Method and device for inserting k pair of reference signal

Abstract

The invention relates to inserting a first and a second Reference Signals in a radio signal to be transmitted over a wireless communication system, the radio signal being emitted according to a specific SC-SFBC scheme, the method comprising: determining K integers such as $\left\{n_i|i\in 〚1;K〛,0\le n_i\le \frac{M}{2}-1,\forall i,j\in {〚1;K〛}^2,i<j.Math.n_i<n_j\right\};$ and for each pair i of first and second Reference Signals: inserting the first Reference Signal in the radio signal, such as samples of the first Reference Signal are in time periods in the radio signal, said time periods being dependent on a first position in the block of symbols processed by the specific SC-SFBC scheme; inserting the second Reference Signal in the radio signal, such as samples of the second Reference Signal are in time periods in the radio signal, said time periods being dependent on a specific position according to the first position in the same block of symbols.

Claims

1. A method for inserting K pairs of a first and a second Reference Signals in a radio signal to be transmitted over a wireless communication system, said radio signal being intended to be emitted by an emitter comprising at least two transmit antennas, each antenna being configured for transmitting on at least an even number M, strictly greater than 4, of different frequencies, and K being a strictly positive integer strictly smaller than M/2, said radio signal being provided by: applying an M size DFT to a block of symbols X=(X.sub.0, . . . X.sub.M−1), and obtaining for each k.sup.th frequency, with k=0 to M−1, a complex symbol S.sub.k in the frequency domain; obtaining, at an output of an IDFT module corresponding to a first transmit antenna, during a given time interval a first signal representing, in the frequency domain, the complex symbols S.sub.k for each k.sup.th frequency with k=0 to M−1; obtaining, at an output of an IDFT module corresponding to a second transmit antenna, during the given time interval a second signal representing, in the frequency domain, the complex symbols (−1).sup.k+1 εS*.sub.(p-1-k)[M] for each k.sup.th frequency, with k=0 to M−1 and p a predefined even integer lower or equal to M−1 and higher or equal to 0 and ε is 1 or −1 and S.sub.k* being the complex conjugate of S.sub.k; emitting the radio signal corresponding to the first and second signal; said method comprising: determining K integers, where $\left\{n_i\text{|}i\in 〚1;K〛,0\le n_i\le \frac{M}{2}-1,\forall i,j\in {〚1;K〛}^2,i<j.Math.n_i<n_j\right\};$ and for each pair i of first and second Reference Signals: inserting the first Reference Signal in the radio signal, where samples of the first Reference Signal are in time periods in the radio signal, said time periods being dependent on a position n.sub.i of the symbol X.sub.n.sub.i in the block of symbols; inserting the second Reference Signal in the radio signal, where samples of the second Reference Signal are in time periods in the radio signal, said time periods being dependent on the position (n.sub.i+M/2) of the symbol X.sub.(n.sub.i.sub.+M/2) in the block of symbols.

2. The method according to claim 1, wherein for each pair i, inserting the first and second Reference Signals is done by setting values of symbol X.sub.n.sub.i and symbol X.sub.(n.sub.i.sub.+M/2) to respectively values representing the first and second Reference Signals of pair i, before applying the DFT.

3. The method according to claim 1, said method further comprising setting the values of the symbols X.sub.n.sub.i and of the symbols X.sub.(n.sub.i.sub.+M/2) to 0, with i ∈ 1;K, before applying the DFT and obtaining subsequent first and second signals at the output of the respective IDFT modules; and wherein for each pair i, inserting the first and second Reference Signals is done by adding the samples of the first Reference Signal and the samples of the second Reference Signal to said subsequent signals at the output of the respective IDFT modules.

4. The method according to claim 3, said method further comprising, to set the subsequent signals to 0 at least during one of the time periods dependent on the position n.sub.i, for at least one i ∈ 1;K, and/or to 0 at least during one of the time periods dependent on the position n.sub.i+M/2, for at least one i ∈ 1;K before inserting the samples of the first and second Reference Signals of pair i.

5. The method according to claim 1, said method further comprising determining: a number L of pairs of positive integer k.sub.l and k′.sub.l with l ∈ 1;L, with L strictly greater than 1, where:
K.sub.1=1,k′.sub.L=K,
∀l∈1;L−1,k.sub.l<k′.sub.l<k.sub.l+1<k′.sub.l+1, a positive integer d strictly greater than 1; the K integers n.sub.i, with i ∈ 1;K where:
∀l∈1;L−1,n.sub.k.sub.l+1−n.sub.k′.sub.l≥d
∀l∈1;L,n.sub.k′.sub.l−n.sub.k.sub.l=k′.sub.l−k.sub.l.

6. The method according to claim 1, said method further comprising determining: a positive integer d strictly greater than 1; the K integers n.sub.i, with i ∈ 1;K where:
n.sub.i+1−n.sub.i≥d.

7. The method according to claim 1, wherein n.sub.K−n.sub.1=K−1.

8. The method according to claim 7, further comprising: determining a positive integer K.sub.CP where K.sub.CP≤└K/2┘; wherein for each pair i where i ∈ 1;K.sub.CP, a value p.sub.n.sub.i, where the samples in the radio signal that are obtained from the symbol X.sub.n.sub.i whose value is set to p.sub.n.sub.i are equal to the samples of the first reference signal of pair i in the radio signal, is equal to a value ${\rho }_{\left(n_{i+K-K_{\mathrm{CP}}}\right)},$ where the samples in the radio signal that are obtained from the symbol $X_{\left(n_{i+K-K_{\mathrm{CP}}}\right)}$ whose value is set to ${\rho }_{\left(n_{i+K-K_{\mathrm{CP}}}\right)}$ are equal to the samples of the first reference signal of pair i+K−K.sub.CP in the radio signal; and/or a value p.sub.(n.sub.i.sub.+M/2), where the samples in the radio signal that are obtained from the symbol X.sub.(n.sub.i.sub.+M/2) whose value is set to p.sub.(n.sub.i.sub.+M/2) are equal to the samples of the second reference signal of pair i in the radio signal, is equal to a value ${\rho }_{\left(n_{i+K-K_{\mathrm{CP}}}+M/2\right)},$ where the samples in the radio signal that are obtained from the symbol $X_{\left(n_{i+K-K_{\mathrm{CP}}}+M\text{/}2\right)}$ whose value is set to ${\rho }_{\left(n_{i+K-K_{\mathrm{CP}}}+M/2\right)}$ are equal to the samples of the second reference signal of pair i+K−K.sub.CP in the radio signal.

