CN1838655A - MIMO-OFDM Receiver - Google Patents

Abstract

A MIMO-OFDM receiver, including: multiple serial-to-parallel converters to organize data from multiple antennas; two groups of OFDM demodulators to receive signals from the serial-to-parallel converters and demodulate the received signals and fast Fourier transform; two channel estimators with the same structure receive data from two groups of OFDM demodulators and estimate the received data; a space-time decoder receives signals from all OFDM demodulators and channel estimators data to perform space-time decoding on the received data. The present invention improves the performance of the receiver by making full use of the additional information provided by the cyclic prefix under the condition that the transmitter does not make any changes, and the complexity of the receiver does not increase much during implementation. The invention has high utilization value in the field of broadband wireless communication.

Description

The MIMO-OFDM receiver

Technical field

The invention belongs to the message pick-up field in duplicating multi-antenna orthogonal frequency division (MIMO-OFDM) communication system, relate generally to channel estimation technique and Data Detection Technology etc.

Background technology

Duplicating multi-antenna orthogonal frequency division (MIMO-OFDM) system is the hot issue that current industrial circle and academia pay close attention to, there have been many wireless communication systems to be based on MIMO-OFDM's, as IEEE802.16e, IEEE802.11n etc., and MIMO-OFDM dominate in the wireless communication system in future very likely.

Adopt the main cause of MIMO-OFDM to be that the MIMO-OFDM system can provide the ability of broadband wireless access, and in broadband access, channel estimating and Data Detection all are very challenging problems.

(Y.Li in background document one, N.Seshadri, and S.Ariyavisitakul, " Channelestimation for OFDM systems with transmitter diversity in mobile wirelesschannels, " IEEE J.S.A.C., vol.17, no.3, pp.461-471, Mar.1999.), proposed a kind of simple and effective channel estimation methods, briefly be described below.

The MIMO-OFDM system as shown in Figure 2, data are at first passed through Space Time Coding device (202) and are carried out Space Time Coding, form the data flow of multidiameter delay, each data flow is passed through OFDM modulator (204) respectively and is carried out the OFDM modulation, comprise operations such as invert fast fourier transformation (IFFT) and interpolation Cyclic Prefix, data after the modulation are launched simultaneously by different antennas, suppose that here transmitting terminal has M antenna.

Signal arrives reception antenna by after the fading channel, and the signal that each reception antenna receives has all comprised the signal and the noise of each transmission antennas transmit that is declined.Here suppose that receiving terminal has N reception antenna.The signal that receives from each reception antenna carries out the demodulation of signal by ofdm demodulator (206), comprise operations such as removing Cyclic Prefix and fast Fourier transform (FFT), the decoder when data after the demodulation send to sky simultaneously (208) and channel estimator (210), after channel estimator (210) is finished channel estimating, decoder (208) when the channel parameter of estimating to obtain is sent to sky, decoded operation when decoder (208) carries out sky when empty then.

Supposing has K subcarrier in the OFDM modulator, the transmission data on i transmitting antenna, a k subcarrier are C _i[k], optional reception antenna, the frequency response of channel on k subcarrier between i transmitting antenna and the reception antenna is H _i[k], the reception data on k subcarrier are x[k], if the length of Cyclic Prefix greater than the maximum delay of channel, the transmission equation that can obtain system so is

$x\left[k\right]=\underset{i=1}{\overset{M}{\mathrm{\Σ}}}{H}_i\left[k\right]C_i\left[k\right]+\mathrm{\η}\left[k\right],$ k＝0，1，…，K-1 (1)

Wherein η [k] is illustrated in k the additive noise on the subcarrier.

