CN101359982B - Space frequency group code detection method and apparatus - Google Patents

Abstract

The invention discloses a method for testing space-frequency block codes, which comprises: calculation in caussian approximation is performed to the interference distribution of received signal variable to obtain the initial probability, the typical values and the variances of sent symbols; the initial probability of the send symbols is updated, and the probability of the sent symbols is obtained; the maximum likehood ratio of the sent symbols is obtained according to the probability of the sent symbols; and the sent symbols is tested according to the maximum likehood ratio of the sent symbols. The invention also provides a space-frequency block code testing device to reduce the error floor in ultra-wideband channel, so that the system performance is improved.

Description

Space frequency group code detection method and device

Technical field

The present invention relates to wireless communication technology field, especially refer to a kind of space frequency group code detection method and device.

Background technology

At present, in broadband wireless communications, in most cases, all there is the phenomenon of multidiameter delay expansion in the channel.And in multipath MIMO (Multiple Input Multiple Output, the multiple-input and multiple-output) channel, because the influence of ISI (Intersymbol Interference, intersymbol interference) causes multipath mimo channel technology to be difficult to be suitable for.At time-domain countermeasures ISI, be in the prior art at the employing Space Time Coding of making a start, perhaps balanced when receiving terminal adopts sky, but the Time-Domain Technique complexity is higher.OFDM (OrthogonalFrequency Division Multiplexing, OFDM) is the effective technology of a kind of ISI of antagonism of growing up in a single aerial system, and OFDM is applied in the mimo system, can reduce the complexity of system.

In the multipath mimo channel; In order to obtain branch collection as much as possible; Need in MIMO-OFDM (Multiple Input Multiple Output-Orthogonal Frequency Division Multiplexing, multi-input multi-output-orthogonal frequency division multiplexing) system, adopt Space Time Coding, space-frequency coding or space-time frequency coding technology.Wherein, empty sign indicating number frequently is on space and frequency, signal to be encoded simultaneously, and the coding of use can adopt block code.

In the multipath MIMO-OFDM that adopts space-time/frequency block code, particularly among the UWB of high-speed communication (Ultra Wide Band, ultra broadband), because distinguishable footpath is more, the time delay expansion of channel is relatively large, so the correlation between the frequency domain adjacent sub-carrier dies down.And detection algorithm of the prior art supposes that the channel in the same grouping is relevant fully constant channel, can occur error floor (error is flat) in the detection, poor-performing.

The orthogonal space frequency block code is applied to the MIMO-OFDM system in the prior art, the q that establishes sub-carrier number P and be p doubly, P=pq, each OFDM symbol sends P=Pq symbol.P symbol is divided into q group, one group of p symbol.Every group of matrix that all adopts orthogonal code to be encoded to p * m; Q the continuous mutually empty matrix frequently that is spliced into P * m of encoder matrix; Each row IFFT (Inverse Fast FourierTransform, inverse fast Fourier transform) of then empty matrix is frequently sending on corresponding antenna behind the interpolation Cyclic Prefix.For example: number of transmit antennas m=4, then orthogonal matrix is:

$\left[\begin{array}{cccc}{x}_1& x_2& x_3& 0\\ -x_2^{*}& x_1^{*}& 0& x_3\\ -x_3^{*}& 0& x_1^{*}& -x_2\\ 0& -x_3^{*}& x_2^{*}& x_1\end{array}\right]$

The SFBC of 4 antenna transmission (Space-Frequency Block Coding; Space-time/frequency block code) as shown in Figure 1; 4 symbols of each row all send on 4 adjacent number of sub-carrier, and different lines is sent on different antennae, and this mode needs the channel of 4 adjacent sub-carriers to keep constant basically.

Space-time/frequency block code of the prior art adopts the simplification maximum likelihood algorithm with linear complexity of Almouti, detects after the branch set also.SFBC-OFDM system with two transmitting antennas is example, and is as shown in Figure 2, and Fig. 2 is the structural representation of prior art SFBC-OFDM system transmitting terminal.At first, adopt convolution code that signal is carried out chnnel coding, interweaves, carry out the SFBC coding after interweaving again, purpose is to divide collection in order fully to obtain all.If because do not carry out chnnel coding, interweave, and only carry out the SFBC coding, then can only obtain space diversity; If adopt convolution code to carry out chnnel coding, interweave, the symbol after the modulation distributes on the OFDM subcarrier, can obtain frequency diversity.So just obtained intact full-diversity in conjunction with chnnel coding and SFBC coding.

Transmitting terminal adopts 2 transmitting antennas, the OFDM of 2K number of sub-carrier, so the interior at interval symbol that sends of OFDM is the matrix S of a 2 * 2K, wherein corresponding 2 transmitting antennas of row are listed as corresponding 2K number of sub-carrier.Transmitting terminal is divided into K subband with the 2K number of sub-carrier, and each subband comprises two adjacent number of sub-carrier, so S is divided into the matrix of K individual 2 * 2

S＝[S ₁，…，S _K] (1)

If the bandwidth of two number of sub-carrier is much smaller than the correlation bandwidth of channel in each subband, the channel coefficients in the so same subband on the subcarrier has strong correlation.

Be divided into the K sub-block behind the modulated data stream demultiplexing, every two modulation symbols.The symbol of k piece is x _k=[x _k(1), x _k(2)] ^TSend on k subband through SFBC coding back.Like this, x _kDraw the k sub-block S of s-matrix behind the coding _k

$S_k=\left[\begin{array}{cc}{x}_k\left(1\right)& -x_k{\left(2\right)}^{*}\\ x_k\left(2\right)& x_k{\left(1\right)}^{*}\end{array}\right]---\left(2\right)$

Data behind the SFBC coding are through framing, and send on different antennas through IFFT (Inverse Fast Fourier Transform, inverse fast Fourier transform) back.