9. The method according to claim 1, wherein a value p.sub.n.sub.i, where the samples in the radio signal that are obtained from the symbol X.sub.n.sub.i whose value is set to p.sub.n.sub.i are equal to the samples of the first reference signal of pair i in the radio signal, is a component of a CAZAC sequence; and/or wherein a value p.sub.(n.sub.i.sub.+M/2), where the samples in the radio signal that are obtained from the symbol X.sub.(n.sub.i.sub.+M/2) whose value is set to p.sub.(n.sub.i.sub.+M/2) are equal to the samples of the second reference signal of pair i in the radio signal, is a component of a CAZAC sequence.

10. The method according to claim 1, wherein a value p.sub.n.sub.i, where the samples in the radio signal that are obtained from the symbol X.sub.n.sub.i whose value is set to p.sub.n.sub.i are equal to the samples of the first reference signal of pair i in the radio signal, is equal to a value p.sub.(n.sub.i.sub.+M/2), where the samples in the radio signal that are obtained from the symbol X.sub.(n.sub.i.sub.+M/2) whose value is set to p.sub.(n.sub.i.sub.+M/2) are equal to the samples of the second reference signal of pair i in the radio signal.

11. The method according to claim 1, wherein a maximum module among the modules of values p.sub.n.sub.i and p.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K, each p.sub.n.sub.i, respectively p.sub.(n.sub.i.sub.+M/2), where for the samples in the radio signal that are obtained from the symbol X.sub.n.sub.i, respectively X.sub.(n.sub.i.sub.+M/2), whose value is set to p.sub.n.sub.i, respectively p.sub.(n.sub.i.sub.+M/2), are equal to the samples of the first reference signal of pair i, respectively the second reference signal of pair i, in the radio signal, is equal to or smaller than a maximum module among modules of modulation symbols of a digital modulation scheme used to obtain said block of symbols.

12. A non-transitory computer readable medium having stored thereon a computer program product comprising code instructions to perform the method according to claim 1, when said instructions are run by a processor.

13. A device for inserting K pairs of a first and a second Reference Signals in a radio signal to be transmitted over a wireless communication system, said radio signal being intended to be emitted by an emitter comprising at least two transmit antennas, each antenna being configured for transmitting on at least an even number M, strictly greater than 4, of different frequencies, and K being a strictly positive strictly smaller than M/2, said radio signal being processed by: applying a M size DFT to a block of symbols X=(X.sub.0, . . . X.sub.M−1), and obtaining for each k.sup.th frequency, with k=0 to M−1, a complex symbol S.sub.k in the frequency domain; obtaining, at an output of an IDFT module corresponding to a first transmit antenna, during a given time interval a first signal representing, in the frequency domain, the complex symbols S.sub.k for each k.sup.th frequency with k=0 to M−1; obtaining, at an output of an IDFT module corresponding to a second transmit antenna, during the given time interval a signal representing, in the frequency domain, the complex symbol (−1).sup.k+1 εS*(.sub.p-1-k).sub.[M] for each K.sup.th frequency, with k=0 to M−1 and p a predefined even integer lower or equal to M−1 and higher or equal to 0 and ε is 1 or −1 and S.sub.k* being the complex conjugate of S.sub.k; emitting the radio signal corresponding to the first and second signal; said device being configured to: determine K integers where $\left\{n_i\text{|}i\in 〚1;K〛,0\le n_i\le \frac{M}{2}-1,\forall i,j\in {〚1;K〛}^2,i<j.Math.n_i<n_j\right\};$ and for each pair i of first and second Reference Signals: insert the first Reference Signal in the radio signal, where samples of the first Reference Signal are in time periods in the radio signal, said time periods being dependent on a position n.sub.i of the symbol X.sub.n.sub.i in the block of symbols; insert the second Reference Signal in the radio signal, where samples of the second Reference Signal are in time periods in the radio signal, said time periods being dependent on the position n.sub.i+M/2 of the symbol X.sub.(n.sub.i.sub.+M/2) in the block of symbols.

14. A method for extracting K pairs of a first and a second Reference Signals in a radio signal receivable over a wireless communication system, said radio signal being emitted by an emitter comprising at least two transmit antennas, each antenna being configured for transmitting on at least an even number M, strictly greater than 4, of different frequencies, and K being a strictly positive integer strictly smaller than M/2, the emission of the radio signal being processed by: applying a M size DFT to a block of symbols X=(X.sub.0, . . . X.sub.M−1), and obtaining for each k.sup.th frequency, with k=0 to M−1, a complex symbol S.sub.k in the frequency domain; obtaining, at an output of an IDFT module corresponding to a first transmit antenna, during a given time interval a first signal representing, in the frequency domain, the complex symbols S.sub.k for each k.sup.th frequency with k=0 to M−1; obtaining, at an output of an IDFT module corresponding to a second transmit antenna, during the given time interval a signal representing, in the frequency domain, the complex symbol (−1).sup.k+1 εS*.sub.(p-1-k)[M] for each k.sup.th frequency with k=0 to M−1 and p a predefined even integer lower or equal to M−1 and higher or equal to 0 and ε is 1 or −1 and S.sub.k* being the complex conjugate of S.sub.k; emitting the radio signal corresponding to the first and second signal; said method comprising: extracting, before applying DFT modules on the received radio signal, parts of the radio signal, each part being received in one time window among time windows, said time windows being strictly included in the given time interval; processing said extracted parts independently from other parts of the received radio signal; wherein said time windows being defined according to time periods dependent on a position n.sub.i of the symbol X.sub.n.sub.i in the block of symbols.

15. The method according to claim 14, wherein each time window among the time windows strictly includes at least one time period among the time periods.

16. The method according to claim 14, wherein each time window is strictly included in at least one time period among the time periods.