If the maximum delay of the channel between i transmitting antenna and the reception antenna is (K ₀-1) the individual sampling interval, channel can be expressed as h on time domain so _i=(h _i[0] h _i[1] ... h _i[K ₀-1]) ^T, the frequency response on k subcarrier is

$H_i\left[k\right]=\underset{l=0}{\overset{K_0-1}{\mathrm{\&Sigma;}}}{h}_i\left[l\right]e^{-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">j</figure-callout>}\frac{2\mathrm{\&pi;}}{K}\mathrm{kl}}---\left(2\right)$

Because MIMO-OFDM generally is applied in the broadband wireless communications, it is very high directly to utilize data in the time domain to carry out the channel estimating complexity, so the channel estimating in the frequency domain is just attractive especially.Channel estimator structure in the background document one as shown in Figure 3.Data after the OFDM demodulation send to pilot frequency locations received signal selector (302), the effect of pilot frequency locations received signal selector (302) is that the reception data that are on the pilot sub-carrier are chosen, if what transmitting terminal sent is targeting sequencing, what choose so is exactly reception data on all subcarriers, if what transmitting terminal sent is to disperse sequence, what choose so just only is the reception data on the pilot sub-carrier.The reception data that pilot frequency locations received signal selector (302) is selected send to correlator (308), correlator (308) takes out the transmission data on the corresponding pilot sub-carrier simultaneously from pilot frequency sequence buffer (304), and with receive data and do special associative operation, suppose and adopt the targeting sequencing mode to send out pilot tone, transmitting terminal has 2 antennas, (can directly be generalized under scattered pilot, the multiple transmit antennas situation, consider various situations in the background document 2), correlator output can be expressed as

$p=\left(\begin{array}{c}{p}_1\\ {\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">p</figure-callout>}}_2\end{array}\right),p_j={\left(\begin{array}{cccc}{p}_j\left[0\right]& p_j\left[1\right]& \&CenterDot;\&CenterDot;\&CenterDot;& p_j[K_0-1]\end{array}\right)}^T,p_j\left[l\right]=\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}x\left[k\right]C_j^{*}\left[k\right]{\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">e</figure-callout>}}^j---\left(3\right)$

The data of correlator (308) output send to matrix multiplier (310), and matrix multiplier (310) takes out constant matrices Q simultaneously from Q matrix buffer (306) ^-1, the multiply operation below carrying out then obtains the estimated value of channel:

$\left(\begin{array}{c}{\hat{h}}_1\\ {\hat{h}}_2\end{array}\right)=Q^{-1}p---\left(4\right)$

Wherein, the matrix Q in the Q matrix buffer (306) ^-1Be to calculate by following matrix inversion:

$Q=\left(\begin{array}{cc}{Q}_{11}& Q_{21}\\ Q_{12}& Q_{22}\end{array}\right)$

$q_{i,j}\left[l\right]=\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{C}_i\left[k\right]C_j^{*}\left[k\right]{\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">e</figure-callout>}}^j$

More than shown in channel estimation methods flow process as shown in Figure 4.

The mean square error (MSE) of described channel estimation methods above background document one has provided:

$\mathrm{MSE}=\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">n</figure-callout>}}^2}{2{\mathrm{<figure-callout\; id="0"\; label="K"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">K</figure-callout>}}_0}\mathrm{Tr}\left(Q^{-1}\right)---\left(5\right)$

Wherein, σ _n ²The variance of expression additive noise.Pattern for even scattered pilot, if a frequency pilot sign is arranged in every m symbol, so mean square error (MSE) just become (5) m doubly, background document two (Bai Wei, " channel estimating in the new generation of wireless communication system ", Shanghai Communications University's doctorate paper, 2003) in detailed derivation is arranged.

The accuracy of channel estimating is most important for the reliability of MIMO-OFDM communication system, and simulation result shows that the channel estimation methods described in the background document 1 needs further to improve.According to the conclusion in the background document 2, the channel estimation methods described in the background document 1 is to be optimum under white Gauss's condition at additive noise, and this is explanation just, if there is not new information source, the performance of channel estimation methods can't be improved.So, need find new information source, to improve performance for estimating channel.

Summary of the invention

The purpose of this invention is to provide a kind of Novel MIM O-OFDM receiver.

