,r}{\mathrm{\&Lambda;}}_1^{i,r}\&CenterDot;\&CenterDot;\&CenterDot;{\mathrm{\&Lambda;}}_m^{i,r}\&CenterDot;\&CenterDot;\&CenterDot;{\mathrm{\&Lambda;}}_{N-1}^{i,r}]}^T,$ I=1,2; R=1,2 ..., Nr calculates diagonal matrix $\widetilde{\mathrm{\&Lambda;}}=\mathrm{Diag}\{\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2,\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2,...,\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\},$ Wherein diag{A} representes with vector A to be the diagonal matrix of diagonal element;

(2) according to the equalizing coefficient [ω that sets ₀ω ₁ω _mω _N-1] ^T, calculate diagonal matrix

${\mathrm{\&Lambda;}}_N=\widetilde{\mathrm{\&Lambda;}}W_H\widetilde{\mathrm{\&Lambda;}}=\mathrm{diag}\{{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_0,{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_1,...,{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_{N-1}\};$

(3) compute matrix $h=F_N^H{\mathrm{\&Lambda;}}_N{F}_N,$ Wherein,

_NBe diagonal matrix, h is a circulation symmetrical matrix, and the h first column element vector does

${\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0=\frac{1}{\sqrt{N}}{F}_N^H{[{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_0{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_1\&CenterDot;\&CenterDot;\&CenterDot;{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_{N-1}]}^T,$ All the other column vectors of h can be by h ₀Cyclic shift obtains;

(4) the capable and preceding L row of the preceding L in the intercepting matrix h upper left corner, the L of formation * L matrix $\tilde{h}={\mathrm{\&Omega;}}_{N,L}^{H}h{\mathrm{\&Omega;}}_{N,L};$

(5) matrix is divided into the block matrix of (L/M) * (L/M); The size of each matrix-block is M * M, and wherein the matrix-block of the capable and preceding j row of i does

H wherein _{I, j}In element

Be matrix

The capable jM column element of iM;

(6) calculate P respectively _ih _{I, j}P _jFirst column vector $b_{i,j}=P_i{h}_{i,\mathrm{<figure-callout\; id="1"\; label="j\; P\; j"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">j</figure-callout>}}{P}_j{}_{}\left[\begin{array}{c}1\\ 0_{(M-1)\&times;1}\end{array}\right],$ P wherein _iBe M * M diagonal matrix, its diagonal element is { f _p(i * M), f _p(i * M+1) ..., f _p(i * M+M-1) }, f _p(t), t=0,1 ..., L-1 is many Methods of Subband Filter Banks prototype filter coefficient, 0 _{(M-1) * 1}Be (M-1) * 1 zero column vector.

(7) stack b _{I, j}, and carry out M point DFT conversion to get column vector $B=\sqrt{M}{F}_M\left\{\underset{i,j=0}{\overset{L/M-1}{\mathrm{\&Sigma;}}}{b}_{i,j}\right\};$

(8), extract and take the signal component on the subband through the subband solutions mapping ${\left[\begin{array}{cccc}{\tilde{H}}_0& {\tilde{H}}_1& \&CenterDot;\&CenterDot;\&CenterDot;& {\tilde{H}}_{K-1}\end{array}\right]}^T=T_{M,K}^{H}B;$

(9) calculate the useful signal average power $E_s^{\&prime;}={\left|\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\tilde{H}}_k\right|}^2.$

The average power calculating formula of disturbing between said receiving end signal is: ${\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="ISI"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">ISI</figure-callout>}}^2=\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\left|{\tilde{H}}_k\right|}^2-{\left|\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\tilde{H}}_k\right|}^2.$

Said step 5) comprises step:

(1) by the channel frequency response of i transmit antennas to the balanced subcarrier of multipath channel frequency domain the r root reception antenna

${\left[\begin{array}{cccccc}{\mathrm{\&Lambda;}}_0^{i,r}& {\mathrm{\&Lambda;}}_1^{i,r}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{\&Lambda;}}_m^{i,r}& \&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{\&Lambda;}}_{N-1}^{i,r}\end{array}\right]}^T,$ I=1,2; R=1,2 ..., Nr calculates diagonal matrix $\widetilde{\mathrm{\&Lambda;}}=\mathrm{Diag}\{\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2,\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2,...,\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\},$ Wherein, diag{A} representes with vector A to be the diagonal matrix of diagonal element;

(2) according to the equalizing coefficient [ω that sets ₀ω ₁ω _mω _N-1] ^T, calculate diagonal matrix:

${\mathrm{\&Lambda;}}_N^{\&prime;}=\widetilde{\mathrm{\&Lambda;}}W_H\widetilde{\mathrm{\&Lambda;}}W\widetilde{\mathrm{\&Lambda;}}=\mathrm{diag}\{{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_0^2,{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_1^2,...,{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_{N-1}^2\};$

(3) compute matrix $h\&prime;=F_N^H{\mathrm{\&Lambda;}}_N^{\&prime;}{F}_N,$ Wherein,

Be diagonal matrix, h ' is a circulation symmetrical matrix, and the h ' first column element vector does ${\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0^{\&prime;}=\frac{1}{\sqrt{N}}{F}_N^H{[{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_0^2{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_1^2\&CenterDot;\&CenterDot;\&CenterDot;{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_{N-1}^2]}^T,$ All the other column vectors of h can by

Cyclic shift obtains;

(4) the capable and preceding L row of the preceding L in intercepting matrix h ' upper left corner, the L of formation * L matrix $\tilde{h}\&prime;={\mathrm{\&Omega;}}_{N,L}^{H}h\&prime;{\mathrm{\&Omega;}}_{N,L};$

(5) with matrix Be divided into the block matrix of (L/M) * (L/M), the size of each matrix-block is M * M, and wherein the matrix-block of the capable and preceding j row of i can be expressed as

H wherein _{I, j}Be matrix

The capable j column element of i;

(6) calculate respectively

First column vector $b_{i,j}^{\&prime;}=P_i{h}_{i,j}^{\&prime;}{\mathrm{<figure-callout\; id="1"\; label="P\; j"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">P</figure-callout>}}_j\left[\begin{array}{c}1\\ 0_{(M-1)\&times;1}\end{array}\right],$ P wherein _iBe M * M diagonal matrix, its diagonal element is { f _p(i * M), f _p(i * M+1) ..., f _p(i * M+M-1) }, f _p(t), t=0,1 ..., L-1 is many Methods of Subband Filter Banks prototype filter coefficient, 0 _{(M-1) * 1}Be (M-1) * 1 zero column vector;

(7) stack

And carry out M point DFT conversion to obtain $B\&prime;=\sqrt{M}{F}_M\left\{\underset{i,j=0}{\overset{L/M-1}{\mathrm{\&Sigma;}}}{b}_{i,j}^{\&prime;}\right\};$