17. The method according to claim 14, wherein each time window is equal to one time period among the time periods.

18. A device for extracting K pairs of a first and a second Reference Signals in a radio signal receivable over a wireless communication system, said radio signal being emitted by an emitter comprising at least two transmit antennas, each antenna being configured for transmitting on at least an even number M, strictly greater than 4, of different frequencies, and K being a strictly positive integer strictly smaller than M/2, the emission of the radio signal being processed by: applying a M size DFT to a block of symbols X=(X.sub.0, . . . X.sub.M−1), and obtaining for each k.sup.th frequency, with k=0 to M−1, a complex symbol S.sub.k in the frequency domain; obtaining, at an output of an IDFT module corresponding to a first transmit antenna, during a given time interval a signal representing, in the frequency domain, the complex symbols S.sub.k for each k.sup.th frequency with k=0 to M−1; obtaining, at an output of an IDFT module corresponding to a second transmit antenna, during the given time interval a signal representing, in the frequency domain, the complex symbol (−1).sup.k+1 εS*.sub.(p-1-k) for each k.sup.th frequency with k=0 to M−1 and p a predefined even integer lower or equal to M−1 and higher or equal to 0 and ε is 1 or −1 and S.sub.k* being the complex conjugate of S.sub.k; emitting the radio signal corresponding to the first and second signal; said device being configured to: extract, before applying DFT modules on the received radio signal, parts of the radio signal, each part being received in one time window among time windows, said time windows being strictly included in the given time interval; process said extracted parts independently from other parts of the received radio signal; wherein said time windows being defined according to time periods dependent on a position n.sub.i of the symbol X in the block of symbols.

Description

BRIEF DESCRIPTION OF DRAWINGS

(1) FIG. 1 illustrates a SC-SFBC type transmitter and receiver.

(2) FIG. 2 schematizes a block diagram of a classical SFBC transmitter.

(3) FIG. 3.1 schematizes a block diagram of a PAPR-preserving SFBC transmitter.

(4) FIG. 3.2 details the PAPR-preserving SFBC logical functioning.

(5) FIG. 3.3 schematizes an equivalent block diagram of a PAPR-preserving SFBC transmitter.

(6) FIG. 3.4 schematizes a block diagram of a PAPR-preserving SFBC receiver.

(7) FIG. 4.1 schematizes a block diagram of Pre-DFT insertion of RS according to the invention.

(8) FIG. 4.2 schematizes a block diagram of Post-IDFT insertion of RS according to the invention.

(9) FIG. 4.3 schematizes a block diagram of RS extraction and DATA decoding according to the invention.

(10) FIG. 5.1 illustrates a flowchart representing the steps of pre-DFT inserting reference signals in the radio signal according to the invention.

(11) FIG. 5.2 illustrates a flowchart representing the steps of post-IDFT inserting reference signals in the radio signal according to the invention.

(12) FIG. 5.3 illustrates a flowchart representing the steps of extracting reference signals in the radio signal according to the invention.

DESCRIPTION OF EMBODIMENTS

(13) Referring to FIG. 1, there is shown a transmitter 1.1 transmitting a radio signal to a receiver 1.2. The transmitter 1.1 is in the cell of the receiver 1.2. This transmission may be a SC-SFBC based transmission in the context of OFDM based system. In this example the transmitter 1.1 is a mobile terminal and the receiver 1.2 is a fixed station which in the context of LTE is a base station. The transmitter 1.1 can as well be the fixed station and the receiver 1.2 a mobile terminal.

(14) The transmitter 1.1 comprises one communication module (COM_trans) 1.3, one processing module (PROC_trans) 1.4 and a memory unit (MEMO_trans) 1.5. The MEMO_trans 1.5 comprises a non-volatile unit which retrieves the computer program and a volatile unit which retrieves the reference signal parameters. The PROC_trans 1.4 which is configured to insert the reference signals according to the invention. The COM_trans is configured to transmit to the receiver 1.2 the radio signal. The processing module 1.4 and the memory unit 1.5 may constitute the device for inserting the reference signals, as previously described. The processing module 1.4 and the memory unit 1.5 can be dedicated to this device or also used for other functions of the transmitter like for processing the radio signal.

(15) The receiver 1.2 comprises one communication module (COM_recei) 1.6, one processing module (PROC_recei) 1.7 and a memory unit (MEMO_recei) 1.8. The MEMO_recei 1.8 comprises a non-volatile unit which retrieves the computer program and a volatile unit which retrieves the reference signal parameters. The PROC_recei 1.7 is configured to extract the reference signals from the radio signal. The COM_recei 1.6 is configured to receive from the transmitter the radio signal. The processing module 1.7 and the memory unit 1.8 may constitute the device for extracting the reference signals, as previously described. The processing module 1.7 and the memory unit 1.8 can be dedicated to this device or also used for other functions of the receiver like for processing the receiving scheme on the radio signal.

(16) Referring to FIG. 2, there is shown a block diagram of a classical SFBC transmitter. Such SFBC transmitters apply Alamouti precoding on the bases of a DFTsOFDM system. This ensures full diversity for a rate of one symbol per channel use. An SFBC transmitter emits a radio signal by emitting on at least two transmit antennas Tx1 2.1 and Tx2 2.2.

(17) The radio signal being provided by applying a M size DFT 2.3 to a block of symbols X=(X.sub.0, . . . X.sub.M−1), for example a block of symbols obtained by a QPSK digital modulation scheme or any other digital modulation scheme as QAM. M is the number of allocated subcarriers. In such SFBC scheme, M is even.

(18) Therefore, in the frequency domain, the DFT 2.3 outputs M complex symbols {S.sub.k}, (k=0 to M−1) one complex symbol for each k-th subcarrier among the M allocated subcarriers. The vector S=(S.sub.0, . . . S.sub.M−1) represents the M-point DFT of the block of modulation symbols X.

(19) The “Alamouti pre-coding” 2.4 is applied to adjacent subcarriers in DFTsOFDM. That is for each M/2 pairs (S.sub.k; S.sub.k+1) formed by the M outputs of the DFT of the same data block with k even (0 included) the Alamouti pre-coding is applied with the pre-coding matrix:

(20) $\hspace{1em}\left[\begin{array}{cc}{S}_k& -S_{k+1}^{*}\\ S_{k+1}& S_k^{*}\end{array}\right]$ where symbols on 1.sup.st and respectively 2.sup.nd column represent the symbols transmitted from antennas Tx1 and respectively Tx2, and symbols on 1.sup.st and respectively 2.sup.nd row represent symbols to be transmitted onto the k-th and respectively (k+1)-th allocated subcarriers.

