RECEIVER FOR ALAMOUTI TYPE SPACE-TIME BLOCK CODING FBMC SYSTEM

Abstract

A method of reception of signals transmitted by a FBMC transmitter using a block Alamouti coding. After demodulation in a base band, the received signal is sampled, with the sample blocks undergoing a sliding FFT before being de-multiplexed towards a first path during a first use of the channel and a second path during a second use of the channel. The vectors received on the first path are multiplied by a first and a second transfer matrix, conjugated to provide first and second vectors. The vectors received on a second path undergo time-reversal and complex conjugation and, if appropriate, multiplication by an imaginary factor, depending on the size of the blocks. The vectors thus obtained are multiplied by first and second transfer matrices to provide third and fourth vectors. The first and fourth (second and third vectors) are then combined and the combined vector is filtered and spectrally de-spread to give an estimate of the block transmitted by the first (second) antenna of the transmitter during the first use of the channel.

Claims

1. A method for receiving signals transmitted by an FBMC transmitter using block Alamouti coding, where the FBMC transmitter uses a plurality N of sub-carriers and an overlap factor K of prototype filters, where the signal received by the receiver is translated to a base band, sampled at a frequency Nf where f is the half-frequency of the FBMC symbols, then undergoes a sliding FFT of size KN to provide sample vectors, wherein said sample vectors are received on a first path during a first channel use and on a second path during a second channel use, the channel comprising a first elementary channel between a first antenna of the transmitter and a first antennae of the receiver and a second elementary channel between a second antenna of the transmitter and a second antenna of the receiver, said first and second elementary channels being characterized, respectively, by a first and a second transfer matrix (H.sub.0,H.sub.1), each vector (W.sub.0.sup.m) in a sequence of vectors received on the first path being multiplied by the conjugate of the first transfer matrix in order to provide a first vector and by the conjugate of the second transfer matrix in order to provide a second vector, where a sequence of vectors received on the second path is time-reversed and each vector (W.sub.1.sup.L−1−m) of the second sequence is conjugated before being multiplied by the first transfer matrix to provide a third vector and by the second transfer matrix to provide a fourth vector, where the first and fourth vectors are combined to provide a first combined vector, the second and third vectors are combined to provide a second combined vector, where the first and second combined vectors are spatially de-spread and filtered in the frequency domain by the prototype filters to provide, respectively, an estimate of a first data vector ({circumflex over (X)}.sub.0.sup.m) transmitted via the first antenna of the transmitter and of a second data vector ({circumflex over (X)}.sub.1.sup.m) transmitted via the second transmitter antenna.

2. The method of reception of signals transmitted by an FBMC transmitter according to claim 1, wherein the first and second antennas of the receiver form a single antenna, and in that said sample vectors are demultiplexed on the first path during the first use of the channel and on the second path during the second use of the channel.

3. The method of reception of signals transmitted by a FBMC transmitter according to claim 1, wherein the first and second antennas are distinct, with the first path being associated with the first antenna of the receiver and the second path being associated with the second antenna of the receiver.

4. The method of reception of signals transmitted by a FBMC transmitter according to claim 1, where the transmitter uses the following matrix as a matrix for block Alamouti coding: $\stackrel{\_}{C}=\left(\begin{array}{cc}{\stackrel{\_}{X}}_0& {\stackrel{\_}{X}}_1\\ -{\stackrel{\_}{X}}_1.Math.T& {\stackrel{\_}{X}}_0.Math.T\end{array}\right)$ where X.sub.0 and X.sub.1 are first and second blocks of input data vectors transmitted during the first use of the channel via the first antenna and the second antenna of the receiver respectively, X.sub.1T is a first transformed block obtained by time-reversal and of the second block, X.sub.0T is a second transformed block obtained by time-reversal of the first block, where the blocks −X.sub.1T and X.sub.0T are transmitted during the second use of the channel via the first antenna and the second antenna of the transmitter respectively, wherein the vectors of the second sequence are multiplied by a factor (j.sup.L−1) where L is the size of the first and second blocks of input data vectors, after conjugation and before multiplication by the first and second transfer matrices.