(8), extract and take the noise component(s) on the subband through the subband solutions mapping ${\left[\begin{array}{cccc}{\tilde{H}}_0^{\&prime;}& {\tilde{H}}_1^{\&prime;}& \&CenterDot;\&CenterDot;\&CenterDot;& {\tilde{H}}_{K-1}^{\&prime;}\end{array}\right]}^T=T_{M,K}^{T}B\&prime;;$

(9) calculating noise variance ${\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">n</figure-callout>}}^2=\frac{{\mathrm{\&sigma;}}^2}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\tilde{H}}_k^{\&prime;}.$

$\mathrm{SINR}=\frac{E_s^{\&prime;}}{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">n</figure-callout>}}^2+{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="ISI"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">ISI</figure-callout>}}^2}$

Said SINR does $=\frac{{\left|\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\tilde{H}}_k\right|}^2}{\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\stackrel{\&OverBar;}{H}}_k{\mathrm{\&sigma;}}^2+\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\left|{\tilde{H}}_k\right|}^2-{\left|\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\tilde{H}}_k\right|}^2}.$

In sum; SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme of the present invention is directed against the generalized multi-carrier transmission plan based on the DFT spread spectrum; Effective Signal to Interference plus Noise Ratio (SINR) method of estimation based on the multi-antenna diversity transmission plan of this system has been proposed; Realized effective Signal to Interference plus Noise Ratio (SINR) is estimated, satisfied adaptive needs.

Description of drawings

Fig. 1 is the structural representation of the transmitter of DFT-S-GMC system.

Fig. 2 is the structural representation of the receiver of DFT-S-GMC system.

Fig. 3 divides the emitting structural sketch map of collection emission for the DFT-S-GMC system.

Fig. 4 is imitated Signal to Interference plus Noise Ratio mapping performance sketch map for 1 son of the SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme of the present invention has.

Fig. 5 is imitated Signal to Interference plus Noise Ratio mapping performance sketch map for 8 sons of the SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme of the present invention have.

Embodiment

In the SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme of the present invention; Mainly the DFT-S-GMC system with 2 transmitting antenna Nr reception antennas is that example is elaborated, and said method mainly may further comprise the steps: the first step: set up the Mathematical Modeling between the signal input and output of said DFT-S-GMC system

Can know that by existing DFT-S-GMC transmission plan the data block of each reception is to be added up (multiplexing) and got (seeing 6 formulas) by the displacement of several time domain waveform symbols.But because prototype filter satisfies shift-orthogonal property (seeing 4 formulas), think that each multiplexing time domain waveform symbol is non-interfering in the balanced data piece so can be similar to, this can be able to confirm in subsequent simulation.Therefore, easy for analyzing, only consider a situation that time domain waveform is multiplexing at this.

Suppose that at transmitting terminal, length is the modulation symbol vector D of K _K, can be expressed as:

$D_K=\left[\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="d"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">d</figure-callout>}}_1\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ d_K\end{array}\right]---\left(10\right)$

So D _KThrough K point DFT spread spectrum, the subband mapping, after the M point inverse filterbank conversion (IFBT), forming length is the parallel sequence that L is ordered

Here, F _KBe K * KDFT spread spectrum matrix, and

$F_K=\frac{1}{\sqrt{K}}{\left(\begin{array}{cccc}1& {\mathrm{<figure-callout\; id="0"\; label="W\; K"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">W</figure-callout>}}_K^{}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; id="0"\; label="W\; K"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">W</figure-callout>}}_K^{}\\ 1& {\mathrm{<figure-callout\; id="1"\; label="W\; K"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">W</figure-callout>}}_K^{}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; id="1"\; label="W\; K"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">W</figure-callout>}}_K^{}\\ \&CenterDot;\&CenterDot;\&CenterDot;& \&CenterDot;\&CenterDot;\&CenterDot;& \&CenterDot;\&CenterDot;\&CenterDot;& \&CenterDot;\&CenterDot;\&CenterDot;\\ 1& W_K^{(K-1)\&CenterDot;1}& \&CenterDot;\&CenterDot;\&CenterDot;& W_K^{(K-1)\&CenterDot;(K-1)}\end{array}\right)}_{K\&times;K}---\left(12\right)$

Wherein $W_k^x=\mathrm{Exp}(-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;\; x}/K),$ $F_K^H{F}_K=F_K{F}_K^H=I_K,$ I _KBe the unit matrix of K * K, subscript " H " expression conjugate transpose;

T _{M, K}Be M * K subband mapping matrix, have only K element to be " 1 " in its M * K element, all the other are " 0 ".If hope to pass through k element map that K point FFT conversion exports upload to m subband defeated, then with T _{M, K}The element of the capable k of m row be changed to " 1 ";

is M point inverse filterbank conversion (IFBT) matrix, wherein

F _MBe M point FFT conversion unitary matrice, and $F_M{F}_M^H=I_M;$

Γ _{L, M}Be the cascade extended matrix of L * M, and Γ _{L, M}=[I _M, I _M..., I _M] ^T, I _MBe the unit matrix of M * M, L is the integral multiple of M;

_LFor L * L is a diagonal matrix, its diagonal element is many Methods of Subband Filter Banks prototype filter L dot factor f _p(t), t=0,1 ..., L-1.

Because only considering a situation that time domain waveform is multiplexing, the data vector of output is that N-L of IFBT conversion dateout vector afterbody interpolation is zero, is the parallel data vector of N to form length

Wherein, ${\mathrm{\&Omega;}}_{N,L}=\left[\begin{array}{c}{I}_L\\ 0_{(N-L)\&times;L}\end{array}\right],$ I _LBe the unit matrix of L * L, 0 _{(N-L) * L}Be (N-L) * L null matrix.Final nucleotide sequence x is emission output after (9) formula is carried out Space Time Coding and added Cyclic Prefix through the encoding scheme that class Aloumouti sends structure.