The UWB channel length of supposing the signal process is L, if the situation that channel length does not wait occurs, but then zero padding makes it equal.J transmitting antenna is expressed as the discrete time vector to the equivalent low pass impulse response of i reception antenna

J transmitting antenna channel gain H on i the reception antenna l number of sub-carrier so _{I, j} ^(l)For

$H_{i,j}^{\left(l\right)}=\underset{k=0}{\overset{L-1}{\mathrm{\&Sigma;}}}{h}_{i,j}\left(l\right)e^{-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;kl}/2K}---\left(3\right)$

Frequency response on the l number of sub-carrier can show to be matrix H ^(l)If the reception antenna number is 1, the signal that receives of k subband does so

$r_k\left(1\right)=H_1^{(2k-1)}{x}_k\left(1\right)+H_2^{(2k-1)}{x}_k\left(1\right)+n_k\left(1\right)---\left(4\right)$

$r_k\left(2\right)=-H_1^{\left(2k\right)}{x}_k{\left(2\right)}^{*}+H_2^{\left(2k\right)}{x}_k{\left(2\right)}^{*}+n_k\left(2\right)---\left(5\right)$

N wherein _k(1), n _k(2) be that real part imaginary part variance all is σ ²Multiple noise.(5) formula got complex conjugate and close with (4) formula write as vector form and obtain

$r_k=\left[\begin{array}{c}{r}_k\left(1\right)\\ r_k^{*}\left(2\right)\end{array}\right]=\left[\begin{array}{cc}{H}_1^{(2k-1)}& H_2^{(2k-1)}\\ H_2^{{\left(2k\right)}^{*}}& -H_1^{{\left(2k\right)}^{*}}\end{array}\right]\left[\begin{array}{c}{x}_k\left(1\right)\\ x_k\left(2\right)\end{array}\right]+\left[\begin{array}{c}{n}_k\left(1\right)\\ n_k^{*}\left(2\right)\end{array}\right]$

$=H_k{x}_k+n_k---\left(6\right)$

Remain unchanged in a subband as if channel, so H _kIt is orthogonal matrix

$H_k^H{H}_k=\left[\begin{array}{cc}{\left|H_1^{(2k-1)}\right|}^2+{\left|H_2^{\left(2k\right)}\right|}^2& 0\\ 0& {\left|H_2^{(2k-1)}\right|}^2+{\left|H_1^{\left(2k\right)}\right|}^2\end{array}\right]=c_k{I}_2---\left(7\right)$

(7) in the formula $c_k={\left|H_1^{(2k-1)}\right|}^2+{\left|H_2^{\left(2k\right)}\right|}^2={\left|H_2^{(2k-1)}\right|}^2+{\left|H_1^{\left(2k\right)}\right|}^2.$ The detection of this transmission symbol can be adopted the Almouti detection algorithm that symbol is separated and detect.But in the UWB channel, because channel has bigger time delay expansion, can there be variation to a certain degree in channel in a subband, so have

$H_k^H{H}_k=\left[\begin{array}{cc}{\left|H_1^{(2k-1)}\right|}^2+{\left|H_2^{\left(2k\right)}\right|}^2& {H_1^{(2k-1)}}^{*}{H}_2^{(2k-1)}-{H_1^{\left(2k\right)}}^{*}{H}_2^{\left(2k\right)}\\ H_1^{(2k-1)}{H_2^{(2k-1)}}^{*}-H_1^{\left(2k\right)}{H_2^{\left(2k\right)}}^{*}& {\left|H_2^{(2k-1)}\right|}^2+{\left|H_1^{\left(2k\right)}\right|}^2\end{array}\right]=\left[\begin{array}{cc}{c}_{k1}& e_{k1}\\ e_{k2}& {\mathrm{<figure-callout\; id="2"\; label="c\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">c</figure-callout>}}_k\end{array}\right]\&NotEqual;c_k{I}_2---\left(8\right)$

In the formula $e_k$ $1_{}={H_1^{(2k-1)}}^{*}{H}_2^{(2k-1)}-{H_1^{\left(2k\right)}}^{*}{H}_2^{\left(2k\right)},$ ${\mathrm{<figure-callout\; id="2"\; label="e\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">e</figure-callout>}}_k={{H_1^{(2k-1)}H}_2^{(2k-1)}}^{*}-{{H_1^{\left(2k\right)}H}_2^{\left(2k\right)}}^{*}.$ Can find out H thus _kNo longer be orthogonal matrix, can bring performance loss, cause bit error rate error floor to occur if adopt the Almouti detection algorithm that symbol is separated to detect.Because; The Almouti algorithm needs the adjacent sub-carrier channel that symbol is corresponding in the same block code to remain unchanged; But in the serious UWB channel of multidiameter delay expansion, the respective channels of adjacent sub-carrier normally changes, and the frequency selectivity of UWB channel is very serious; Channel frequency response between the adjacent sub-carrier alters a great deal, so if in the UWB channel, adopt the Almouti algorithm can cause serious error floor.A kind ofly in the prior art adopt the parallel method that detects with serial interference elimination; The output signal is carried out interference eliminated, decipher again, therefore reduced the interference in the decoding symbol; Reduction error floor that can be bigger; Improve and detect performance, but its testing result is directly carried out hard decision to signal, has lost the reliability information that comprises in the signal amplitude.

Summary of the invention

The embodiment of the invention provides a kind of space frequency group code detection method and device, when detecting to solve the available technology adopting space-time/frequency block code, can cause the problem of serious error floor.