5. The method of reception of signals transmitted by a FBMC receiver according to claim 4, wherein the input data vector of rank m in the first block of input data vectors, X.sub.0.sup.m, and the input data vector of rank m in the second block of input data vectors, X.sub.1.sup.m, are estimated from:
{circumflex over (X)}.sub.0.sup.m=μG(H.sub.0*W.sub.0.sup.m+j.sup.L−1H.sub.1W.sub.1.sup.L−1−m*)⊙M.sup.m*
{circumflex over (X)}.sub.1.sup.m=μG(H.sub.1*W.sub.0.sup.m−j.sup.L−1H.sub.0W.sub.1.sup.L−1−m*)⊙M.sup.m* where {circumflex over (X)}.sub.0.sup.m and X.sub.1.sup.m are respectively the estimates of the vectors X.sub.0.sup.m and X.sub.1.sup.m, H.sub.0, H.sub.1, are respectively the first and second transfer matrices, W.sub.0.sup.m the vector of rank m received on the first path, W.sub.1.sup.L−1−m the vector of rank L−1−m received on the second path, M.sup.m a vector which represents a OQAM coding of vectors X.sub.0.sup.m and X.sub.1.sup.m, G a matrix which represents, in the frequency domain, a spectral de-spreading and filtering by prototype filters, μ is a coefficient of normalisation and ⊙ represents the Hadamard product.

6. The method of reception of the signals transmitted by a FBMC transmitter according to claim 1, wherein the transmitter uses the following matrix as a matrix for block Alamouti coding: ${\stackrel{\_}{C}}^{\prime }=\left(\begin{array}{cc}{\stackrel{\_}{X}}_0& {\stackrel{\_}{X}}_1\\ -\left(j^{L-1}\right).Math.{\stackrel{\_}{X}}_1.Math.T& \left(j^{L-1}\right).Math.{\stackrel{\_}{X}}_0.Math.T\end{array}\right)$ where X.sub.0 and X.sub.1 are the first and second blocks of input data vectors transmitted during the first use of the channel via the first antenna and the second antenna, respectively, of the transmitter, X.sub.1T is a first transformed block obtained by time-reversal and of the second block, X.sub.0T is a second transformed block obtained by time reversal of the first block, wherein the vectors of the second sequence, after conjugation, are multiplied directly by the first and second transfer matrices.

7. The method of reception of signals transmitted by a FBMC transmitter according to claim 6, wherein the input data vector of rank m in the first block of input data vectors, X.sub.0.sup.m, and the input data vector of rank m in the second block of input data vectors, X.sub.1.sup.m, are estimated from:
{circumflex over (X)}.sub.0.sup.m=μG(H.sub.0*W.sub.0.sup.m+H.sub.1W.sub.1.sup.L−1−m*)⊙M.sup.m*
{circumflex over (X)}.sub.1.sup.m=μG(H.sub.1*W.sub.0.sup.m−H.sub.0W.sub.1.sup.L−1−m*)⊙M.sup.m* where {circumflex over (X)}.sub.0.sup.m and {circumflex over (X)}.sub.1.sup.m are respectively the estimates of the vectors X.sub.0.sup.m and X.sub.1.sup.m, H.sub.0, H.sub.1 are respectively the first and second transfer matrices, W.sub.0.sup.m the vector of rank m received on the first path, W.sub.1.sup.L−1−m the vector of rank L−1−m received on the second path, M.sup.m a vector which represents a OQAM coding of vectors X.sub.0.sup.m and X.sub.1.sup.m, G a matrix which represents, in the frequency domain, a spectral de-spreading and filtering by prototype filters, μ is a coefficient of normalisation and ⊙ represents the Hadamard product.

8. The method of reception of signals transmitted by an FBMC transmitter according to claim 4, where the first and second blocks of input data vectors are preceded, respectively, by first and second preambles, a first guard block made up of null vectors separating the first block of data vectors and the first transformed block, a second guard block made up of null vectors separating the second block of data vectors and the second transformed block, wherein the first and second preambles are known to the receiver and that on the first path, at the output of the sliding FFT, elimination of the interference affecting the vectors received on the first path is carried out by subtraction at this path of the contribution due to the first and second preambles.