Through after the multipath channel, at receiving terminal, at first the data that receive are removed Cyclic Prefix after, for reception antenna r (r=1,2 ..., Nr)

${y_1}^{\left(j\right)}=H_{1r}^{\left(j\right)}{x}_1^{\left(j\right)}+H_{2r}^{\left(j\right)}{x}_2^{\left(j\right)}+z_r^{\left(j\right)},j=k,k+1---\left(14\right)$

Wherein, H _1r ^(j)And H _2r ^(j)Being respectively piece j goes up from transmitting antenna ₁With ₂To reception antenna _rChannel matrix.To y _r ^(j)Pass through N point DFT conversion, obtain the frequency domain data vector

Because $H=F_N^H\mathrm{\&Lambda;}{F}_N,$ So can be expressed as:

${Y_r}^{\left(j\right)}=F_N{y_r}^{\left(j\right)}={\mathrm{\&Lambda;}}_{1r}^{\left(j\right)}{X}_1^{\left(j\right)}+{\mathrm{\&Lambda;}}_{2r}^{\left(j\right)}{X}_2^{\left(j\right)}+Z_r^{\left(j\right)}---\left(15\right)$

Wherein, ${X_i}^{\left(j\right)}=F_N{x_i}^{\left(j\right)},i=\mathrm{1,2},Z_r^{\left(j\right)}=F_N{z}_r^{\left(j\right)},$ F _NBe N point D _F ^TTransformation matrix, and $F_N{F}_N^H=I_N,$ IN is the unit matrix of N * N,

Be channel frequency response;

Because send the coding criterion of structure, can get:

X ₁ ^(k+1)(n)=-X ₂ ^(k)(n) and X ₂ ^(k+1)(n)=X ₁ ^(k)(n), n=0,1 ..., N-1; K=0,2,4... (16)

In slow time varying channel, suppose that channel matrix is constant in two continuous blocks, that is:

${H_{\mathrm{ir}}}^{(k+1)}={H_{\mathrm{ir}}}^{\left(k\right)}=H_{\mathrm{ir}},i=\mathrm{1,2}\&DoubleLeftRightArrow;{{\mathrm{\&Lambda;}}_{\mathrm{ir}}}^{(k+1)}={{\mathrm{\&Lambda;}}_{\mathrm{ir}}}^{\left(k\right)}={\mathrm{\&Lambda;}}_{\mathrm{ir}},i=\mathrm{1,2}---\left(17\right)$

In conjunction with (15)-(17) formula, obtain:

$Y_r=\left[\begin{array}{c}{Y_r}^{\left(k\right)}\\ {{\stackrel{\&OverBar;}{Y}}_r}^{(k+1)}\end{array}\right]=\left[\begin{array}{cc}{\mathrm{\&Lambda;}}_{1r}& {\mathrm{\&Lambda;}}_{2r}\\ {{\mathrm{\&Lambda;}}_{2r}}^{*}& -{{\mathrm{\&Lambda;}}_{1r}}^{*}\end{array}\right]\left[\begin{array}{c}{X_1}^{\left(k\right)}\\ {X_2}^{\left(k\right)}\end{array}\right]+\left[\begin{array}{c}{Z_r}^{\left(k\right)}\\ {{\stackrel{\&OverBar;}{Z}}_r}^{(k+1)}\end{array}\right]---\left(18\right)$

So represent altogether just can be write as following formula for Nr root reception antenna:

$Y=\left[\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="Y"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">Y</figure-callout>}}_1\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ Y_{\mathrm{Nr}}\end{array}\right]=\left[\begin{array}{c}{Y_1}^{\left(k\right)}\\ {{\stackrel{\&OverBar;}{Y}}_1}^{(k+1)}\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ {Y_{\mathrm{Nr}}}^{\left(k\right)}\\ {{\stackrel{\&OverBar;}{Y}}_{\mathrm{Nr}}}^{(k+1)}\end{array}\right]=\left[\begin{array}{cc}{\mathrm{\&Lambda;}}_{11}& {\mathrm{\&Lambda;}}_{21}\\ {{\mathrm{\&Lambda;}}_{21}}^{*}& -{{\mathrm{\&Lambda;}}_{11}}^{*}\\ \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;\\ {\mathrm{\&Lambda;}}_{1\mathrm{Nr}}& {\mathrm{\&Lambda;}}_{2\mathrm{Nr}}\\ {{\mathrm{\&Lambda;}}_{2\mathrm{Nr}}}^{*}& -{{\mathrm{\&Lambda;}}_{1\mathrm{Nr}}}^{*}\end{array}\right]\left[\begin{array}{c}{X_1}^{\left(k\right)}\\ {X_2}^{\left(k\right)}\end{array}\right]+\left[\begin{array}{c}{Z_1}^{\left(k\right)}\\ {{\stackrel{\&OverBar;}{Z}}_1}^{(k+1)}\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ {Z_{\mathrm{Nr}}}^{\left(k\right)}\\ {{\stackrel{\&OverBar;}{Z}}_{\mathrm{Nr}}}^{(k+1)}\end{array}\right]$

Wherein Z is that variance is σ ²The N point DFT conversion output vector of time domain AWGN noise vector.

So we can multiply by the following formula both sides

_HDecomposite symbol X ₁ ^(k)And X ₂ ^(k), that is:

$\tilde{Y}=\left[\begin{array}{c}{\tilde{Y}}_1\\ {\tilde{Y}}_2\end{array}\right]={\mathrm{\&Lambda;}}^{H}Y=\left[\begin{array}{cc}\widetilde{\mathrm{\&Lambda;}}& 0\\ 0& \widetilde{\mathrm{\&Lambda;}}\end{array}\right]\left[\begin{array}{c}{X_1}^{\left(k\right)}\\ {X_2}^{\left(k\right)}\end{array}\right]+Z\&prime;---\left(20\right)$

Wherein

Be N * N diagonal matrix, its diagonal element vector ${[{\widetilde{\mathrm{\&Lambda;}}}_0{\widetilde{\mathrm{\&Lambda;}}}_1\&CenterDot;\&CenterDot;\&CenterDot;{\widetilde{\mathrm{\&Lambda;}}}_m\&CenterDot;\&CenterDot;\&CenterDot;{\widetilde{\mathrm{\&Lambda;}}}_{N-1}]}^T,$ Wherein ${\widetilde{\mathrm{\&Lambda;}}}_m=\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_m^{i,r}\right|}^2,$ And

Represent the channel frequency response of i transmit antennas to m frequency domain equalization subcarrier between the r root reception antenna.Channel frequency response can be estimated to obtain in the system of reality through pilot tone, and in emulation, thinks known fully, like this $\left[\begin{array}{c}{X_1}^{\left(k\right)}\\ {X_2}^{\left(k\right)}\end{array}\right]$ Just can in frequency domain, recover to come out through the MMSE method.

Through behind the frequency domain equalization, can get the time domain data vector

$Y\&prime;=F_N^H\widetilde{\mathrm{\&Lambda;}}W_H\tilde{Y}---\left(21\right)$

Wherein, W is N * N diagonal matrix, its diagonal element vector [ω ₀ω ₁ω _mω _N-1] ^TBe the frequency domain equalization coefficient.To the data vector behind the frequency domain equalization, L point data before the first intercepting, and after this L point data carried out M point bank of filters conversion (FBT), the subband solutions mapping was passed through K point IDFT despreading, the K point data symbolic vector that can estimate again

First is useful signal and intersymbol interference component in the following formula, and second is noise component(s).

And it is (17) in the formula, corresponding with transmitting terminal

is M point bank of filters conversion (FBT) matrix;

is K * M subband solutions mapping matrix;

is K * KIDFT despreading matrix.