The embodiment of the invention provides a kind of space frequency group code detection method, comprising:

To received signal variable carry out Gaussian approximation in disturbing to distribute, obtain the initial probabilistic information that respectively sends symbol, said initial probabilistic information comprises initial probability, average and variance;

The said initial probability that respectively sends symbol is upgraded, obtain the probability that respectively sends symbol;

Obtain the said maximum likelihood ratio that respectively sends symbol according to the said probability that respectively sends symbol;

Detect based on the said said symbol that respectively sends of maximum likelihood comparison that respectively sends symbol;

Said reception signal is first to n, wherein, and n >=2;

Said initial probability to each transmission symbol upgrades, and obtains the probability that respectively sends symbol, may further comprise the steps:

Bring n initial average and variance of sending symbol into first reception signal variable, upgrade the initial probability of the first transmission symbol, obtain the probability of the first transmission symbol;

Preceding n-1 average and n of variance substitution of sending symbol received signal variable, upgrade n initial probability that sends symbol, obtain n probability that sends symbol;

So iteration is upgraded, treat respectively to send the probability of symbol stable after, respectively sent the probability of symbol.

The embodiment of the invention also provides a kind of space-time/frequency block code checkout gear, comprising:

The probability calculation unit, the interference profile that is used for is to received signal carried out Gaussian approximation, obtains the initial probability that respectively sends symbol; The said initial probability that respectively sends symbol is upgraded, obtain the probability that respectively sends symbol; Said reception signal is first to n, wherein, and n >=2;

The maximum likelihood ratio acquiring unit is used for obtaining the maximum likelihood ratio that respectively sends symbol according to the probability of each symbol; With

Detecting unit is used for the maximum likelihood ratio that respectively sends symbol according to said, and the said symbol that respectively sends is detected;

Said initial probability to each transmission symbol upgrades, and obtains the probability that respectively sends symbol, may further comprise the steps:

Bring n initial average and variance of sending symbol into first reception signal variable, upgrade the initial probability of the first transmission symbol, obtain the probability of the first transmission symbol;

Preceding n-1 average and n of variance substitution of sending symbol received signal variable, upgrade n initial probability that sends symbol, obtain n probability that sends symbol;

So iteration is upgraded, treat respectively to send the probability of symbol stable after, respectively sent the probability of symbol.

Compared with prior art; The embodiment of the invention is respectively sent the initial probabilistic information of symbol through adopting Gaussian approximation, and each initial probability that sends symbol is upgraded, and obtains the probability that respectively sends symbol; Respectively send the maximum likelihood ratio of symbol again according to the probability calculation of each transmission symbol; Each is sent symbol adjudicate, compared with prior art, the embodiment of the invention has reduced the error floor in the UWB channel; Improved systematic function, and the detection method of the embodiment of the invention has lower complexity.

Description of drawings

Fig. 1 is the sketch map that sends SFBC in the prior art;

Fig. 2 is the structural representation of SFBC-OFDM system transmitting terminal in the prior art;

Fig. 3 is the schematic flow sheet of a kind of space frequency group code detection method of the embodiment of the invention;

Fig. 4 is the planisphere of embodiment of the invention gray mappings;

Fig. 5 is the planisphere of the non-gray mappings of the embodiment of the invention;

Fig. 6 is the structural representation of embodiment of the invention SFBC-OFDM system receiving terminal;

Fig. 7 is the structural representation of a kind of space-time/frequency block code checkout gear of the embodiment of the invention;

Fig. 8 is that the performance under the embodiment of the invention CM1 channel compares sketch map;

Fig. 9 is that the performance under the embodiment of the invention CM2 channel compares sketch map;

Figure 10 is that the performance under the embodiment of the invention CM3 channel compares sketch map;

Figure 11 is that the performance under the embodiment of the invention CM4 channel compares sketch map;

Figure 12 is that the iteration performance under the embodiment of the invention CM1 channel compares sketch map;

Figure 13 is that the iteration performance under the embodiment of the invention CM2 channel compares sketch map;

Figure 14 is that the iteration performance under the embodiment of the invention CM3 channel compares sketch map;

Figure 15 is that the iteration performance under the embodiment of the invention CM4 channel compares sketch map.

Embodiment

Be elaborated below in conjunction with accompanying drawing and specific embodiment.

Fig. 3 is the schematic flow sheet of a kind of space frequency group code detection method of the embodiment of the invention, and is as shown in Figure 3, may further comprise the steps:

Step 301, the interference profile of variable is carried out Gaussian approximation to received signal, obtains the probabilistic information that respectively sends symbol, and this probabilistic information comprises initial probability, average and variance.

To the reception vector r in (6) formula _kHandle as follows, multiply by H on (6) formula both sides together _k ^H

$y_k=H_k^H{r}_k=\left[\begin{array}{cc}{c}_{k1}& e_{k1}\\ e_{k2}& c_{k2}\end{array}\right]x_k+H_k^H{n}_k---\left(9\right)$

(9) formula can be divided and write as two formulas and obtain

$y_k\left(1\right)=c_{k1}{x}_k\left(1\right)+e_{k1}{x}_k\left(2\right)+{\tilde{n}}_{k1}---\left(10\right)$

$y_k\left(2\right)=e_{k2}{x}_k\left(1\right)+c_{k2}{x}_k\left(2\right)+{\tilde{n}}_{k2}---\left(11\right)$

(10) in the formula

Real part imaginary part variance all be (| H ₁ ^(2k-1)| ²+ | H ₂ ^(2k-1)| ²) σ ²(11) in the formula Real part, imaginary part variance all be (| H ₁ ^(2k)| ²+ | H ₂ ^(2k)| ²) σ ²If adopt maximum-likelihood criterion to detect, then need associating (10), (11) two formulas, adopt the method for exhaustion to x _k(1) and x _k(2) detect simultaneously, the complexity that when constellation is big, detects is higher.But because e generally _K1＜＜c _K1, e _K2＜＜c _K2, therefore in (10) formula with e _K1x _k(2) be regarded as noise to x _k(1) detect, equally in (11) formula with e _K2x _k(1) is regarded as noise to x _k(2) detect, thus, to x _k(1) and x _k(2) then separable the coming of detection carried out, and when constellation is big, reduces the complexity that detects.