For achieving the above object, a kind of MIMO-OFDM receiver comprises:

A plurality of deserializers will be organized from the data of a plurality of antennas;

Two groups of ofdm demodulators are accepted the signal from deserializer, and the signal that receives is separated the mediation fast fourier transform;

The channel estimator of two same structures receives the data from two groups of ofdm demodulators, and the data that receive are estimated;

Decoder when empty receives the data from all ofdm demodulators and channel estimator, the decoding when data that receive are carried out sky.

The present invention does not do under any change condition at transmitter, and receiver improves the performance of receiver by making full use of the additional information that Cyclic Prefix provides, and the receiver complexity increases few when implementing.The present invention has very high value in the broadband wireless communications field.

Description of drawings

The receiver system block diagram that Fig. 1 the present invention proposes;

Fig. 2 is the MIMO-OFDM system block diagram in the background document 1;

Fig. 3 is the block diagram of channel estimation methods in the background document 1;

Fig. 4 is the flow chart of channel estimation methods in the background document 1;

Fig. 5 is the channel estimator block diagram that the present invention proposes;

Fig. 6 be the present invention propose empty the time decoder block diagram;

Fig. 7 be the channel estimation methods that proposes among the present invention and conventional method the MSE performance relatively;

Fig. 8 be the receiver scheme method that proposes among the present invention and traditional scheme performance relatively.

Embodiment

The receiver structure that proposes about the present invention as shown in Figure 1, detailed implementing procedure is as follows:

Here suppose that the length that comprises M transmitting antenna, a K subcarrier, Cyclic Prefix in the MIMO-OFDM system is L _Cp, lock unit estimated channel length is K ₀-1, receiving terminal has N reception antenna.The signal that receives from each reception antenna obtains K+L by two deserializers (102) _CpIndividual data, L _Cp+ 1 to L _Cp+ K data send to one group of ofdm demodulator (104) and carry out the demodulation of signal, carry out the fast Fourier transform (FFT) operation, select parameter J:0≤J≤L _Cp-K ₀+ 1, L _Cp+ 1 to L _Cp+ K-J data and L _Cp-J+1 is to L _CpIndividual data send to another group ofdm demodulator (104) in order and carry out the demodulation of signal, carry out the fast Fourier transform (FFT) operation.Decoder (108) and two channel estimators (106) when the data after two groups of ofdm demodulators (104) demodulation send to sky simultaneously, after channel estimator (106) is finished channel estimating, decoder (108) when the channel parameter of estimating to obtain is sent to sky, decoded operation when decoder (108) obtains carrying out sky after the data that all ofdm demodulators (104) and all channel estimators (106) send when empty then.

The result of channel estimator (106) as shown in Figure 5, detailed implementation step is as follows:

Data after ofdm demodulator 1 demodulation send to pilot frequency locations received signal selector (502), the effect of pilot frequency locations received signal selector (502) is that the reception data that are on the pilot sub-carrier are chosen, if what transmitting terminal sent is targeting sequencing, what choose so is exactly reception data on all subcarriers, suppose to obtain data and be expressed as x[0], x[1], x[K-1], if what transmitting terminal sent is to disperse sequence, what choose so just only is reception data on the pilot sub-carrier, is example here with the targeting sequencing.Data after ofdm demodulator 2 demodulation send to pilot frequency locations received signal selector (502), pilot frequency locations received signal selector (502) is selected the reception data that are on the pilot sub-carrier according to pilot frequency mode, what suppose the transmitting terminal transmission is targeting sequencing, then obtain data and be expressed as r[0], r[1],, r[K-1].

The reception data that two pilot frequency locations received signal selectors (502) are selected send to correlator (504) respectively, correlator (504) takes out the transmission data on the corresponding pilot sub-carrier simultaneously from pilot frequency sequence buffer (506), suppose that targeting sequencing is expressed as C _i[0], C _i[1] ..., C _i[K-1], correlator (504) is done special associative operation with pilot frequency sequence and reception data.Be without loss of generality, suppose that transmitting terminal has 2 antennas, two correlator outputs are expressed as respectively