(21) Therefore the outputs of the Alamouti pre-coding 2.4 are the vectors S.sup.Tx1 and S.sup.Tx2, with:
S.sup.Tx1=(S.sup.Tx1.sub.k)=(S.sub.k)
S.sup.Tx2=(S.sup.Tx2.sub.k)=((−1).sup.k+1S*.sub.k+(−1).sub.k)

(22) Each vector is mapped to the M allocated subcarriers out of N existing subcarriers via subcarrier mapping modules 2.5 and 2.6. The subcarrier mapping can be for example localized, that is the M elements of each vector S.sup.Tx1,2 are mapped to M consecutive subcarriers among the N existing. The subcarrier mapping can be for example distributed, that is the M elements of each vector S.sup.Tx1,2 are mapped equally distanced over the entire bandwidth with zero occupying the unused subcarriers. Inverse DFT of size N 2.7 and 2.8 are then applied to the two resulting vectors {tilde over (S)}.sup.Tx1 and {tilde over (S)}.sup.Tx2, therefore generating two DFTsOFDM symbols, to be sent in the same time from the two transmit antennas. A cyclic prefix can be optionally appended after IDFT.

(23) The output of each IDFT is emitted on each antenna, the output of the IDFT 2.7, applied on {tilde over (S)}.sup.Tx1, is emitted on Tx1 2.1 and the output of the IDFT 2.8, applied on {tilde over (S)}.sup.Tx2, is emitted on Tx2 2.2. Such a SFBC scheme breaks the PAPR property of the signal transmitted on the second transmit antenna. To preserve the PAPR property evolved SFBC schemes have been developed.

(24) Referring to FIG. 3.1, there is shown a block diagram of a PAPR-preserving SFBC transmitter also called SC-SFBC. Such transmitters apply also an Alamouti pre-coding at subcarrier level, after an M sized DFT. In contrary to a classical SFBC scheme of FIG. 2, a PAPR-preserving SFBC 3.1.4 applies Alamouti pre-coding to non-adjacent subcarriers, more precisely Alamouti pre-coding is applied to each pairs of symbols (S.sub.k; S.sub.(p-1-k)[M]) of the M outputs of the DFT 3.1.3 of the same data block, and mapped onto subcarriers k-th and (p-1-k)[M]-th allocated subcarriers, where p is an even integer. Usually p is chosen close to M/2 (that is minimizing |M/2−p|) to minimize the maximum distance between pairs of subcarriers k and (p-1-k)[M]. This PAPR-preserving SFBC logical functioning will be explained more precisely in FIG. 3.2.

(25) Besides this specific applying of the “Alamouti pre-coding” 3.1.4, the PAPR-preserving SFBC scheme is identical to a SFBC scheme. Therefore, the outputs of the specific Alamouti pre-coding 3.1.4, that is the PAPR preserving SFBC module 3.1.4, are the vectors S.sup.Tx1 and S.sup.Tx2 on which we apply subcarrier mapping modules 3.1.5 and 3.1.6. These vectors are mapped to the M allocated subcarriers out of N existing subcarriers. The vectors {tilde over (S)}.sup.Tx1 and {tilde over (S)}.sup.Tx2 respectively resulting from the subcarrier mapping 3.1.5 associated to antenna Tx1 3.1.1 and the subcarrier mapping 3.1.6 associated to antenna Tx2 3.1.2, are inputted in the respective N size IDFT associated with each antenna 3.1.7 and 3.1.8. The same subcarrier mapping is usually applied to signals intended for both transmit antennas.

(26) Referring to FIG. 3.2 there is shown in detail the logical functioning of a PAPR-preserving SFBC module 3.1.4.

(27) The vector S=(S.sub.k), which represents the (output of the) M-point DFT of the block of modulation symbols X, is inputted in the PAPR preserving SFBC module 3.1.4. For each pair (S.sub.k; S.sub.(p-1-k)[M]) of the allocated subcarriers k and (p-1-k)[M], the PAPR preserving SFBC applies the matrices:

(28) $A^{\left(I\right)}=\left[\begin{array}{cc}{S}_k& -S_{\left(p-1-k\right)\left[M\right]}^{*}\\ S_{\left(p-1-k\right)\left[M\right]}& S_k^{*}\end{array}\right]$
when k is even and

(29) 0$A^{\left(\mathrm{II}\right)}=\left[\begin{array}{cc}{S}_k& S_{\left(p-1-k\right)\left[M\right]}^{*}\\ S_{\left(p-1-k\right)\left[M\right]}& -S_k^{*}\end{array}\right]$
when k is odd.

(30) Note that changing the sign (+/−) of the signal related to the second antenna does not change the method.

(31) For both matrices symbols on 1.sup.st and respectively 2.sup.nd column represent the symbols transmitted from antennas Tx1 and respectively Tx2, and symbols on 1.sup.st and respectively 2.sup.nd row represent symbols to be transmitted onto the k-th and respectively (p-1-k)[M]-th allocated subcarriers. Therefore the outputs of the PAPR preserving SFBC 3.1.4 is the vector S.sup.Tx1 related to antenna Tx1 3.1.7 and S.sup.Tx2 related to antenna Tx2 3.1.8, with:
S.sup.Tx1=(S.sup.Tx1.sub.k)=(S.sub.k)
S.sup.Tx2=(S.sup.Tx2.sub.k)=ε((−1).sup.k+1S*.sub.(p-1-k)[M])
With ε the value 1 or −1. When not stated otherwise, in the following we consider ε=1. Indeed changing the sign (+/−) of the signal related to the second antenna does not change the method.