Description

BRIEF DESCRIPTION OF THE ILLUSTRATIONS

[0056] Other characteristics and advantages of the invention will appear on reading the preferential embodiments of the invention made in reference to the attached figures, among which:

[0057] FIG. 1 schematically shows a FS-FBMC telecommunication system known to the prior art;

[0058] FIG. 2A shows the spectral spreading undertaken upstream of the IFFT module of FIG. 1;

[0059] FIG. 2B shows the spectral de-spreading undertaken downstream of the IFFT module of FIG. 1;

[0060] FIG. 3 shows the combination of FBMC symbols in FIG. 1;

[0061] FIG. 4 schematically shows the transmission of two sequences of blocks of symbols by an FBMC transmitter using a block Alamouti coding known in the prior art;

[0062] FIG. 5 schematically shows the architecture of an FBMC receiver used to receive sequences of blocks of symbols transmitted by the transmitter in FIG. 4;

[0063] FIG. 6 schematically shows the architecture of an FS-FBMC receiver, according to a first embodiment of the invention, used to receive sequences of blocks of symbols coded by a block Alamouti coding;

[0064] FIG. 7 schematically shows the architecture of a FS-FBMC receiver according to one alternative of the first embodiment of the invention.

[0065] FIG. 8A schematically shows a first example of transmission of two sequences of blocks of symbols by an FBMC transmitter using a first block Alamouti coding;

[0066] FIG. 8B schematically shows a second example of transmission of two sequences of blocks of symbols by an FBMC transmitter using a second block Alamouti coding;

[0067] FIG. 9 schematically shows the architecture of an FS-FBMC receiver, according to a second embodiment of the invention, used to receive sequences of blocks transmitted according to the schematic representation in FIG. 8B.

DETAILED DESCRIPTION OF PARTICULAR EMBODIMENTS

[0068] In order to facilitate understanding of the notations, we will first of all consider an FS-FBMC transmitter as described in relation to FIG. 1.

[0069] Unlike preceding notations, the column vectors X.sup.m, m=0, . . . , L−1, of size N, will, in what follows, represent the input data vectors, in other words the data at the input to the OQAM modulator. The elements of these vectors are therefore real value elements.

[0070] The signal transmitted by the transmitter at the instant m can be represented by a column vector Z.sup.m of size KN whose elements are samples at a frequency Nf. The vector Z.sup.m can be expressed as a function of the input data vectors X.sup.m−(K−1), . . . , X.sup.m, . . . , X.sup.m+(K−1), that is:

[00008]$\begin{array}{cc}{Z}^m=F^H.Math.G\left(X^m\odot M^m\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.Q_{\frac{\mathrm{pN}}{2}}.Math.F^H.Math.G\left(X^{m-p}\odot M^{m-p}\right)+Q_{\mathrm{KN}-\frac{\mathrm{pN}}{2}}.Math.F^H.Math.G\left(X^{m+p}\odot M^{m+p}\right)& \left(8\right)\end{array}$

where ⊙ is the Hadamard product, F is the discrete Fourier transform matrix of size KN×KN, G is a matrix of size KN×N representing the spectral spreading and the transfer function of the prototype filter in the frequency domain, that is:

[00009]$\begin{array}{cc}G=\left(\begin{array}{cccc}{G}_{K-1}& 0& \mathrm{.Math.}& 0\\ \mathrm{.Math.}& G_{K-1}& \ddots & \mathrm{.Math.}\\ G_0& \mathrm{.Math.}& \ddots & 0\\ \mathrm{.Math.}& G_0& \ddots & G_{K-1}\\ G_{-K+1}& \mathrm{.Math.}& \ddots & \mathrm{.Math.}\\ 0& G_{-K+1}& \ddots & G_0\\ \mathrm{.Math.}& \ddots & \ddots & \mathrm{.Math.}\\ 0& \mathrm{.Math.}& 0& G_{-K+1}\end{array}\right)& \left(9\right)\end{array}$

[0071] M.sup.m is a column vector of size N which expresses the OQAM modulation, namely a vector whose elements are given by:

M.sup.m[k]=j.sup.m+k(−1).sup.km  (10)

and Q.sub.l is an offset matrix of l samples, of size KN×KN defined by:

[00010]$\begin{array}{cc}{Q}_l=\left(\begin{array}{cc}{0}_{l\times \left(\mathrm{KN}-l\right)}& 0_{l\times l}\\ I_{\mathrm{KN}-l}& 0_{\left(\mathrm{KN}-l\right)\times l}\end{array}\right)& \left(11\right)\end{array}$

where T .sub.KN−is the identity matrix of size (KN−l)×(KN−)

[0072] It will be understood that the terms beneath the summation sign in expression (8) represent the 2K−1 FBMC symbols which are combined in FIG. 3.