Second step:, set the equalizing coefficient of the frequency domain equalization subcarrier of said DFT-S-GMC system according to channel frequency response, interchannel noise variance and the equalization methods of said DFT-S-GMC system

For compeling zero (ZF) equilibrium, m the equalizing coefficient ω that the frequency domain equalization subcarrier is corresponding _mCan be made as

${\mathrm{\&omega;}}_m=\frac{1}{{\left|{\widetilde{\mathrm{\&Lambda;}}}_m\right|}^2}=\frac{1}{{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_m^{i,r}\right|}^2\right|}^2}---\left(22\right)$

For least mean-square error (MMSE) equilibrium, m the equalizing coefficient ω that the frequency domain equalization subcarrier is corresponding _mCan be made as

${\mathrm{\&omega;}}_m=\frac{1}{{\left|{\widetilde{\mathrm{\&Lambda;}}}_m\right|}^2+\frac{1}{\mathrm{SNR}}}=\frac{1}{{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_m^{i,r}\right|}^2\right|}^2+{\mathrm{\&sigma;}}^2}---\left(23\right)$

Wherein, σ ²Noise variance for the frequency domain equalization subcarrier.

The 3rd step: the average power according to said Mathematical Modeling and the said DFT-S-GMC system receiving terminal of said frequency domain equalization coefficient calculations useful signal can be got by first of (24) formula

It is following that the useful signal average power is calculated step:

(1) by the channel frequency response of i transmit antennas to the balanced subcarrier of multipath channel frequency domain the r root reception antenna

${[{\mathrm{\&Lambda;}}_0^{i,r}{\mathrm{\&Lambda;}}_1^{i,r}\&CenterDot;\&CenterDot;\&CenterDot;{\mathrm{\&Lambda;}}_m^{i,r}\&CenterDot;\&CenterDot;\&CenterDot;{\mathrm{\&Lambda;}}_{N-1}^{i,r}]}^T,i=\mathrm{1,2};r=\mathrm{1,2},...,\mathrm{Nr},$ Calculate diagonal matrix

$\widetilde{\mathrm{\&Lambda;}}=\mathrm{diag}\{\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2,\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2,...,\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\}---\left(26\right)$

Diag{A} representes with vector A to be the diagonal matrix of diagonal element.

(2) by the corresponding equalizing coefficient [ω of frequency domain equalization subcarrier ₀ω ₁ω _mω _N-1] ^T, calculate diagonal matrix

${\mathrm{\&Lambda;}}_N=\widetilde{\mathrm{\&Lambda;}}W_H\widetilde{\mathrm{\&Lambda;}}=\mathrm{diag}\{{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_0,{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_1,...,{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_{N-1}\}---\left(27\right)$

(3) compute matrix $h=F_N^H{\mathrm{\&Lambda;}}_N{F}_N---\left(28\right)$

Because

_NBe diagonal matrix, so h is a circulation symmetrical matrix, and the h first column element vector does

${\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0=\frac{1}{\sqrt{N}}{F}_N^H{[{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_0{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_1\&CenterDot;\&CenterDot;\&CenterDot;{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_{N-1}]}^T$

All the other column vectors of h can be by h ₀Cyclic shift obtains.

(4) the capable and preceding L row of the preceding L in the intercepting matrix h upper left corner, the L of formation * L matrix

$\tilde{h}={\mathrm{\&Omega;}}_{N,L}^{H}h{\mathrm{\&Omega;}}_{N,L}---\left(29\right)$

(5) matrix

is divided into the block matrix of (L/M) * (L/M); The size of each matrix-block is M * M, and wherein the matrix-block of the capable and preceding j row of i can be expressed as

H wherein _{I, j}In element

Be matrix

The capable jM column element of iM;

(6) calculate P respectively _ih _{I, j}P _jFirst column vector

$b_{i,j}=P_i{h}_{i,\mathrm{<figure-callout\; id="1"\; label="j\; P\; j"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">j</figure-callout>}}{P}_j{}_{}\left[\begin{array}{c}1\\ 0_{(M-1)\&times;1}\end{array}\right]---\left(31\right)$

P wherein _iBe M * M diagonal matrix, its diagonal element is { f _p(i * M), f _p(i * M+1) ..., f _p(i * M+M-1) }, f _p(t), t=0,1 ..., L-1 is many Methods of Subband Filter Banks prototype filter coefficient, 0 _{(M-1) * 1}Be (M-1) * 1 zero column vector.

(7) stack b _{I, j}, and carry out M point DFT conversion, can get column vector

$B=\sqrt{M}{F}_M\left\{\underset{i,j=0}{\overset{L/M-1}{\mathrm{\&Sigma;}}}{b}_{i,j}\right\}---\left(32\right)$

(8), extract and take the signal component on the subband through the subband solutions mapping

${\left[\begin{array}{cccc}{\tilde{H}}_0& {\tilde{H}}_1& \&CenterDot;\&CenterDot;\&CenterDot;& {\tilde{H}}_{K-1}\end{array}\right]}^T=T_{M,K}^{H}B---\left(33\right)$

(9) calculate the useful signal average power

$E_s^{\&prime;}={\left|\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\tilde{H}}_k\right|}^2---\left(34\right)$

In fact; Matrix can be approximately circular matrix; Then

is M * M diagonal matrix; Its diagonal element vector is identical with B, and

$T_{M,K}^T{F}_M\left\{\underset{i,j=0}{\overset{L/M-1}{\mathrm{\&Sigma;}}}{P}_i{h}_{i,j}{P}_j\right\}{F}_M^H{T}_{M,K}=\mathrm{diag}({\tilde{H}}_0{\tilde{H}}_1,\&CenterDot;\&CenterDot;\&CenterDot;,{\tilde{H}}_{K-1})={\mathrm{\&Lambda;}}_K---\left(35\right)$

Can know that by DFT conversion character

is circular matrix.Order

$F_K^H{\mathrm{\&Lambda;}}_K{F}_K=\left(\begin{array}{cccc}{h}_0& h_{K-1}& \&CenterDot;\&CenterDot;\&CenterDot;& h_1\\ h_1& {\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; id="2"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_2\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;\&CenterDot;\&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ h_{K-1}& h_{K-2}& \&CenterDot;\&CenterDot;\&CenterDot;& h_0\end{array}\right)---\left(36\right)$

Wherein

${\left[\begin{array}{cccc}{h}_0& {\mathrm{<figure-callout\; id="1"\; label="h"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">h</figure-callout>}}_1& \&CenterDot;\&CenterDot;\&CenterDot;& h_{K-1}\end{array}\right]}^T=\frac{1}{\sqrt{K}}{F}_K^H{\left[\begin{array}{cccc}{\tilde{H}}_0& {\tilde{H}}_1& \&CenterDot;\&CenterDot;\&CenterDot;& {\tilde{H}}_{K-1}\end{array}\right]}^T---\left(37\right)$