If x _kIn symbol adopt the QPSK (Quaternary Phase Shift Keying, quaternary PSK) of gray mappings.Fig. 4 is the planisphere of embodiment of the invention gray mappings, and is as shown in Figure 4, and real part imaginary part amplitude all is ± 1.(10) formula of employing detects x _k(1), (11) formula detects x _k(2).And, can further separate x again for gray mappings mode shown in Figure 4 _k(1), x _k(2) real part imaginary part detects.By y _k(1) real part detects x _k(1) first bit, y _k(1) imaginary part detects x _k(1) second bit; By y _k(2) real part detects x _k(2) first bit, y _k(2) imaginary part detects x _k(2) second bit.

In (10) formula, to x _k(1) forming the effective noise that disturbs does ${\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">N</figure-callout>}}_1=e_{k1}{x}_k\left(2\right)+{\tilde{n}}_1.$ If x _k(2) average is E (x _k(2))=ER (x _k(2))+jEI (x _k(2)), ER (x wherein _k(2)) be the average of real part, EI (x _k(2)) be the average of imaginary part, with ER (x _kAnd EI (x (2)) _k(2)) initial value all is made as 0; x _k(2) variance is Var (x _k(2))=VarR (x _k(2))+VarI (x _k(2)), same, VarR (x _k(2)) be the variance of real part, VarI (x _k(2)) be the variance of imaginary part, with VarR (x _kAnd VarI (x (2)) _k(2)) initial value all is made as 1.With x _k(2) real part all is approximately the Gaussian distribution with identical average and variance with imaginary part.So x _k(1) first bit is that 0 and 1 probability can be by y _k(1) real part calculates, as follows:

${\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(0\right)=\frac{{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,0}}\left(\mathrm{Re}\left(y_k\left(1\right)\right)\right)}{{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,0}}\left(\mathrm{Re}\left(y_k\left(1\right)\right)\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,1}}\left(\mathrm{Re}\left(y_k\left(1\right)\right)\right)}$

${\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(1\right)=1-{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(0\right)$

$=1-\frac{{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,0}}\left(\mathrm{Re}\left(y_k\left(1\right)\right)\right)}{{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,0}}\left(\mathrm{Re}\left(y_k\left(1\right)\right)\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,1}}\left(\mathrm{Re}\left(y_k\left(1\right)\right)\right)}$

Wherein,

For first bit sends 0 o'clock y _k(1) real part is the probability density of x;

For first bit sends 1 o'clock y _k(1) real part is the probability density of x.

Because x _k(2) probability is unknown, x _k(2) the initial average of real part, imaginary part all is 0, and variance all is 1.Therefore, can obtain x _k(1) second bit is that 0 and 1 probability is following:

${\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(1\right)=1-{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(0\right)$

After obtaining the probability of above-mentioned (12), (13), (16), (17) formula, the average and the variance that then can calculate first symbol are respectively:

E(x _k(1))＝1-2p _k1，1(0)+j(1-2p _k1，2(0)) (18)

VarR(x _k(1))＝4p _k1，1(0)(1-p _k1，1(0)) (19)

VarI(x _k(1))＝4p _k1，2(0)(1-p _k1，2(0)) (20)

After obtaining the mean variance of first symbol real part and imaginary part, can this average and variance substitution (11) formula be calculated x _k(2) probability goes out x according to this probability calculation again _k(2) average and variance, x _k(2) probability calculation is similar with (12)～(17) formula, need only be with x _k(2) and x _k(1) exchanges y _k(1) and y _k(2) exchange,

With

Exchange c _K1With c _K2Exchange.Be x _k(2) first bit is that 0 and 1 probability is respectively:

${\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(0\right)=\frac{{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,0}}\left(\mathrm{Re}\left(y_k\left(2\right)\right)\right)}{{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,0}}\left(\mathrm{Re}\left(y_k\left(2\right)\right)\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,1}}\left(\mathrm{Re}\left(y_k\left(2\right)\right)\right)}$

${\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(1\right)=1-{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(0\right)$

$=1-\frac{{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,0}}\left(\mathrm{Re}\left(y_k\left(2\right)\right)\right)}{{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,0}}\left(\mathrm{Re}\left(y_k\left(2\right)\right)\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,1}}\left(\mathrm{Re}\left(y_k\left(2\right)\right)\right)}$

Wherein,

Be that second bit sends 0 o'clock y _k(2) real part is the probability density of x;

Be that second bit sends 1 o'clock y _k(2) real part is the probability density of x.

Thereby obtain x _k(2) second bits are that 0 and 1 probability is respectively:

${\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(1\right)=1-{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(0\right)$

Step 302 is upgraded each initial probability that sends symbol, obtains the probability that respectively sends symbol.

Obtain x in the formula of above-mentioned (21)～(26) _k(2) after probability, average and the variance, again with x _k(2) further upgrade x in average and variance substitution (10) formula _k(1) probability, average and variance will be upgraded back x _k(1) average and variance substitution (11) formula are upgraded x _k(2) probability, average and variance.So iteration is upgraded several times, treats x _k(1) and x _k(2) after probability is stablized, can obtain x _k(1) and x _k(2) each bit is respectively 0 and 1 probability.If all pressing the arrangement of s-matrix in (1) formula, each reception signal is positioned at adjacent sub-carrier, because e _K1, e _K2All very little, algorithm can be restrained soon, and common twice iteration just can make probability reach stable, and lower complexity is arranged.