[0073] The signal received by the FBMC receiver at the instant in may similarly be expressed in the form of a data vector at the output of the OQAM demodulator, here referred to as Y.sup.m, of size KN. The vector Y.sup.m can be expressed as a function of the vector Z.sup.m which represents the transmitted signal, either by carrying out abstraction of the noise term:

Y.sup.m=(G.sup.HFH.sub.0Z.sup.m)⊙M.sup.m*  (12)

or, given that G.sup.HFF.sup.HG=I.sub.N and that (X.sup.m ⊙M.sup.m)⊙M.sup.m*=X.sup.m:

[00011]$\begin{array}{cc}{Y}^m=H_0\left(X^m+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.U^p\left(X^{m-p}\odot M^{m-p}\right)\odot M^{m^{*}}+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.V^p\left(X^{m+p}\odot M^{m+p}\right)\odot M^{m^{*}}\right).Math..Math.\text{where:}& \left(13\right)\\ U^p=G^H.Math.{\mathrm{FQ}}^{\frac{\mathrm{pN}}{2}}.Math.F^H.Math.G.Math..Math.\mathrm{and}.Math..Math.V^p=G^H.Math.{\mathrm{FQ}}^{\mathrm{KN}-\frac{\mathrm{pN}}{2}}.Math.F^H.Math.G& \left(14\right)\end{array}$

[0074] It will be seen that G.sup.H=G.sup.T given that the coefficients of the filter transfer matrix are real.

[0075] It is now assumed that a block Alamouti coding is carried out, with a coding matrix defined by:

[00012]$\begin{array}{cc}\stackrel{\_}{C}=\left(\begin{array}{cc}{\stackrel{\_}{X}}_0& {\stackrel{\_}{X}}_1\\ -{\stackrel{\_}{X}}_1.Math.T& {\stackrel{\_}{X}}_0.Math.T\end{array}\right)& \left(15\right)\end{array}$

[0076] The basic idea of the invention is to use a receiver implemented in the frequency domain (FS-FBMC receiver) and to combine the two blocks at the output of the FFT module (module 170 in FIG. 1), during the first and second use of the channel respectively.

[0077] Then X.sub.0.sup.m is the m.sup.th input data vector of the first block X.sub.0 and X.sub.1.sup.m the m.sup.th input data vector of the second block X.sub.1 respectively. Furthermore W.sub.0.sup.m is the m.sup.th sample vector at the output to the FFT module, before de-spreading and filtering, during the first use of the channel. Similarly, W.sub.1.sup.m is the m.sup.th sample vector at the output of the FFT module, before de-spreading and filtering, during the second use of the channel.

[0078] During the first use of the channel the vector W.sub.0.sup.m can be expressed as follows:

[00013]$\begin{array}{cc}{W}_0^m=H_0\left(G\left(X_0^m\odot M^m\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.A^p\left(X_0^{m-p}\odot M^{m-p}\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.B^p\left(X_0^{m+p}\odot M^{m+p}\right)\right)+H_1\left(G.Math.\left(X_1^m\odot M^m\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.A^p\left(X_1^{m-p}\odot M^{m-p}\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.B^p\left(X_1^{m+p}\odot M^{m+p}\right)\right)& \left(16\right)\\ \mathrm{where}.Math..Math.A^p={\mathrm{FQ}}^{\frac{\mathrm{pN}}{2}}.Math.F^H.Math..Math.\mathrm{and}.Math..Math.B^p={\mathrm{FQ}}^{\mathrm{KN}-\frac{\mathrm{pN}}{2}}.Math.F^H.& \left(17\right)\end{array}$

[0079] Similarly, during the second use of the channel the vector W.sub.1.sup.m can be expressed as follows:

[00014]$\begin{array}{cc}{W}_1^m=-H_0\left(G\left(X_1^{L-1-m}\odot M^m\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.A^p\left(X_1^{L-1-m+p}\odot M^{m-p}\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.B^p\left(X_1^{L-1-m-p}\odot M^{m+p}\right)\right)+H_1\left(G.Math.\left(X_0^{L-1-m}\odot M^m\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.A^p\left(X_0^{L-1-m+p}\odot M^{m-p}\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.B^p\left(X_0^{L-1-m-p}\odot M^{m+p}\right)\right)& \left(18\right)\end{array}$

an expression in which good use has been made of the fact that the input data vectors were real value vectors.
It will be seen that the transfer matrices for the elementary channels, H.sub.0 and H.sub.1 are here of size KN×KN due to the spectral spreading.