$p=\left(\begin{array}{c}{p}_1\\ {\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">p</figure-callout>}}_2\end{array}\right),p_j={\left(\begin{array}{cccc}{p}_j\left[0\right]& p_j\left[1\right]& \&CenterDot;\&CenterDot;\&CenterDot;& p_j[K_0-1]\end{array}\right)}^T,p_j\left[l\right]=\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}x\left[k\right]C_j^{*}\left[k\right]{\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">e</figure-callout>}}^j$

$s=\left(\begin{array}{c}{s}_1\\ {\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">s</figure-callout>}}_2\end{array}\right),s_i={\left(\begin{array}{cccc}{s}_i\left[0\right]& s_i\left[1\right]& \&CenterDot;\&CenterDot;\&CenterDot;& s_i[K_0-1]\end{array}\right)}^T,s_j\left[l\right]=\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}x\left[k\right]C_j^{*}\left[k\right]{\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">e</figure-callout>}}^j$

The data of two correlators (504) output send to adder (508), carry out add operation, obtain

P＝p+s

The data of adder (508) output send to matrix multiplier (512), and matrix multiplier (512) takes out constant matrices (2Q) simultaneously from Q matrix buffer (510) ^-1, the multiply operation below carrying out then obtains the estimated value of channel:

$\left(\begin{array}{c}{\hat{h}}_1\\ {\hat{h}}_2\end{array}\right)={\left(2Q\right)}^{-1}P$

Wherein, the matrix (2Q) in the Q matrix buffer (510) ^-1Be to calculate by following matrix inversion:

$Q=\left(\begin{array}{cc}{Q}_{11}& Q_{21}\\ Q_{12}& Q_{22}\end{array}\right)$

$q_{i,j}\left[l\right]=\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{C}_i\left[k\right]C_j^{*}\left[k\right]{\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">e</figure-callout>}}^j$

h _i＝(h _i[0] h _i[1]…h _i[K ₀-1]) ^T

The mean square error of channel estimation methods of the present invention (MSE):

$\mathrm{MSE}=(1-\frac{J}{2K})\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">n</figure-callout>}}^2}{2{\mathrm{<figure-callout\; id="0"\; label="K"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">K</figure-callout>}}_0}\mathrm{Tr}\left(Q^{-1}\right)$

Wherein, σ _n ²The variance of expression additive noise.For the pattern of even scattered pilot, if a frequency pilot sign is arranged in every m symbol, so mean square error (MSE) just become following formula m doubly.

The result of decoder (108) as shown in Figure 6 when empty, the basic skills of implementing be the data that different ofdm demodulators come out regard that different reception antennas receive as data, but the channel fading of their experience is identical, Data Detection Algorithm when using use always empty then, as zero forcing algorithm, MMSE algorithm and maximum likelihood algorithm etc., carry out Data Detection.

Embodiment

A wireless communication system that adopts multi-antenna technology, transmitting terminal have two antennas, receiving terminals to have that the length of two antennas, 1024 subcarriers, Cyclic Prefix is 256, lock unit estimated channel length is 7.Transmitting terminal obtains two channel parallel datas through 4QAM modulation and Alamouti space-frequency coding, then through sending after the OFDM modulation.

Suppose channel for adding up independently Rayleigh decline, maximum delay is 7 sampling intervals, and interchannel noise is modeled as additive white Gaussian noise.The signal that receives from each reception antenna obtains 1280 data by deserializer (102), the the 257th to the 1280th data send to an ofdm demodulator (104) and carry out the demodulation of signal, the operation of execution fast Fourier transform (FFT), select parameter J to be respectively 0,100 and 250, here we claim the J value to be exchange length, the 257th to 1280-J data and 256 data of 257-J to the send to another ofdm demodulator (104) in order and carry out the demodulation of signal, carry out the fast Fourier transform (FFT) operation.Decoder (108) and channel estimator (106) when the data after two ofdm demodulators (104) demodulation send to sky simultaneously.Supposing the system at first sends the targeting sequencing of a frequency domain before sending data.Data after ofdm demodulator 1 demodulation send to pilot frequency locations received signal selector (502), the effect of pilot frequency locations received signal selector (502) is that the reception data that are on the pilot sub-carrier are chosen, obtain data x[0], x[1], x[K-1], data after ofdm demodulator 2 demodulation send to pilot frequency locations received signal selector (502), pilot frequency locations received signal selector (502) is selected the reception data that are on the pilot sub-carrier according to pilot frequency mode, obtain data r[0], r[1] ..., r[K-1].The reception data that two pilot frequency locations received signal selectors (502) are selected send to correlator (504) respectively, correlator (504) takes out the transmission data on the corresponding pilot sub-carrier simultaneously from pilot frequency sequence buffer (506), suppose that targeting sequencing is expressed as C _i[0], C _i[1] ..., C _i[K-1], correlator (504) is done special associative operation with pilot frequency sequence and reception data.Two correlator outputs are expressed as respectively