The foregoing description is that signal constellation (in digital modulation) adopts the iterative calculation method in the testing process of gray mappings, if signal constellation (in digital modulation) adopts non-gray mappings mode shown in Figure 5, the method for calculating probability that then real part and imaginary component leave in the foregoing description is no longer set up.Because gray mappings planisphere shown in Figure 4; First bit is by the real part decision that receives signal; Second bit is by the imaginary part decision that receives signal; Therefore calculate the probabilistic information that sends bit, can be undertaken by the method that real part, imaginary component leave, and in the non-gray mappings planisphere shown in Figure 5; Different transmission bits and receive the real part of signal, the mapping relations between the imaginary part are not fixed, thereby just can't adopt real part and imaginary component leave in the foregoing description computational methods to calculate the probabilistic information of non-gray mappings yet.Need revise as follows: with x _k(2) probability is initialized as

p ₂(00)＝p ₂(01)＝p ₂(10)＝p ₂(11)＝1/4 (27)

So initial average and variance do

E(x _k2(2))＝0+j0 (28)

VarR(x _k2(2))＝VarI(x _k2(2))＝1 (29)

The x that is sending _k(1) corresponding bit is under 00,01,10,11 the condition, to receive y _k(1) probability density does

Wherein the v span is 00,01,10,11,

the corresponding constellation point of expression v.So receiving y _k(1) under the condition, the x of transmission _k(1) corresponding bit is that the probability of v is:

${\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>}}_{1,v}=\frac{{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{1,v}\left(y_k\left(1\right)\right)}{\underset{s}{\mathrm{\&Sigma;}}{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{1,s}\left(y_k\left(1\right)\right)}---\left(31\right)$

(31) in the formula

Be the x that sends _k(1) corresponding bit is respectively under 00,01,10,11 the condition, receives y _k(1) probability density summation also is $\underset{s}{\mathrm{\&Sigma;}}{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{1,s}\left(y_k\left(1\right)\right)={\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,00}}\left(y_k\left(1\right)\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,01}}\left(y_k\left(1\right)\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,10}}\left(y_k\left(1\right)\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,11}}\left(y_k\left(1\right)\right),$ X then _k(1) average and variance are respectively:

Var(x _k(1))＝S-|E(x _k(1))| ² (33)

S in the formula (33) is the average energy of constellation.Obtain x _k(1) can be after probability, average and the variance with calculating x in these probability, average and variance substitution (11)-(17) formula _k(2) probability, average and variance.

Must be in same the calculating with x _k(2) and x _k(1), y _k(1) and y _k(2), With Exchange and c _K1With c _K2Exchange the x that is promptly sending _k(2) corresponding bit is to receive y under 00,01,10,11 the condition _k(2) probability density does

So receiving y _k(2) under the condition, the x of transmission _k(2) corresponding bit is that the probability of v is:

${\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>}}_{2,v}=\frac{{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{2,v}\left(y_k\left(2\right)\right)}{\underset{s}{\mathrm{\&Sigma;}}{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{2,s}\left(y_k\left(2\right)\right)}---\left(35\right)$

X then _k(2) average and variance are respectively:

Var(x _k(2))＝S-|E(x _k(2))| ² (37)

Obtain x _k(2) after probability, average and the variance, again with further upgrading x in those probability, average and variance substitution (10) formula and (12)-(17) formula _k(1) probability, average and variance after so iteration is upgraded and treated that several times probability is stable, can obtain each bit and be respectively 0 and 1 probability.

After above-mentioned probability is stable, then can try to achieve x _k(1) probability distribution of corresponding bit, the individual bit of m (m=1,2) are that the probability of b (b=0,1) is:

${\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(b\right)=\frac{\underset{v:v_m=\mathrm{<figure-callout\; id="1"\; label="b\; p"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">b</figure-callout>}}{\mathrm{\&Sigma;}}{p}_{}{}_{1,v}}{\underset{\mathrm{<figure-callout\; id="1"\; label="v\; p"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">v</figure-callout>}}{\mathrm{\&Sigma;}}{p}_{}{}_{1,v}}---\left(38\right)$

The v that summation in the molecule is b to m bit value carries out.x _k(2) the probability calculation formula of corresponding bit is:

${\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(b\right)=\frac{\underset{v:v_m=\mathrm{<figure-callout\; id="2"\; label="b\; p"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">b</figure-callout>}}{\mathrm{\&Sigma;}}{p}_{}{}_{2,v}}{\underset{\mathrm{<figure-callout\; id="2"\; label="v\; p"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">v</figure-callout>}}{\mathrm{\&Sigma;}}{p}_{}{}_{2,v}}---\left(39\right)$

Above-mentioned computational methods can be generalized to any non-gray mappings mode of any constellation.If corresponding q the bit of constellation symbol, so the span of v be q bit all 2 ^qPlant possible combination, x _k(2) probability is initialized as:

p ₂(v)＝1/2 ^q (40)

Initial average and variance are respectively:

E(x _k2(2))＝0+j0 (41)

Var(x _k2(2))＝S (42)

And (39) the possible value of m is 1 in the formula, 2..., q.

In the system that adopts chnnel coding, decipher sending into decoder after the above-mentioned probabilistic information deinterleaving.As shown in Figure 6, Fig. 6 is the structural representation of embodiment of the invention SFBC-OFDM system receiving terminal.If at system's transmitting terminal shown in Figure 2, carry out also having adopted chnnel coding before the SFBC coding to sending signal, then at system receiving terminal shown in Figure 6, just need after carrying out SFBC decoding to received signal, send into decoder and carry out channel decoding.If adopt SISO (Soft-input and soft-output in the decoder; Soft inputting and soft output) decoding; Decoder provides the probabilistic information of each bit; Because it is more reliable that new probabilistic information is compared original input probability information, therefore, feeds back to SFBC after can new probabilistic information being interweaved and carry out iteration.Interleaver can guarantee the approximate independence of the feedback information and the information of reception.