[0080] If the block of vectors W.sub.1.sup.m, m=0, . . . , L−1, is transformed by time-reversal and complex conjugation of the block, the m.sup.th vector of the block transformed in this way may be written, from (18):

[00015]$\begin{array}{cc}{W}_1^{L-m-1^{*}}=-H_0^{*}\left(G\left(X_1^m\odot M^{L-1-m^{*}}\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.A^{p^{*}}\left(X_1^{m+p}\odot M^{L-1-m-p^{*}}\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.B^{p^{*}}\left(X_1^{m-p}\odot M^{L-1-m+p^{*}}\right)\right)+H_1^{*}\left(G\left(X_0^m\odot M^{L-1-m^{*}}\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.A^{p^{*}}\left(X_0^{m+p}\odot M^{L-1-m-p^{*}}\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.B^{p^{*}}\left(X_0^{m-p}\odot M^{L-1-m+p^{*}}\right)\right)& \left(19\right)\end{array}$

That is by taking into consideration that:

M.sup.L−1−m*=−M.sup.mj.sup.L−1; M.sup.L−1−m-p*=−M.sup.m+pj.sup.L−1; M.sup.L−1−m+p*=−M.sup.m−pj.sup.L−1

where it is assumed that the size L of the block was an even number, and that:

A.sup.p*=B.sup.p; B.sup.p*=A.sup.p

the vector W.sub.1.sup.L−m−1* of the reversed block can finally be written as:

[00016]$\begin{array}{cc}{W}_1^{L-m-1^{*}}=H_0^{*}.Math.j^{L-1}\left(G\left(X_1^m\odot M^m\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.B^p\left(X_1^{m+p}\odot M^{m+p}\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.A^p\left(X_1^{m-p}\odot M^{m-p}\right)\right)-H_1^{*}.Math.j^{L-1}\left(G\left(X_0^m\odot M^m\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.B^p\left(X_0^{m+p}\odot M^{m+p}\right)+\underset{p=1}{\overset{K-1}{.Math.}}.Math..Math.B^p\left(X_0^{m-p}\odot M^{m-p}\right)\right)& \left(20\right)\end{array}$

[0081] The vectors of the transmitted data X.sub.0.sup.m, X.sub.1.sup.m can be estimated by undertaking a combination of vectors W.sub.0.sup.m and W.sub.1.sup.L−m−1*

{hacek over (X)}.sub.0.sup.m=μ(H.sub.0*W.sub.0.sup.m+j.sup.L−1H.sub.1W.sub.1.sup.L−1−m*)  (21-1)

{hacek over (X)}.sub.1.sup.m=μ(H.sub.1*W.sub.0.sup.m+j.sup.L−1H.sub.0W.sub.1.sup.L−1−m*)  (21-2)

where

[00017]$\mu =\frac{1}{\mathrm{Tr}\left(H_0^H.Math.H_0+H_1^H.Math.H_1\right)},$

then filtering and spectral de-spreading and finally an OQAM demodulation:

{circumflex over (X)}.sub.0.sup.m=μG(H.sub.0*W.sub.0.sup.m+j.sup.L−1H.sub.1W.sub.1.sup.L−1−m*)⊙M.sup.m*  (22-1)

{circumflex over (X)}.sub.1.sup.m=μG(H.sub.1*W.sub.0.sup.m−j.sup.L−1H.sub.0W.sub.1.sup.L−1−m*)⊙M.sup.m*  (22-2)

[0082] FIG. 6 schematically shows the architecture of an FS-FBMC receiver, according to a first embodiment of the invention, used to receive sequences of blocks of symbols coded by a block Alamouti coding.

[0083] The receiver comprises a sampling module 610 for sampling the signal received in a base band at the rate of Nf where N is the number of sub-carriers and f is the frequency of the FBMC symbols. The samples are grouped together in the form of blocks of size KN by a serial-parallel converter 620.