Like this, (18) formula can be expressed as

$D\&prime;=\left(\begin{array}{cccc}{h}_0& h_{K-1}& \&CenterDot;\&CenterDot;\&CenterDot;& h_1\\ h_1& {\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; id="2"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_2\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;\&CenterDot;\&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ h_{K-1}& h_K& \&CenterDot;\&CenterDot;\&CenterDot;& h_0\end{array}\right)\left[\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="d"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">d</figure-callout>}}_1\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ d_K\end{array}\right]=\left[\begin{array}{c}{h}_0{\mathrm{<figure-callout\; id="1"\; label="d"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">d</figure-callout>}}_1+{\mathrm{ISI}}_1\\ h_0{\mathrm{<figure-callout\; id="2"\; label="d"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">d</figure-callout>}}_2+{\mathrm{<figure-callout\; id="2"\; label="ISI"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">ISI</figure-callout>}}_2\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ h_0{d}_K+{\mathrm{ISI}}_K\end{array}\right]---\left(38\right)$

The energy normalization of each modulation symbol element, i.e. its average power E in the signal phasor of supposing to launch _s=E [| d _k| ²]=1, k=1,2 ..., K, like this, the average energy of useful signal does

$E_s^{\&prime;}={\left|{\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0\right|}^2{E}_s={\left|{\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0\right|}^2---\left(39\right)$

By (29) Shi Kede

${\left|{\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0\right|}^2={\left|\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\tilde{H}}_k\right|}^2---\left(40\right)$

The 4th step: calculate the average power of disturbing between said receiving end signal according to the average power of said receiving terminal useful signal, the intersymbol interference average power can be estimated as

${\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="ISI"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">ISI</figure-callout>}}^2=\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\left|{\tilde{H}}_k\right|}^2-{\left|\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\tilde{H}}_k\right|}^2---\left(41\right)$

In fact, can know by (30) formula, for independent same distribution, energy normalized modulation symbol vector D _K, interference components is identical between the average symbol on all demodulation symbols, and the intersymbol interference energy does

${\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="ISI"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">ISI</figure-callout>}}^2={\left|{\mathrm{<figure-callout\; id="1"\; label="h"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">h</figure-callout>}}_1\right|}^2+{\left|{\mathrm{<figure-callout\; id="2"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_2\right|}^2+\&CenterDot;\&CenterDot;\&CenterDot;{\left|h_{K-1}\right|}^2---\left(42\right)$

Again by (29) Shi Kede

${\left|{\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0\right|}^2+{\left|{\mathrm{<figure-callout\; id="1"\; label="h"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">h</figure-callout>}}_1\right|}^2+{\left|{\mathrm{<figure-callout\; id="2"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_2\right|}^2\&CenterDot;\&CenterDot;\&CenterDot;{\left|h_{K-1}\right|}^2=\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\left|{\tilde{H}}_k\right|}^2---\left(43\right)$

Promptly

${\left|{\mathrm{<figure-callout\; id="1"\; label="h"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">h</figure-callout>}}_1\right|}^2+{\left|{\mathrm{<figure-callout\; id="2"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_2\right|}^2+\&CenterDot;\&CenterDot;\&CenterDot;{\left|h_{K-1}\right|}^2=\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\left|{\tilde{H}}_k\right|}^2-{\left|{\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0\right|}^2---\left(44\right)$

The 5th step: according to the noise variance of said Mathematical Modeling and the corresponding noise of the said receiving terminal of said frequency domain equalization coefficient calculations

Wherein z is the time domain noise vector, and its variance is σ ²

So the noise vector covariance matrix does

The noise variance calculation procedure is following:

(1) by the channel frequency response of i transmit antennas to the balanced subcarrier of multipath channel frequency domain the r root reception antenna

${[{\mathrm{\&Lambda;}}_0^{i,r}{\mathrm{\&Lambda;}}_1^{i,r}\&CenterDot;\&CenterDot;\&CenterDot;{\mathrm{\&Lambda;}}_m^{i,r}\&CenterDot;\&CenterDot;\&CenterDot;{\mathrm{\&Lambda;}}_{N-1}^{i,r}]}^T,i=\mathrm{1,2};r=\mathrm{1,2},...,\mathrm{Nr},$ Calculate diagonal matrix

$\widetilde{\mathrm{\&Lambda;}}=\mathrm{diag}\{\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2,\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2,...,\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\}---\left(47\right)$

Diag{A} representes with vector A to be the diagonal matrix of diagonal element.

(2) by the corresponding equalizing coefficient [ω of frequency domain equalization subcarrier ₀ω ₁ω _mω _N-1] ^T, calculate diagonal matrix

${\mathrm{\&Lambda;}}_N^{\&prime;}=\widetilde{\mathrm{\&Lambda;}}W_H\widetilde{\mathrm{\&Lambda;}}W\widetilde{\mathrm{\&Lambda;}}=\mathrm{diag}\{{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_0^2,{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_1^2,...,{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_{N-1}^2\}---\left(48\right)$

(3) compute matrix

$h\&prime;=F_N^H{\mathrm{\&Lambda;}}_N^{\&prime;}{F}_N---\left(49\right)$

Because

is diagonal matrix; So h ' is a circulation symmetrical matrix, and the h ' first column element vector does

${\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0^{\&prime;}=\frac{1}{\sqrt{N}}{F}_N^H{[{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_0^2{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_1^2\&CenterDot;\&CenterDot;\&CenterDot;{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_{N-1}^2]}^T$

All the other column vectors of h ' can be obtained by

cyclic shift.

(4) the capable and preceding L row of the preceding L in intercepting matrix h ' upper left corner, the L of formation * L matrix

$\tilde{h}\&prime;={\mathrm{\&Omega;}}_{N,L}^{H}h\&prime;{\mathrm{\&Omega;}}_{N,L}---\left(50\right)$

(5) matrix

is divided into the block matrix of (L/M) * (L/M); The size of each matrix-block is M * M, and wherein the matrix-block of the capable and preceding j row of i can be expressed as

Where is the matrix

i-th row and j column element;

(6) calculate

first column vector respectively

$b_{i,j}^{\&prime;}=P_i{h}_{i,j}^{\&prime;}{\mathrm{<figure-callout\; id="1"\; label="P\; j"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">P</figure-callout>}}_j\left[\begin{array}{c}1\\ 0_{(M-1)\&times;1}\end{array}\right]---\left(52\right)$

P wherein _iBe M * M diagonal matrix, its diagonal element is { f _p(i * M), f _p(i * M+1) ..., f _p(i * M+M-1) }, f _p(t), t=0,1 ..., L-1 is many Methods of Subband Filter Banks prototype filter coefficient, 0 _{(M-1) * 1}Be (M-1) * 1 zero column vector.