(32) Such PAPR preserving SFBC transmitter is equivalent to sending to two independent transmit antennas, in the same time interval representing the duration of a DFTsOFDM symbol, a DFTsOFDM symbol corresponding to the block of symbols eqX.sup.Tx1=(eqX.sup.Tx1.sub.k)=(X.sub.k) on the first transmit antenna and a DFTsOFDM symbol corresponding to the block of symbols eqX.sup.Tx2=(eqX.sup.Tx2.sub.k)=(e.sup.j2π(p−1)k/MX*.sub.(k+M/2)[M]) on the second transmit antenna, as shown in FIG. 3.3. This equivalence is shown in the literature (Cristina Ciochina et al.: “Single-Carrier Space-Frequency Block Coding: Performance Evaluation”, Vehicular Technology Conference, 2007. VTC-2007 Fall. 2007 IEEE 66th, IEEE, P1, Sep. 1, 2007, pp. 715-719). Since the PAPR of eqX.sup.Tx1.sub.k is equal to the PAPR of eqX.sup.Tx2.sub.k the signal sent on Tx1 and on Tx2, obtained through DFTsOFDM modulation applied respectively to the symbol blocks eqX.sup.Tx1 and eqX.sup.Tx2, have the same PAPR. Thus both have SC-type envelope fluctuations, leading to low PAPR. Therefore, a PAPR-preserving SFBC system preserves the single carrier property or the PAPR property.

(33) Referring to FIG. 3.4, there is shown a block diagram of a PAPR-preserving SFBC receiver. Such a receiver is configured to decode a radio signal emitted by a PAPR-preserving SFBC. This example shows 2 receive antennas but such receiver can have only one antenna (MISO) or more antennas (MIMO). In this example said radio signal is received on two antennas Rx1 3.4.1 and Rx2 3.4.2. The radio signal received by each antennas differs, and the more the two antennas are spaced from each other the more the radio signal received on each antennas is likely to be different, which introduces receive diversity. After an optional guard removal the resulting {tilde over (y)}.sup.Rx1 and {tilde over (y)}.sup.Rx2 are inputted into two N size DFT (3.4.3 and 3.4.4) and then in subcarriers de-mapping modules (3.4.5 and 3.4.6), one associated to Rx1 3.4.1 one associated with Rx2 3.4.2. The result in the frequency domain is two vectors T.sup.Rx1 and T.sup.Rx2. Before inputting T.sup.Rx1 and T.sup.Rx2 in the PAPR preserving SFBC de-combiner 3.4.8, the PAPR preserving SFBC de-combiner 3.4.8 can be adjusted based on the channel estimation, channel estimation which is computed for example on the basis of received DMRS (demodulation reference signal). Afterwards, the T.sup.Rx1 and T.sup.Rx2 are inputted in the PAPR preserving SFBC de-combiner 3.4.8 which outputs one M-size vector T. A M size DFT is applied to T to obtain a block of symbols Y=(Y.sub.k) in the time domain. The vector Y once completely decoded, that is for example after constellation de-mapping and error correction, enables estimating the digital data at the origin of X.

(34) If reference signals are pre-DFT inserted in random position and multiplexed with the data modulation symbols at the DFT input, the samples of the reference signals received cannot be extracted from the received signal and processed before obtaining at first vector Y at the output of the IDFT module 3.4.9. Therefore, the Alamouti de-combining 3.4.8 will de-combine T.sup.Rx1 and T.sup.Rx2 without taking into account the information conveyed by the reference signals, which can lead to strongly degraded performance of the Alamouti decombining module 3.4.8.

(35) Referring to FIG. 4.1, there is shown a block diagram of Pre-DFT insertion of reference signals according to the invention. Regarding the transmitter, the scheme applied is the same than the PAPR-preserving SFBC scheme described in FIGS. 3.1 and 3.2. Therefore, a M size DFT 4.1.3, a PAPR preserving SFBC pre-coder 4.1.4, subcarrier mapping modules 4.1.5 and 4.1.6 and N size IDFT modules are successively applied to the block of symbols X=(X.sub.0, . . . X.sub.M−1) to obtain the radio signal emitted by Tx1 4.1.1 and Tx2 4.1.2.

(36) In this embodiment, reference signal are inserted pre-DFT, that is by setting values of the symbols X.sub.n which are chosen to be reference signals. Therefore, when inserting 2K reference signals, K being a positive integer smaller or equal to M/2 (it may be set strictly smaller than M/2 to avoid the block of symbols X to be a full block of reference signals), K integers n.sub.i are to be determined such as

(37) $\left\{n_i\text{|}i\in 〚1;K〛,0\le n_i\le \frac{M}{2}-1,\forall i,j\in {〚1;K〛}^2,i<j.Math.n_i<n_j\right\}.$

(38) Then reference signals are inserted directly in the block of symbol X at the positions n.sub.i and n.sub.i+M/2. The RS insertion module 4.1.9 inserts the reference signals by setting each value of the symbols X.sub.n which are at positions n.sub.i or n.sub.i+M/2 with i ∈ 1;K at a value of a reference signal. The RS insertion module 4.1.9 may be configured in a static way by previously configuring the positions n.sub.i or n.sub.i+M/2 with i ∈ 1;K. Several configurations may also be previously programmed, for example one configuration for each number K. Exact values of K and ni can be either fixed, or configurable. Configuration can be done in an implicit manner (based on other parameters known by the transmitter), or in an explicit manner (based on instructions that the receiver is returning to the transmitter via, e.g., a control channel). The Data modulator module 4.1.10 may be configured to insert modulation symbols into the block of symbols in positions that do not conflict with the positions n.sub.i or n.sub.i+M/2 with i ∈ 1;K of the reference signals. The RS insertion module 4.1.9 may inform the Data modulator module 4.1.10 of the chosen configuration.

(39) The positions

(40) $\left\{n_i\text{|}i\in 〚1;K〛,0\le n_i\le \frac{M}{2}-1,\forall i,j\in {〚1;K〛}^2,i<j.Math.n_i<n_j\right\}$
can be advantageously chosen. For example, the RS insertion module 4.1.9 may be configured with the positions n.sub.i consecutive, that is with n.sub.K−n.sub.1=K−1. Grouping the RS on consecutive positions enables to reduce the interference of the RS with other symbols in the radio signal.

(41) In another example, the positions {n.sub.i|i ∈ 1;K} of the RS can be divided in groups of consecutive positions. That is for example, 3 groups of RS of consecutive positions n.sub.1 to n.sub.k′.sub.1, n.sub.K.sub.2 (n.sub.k.sub.2≠n.sub.k′.sub.1) to n.sub.k′.sub.2 and n.sub.k.sub.3 (n.sub.k.sub.3≠n.sub.k′.sub.2) to n.sub.k, where n.sub.k′.sub.1−n.sub.1=k′.sub.1−1, n.sub.k′.sub.2−n.sub.k.sub.2=k′.sub.2−k.sub.2 and n.sub.k−n.sub.k.sub.3=K−k.sub.3. Having several groups which are separated by other symbols in the block of symbols X, enables to track fast phase variations at a lower level than the time duration of a block of symbol.