[0084] The receiver is assumed to be synchronised on the FBMC symbols, in other words, the start of an FFT window coincides with the first sample of an FBMC symbol (transmitted by one or the other of the transmission antennae). Moreover, the receiver is assumed to be synchronised on the instants of use of the channel so that it knows the instants at which the first and second blocks are received.

[0085] The blocks of samples undergo a FFT of size KN in the FFT module 630.

[0086] A demultiplexer 640 supplies the vectors at the output from the FFT at a first output 641 during the first use of the channel and at a second output 642 during the second use of the channel. The L vectors (of size KN) generated sequentially at the first output are stored in a first buffer memory 651, configured in the form of a FIFO (first-in first-out) buffer. The L vectors generated sequentially at the second output are also stored in a second buffer memory 652 configured in the form of a LIFO (last-in first-out) buffer. The module 660 thus reads the L vectors in the reverse order to that in which they are stored (LIFO), so as to perform a time-reversal and furthermore undertakes a complex conjugation of each of these vectors. A multiplier 670 multiplies the elements of the vectors at the output from the module 660 by (j).sup.L−1, in other words by j if L is an even number.

[0087] Each element of a vector generated at the first output is multiplied in 681 by the complex conjugate of the coefficient of the first elementary channel between the first transmission antenna and the reception antenna, at the frequency of the sub-carrier carrying the element in question (the operation is symbolised here by a multiplication of the vector at the output from the buffer memory by the matrix H.sub.0*) and in 683 by the complex conjugate of the coefficient of the second elementary channel between the second transmission antenna and the reception antenna, at the same sub-carrier frequency (the operation is symbolised here by a multiplication of the vector of samples at the FFT output by the matrix H.sub.1*). The matrices H.sub.0 and H.sub.1 are here meant to be of size KN×KN and here represent the coefficients of the elementary channels for the KN spectrally spread sub-carriers. An identical channel coefficient can be chosen for the K frequencies produced by a given sub-channel. It is assumed that the matrices H.sub.0 and H.sub.1 are constant over the duration of the sequence (flat fading assumption).

[0088] Similarly, each element in a vector generated at the second output is multiplied in 682 by the coefficient for the channel between the first transmission antenna and the reception antenna at the frequency of the sub-carrier carrying the element in question (operation symbolised by a multiplication of the vector at the output of the FFT by the matrix H.sub.0) and in 684 by the coefficient of the channel between the second transmission antenna and the reception antenna at the frequency of the same sub-carrier (operation symbolised by a multiplication of the vector at the output of the FFT by the matrix H.sub.1).

[0089] The vectors at the output of the multiplier 681 are summed, element by element, with those at the output of the multiplier 684, in the summer 691. The successive vectors of size N at the output of the summer 691 are then supplied to a first spectral de-spreading and filtering module 695.

[0090] Similarly, the vectors at the output of the multiplier 682 are subtracted, element by element, from those at the output of the multiplier 683, in the summer 692. The successive vectors of size N at the output of the summer 692 are then supplied to a second spectral de-spreading and filtering module 696.

[0091] The vectors obtained by the first and second modules 695 and 696 then undergo OQAM demodulation (not shown) in order to obtain the estimated data vectors {circumflex over (X)}.sub.0.sup.m and {circumflex over (X)}.sub.1.sup.m, m=0 . . . L−1.

[0092] FIG. 7 schematically shows the architecture of a FS-FBMC receiver according to one alternative of the first embodiment of the invention.

[0093] Unlike the FS-FBMC receiver of FIG. 6, the receiver here comprises two reception antennas. The signal received on the first antenna during the first use of the channel is demodulated into a base band then sampled at a rate Nf in the sampler 711. The samples are grouped together in the form of blocks of size KN by the serial-parallel converter 721 before undergoing a sliding FFT of size KN in 731. The vectors of the samples at the output of the FFT are then processed on a first path.

[0094] Similarly the signal received on the second antenna during the second use of the channel is demodulated into a base band then sampled at a rate Nf in the sampler 712. The samples are grouped together in the form of blocks of size KN by the serial-parallel converter 722 before undergoing a sliding FFT of size KN in 732. The vectors of the samples at the output of the FFT are then processed on a second path.

[0095] The remaining elements 751-796 are, respectively, identical to elements 651-696 of FIG. 6.