Be example still with mapping mode shown in Figure 5, under the condition that the feedback prior information is arranged, x _k(1) first bit is that 0 and 1 probability is respectively:

${\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(0\right)=\frac{{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,00}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,2}}\left(0\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,01}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,2}}\left(1\right)}{{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,00}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,2}}\left(0\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,01}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,2}}\left(1\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,10}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,2}}\left(0\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,11}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,2}}\left(1\right)}---\left(43\right)$

${\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(1\right)=\frac{{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,10}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,2}}\left(0\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,11}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,2}}\left(1\right)}{{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,00}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,2}}\left(0\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,01}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,2}}\left(1\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,10}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,2}}\left(0\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,11}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,2}}\left(1\right)}---\left(44\right)$

x _k(1) second bit is that 0 and 1 probability is respectively:

${\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(0\right)=\frac{{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,00}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,1}}\left(0\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,10}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,1}}\left(1\right)}{{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,00}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,1}}\left(0\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,01}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,1}}\left(0\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,10}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,1}}\left(1\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,11}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,1}}\left(1\right)}---\left(45\right)$

${\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(1\right)=\frac{{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,01}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,1}}\left(0\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,11}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,1}}\left(1\right)}{{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,00}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,1}}\left(0\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,01}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,1}}\left(0\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,10}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,1}}\left(1\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,11}}{\mathrm{<figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{1,1}}\left(1\right)}---\left(46\right)$

F in the formula _{1, v}Shown in (30), p _{K1, l} ^Pri(b) the individual bit of expression the 1st symbol l of k subband (l=1,2) is the prior probability of b (b=0,1), and does not consider the prior probability of bit in the calculating of (37).

x _k(2) x can be similarly tried to achieve in the calculating of corresponding bit _k(2) first bit is that 0 and 1 probability is respectively:

${\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(0\right)=\frac{{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,00}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,2}}\left(0\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,01}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,2}}\left(1\right)}{{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,00}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,2}}\left(0\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,01}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,2}}\left(1\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,10}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,2}}\left(0\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,11}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,2}}\left(1\right)}---\left(47\right)$

${\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(1\right)=\frac{{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,10}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,2}}\left(0\right)+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{1,11}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,2}}\left(1\right)}{{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,00}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,2}}\left(0\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,01}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,2}}\left(1\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,10}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,2}}\left(0\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,11}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,2}}\left(1\right)}---\left(48\right)$

x _k(2) second bits are that 0 and 1 probability is respectively:

${\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(0\right)=\frac{{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,00}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,1}}\left(0\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,10}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,1}}\left(1\right)}{{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,00}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,1}}\left(0\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,01}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,1}}\left(0\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,10}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,1}}\left(1\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,11}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,1}}\left(1\right)}---\left(49\right)$

${\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">p</figure-callout>}}_k\left(1\right)=\frac{{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,01}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,1}}\left(0\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,11}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,1}}\left(1\right)}{{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,00}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,1}}\left(0\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,01}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,1}}\left(0\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,10}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,1}}\left(1\right)+{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">f</figure-callout>}}_{\mathrm{2,11}}{\mathrm{<figure-callout\; id="2"\; label="p\; k"\; filenames="H071E1234820070815E000011.png,H071E1234820070815E000072.png"\; state="\{\{state\}\}">\; <figure-callout\; id="1"\; label="p\; k"\; filenames="H071E1234820070815E000081.png"\; state="\{\{state\}\}">p</figure-callout>\; </figure-callout>}}_k^{\mathrm{2,1}}\left(1\right)}---\left(50\right)$

p _{K1, l} ^Pri(b) when calculating for the first time, be initialized as 0.5, the feedback information by the SISO decoder in the later calculating provides.For any non-gray mappings mode of any constellation, after trying to achieve the posterior probability of each constellation symbol, i symbol x of k subband _k(i) probability of (i=1,2) m bit transmission b (b=0,1) is:

$p_{\mathrm{ki},m}\left(b\right)=\frac{\underset{v:v_m=b}{\mathrm{\&Sigma;}}{f}_{i,v}\underset{n\&NotEqual;m}{\mathrm{\&Pi;}}{p}_{\mathrm{ki},n}^{\mathrm{pri}}\left(v_n\right)}{\underset{v}{\mathrm{\&Sigma;}}{f}_{i,v}\underset{n\&NotEqual;m}{\mathrm{\&Pi;}}{p}_{\mathrm{ki},n}^{\mathrm{pri}}\left(v_n\right)},(i=\mathrm{1,2},b=\mathrm{0,1})---\left(51\right)$

p _{Ki, n} ^Pri(v _n) represent that n bit of i symbol of k subband is v _nPrior probability.

Step 303 is obtained the maximum likelihood ratio that respectively sends symbol according to each probability that sends symbol.

According to the probabilistic information that obtains in the step 302, symbol x _k(1) first bit is that 0 and 1 probability is respectively p _K1,1(0), p _K1,1(1), second bit is that 0 and 1 probability is respectively p _K1,2(0), p _K1,2(1), symbol x _k(2) first bit is that 0 and 1 probability is respectively p _K2,1(0), p _K2,1(1), second bit is that 0 and 1 probability is respectively p _K2,2(0), p _K2,2(1), i symbol x in k subband then _k(i) maximum likelihood ratio of m the bit of (i=1,2) does

${\mathrm{\&Lambda;}}_{\mathrm{ki},m}=\log \frac{p_{\mathrm{ki},m}\left(1\right)}{p_{\mathrm{ki},m}\left(0\right)}---\left(52\right)$

Step 304 is respectively sent symbol according to the maximum likelihood comparison of each transmission symbol and is detected.

Send symbol according to the maximum likelihood comparison of calculating and detect, if Λ _{Ki, m}＞0, then adjudicate i symbol x in k the subband _k(i) m bit is 1; If Λ _{Ki, m}＜0, then adjudicate i symbol x in k the subband _k(i) m bit is 0.

The embodiment of the invention is calculated the maximum likelihood comparison according to probabilistic information and is sent the detection that symbol is adjudicated; Compare and directly carry out the detection of hard decision in the prior art to received signal; The testing result of the embodiment of the invention is more reliable, because comprised the reliability information that receives signal in the maximum likelihood ratio that calculates.