(42) For the same reasons it may be advantageous to set the values of the positions on one extreme (or both extremes) of each group as protection. For example, by defining respectively the values of the symbols

(43) $X_{n_{k_1^{\prime }}},X_{n_{k{\hspace{0.17em}}_2^{\prime }}},X_{n_{K\hspace{0.17em}}}$
to the values of the symbols

(44) $X_{n_1},X_{n_{k_2}},X_{n_{k_3}}.$
Or by setting the values of the symbols

(45) $X_{n_{k_1^{\prime }}},X_{n_{k{\hspace{0.17em}}_2^{\prime }}},X_{n_{K\hspace{0.17em}}}$
to 0. The protection can be enhanced by defining more than the last symbol of the group as a protection symbol.

(46) Regarding the values of the symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K, that is the symbols of X positioned at n.sub.i and n.sub.i+M/2 for i ∈ 1;K, they may be set as components of a CAZAC sequence. More specifically the values of the symbols X.sub.n.sub.i with i ∈ 1;K can be derived from values of a first CAZAC sequence and/or the values of the symbols X.sub.(n.sub.i.sub.+M/s) with i ∈ 1;K can be derived from values of a second CAZAC sequence. The CAZAC sequences may be for example Zadoff-Chu sequences. That is for example with the values of X.sub.n.sub.i with i ∈ 1;K, can be set to the values of a CAZAC sequence of length K, or can be obtained by truncating a CAZAC sequence of length superior to K, or can be obtained by cyclic extension from a CAZAC sequence of length inferior to K.

(47) It is advantageous to set equal values for X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K. Indeed, at the output of the DFT module 4.2.3 the common value of X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) only contributes to values S.sub.2k of vector S. Therefore, on the first transmit antenna, only one out of two occupied subcarriers (0-th, 2.sup.nd, etc) carry information (S.sup.Tx1.sub.2k) relative to the reference signals, while on the second transmit antenna, only the other one out of two occupied subcarriers (1st, 3rd, etc) carry information (S.sup.Tx2.sub.2k+1) relative to the reference signals. Therefore, the samples of reference signals received from each antenna are orthogonal to each other and thus enabling to distinguish samples of reference signals emitted by the first antenna and the ones emitted by the second antenna.

(48) It is advantageous to set values for X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K, such as the maximum absolute values modules are equal or smaller than the maximum absolute values of the modulation symbols of the digital modulation scheme used for modulation. For example, the digital modulation scheme can be QPSK (quadrature phase-shift keying) or other PSK (phase-shift keying) whose values are all of module equal to 1, in this example the values of |X.sub.n.sub.i| and |X.sub.(n.sub.i.sub.+M/2)| are chosen as to be smaller or equal to 1.

(49) Referring to FIG. 4.2, there is shown a block diagram of Post-IDFT insertion of reference signals according to the invention. In this embodiment, the reference signals are not inserted pre-DFT (that is by setting the values of the symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K, to non-null values known by the receiver as shown in FIG. 4.1). The insertion of the reference signals is done post IDFT. For this, the data modulator module 4.2.1 is configured to set the values of the symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K to 0. The configuration of the data modulator module 4.2.1 may be made by the RS insertion module 4.2.2 which can send the position configuration to the data modulator module 4.2.1. On this incomplete block of symbols X.sub.DATA, the same scheme of the previous embodiment of FIG. 4.1 is applied, starting by a M size DFT 4.2.3. At the respective IDFT outputs subsequent signals are obtained, that is {tilde over (x)}.sub.DATA.sup.Tx1 at the output of the IDFT related to the antenna Tx1 and {tilde over (x)}.sub.DATA.sup.Tx2 at the output of the IDFT related to the antenna Tx2. The RS insertion module 4.2.2 adds respectively to each of the output signals of the IDFT modules (4.2.4 and 4.2.5), which are {tilde over (x)}.sub.DATA.sup.Tx1 corresponding to the antenna Tx1 and {tilde over (x)}.sub.DATA.sup.Tx2 corresponding to the antenna Tx2, the signals {tilde over (x)}.sub.RS.sup.Tx1 and the signal {tilde over (x)}.sub.RS.sup.Tx2 respectively. The signal {tilde over (x)}.sub.RS.sup.Tx1 and {tilde over (x)}.sub.RS.sup.Tx2 are pre-computed samples of the corresponding symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K. That is rather than setting the pre-DFT values of the symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K in the block of symbol X, the samples of the reference signals are previously computed to obtain samples identical or at least equivalent to those that would have been obtain, at the output of the IDFT, by setting the values of the symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K to non-null values known by the receiver, as previously described. For example, the samples are obtained by applying the same scheme (SC-SFBC) to a block symbol where the values of the symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K are set respectively to the values representing the reference signals and setting the values of the other symbols to 0 (that is by not introducing other symbols). Therefore, when applying the same scheme to the incomplete block of symbols X.sub.RS, at the output of the IDFT modules (4.2.4 and 4.2.5) we obtain {tilde over (x)}.sub.RS.sup.Tx1 and {tilde over (x)}.sub.RS.sup.Tx2 that are the samples in the signals outputted by the IDFT of the corresponding symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K, pre-computed.

(50) In the embodiments of the FIG. 4.2 where the reference signals are inserted post-IDFT the signals obtained at the output of the adders, {tilde over (x)}.sup.Tx1 and {tilde over (x)}.sup.Tx2, are equivalent to the signals at the outputs of the IDFT modules when pre-DFT inserting the reference signals. Therefore, all the features shown in relation with pre-DFT insertion can be applied to post-IDFT insertion.

(51) For example, the positions advantageously chosen in pre-DFT insertion can be applied, by setting to 0 the symbols at those positions and then post-IDFT inserting in those positions the samples of the RS previously computed corresponding to those symbols.