(7) stack

and carry out M point DFT conversion can get

$B\&prime;=\sqrt{M}{F}_M\left\{\underset{i,j=0}{\overset{L/M-1}{\mathrm{\&Sigma;}}}{b}_{i,j}^{\&prime;}\right\}---\left(53\right)$

(8), extract and take the noise component(s) on the subband through the subband solutions mapping

${\left[\begin{array}{cccc}{\tilde{H}}_0^{\&prime;}& {\tilde{H}}_1^{\&prime;}& \&CenterDot;\&CenterDot;\&CenterDot;& {\tilde{H}}_{K-1}^{\&prime;}\end{array}\right]}^T=T_{M,K}^{T}B\&prime;---\left(54\right)$

(9) calculating noise variance

${\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">n</figure-callout>}}^2=\frac{{\mathrm{\&sigma;}}^2}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\tilde{H}}_k^{\&prime;}---\left(55\right)$

In fact; Matrix

can be approximately circular matrix; Then

is M * M diagonal matrix; Its diagonal element vector is identical with B ', and

$T_{M,K}^T{F}_M\left\{\underset{i,j=0}{\overset{L/M-1}{\mathrm{\&Sigma;}}}{P}_i{h}_{i,j}^{\&prime;}{P}_j\right\}{F}_M^H{T}_{M,K}=\mathrm{diag}({\tilde{H}}_0^{\&prime;},{\tilde{H}}_1^{\&prime;},\&CenterDot;\&CenterDot;\&CenterDot;,{\tilde{H}}_{K-1}^{\&prime;})={\mathrm{\&Lambda;}}_K^{\&prime;}---\left(56\right)$

Can know that by DFT conversion character

is circular matrix.Order

$F_K^H{\mathrm{\&Lambda;}}_K^{\&prime;}{F}_K=\left(\begin{array}{cccc}{\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0^{\&prime;}& h_{K-1}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; id="1"\; label="h"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">h</figure-callout>}}_1^{\&prime;}\\ {\mathrm{<figure-callout\; id="1"\; label="h"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">h</figure-callout>}}_1^{\&prime;}& {\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0^{\&prime;}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; id="2"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_2^{\&prime;}\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;\&CenterDot;\&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ h_{K-1}^{\&prime;}& h_{K-2}^{\&prime;}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0^{\&prime;}\end{array}\right)---\left(57\right)$

Wherein ${\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0^{\&prime;}=\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\tilde{H}}_k^{\&prime;}.$

Like this

$E\left(\tilde{Z}{\tilde{Z}}^H\right)={\mathrm{\&sigma;}}^2\left(\begin{array}{cccc}{\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0^{\&prime;}& h_{K-1}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; id="1"\; label="h"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">h</figure-callout>}}_1^{\&prime;}\\ {\mathrm{<figure-callout\; id="1"\; label="h"\; filenames="S07142978420070731E000012.png,S07142978420070731E000013.png"\; state="\{\{state\}\}">h</figure-callout>}}_1^{\&prime;}& {\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0^{\&prime;}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; id="2"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_2^{\&prime;}\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;\&CenterDot;\&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ h_{K-1}^{\&prime;}& h_{K-2}^{\&prime;}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">h</figure-callout>}}_0^{\&prime;}\end{array}\right)---\left(58\right)$

Noise vector covariance matrix diagonal element is noise variance.

The 6th step: calculate said SINR according to the average power of disturbing between the average power of said receiving terminal useful signal, said receiving end signal, said receiving terminal noise variance, effectively the SINR expression formula does

$\mathrm{SINR}=\frac{E_s^{\&prime;}}{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">n</figure-callout>}}^2+{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="ISI"\; filenames="S07142978420070731E000013.png,S07142978420070731E000022.png"\; state="\{\{state\}\}">ISI</figure-callout>}}^2}$

$=\frac{{\left|\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\tilde{H}}_k\right|}^2}{\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\stackrel{\&OverBar;}{H}}_k{\mathrm{\&sigma;}}^2+\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\left|{\tilde{H}}_k\right|}^2-{\left|\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\tilde{H}}_k\right|}^2}.$

According to the simulation parameter of table 1 the SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme of the present invention is carried out the resulting result of emulation and see also Fig. 4 and Fig. 5; It has compared difference and has taken number of sub-bands (1 and 8); Adopt the branch collection transmission plan of two transmitting antennas and two reception antennas (2x2) configuration, DFT-S-GMC under the PB-3km/h channel to the Block Error Rate (BLER) and its BLER performance under additive white Gaussian noise (AWGN) channel of effective signal-to-noise ratio (effectively Eb/NO) to signal to noise ratio (Eb/NO).The effective Signal to Interference plus Noise Ratio that need to prove the DFT-S-GMC multi-antenna diversity scheme system that the inventive method is estimated is equivalent to the symbol signal to noise ratio, that is, and and Es/NO.Therefore can directly obtain corresponding bit signal to noise ratio Eb/NO by modulation coding mode.Can know by simulation result; The effective Signal to Interference plus Noise Ratio of DFT-S-GMC multi-antenna diversity scheme system that utilizes the inventive method to estimate; Performance curve under its multipath channel can mate its performance curve under the white Gaussian noise channel well, and both signal to noise ratio errors are about about 0.1dB.

Table 1 simulation system parameters

Claims (8)

1. SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme, wherein, the DFT-S-GMC system is the generalized multi-carrier transmission system of DFT spread spectrum, SINR is effective Signal to Interference plus Noise Ratio, it is characterized in that comprising step:

1) sets up Mathematical Modeling between the signal input and output of said DFT-S-GMC system;

2), set the equalizing coefficient of the frequency domain equalization subcarrier of said DFT-S-GMC system according to channel frequency response, interchannel noise variance and the equalization methods of said DFT-S-GMC system;

3) according to the average power of said Mathematical Modeling and the said DFT-S-GMC system receiving terminal of said frequency domain equalization coefficient calculations useful signal;

4) calculate the average power of disturbing between said receiving end signal according to the average power of said receiving terminal useful signal;

5) according to the noise variance of said Mathematical Modeling and the corresponding noise of the said receiving terminal of said frequency domain equalization coefficient calculations;

6) calculate said SINR according to the average power of disturbing between the average power of said receiving terminal useful signal, said receiving end signal, said receiving terminal noise variance;

Said Mathematical Modeling is:

$\hat{D}=F_K^H{T}_{M,K}^T{F}_M{\mathrm{\&Gamma;}}_{L,M}^T{\mathrm{\&gamma;}}_L^H{\mathrm{\&Omega;}}_{N,L}^H{F}_N^H{\widetilde{\mathrm{\&Lambda;}}}^{H}W{\mathrm{\&Lambda;}}^{H}Y$