$p=\left(\begin{array}{c}{p}_1\\ {\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">p</figure-callout>}}_2\end{array}\right),p_j={\left(\begin{array}{cccc}{p}_j\left[0\right]& p_j\left[1\right]& \&CenterDot;\&CenterDot;\&CenterDot;& p_j[K_0-1]\end{array}\right)}^T,p_j\left[l\right]=\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}x\left[k\right]C_j^{*}\left[k\right]{\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">e</figure-callout>}}^j$

$s=\left(\begin{array}{c}{s}_1\\ {\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">s</figure-callout>}}_2\end{array}\right),s_i={\left(\begin{array}{cccc}{s}_i\left[0\right]& s_i\left[1\right]& \&CenterDot;\&CenterDot;\&CenterDot;& s_i[K_0-1]\end{array}\right)}^T,s_j\left[l\right]=\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}x\left[k\right]C_j^{*}\left[k\right]{\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">e</figure-callout>}}^j$

The data of two correlators (504) output send to adder (508), carry out add operation, obtain P=p+s.

The data of adder (508) output send to matrix multiplier (512), and matrix multiplier (512) takes out constant matrices (2Q) simultaneously from Q matrix buffer (510) ^-1, the multiply operation below carrying out then obtains the estimated value of channel:

$\left(\begin{array}{c}{\hat{h}}_1\\ {\hat{h}}_2\end{array}\right)={\left(2Q\right)}^{-1}P$

Wherein, the matrix (2Q) in the Q matrix buffer (510) ^-1Be to calculate by following matrix inversion:

$Q=\left(\begin{array}{cc}{Q}_{11}& Q_{21}\\ Q_{12}& Q_{22}\end{array}\right)$

$q_{i,j}\left[l\right]=\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{C}_i\left[k\right]C_j^{*}\left[k\right]{\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="A20051005919000181.png"\; state="\{\{state\}\}">e</figure-callout>}}^j$

h _i＝(h _i[0] h _i[1]…h _i[K ₀-1]) ^T

After decoder (108) obtains the data of the dateout of all ofdm demodulators and all channel estimators when empty, be to detect data d[2k], d[2k+1], need to take out the dateout x of ofdm demodulator ₁[2k], x ₁[2k+1], x ₂[2k], x ₂[2k+1], r ₁[2k], r ₁[2k+1], r ₁[2k], r ₁[2k+1] obtains column vector R, R (k)=(x ₁[2k] x ₁ ^*[2k+1] x ₂[2k] x ₂ ^*[2k+1] r ₁[2k] r ₁ ^*[2k+1] r ₂[2k] r ₂ ^*[2k+1]) ^TK=0 wherein, 1 ..., 511, subscript is represented reception antenna, () ^*Common strategic point is got in expression.

Also to take out the frequency domain channel data that estimation obtains, obtain channel vector H (k),

$H\left(k\right)=\left[\begin{array}{cc}{H}_{11}\left(2k\right)& H_{21}\left(2k\right)\\ {H_{21}}^{*}(2k+1)& -{H_{11}}^{*}(2k+1)\\ H_{12}\left(2k\right)& H_{22}\left(2k\right)\\ {{H_{22}}^{*}(2k+1)}^{}& -{H_{12}}^{*}(2k+1)\\ H_{11}\left(2k\right)& H_{21}\left(2k\right)\\ {H_{21}}^{*}(2k+1)& {{-H}_{11}}^{*}(2k+1)\\ H_{12}\left(2k\right)& H_{22}\left(2k\right)\\ {H_{22}}^{*}(2k+1)& -{H_{12}}^{*}(2k+1)\end{array}\right]$

H wherein _Ij(k) value of channel on k subcarrier between i transmitting antenna of expression and j the reception antenna.