(52) In the embodiments that set the values of the symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K, these embodiments can be applied in the case of post-IDFT insertion, by adding at the IDFT outputs the samples of the reference signals that have previously been computed with the incomplete block of symbols X.sub.RS in which the values of the symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K are the same as the corresponding values of the corresponding symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K set pre-DFT and all other symbols are set to 0.

(53) Before adding the signal {tilde over (x)}.sub.RS.sup.Tx1 and the signal {tilde over (x)}.sub.RS.sup.Tx2, it is advantageous to filter the signals {tilde over (x)}.sub.DATA.sup.Tx1 and {tilde over (x)}.sub.DATA.sup.Tx2 to ensure that the samples in the signal {tilde over (x)}.sub.DATA.sup.Tx1 and the signal {tilde over (x)}.sub.DATA.sup.Tx2 of the corresponding symbols X.sub.n.sub.i and symbols X.sub.(n.sub.i.sub.+M/2)[M], whose values have been set to 0, are also strictly equal to 0 for the time periods in which are inserted the samples of the reference signal or at least time periods in which are inserted the parts of the samples of high energy. Therefore, this enables to reduce the interference of signals {tilde over (x)}.sub.DATA.sup.Tx1 and {tilde over (x)}.sub.DATA.sup.Tx2 onto at least the high energy part of signals {tilde over (x)}.sub.RS.sup.Tx1 and {tilde over (x)}.sub.RS.sup.Tx2.

(54) Referring to FIG. 4.3, there is shown a block diagram of reference signals extraction according to the invention. The radio signal emitted according to the previous embodiments is received by the receiver 1.2, after having crossed a multipath channel and suffered noise and phase noise effects. The receiver can receive the radio signal on one antenna, which is the case in MISO telecommunication system, or on several antennas Rx1, Rx2 . . . RxQ as shown in FIG. 4.3.

(55) After the Analogue to Digital converters (ADC) 4.3.1 have been applied to the radio signal received by each antenna, the reference signals are extracted. Here, by reference signal extraction we understand separating time domain portions of the received radio signal containing part or all of the information relative to the sent reference signals, corrupted by the channel and the noise/phase noise during the transmission process.

(56) This is possible since in the time domain samples in the radio signal of the reference signals are superposed, for example when the samples in the radio signal of the reference signal corresponding to the symbols X.sub.n.sub.i are emitted by Tx1 4.1.1 there are samples in the radio signal of the reference signal corresponding to the symbols X.sub.(n.sub.i.sub.+M/2)[M] that are emitted by Tx2 4.1.2. This is shown by the fact that the PAPR preserving SFBC transmitter is equivalent to independently sending onto the two antennas signals obtained by applying DFTsOFDM scheme on each of the block of symbols:
eqX.sup.Tx1=(eqX.sup.Tx1.sub.k)=(X.sub.k)
eqX.sup.Tx2=(eqX.sup.Tx2.sub.k)=(e.sup.j2π(p−1)k/MX*.sub.(k+M/2)[M])

(57) The symbol eqX.sup.Tx1.sub.n.sub.i=(X.sub.n.sub.i) is processed at the same time than eqX.sup.Tx2.sub.n.sub.i=e.sup.j2π(p−1)n.sup.i.sup./MX*.sub.(n.sub.i.sub.+M/2)[M], the first symbol giving part of the samples in the radio signal of the reference signal corresponding to the symbols X.sub.n.sub.i and the second one giving part of the samples in the radio signal of the reference signal corresponding to the symbols X.sub.(n.sub.i.sub.+M/2)[M]. Symmetrically, the symbol eqX.sup.Tx1.sub.n.sub.i.sub.+M/2 and eqX.sup.Tx2.sub.n.sub.i+M/2 are processed at the same time, these symbols giving respectively the other part of the samples in the radio signal of the reference signals corresponding to the symbols X.sub.(n.sub.i.sub.+M/2)[M] and to the symbols X.sub.n.sub.i. Therefore, samples of the reference signals corresponding to the symbol X.sub.n.sub.i and samples of the reference signals corresponding to the symbol X.sub.(n.sub.i.sub.+M/2)[M] are emitted at the same time, that is at time periods being dependent to the position n.sub.i of the symbol X.sub.n.sub.i and no parts of high energy of samples corresponding to symbol which are at other position than position n.sub.i are emitted at this same time, except for the samples corresponding to the symbol X.sub.(n.sub.i.sub.+M/2).

(58) Therefore, by extracting the parts of the signals outputted by the ADCs 4.3.1 during the time periods dependent to the position n.sub.i, both the samples corresponding to the symbols X.sub.n.sub.i and the samples corresponding to the symbols X.sub.(n.sub.i.sub.+M/2)[M] are extracted. It is advantageous to take into account only the time periods corresponding to parts of the samples with high energy to avoid extracting unusable samples. These time periods depend on the type of subcarrier mapping that is implemented. For example a localized or a distributed implementation give completely different time periods.

(59) Each type of subcarrier mapping has its own distribution, across the time domain, of the samples in the radio signal. These distribution are well known by the person skilled in the art and shown in the literature (for example: Cf. reference Hyung G. Myung Single Carrier Orthogonal Multiple Access Technique for Broadband WirelessCommunications Ph.D. Thesis Defense|2006 Dec. 18) which can easily transpose the teaching of the invention to other subcarrier mapping, thus the invention is not limited to a specific subcarrier mapping.

(60) The extractor 4.3.2 may be configured to apply time-domain windows for extracting the reference signals according to the time periods of the received samples of the reference signals inserted (pre-DFT or post-DFT). A first configuration is to apply time domain windows, each window being equal to one time period among the time periods of the received samples of the reference signals. The sizes of the windows may as well exceed the size of the windows of the first configuration each window being positioned to include a window of the first configuration. This enables to extract a slightly wider part of the received samples corresponding to the reference signals which is advantageous when the receiver 1.2 is capable of interference mitigation. The size of the windows may be taken smaller than the size of the windows of the first configuration and each window may be positioned to be included in a window of the first configuration, enabling to limit the extraction of the received samples of non-reference signals which may cause interference with the samples of the reference signal, which is advantageous when the receiver 1.2 is of low performance regarding interference mitigation.