In addition; The space-time/frequency block code that embodiments of the invention are not limited to two transmitting antennas detects; The space-time/frequency block code that more comprises a plurality of transmitting antennas detects, and for the space-time/frequency block code detection of a plurality of transmitting antennas, the reception signal variable is that n (n >=2) is individual; The interference profile of variable is carried out Gaussian approximation and is obtained the initial probabilistic information that respectively sends symbol to received signal, may further comprise the steps:

Interference profile to the first reception signal variable is carried out Gaussian approximation, obtains initial probability, average and the variance of the first transmission symbol;

Preceding n-1 average and n of variance substitution of sending symbol received signal variable, obtain n initial probability, average and variance of sending symbol;

The above-mentioned initial probabilistic information that respectively sends symbol is upgraded, obtains the probability that respectively sends symbol, may further comprise the steps:

N initial average and variance substitution first of sending symbol received signal variable, upgrade the initial probability of the first transmission symbol, obtain the probability of the first transmission symbol;

Preceding n-1 average and n of variance substitution of sending symbol received signal variable, upgrade n initial probability that sends symbol, obtain n probability that sends symbol.

After respectively being sent the probability of symbol, respectively send the maximum likelihood ratio of symbol again based on this probability calculation, thereby adjudicate, also can carry out iterative detection with the channel decoding algorithm of soft inputting and soft output.Therefore, the space-time/frequency block code detection for a plurality of transmitting antennas is not going to repeat.

Embodiments of the invention also provide a kind of space-time/frequency block code checkout gear, and are as shown in Figure 7, comprising: probability calculation unit 100, maximum likelihood ratio acquiring unit 200 and detecting unit 300.Wherein the probability calculation unit 100, and the interference profile that is used for is to received signal carried out Gaussian approximation, obtains the probability that respectively sends symbol.Maximum likelihood ratio acquiring unit 200 connects probability calculation unit 100, is used for obtaining the maximum likelihood ratio that respectively sends symbol according to the probability of each symbol.Detecting unit 300 connects maximum likelihood ratio acquiring unit 200, is used for sending the maximum likelihood ratio of symbol based on each, each is sent symbol detect.

Wherein, probability calculation unit 100 comprises: Gaussian approximation subelement 110 and information iteration subelement 120.Gaussian approximation subelement 110 is used for the interference profile of the first reception signal variable is carried out Gaussian approximation, obtains initial probability, average and the variance of the first transmission symbol.Information iteration subelement 120 connects Gaussian approximation subelement 110, is used to carry out the iterative operation of initial probability, average and variance, each initial probability, average and variance of sending symbol is upgraded, thereby respectively sent the probability of symbol.

Another embodiment of the present invention has been set up information updating unit 400 on the basis of said apparatus, connect probability calculation unit 100, to adopting the reception signal of chnnel coding, the probabilistic information after the decoding is fed back to the space-time/frequency block code decoder carry out the iteration renewal.

In order to verify the performance of embodiment of the invention algorithm, Fig. 8～Figure 11 is under CM1～CM4 channel, and the performance of the detection method of the embodiment of the invention, Almouti detection method and JK detection method compares.Like Fig. 8～shown in Figure 11, Fig. 8～Figure 11 is respectively the performance sketch map relatively of detection method, Almouti detection method and the JK detection method method of the embodiment of the invention under CM1～CM4 channel.A among the figure, b, c represent the performance curve of the detection method that Almouti detection method, JK detection method and the embodiment of the invention provide respectively; As can be seen from the figure; Among the SFBC-OFDM under the UWB channel; Adopt the Almouti detection method serious error floor phenomenon can occur, along with the increase from CM1 to the CM4 channel delay, error floor is more and more serious.The JK detection method is compared the Almouti detection method, though can suitably improve the detection performance, reduces error floor, and the JK detection method but begins to occur error floor under the condition of CM2 channel.And the detection method of the embodiment of the invention obvious error floor just occurs until the CM4 channel in observation bit error rate scope; Can find out thus; The detection method of the embodiment of the invention can be under abominable channel condition bigger raising performance, and reduce the appearance of error floor greatly.

The iteration performance that Figure 12～Figure 15 is respectively under CM1～CM4 channel compares sketch map.The performance of the detection method that the embodiment of the invention provided when CM1～CM4 channel adopted chnnel coding to carry out repeatedly iterative detection down has been shown among the figure.As a comparison, also show same coding among the figure and still adopt gray mappings, and adopt Almouti to detect the systematic function curve of decoding then, adopt the system of this kind detection not have iteration gain.The performance curve of 10 iteration of detection method that 3 iteration of detection method that 2 iteration of detection method that detection method 1 iteration, embodiment of the invention that on behalf of Almouti detection method, the embodiment of the invention, the A among the figure, B, C, D, E provide respectively provides, the present invention's enforcement provide and the embodiment of the invention provide; As can be seen from the figure; Probably there is the iteration gain of 2dB in system; And 3 times iteration just can obtain all gains basically, and it is very little that 10 iteration obtain gain than 3 iteration.In addition, iterative algorithm is compared the gain that the Almouti detection method has general 1dB under CM1 and CM2 channel condition, along with new time delay increases, under the CM3 channel 10 ^-5Bit error rate the gain of 3dB is probably arranged, and CM4 probably has the gain of 5dB.This is because there is the appearance of error floor in employing Almouti detection method system, and the performance of the detection method that the embodiment of the invention provides does not change with the variation of channel basically, can better work under the channel condition of long time delay expansion.