$=F_K^H{T}_{M,K}^T{F}_M{\mathrm{\&Gamma;}}_{L,M}^T{\mathrm{\&gamma;}}_L^H{\mathrm{\&Omega;}}_{N,L}^H{F}_N^H{\widetilde{\mathrm{\&Lambda;}}}^H{\mathrm{W\&Lambda;}}^H({\mathrm{\&Lambda;F}}_N{\mathrm{\&Omega;}}_{N,L}{\mathrm{\&gamma;}}_L{\mathrm{\&Gamma;}}_{L,M}{F}_M^H{T}_{M,K}{F}_K{D}_K+Z)$

$=F_K^H{T}_{M,K}^T{F}_M{\mathrm{\&Gamma;}}_{L,M}^T{\mathrm{\&gamma;}}_L^H{\mathrm{\&Omega;}}_{N,L}^H{F}_N^H{\widetilde{\mathrm{\&Lambda;}}}^{H}W\widetilde{\mathrm{\&Lambda;}}{F}_N{\mathrm{\&Omega;}}_{N,L}{\mathrm{\&gamma;}}_L{\mathrm{\&Gamma;}}_{L,M}{F}_M^H{T}_{M,K}{F}_K{D}_K$

$+F_K^H{T}_{M,K}^T{F}_M{\mathrm{\&Gamma;}}_{L,M}^T{\mathrm{\&gamma;}}_L^H{\mathrm{\&Omega;}}_{N,L}^H{F}_N^H{\widetilde{\mathrm{\&Lambda;}}}^{H}W{\mathrm{\&Lambda;}}^{H}Z$

Wherein, subscript " T " expression transposition, subscript " H " expression conjugate transpose;

Y receives signal frequency-domain to represent,

$Y=\left[\begin{array}{c}{Y}_1\\ .\\ .\\ .\\ Y_{\mathrm{Nr}}\end{array}\right]=\left[\begin{array}{c}{Y}_1^{\left(k\right)}\\ {\stackrel{\&OverBar;}{Y}}_1^{(k+1)}\\ .\\ .\\ .\\ {Y_{\mathrm{Nr}}}^{\left(k\right)}\\ {{\stackrel{\&OverBar;}{Y}}_{\mathrm{Nr}}}^{(k+1)}\end{array}\right]=\left[\begin{array}{cc}{\mathrm{\&Lambda;}}_{11}& {\mathrm{\&Lambda;}}_{21}\\ {{\mathrm{\&Lambda;}}_{21}}^{*}& -{{\mathrm{\&Lambda;}}_{11}}^{*}\\ .& .\\ .& .\\ .& .\\ {\mathrm{\&Lambda;}}_{1\mathrm{Nr}}& {\mathrm{\&Lambda;}}_{2\mathrm{Nr}}\\ {{\mathrm{\&Lambda;}}_{2\mathrm{Nr}}}^{*}& {{-\mathrm{\&Lambda;}}_{1\mathrm{Nr}}}^{*}\end{array}\right]\left[\begin{array}{c}{X_1}^{\left(k\right)}\\ {X_2}^{\left(k\right)}\end{array}\right]+\left[\begin{array}{c}{Z}_1^{\left(k\right)}\\ {\stackrel{\&OverBar;}{Z}}_1^{(k+1)}\\ .\\ .\\ .\\ {Z_{\mathrm{Nr}}}^{\left(k\right)}\\ {{\stackrel{\&OverBar;}{Z}}_{\mathrm{Nr}}}^{(k+1)}\end{array}\right]$

$=\mathrm{\&Lambda;X}+Z={\mathrm{\&Lambda;F}}_N{\mathrm{\&Omega;}}_{N,L}{\mathrm{\&gamma;}}_L{\mathrm{\&Gamma;}}_{L,M}{F}_M^H{T}_{M,K}{F}_K{D}_K+Z$

Z is that variance is σ ²The N point DFT conversion output vector of time domain AWGN noise vector;

D _KFor the length of the generalized multi-carrier transmission system transmitting terminal of said DFT spread spectrum transmission is the modulation symbol vector of K, the number of sub-bands that K also takies for transmitting terminal;

F _KBe K * KDFT spread spectrum matrix, and

I _KBe K * K unit matrix;

T _{M, K}Be M * K subband mapping matrix, have only K element in its M * K element for " 1 ", all the other be " 0 ", when hope will pass through k element map that K point FFT conversion exports upload to m subband defeated, then with T _{M, K}The element of the capable k of m row be changed to " 1 ";

Be M point inverse filterbank conversion (IFBT) matrix, wherein, F _MBe M point FFT conversion unitary matrice, and

Γ _{L, M}Be the cascade extended matrix of L * M, and Γ _{L, M}=[I _M, I _M..., I _M] ^T, I _MBe the unit matrix of M * M, L is the integral multiple of M, γ _LFor L * L is a diagonal matrix, its diagonal element is many Methods of Subband Filter Banks prototype filter L dot factor f _p(t), t=0,1 ..., L-1;

${\mathrm{\&Omega;}}_{N,L}=\left[\begin{array}{c}{I}_L\\ 0_{(N-L)\&times;L}\end{array}\right],$ I _LBe the unit matrix of L * L, 0 _{(N-L) * L}Be (N-L) * L null matrix;

is N * N diagonal matrix, its diagonal element vector

wherein and

expression i transmit antennas to the channel frequency response of m frequency domain equalization subcarrier between the r root reception antenna;

W is N * N diagonal matrix, its diagonal element vector [ω ₀ω ₁ω _mω _N-1] ^TBe the frequency domain equalization coefficient;

Be M point bank of filters conversion (FBT) matrix;

is K * M subband solutions mapping matrix;

7)

is K * K IDFT despreading matrix.

2. the SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme as claimed in claim 1 is characterized in that: said DFT-S-GMC system is the system that comprises 2 a plurality of reception antennas of transmitting antenna.

3. the SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme as claimed in claim 2 is characterized in that: said DFT-S-GMC system adopts zero forcing equalization, said step 2) in set the equalizing coefficient ω of m frequency domain equalization subcarrier correspondence _mFor

Wherein,

Be the channel frequency response of i transmit antennas to m frequency domain equalization subcarrier between the r root reception antenna.

4. the SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme as claimed in claim 2; It is characterized in that: said DFT-S-GMC system adopts least mean-square error balanced, then said step 2) in set the equalizing coefficient ω of m frequency domain equalization subcarrier correspondence _mFor ${\mathrm{\&omega;}}_m=\frac{1}{{\left|{\widetilde{\mathrm{\&Lambda;}}}_m\right|}^2+\frac{1}{\mathrm{SNR}}}=\frac{1}{{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_m^{i,r}\right|}^2\right|}^2+{\mathrm{\&sigma;}}^2},$ Wherein,

Be the channel frequency response of i transmit antennas to m frequency domain equalization subcarrier between the r root reception antenna, σ ²Noise variance for the frequency domain equalization subcarrier.