(61) Once the received samples of the reference signals are extracted by the extractor 4.3.2 they may be time domain or frequency domain processed. The processing of the samples of the reference signals is a common processing well known by the person skilled in the art. For example, reference signals can serve as base for channel estimation through known channel estimation methods applied in the time or in the frequency domain. For example, reference signals can serve to improve the quality of a channel estimate acquired based on other dedicated reference signals (e.g. dedicated DFTsOFDM symbols carrying only reference symbols such as DMRS). The processing in the frequency domain is especially relevant in the case where the corresponding values of the samples corresponding to the symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2)[M] are equal, as previously mentioned. Once the received samples of the reference signal are processed, the channel estimation module 4.3.3 can compare these reference signals with reference values, as part of the channel estimation process.

(62) The channel estimation may also result from a classical implementation with dedicated DMRS (demodulation reference signal) which occupy a full block of symbols, in this case the reference signal according to the invention can be used to improve the channel estimation quality.

(63) Once the channel estimation module 4.3.3 calculated the estimated channel, the PAPR preserving SFBC de-combiner 4.3.4 may be set to compensate the corruption (phase shift, amplitude . . . ) of the signal in the channel between transmitter and receiver. Enabling to enhance the performance of the processing to obtain the block of symbols Y=(Y.sub.0, . . . Y.sub.M−1).

(64) The extractor 4.3.2 can also be placed after the guard removal modules.

(65) Referring to FIG. 5.1 there is shown a flowchart representing the steps of pre-DFT inserting reference signals in a radio signal according to the invention.

(66) At step S11 the RS insertion module 4.1.9 is configured either in a static way or dynamically, that is that the RS insertion module 4.1.9 is reconfigured depending for example on a feedback from the receiver through a control channel. In the case of a dynamic configuration the RS insertion module 4.1.9 may choose another configuration upon those saved in the MEMO_trans 1.5. Indeed, several configurations may be pre-parametred in the RS insertion module 4.1.9, those configurations can be ordered according to the number of reference signals the configuration provides. A configuration may be defined by the number of reference signals K, by the positions n.sub.i in the block of symbols X of the symbols X.sub.n.sub.i to which corresponds the different reference signals to be inserted.

(67) RS insertion module 4.1.9 may inform the Data modulator module 4.1.10 of the chosen configuration. Enabling the Data modulator module 4.1.10 to insert modulation symbols into the block of symbols in positions that do not conflict with the positions n.sub.i or n.sub.i+M/2 with i ∈ 1;K of the reference signals.

(68) At step S12 the RS insertion module 4.1.9, inserts the reference signals as previously described, by setting each value of the symbols X.sub.n which are at positions n.sub.i or n.sub.i+M/2 with i ∈ 1;K at a value of a reference signal.

(69) At step S13 the signal is processed, that is on the block of symbols X=(X.sub.0, . . . X.sub.M−1) is applied a SC-SFBC-like scheme (DFT-PAPR preserving SFBC-subcarrier mapping-IDFT).

(70) At step S14 the signal is emitted by Tx1 4.1.1 and Tx2 4.1.2.

(71) Referring to FIG. 5.2 there is shown a flowchart representing the steps of post-IDFT inserting reference signals in a radio signal according to the invention.

(72) At step S21 the RS insertion module 4.2.2 may also be configured in a static way or dynamically as in FIG. 5.1. Several configurations may also be pre-parametred in the RS insertion module 4.2.2, those configurations can be ordered according to the number of reference signals the configuration provides. A configuration may be defined by the number of reference signal K, by the positions n, in the block of symbols X of the symbols X.sub.n.sub.i to which corresponds the different reference signals to be inserted. When configured, the RS insertion module 4.2.2 may inform the Data modulator module 4.2.1 of the configuration chosen.

(73) At step S22, based on the configuration of the RS insertion module 4.2.2, the Data modulator module 4.2.1 sets the values of the symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K to 0, as previously described in FIG. 4.2.

(74) At step S23 the signal is processed, that is on the block of symbols X=(X.sub.0, . . . X.sub.M−1), to which the values of the symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K have been set to 0, is applied a SC-SFBC type scheme (DFT-PAPR preserving SFBC-subcarrier mapping-IDFT).

(75) At step S24 the RS insertion module 4.2.2 adds respectively to each of the output signals of the IDFT modules (4.2.4 and 4.2.5), which are {tilde over (x)}.sub.DATA.sup.Tx1 corresponding to the antenna Tx1 4.2.6 and {tilde over (x)}.sub.DATA.sup.Tx2 corresponding to the antenna Tx2 4.2.7, the signal {tilde over (x)}.sub.RS.sup.Tx1 and the signal {tilde over (x)}.sub.RS.sup.Tx2. The signal {tilde over (x)}.sub.RS.sup.Tx1 and {tilde over (x)}.sub.RS.sup.Tx2 are pre-computed based on the symbols X.sub.n.sub.i and X.sub.(n.sub.i.sub.+M/2) with i ∈ 1;K.

(76) At step S25 the signal is emitted by Tx1 4.2.6 and Tx2 4.2.7.

(77) Referring to FIG. 5.3 there is shown a flowchart representing the steps of extracting reference signals in the radio signal according to the invention.

(78) At step S31 the extractor 4.3.2 is configured according to the configuration of the RS insertion module (4.1.9 or 4.2.2). The same configurations pre-parametred in the RS insertion module (4.1.9 or 4.2.2) may be pre-parametred in the extractor 4.3.2. The transmitter 1.1 can optionally send control information to the receiver 1.2 through a control channel, this control information pointing the configuration to set for extracting the reference signal being sent by the transmitter.

(79) At step S32 the extractor 4.3.2 extracts parts of the signals outputted by the ADCs 4.3.1 during the time periods corresponding with the received samples of the reference signals. The extraction is conduct as described in FIG. 4.3.

(80) At step S33 the samples of the reference signals are processed as previously described.

(81) At step S34 the channel estimation module 4.3.3 compares these reference signals with reference values, that is the corresponding values of the emitted samples of the reference signals, to obtain a channel estimation quality. The channel estimation module 4.3.3 may also specify a previously obtained channel estimation quality.

(82) At step S35 the signal received is then processed, using the channel estimation quality to enhance the performance of the processing. For example the PAPR preserving SFBC de-combiner 4.3.4 may be set to compensate the corruption (phase shift, amplitude . . . ) of the signal in the channel between transmitter and receiver.