[0096] It will be understood that unlike the FS-FBMC receiver in FIG. 6, no demultiplexing is carried out at the FFT output since both paths are separated from the reception antennas. It is necessary, however, that the receiver be synchronised with the FBMC symbol and, moreover, that the first path be synchronised with the instants of the first use of the channel and that the second path be synchronised with the instants of second use of the channel.

[0097] The structure of the receiver of FIG. 6 or of FIG. 7 can be simplified when the transmitter, instead of using the coding given by (15), uses the block Alamouti coding defined by:

[00018]$\begin{array}{cc}{\stackrel{\_}{C}}^{\prime }=\left(\begin{array}{cc}{\stackrel{\_}{X}}_0& {\stackrel{\_}{X}}_1\\ -\left(j^{L-1}\right).Math.{\stackrel{\_}{X}}_1.Math.T& \left(j^{L-1}\right).Math.{\stackrel{\_}{X}}_0.Math.T\end{array}\right)& \left(23\right)\end{array}$

[0098] In this case the multiplication by the factor (j.sup.L−1) can be removed at the reception and consequently the multiplier 670 or 770 can be omitted.

[0099] FIG. 8A schematically shows a first example of the transmission of two sequences of blocks of symbols by an FBMC transmitter using a first block Alamouti coding, as given by the coding matrix given by the expression (15).

[0100] The blocks of data to be transmitted are here considered upstream of the OQAM modulation.

[0101] A first sequence of blocks 801 is formed by a first guard block 811, a first block of L input data vectors, X.sub.0, 821, a second guard block, 831, followed by a first transformed block, −X.sub.1T, 841, obtained by time-reversal and change of sign of the first input data block.

[0102] A second sequence of blocks 802 is formed by a first guard block 812, a second block of L input data vectors, X.sub.1, 822, a second guard block, 832, followed by a second transformed block, X.sub.0T, 842, obtained by time-reversal and change of sign of the first input data block.

[0103] The guard blocks are made up of null vectors in order to prevent interference between the data blocks and the transformed blocks.

[0104] The first and second sequences are respectively transmitted by the first and second antennas, 892 and 892, after FBMC modulation.

[0105] FIG. 8B schematically shows a second example of the transmission of two sequences of blocks of symbols by an FBMC transmitter using a second block Alamouti coding.

[0106] The second example is identical to the first except that the first guard block is replaced in the first sequence by a first preamble 811′ and in the second by sequence by a second preamble 812′. The other blocks remain unchanged and are therefore not described again.

[0107] The first and second preambles generate interference which affects the first symbols of the blocks X.sub.0 and X.sub.1, interference which does not symmetrically affect the blocks −X.sub.1T and X.sub.0T. This asymmetry does not allow the interference for the input data vectors X.sub.0.sup.m, X.sub.1.sup.m at the beginning of the block, to be eliminated. The preamble symbols are however known to the receiver and it is possible to eliminate this interference once an estimate of the transmission channel is available.

[0108] FIG. 9 schematically shows the architecture of an FS-FBMC receiver, according to a second embodiment of the invention, used to receive sequences of blocks in FIG. 8B.

[0109] Apart from the interference canceller 945, the elements 910 to 996 are identical to the elements 610 to 696 already described in relation to FIG. 6.

[0110] The interference canceller 945 is located on the first output from the demultiplexer 940 and therefore only operates during the first use of the channel. Its purpose is to eliminate the interference generated at the receiver by the first and second preambles 811′ and 812′, on the payload X.sub.0, X.sub.1. More precisely, the interference canceller receives an estimate of the transmission channel, namely the transfer matrices for the elementary channels H.sub.0 and H.sub.1. Since the preambles 811′ and 812′ are known to the receiver, the latter can reconstitute the contribution of the preambles to the signal received during the reception of blocks X.sub.o and X.sub.1, it being understood that only the first K+E vectors received at the beginning of these blocks are affected by this interference, where K is the length of the prototype filter and E the time-spreading of the channel expressing as number of samples. The contribution of the preambles is subtracted from the received signal at the output of the FFT module 830. Once the interference is eliminated, the transmitted blocks X.sub.0 and X.sub.1 may be estimated in accordance with (22-1) and (22-1), it being understood once more that the term j.sup.L−1 may be omitted if the block Alamouti coding defined by (23) is used. Those skilled in the art will moreover understand that this second embodiment may also take the form of the alternative in FIG. 7 by adding an interference canceller at the output of the FFT module 731.