5. like claim 3 or 4 described SINR methods of estimation, it is characterized in that said step 3) comprises step based on DFT-S-GMC system multi-antenna diversity scheme:

(1) according to the channel frequency response of i transmit antennas to the balanced subcarrier of multipath channel frequency domain between the r root reception antenna ${\left[\begin{array}{cccccc}{\mathrm{\&Lambda;}}_0^{i,r}& {\mathrm{\&Lambda;}}_1^{i,r}& ...& {\mathrm{\&Lambda;}}_m^{i,r}& ...& {\mathrm{\&Lambda;}}_{N-1}^{i,r}\end{array}\right]}^T,$ I=1,2; R=1,2 ..., Nr calculates diagonal matrix $\widetilde{\mathrm{\&Lambda;}}=\mathrm{Diag}\{\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2,\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2,...,\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\},$ Wherein diag{A} representes with vector A to be the diagonal matrix of diagonal element;

(2) according to the equalizing coefficient [ω that sets ₀ω ₁ω _mω _N-1] ^T, calculate diagonal matrix ${\mathrm{\&Lambda;}}_N={\widetilde{\mathrm{\&Lambda;}}}^{H}W\widetilde{\mathrm{\&Lambda;}}=\mathrm{Diag}\{{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_0,{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_1,...,{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_{N-1}\};$

(3) compute matrix

Wherein, Λ _NBe diagonal matrix, h is a circulation symmetrical matrix, and the h first column element vector does

$h_0=\frac{1}{\sqrt{N}}{F}_N^H{[{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_0{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_1...{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\right|}^2{\mathrm{\&omega;}}_{N-1}]}^T,$ All the other column vectors of h can be by h ₀Cyclic shift obtains;

(4) the capable and preceding L row of the preceding L in the intercepting matrix h upper left corner, the L of formation * L matrix

(5) matrix

is divided into the block matrix of (L/M) * (L/M); The size of each matrix-block is M * M, and wherein the matrix-block of the capable and preceding j row of i does

H wherein _{I, j}In element

Be matrix

The capable jM column element of iM;

(6) calculate P respectively _ih _{I, j}P _jFirst column vector $b_{i,j}=P_i{h}_{i,j}{P}_j\left[\begin{array}{c}1\\ 0_{(M-1)\&times;1}\end{array}\right],$ P wherein _iBe M * M diagonal matrix, its diagonal element is { f _p(i * M), f _p(i * M+1) ..., f _p(i * M+M-1) }, f _p(t), t=0,1 ..., L-1 is many Methods of Subband Filter Banks prototype filter coefficient, 0 _{(M-1) * 1}Be (M-1) * 1 zero column vector;

(7) stack b _{I, j}, and carry out M point DFT conversion to get column vector

(8), extract and take the signal component on the subband through the subband solutions mapping ${\left[\begin{array}{cccc}{\tilde{H}}_0& {\tilde{H}}_1& ...& {\tilde{H}}_{K-1}\end{array}\right]}^T=T_{M,K}^{H}B;$

(9) calculate useful signal average power

6. the SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme as claimed in claim 5 is characterized in that: the average power calculating formula of disturbing between said receiving end signal is:

7. the SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme as claimed in claim 6 is characterized in that said step 5) comprises step:

(1) by the channel frequency response of i transmit antennas to the balanced subcarrier of multipath channel frequency domain the r root reception antenna ${\left[\begin{array}{cccccc}{\mathrm{\&Lambda;}}_0^{i,r}& {\mathrm{\&Lambda;}}_1^{i,r}& ...& {\mathrm{\&Lambda;}}_m^{i,r}& ...& {\mathrm{\&Lambda;}}_{N-1}^{i,r}\end{array}\right]}^T,$ I=1,2; R=1,2 ..., Nr calculates diagonal matrix $\widetilde{\mathrm{\&Lambda;}}=\mathrm{Diag}\{\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2,\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2,...,\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\},$ Wherein, diag{A} representes with vector A to be the diagonal matrix of diagonal element;

(2) according to the equalizing coefficient [ω that sets ₀ω ₁ω _mω _N-1] ^T, calculate diagonal matrix: ${\mathrm{\&Lambda;}}_N^{\&prime;}={\widetilde{\mathrm{\&Lambda;}}}^{H}W\widetilde{\mathrm{\&Lambda;}}W\widetilde{\mathrm{\&Lambda;}}=\mathrm{Diag}\{{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_0^2,{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_1^2,...,{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_{N-1}^2\};$

(3) compute matrix Wherein, Λ ' _NBe diagonal matrix, h ' is a circulation symmetrical matrix, and the h ' first column element vector does $h_0^{\&prime;}=\frac{1}{\sqrt{N}}{F}_N^H{[{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_0^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_0^2{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_1^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_1^2...{\left|\underset{r=1}{\overset{\mathrm{Nr}}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&Lambda;}}_{N-1}^{i,r}\right|}^2\right|}^3{\mathrm{\&omega;}}_{N-1}^2]}^T$ , all the other column vectors of h ' can be by h ' ₀Cyclic shift obtains;

(4) the capable and preceding L row of the preceding L in intercepting matrix h ' upper left corner, the L of formation * L matrix

(5) matrix

is divided into the block matrix of (L/M) * (L/M); The size of each matrix-block is M * M, wherein capable the and preceding j of i row matrix-block can be expressed as

wherein

be matrix

the capable j column element of i;

(6) calculate P respectively _iH ' _{I, j}P _jFirst column vector

P wherein _iBe M * M diagonal matrix, its diagonal element is { f _p(i * M), f _p(i * M+1) ..., f _p(i * M+M-1) }, f _p(t), t=0,1 ..., L-1 is many Methods of Subband Filter Banks prototype filter coefficient, 0 _{(M-1) * 1}Be (M-1) * 1 zero column vector;

(7) stack b ' _{I, j}, and carry out M point DFT conversion to obtain

(8), extract and take the noise component(s) on the subband through the subband solutions mapping ${\left[\begin{array}{cccc}{\tilde{H}}_0^{\&prime;}& {\tilde{H}}_1^{\&prime;}& ...& {\tilde{H}}_{K-1}^{\&prime;}\end{array}\right]}^T=T_{M,K}^T{B}^{\&prime;};$

(9) calculating noise variance

8. the SINR method of estimation based on DFT-S-GMC system multi-antenna diversity scheme as claimed in claim 7, its characteristic

$\mathrm{SINR}=\frac{E_s^{\&prime;}}{{\mathrm{\&sigma;}}_n^2+{\mathrm{\&sigma;}}_{\mathrm{ISI}}^2}$

$=\frac{{\left|\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\tilde{H}}_K\right|}^2}{\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\stackrel{\&OverBar;}{H}}_k{\mathrm{\&sigma;}}^2+\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\left|{\tilde{H}}_k\right|}^2-{\left|\frac{1}{K}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{\tilde{H}}_k\right|}^2}.$

Be: said SINR does

Images