Adopt Alamouti maximum-likelihood decoding method to obtain data d[2k then], d[2k+1] estimated value:

$\left(\begin{array}{c}\hat{d}\left[2k\right]\\ \hat{d}[2k+1]\end{array}\right)=\arg \min \{R\left(k\right)-H\left(k\right)\left(\begin{array}{c}d\left[2k\right]\\ d[2k+1]\end{array}\right)\}$

Fig. 7 and Fig. 8 are the receiver scheme that proposes of the present invention and the performance comparative analysis of traditional scheme, and wherein traditional scheme refers to channel estimation methods described in the background document 1 and Data Detection method.Fig. 7 is that the MSE performance of channel estimating compares, and Fig. 8 is the comparison of receiver performance.As can be seen from the figure, with respect to traditional scheme, the receiver scheme that the present invention proposes has certain improvement on performance, and the receiver scheme that the present invention proposes do not need transmitter to do any change, and receiving the complexity increase neither be a lot.

Claims (6)

1. MIMO-OFDM receiver comprises:

A plurality of deserializers will be organized from the data of a plurality of antennas;

Two groups of ofdm demodulators are accepted the signal from deserializer, and the signal that receives is separated the mediation fast fourier transform;

The channel estimator of two same structures receives the data from two groups of ofdm demodulators, and the data that receive are estimated;

Decoder when empty receives the data from all ofdm demodulators and channel estimator, the decoding when data that receive are carried out sky.

2. by the described receiver of claim 1, it is characterized in that:

First data block of one of deserializer output is to first group of ofdm demodulator, and first group of data block is (Lcp+1) data of the time domain OFDM symbol that receives last data to the OFDM symbol;

Second data block of another group deserializer output is to another group ofdm demodulator, second data block is that (Lcp+1) data of the time domain OFDM symbol that receives are to (Lcp+K-J, K is the number of sub carrier wave of OFDM) data, add in the back that then (Lcp-J+1) data of OFDM symbol are to Lcp data.

3. by the described receiver of claim 2, it is characterized in that second data block of deserializer output forms in the following way: be not subjected to the part or all of data that data are harassed in the previous OFDM symbol with in the OFDM symbol cyclic prefix those and replaced J last in OFDM symbol data.

4. by the described receiver of claim 1, it is characterized in that described channel estimator comprises:

Frequency position, two roads received signal selector selects to be in the reception data on the pilot sub-carrier;

Two correlators receive the data from frequency position, road received signal selector, take out the transmission data on the corresponding pilot sub-carrier simultaneously from the pilot frequency sequence buffer;

Adder is carried out add operation to the data from two correlators;

Matrix multiplier receives the data from adder output, takes out constant matrices simultaneously and carry out multiplying from Q matrix buffer.

5. by the described receiver of claim 4, it is characterized in that the matrix (2Q)-the 1st in the Q matrix buffer, calculate by following matrix inversion:

$Q=\left(\begin{array}{cc}{Q}_{11}& Q_{21}\\ Q_{12}& Q_{22}\end{array}\right)$

$q_{i,j}\left[l\right]=\underset{k=0}{\overset{k=1}{\mathrm{\&Sigma;}}}{C}_l\left[k\right]C_l^{*}\left[k\right]e^{j\frac{2\mathrm{\&pi;}}{K}\mathrm{kl}}$

h _l=(h _l[0]?h _l[1]…h _l[k ₀-1] ^T

6. by the described receiver of claim 4, the form that it is characterized in that the Q matrix is by the special pilot frequency sequence C of design _l[0], C _l[1] ..., C _l[K-1] makes that the Q matrix is that a unit matrix is multiplied by an invariant.

Images