In sum; The embodiment of the invention obtains sending the probabilistic information that symbol respectively sends bit through adopting Gaussian approximation; Calculate the maximum likelihood ratio that respectively sends symbol according to this probabilistic information again, respectively send symbol according to the maximum likelihood comparison of each transmission symbol then and detect; And carry out iterative detection with the channel decoding algorithm of soft inputting and soft output, and reduced the error floor in the UWB channel, improved the performance of system, and the detection method of the embodiment of the invention is compared detection method of the prior art and is had lower complexity.

The above only is a preferred implementation of the present invention; Should be pointed out that for those skilled in the art, under the prerequisite that does not break away from the principle of the invention; Can also make some improvement and retouching, these improvement and retouching also should be regarded as protection scope of the present invention.

Claims (9)

1. a space frequency group code detection method is characterized in that, comprising:

The interference profile of variable is carried out Gaussian approximation to received signal, obtains the initial probabilistic information that respectively sends symbol, and said initial probabilistic information comprises initial probability, average and variance;

The said initial probability that respectively sends symbol is upgraded, obtain the probability that respectively sends symbol;

Obtain the said maximum likelihood ratio that respectively sends symbol according to the said probability that respectively sends symbol;

Detect based on the said said symbol that respectively sends of maximum likelihood comparison that respectively sends symbol;

Said reception signal is first to n, wherein, and n >=2;

Said initial probability to each transmission symbol upgrades, and obtains the probability that respectively sends symbol, may further comprise the steps:

N initial average and variance substitution first of sending symbol received signal variable, upgrade the initial probability of the first transmission symbol, obtain the probability of the first transmission symbol;

Preceding n-1 average and n of variance substitution of sending symbol received signal variable, upgrade n initial probability that sends symbol, obtain n probability that sends symbol;

So iteration is upgraded, treat respectively to send the probability of symbol stable after, respectively sent the probability of symbol.

2. space frequency group code detection method according to claim 1 is characterized in that,

The interference profile of said variable is to received signal carried out Gaussian approximation, obtains the initial probabilistic information that respectively sends symbol, comprising:

Interference profile to the first reception signal variable is carried out Gaussian approximation, obtains initial probability, average and the variance of the first transmission symbol;

Preceding n-1 average and n of variance substitution of sending symbol received signal variable, obtain n initial probability, average and variance of sending symbol.

3. space frequency group code detection method as claimed in claim 2 is characterized in that, said reception signal is two, and the interference profile of said variable is to received signal carried out Gaussian approximation, obtains the initial probabilistic information that respectively sends symbol, comprising:

Interference profile to the first reception signal variable is carried out Gaussian approximation, obtains initial probability, average and the variance of the first transmission symbol;

The average and the variance substitution second of the said first transmission symbol are received signal variable, obtain initial probability, average and the variance of the second transmission symbol.

4. space frequency group code detection method as claimed in claim 3 is characterized in that, when said reception signal was two, said initial probability to each transmission symbol upgraded, and obtains the probability that respectively sends symbol, comprising:

Send the average of symbol and average and the variance that variance and said second is sent symbol with said first; Substitution said second receives signal variable and said first and receives signal variable respectively; Upgrade said first and send the initial probability that symbol and said second sends symbol, obtain said first and send the probability that symbol and said second sends symbol.

5. like the said space frequency group code detection method of claim 3, it is characterized in that said interference profile to the first reception signal variable is carried out Gaussian approximation, obtains initial probability, average and the variance of the first transmission symbol, comprising:

If said reception signal is a gray mappings, then the average with interference signal in the said first reception signal variable is approximately the Gaussian distribution with identical real part and imaginary part respectively with variance, obtains the initial probability of the first transmission symbol.

6. space frequency group code detection method according to claim 1; It is characterized in that; Said obtain sending the probability of symbol after; Also comprise: if said reception signal adopts chnnel coding, then said reception signal is carried out channel decoding, the new probabilistic information that decoder is provided feeds back to the space-time/frequency block code decoder to carry out iteration and upgrades.

7. like the said space frequency group code detection method of claim 6, it is characterized in that, the said channel decoding of carrying out to received signal, the new probabilistic information that decoder is provided feeds back to the space-time/frequency block code decoder to carry out iteration and upgrades, and comprising:

Feed back to the space-time/frequency block code decoder after new probabilistic information after the decoding of said reception signaling channel interweaved;

The new probabilistic information of said space-time/frequency block code decoder after according to said channel decoding upgrades the probabilistic information of said reception signal.

8. a space-time/frequency block code checkout gear is characterized in that, comprising:

The probability calculation unit, the interference profile that is used for is to received signal carried out Gaussian approximation, obtains the initial probability that respectively sends symbol; The said initial probability that respectively sends symbol is upgraded, obtain the probability that respectively sends symbol; Said reception signal is first to n, wherein, and n >=2;

The maximum likelihood ratio acquiring unit is used for obtaining the maximum likelihood ratio that respectively sends symbol according to the probability of each transmission symbol; With

Detecting unit is used for the maximum likelihood ratio that respectively sends symbol according to said, and the said symbol that respectively sends is detected;

Said initial probability to each transmission symbol upgrades, and obtains the probability that respectively sends symbol, may further comprise the steps:

N initial average and variance substitution first of sending symbol received signal variable, upgrade the initial probability of the first transmission symbol, obtain the probability of the first transmission symbol;

Preceding n-1 average and n of variance substitution of sending symbol received signal variable, upgrade n initial probability that sends symbol, obtain n probability that sends symbol;

So iteration is upgraded, treat respectively to send the probability of symbol stable after, respectively sent the probability of symbol.

9. like the said space-time/frequency block code checkout gear of claim 8, it is characterized in that said probability calculation unit comprises:

The Gaussian approximation subelement is used for the interference profile of the first reception signal variable is carried out Gaussian approximation, obtains initial probability, average and the variance of the first transmission symbol; With

Information iteration subelement is used for initial probability, average and variance are carried out iterative operation, each initial probability, average and variance of sending symbol is upgraded, thereby respectively sent the probability of symbol.

Images