Receiver with prefiltering for discrete fourier transform-spread-orthogonal frequency division multiplexing (DFT-S-OFDM) based systems

1. A receiver for discrete Fourier transform-spread-orthogonal frequency division multiplexing (DFT-S-OFDM) based systems, comprising:

a prefilter for demodulating one of a received signal codeword and received signal codewords; and

a log-likelihood ratio LLR module responsive to the prefilter,

wherein the prefilter includes a pairing and whitening module that, based on channel estimates and a data rate, enables the LLR module to perform either a Serial-In-Serial-Out (SISO) based log likelihood ratio processing of an output from the pairing and whitening module or a two-symbol max-log soft output demodulator (two-symbol MLSD) based log likelihood ratio processing of the output from the pairing and whitening module, the pairing and whitening module electing not to process an input received from an inverse discrete Fourier transform (IDFT) module when the LLR module is enabled to perform the SISO based log likelihood ratio processing.

2. The receiver of claim 1 , wherein the pairing and whitening module provides an output of the form

s

^

=

[

s

^

1

,

…

⁢

,

s

^

M

]

T

⁢

=

Δ

⁢

F

†

⁢

x

^

,

where {circumflex over (x)} is a per-tone equalized received signal vector in the frequency domain and F † performs an inverse discrete Fourier transform (IDFT) operation on an M×M matrix.

3. The receiver of claim 1 , wherein for received signal codewords the pairing and whitening module provides an output of the form

s

^

(

k

)

⁢

=

Δ

⁢

F

†

⁢

x

^

(

k

)

,

where {circumflex over (x)} (k) is a per-tone equalized received signal vector in the frequency domain corresponding to the k th codeword and F \ performs an inverse discrete fourier transform (IDFT) operation.

4. The receiver of claim 1 , wherein for received signal codewords the pairing procedure is consistent with the following signal relationships

r=F[h 1 (1)† R 1 −1 h 1 (2) , . . . , h M (1)† R M −1 h M (2) ] T

and compute {circumflex over (q)} as

q

^

=

arg

⁢

⁢

max

1

≤

k

≤

M

⁢

{



r

k



}

-

1

,

where h m (1) , h m (2) denote the frequency domain channel response vectors on an m th tone corresponding to codeword 1 and codeword 2, respectively, and R m denotes a covariance matrix for the m th tone and F denotes a discrete Fourier transform matrix, {circumflex over (q)} being determined by subtracting one from an index of a maximum magnitude element of the vector r, an m th symbol of codeword 1 being paired with an (m+{circumflex over (q)}) th symbol of codeword 2 subject to a modulo operation.

5. A receiver for discrete Fourier transform-spread-orthogonal frequency division multiplexing (DFT-S-OFDM) based systems, comprising:

a prefilter for demodulating one of a received signal codeword and received signal codewords; and

a log-likelihood ratio LLR module responsive to the prefilter,

wherein the prefilter includes a pairing and whitening module that, based on channel estimates and a data rate, enables the LLR module to perform either a Serial-In-Serial-Out (SISO) based log likelihood ratio processing of an output from the pairing and whitening module or a two-symbol max-log soft output demodulator (two-symbol MLSD) based log likelihood ratio processing of the output from the pairing and whitening module, the pairing and whitening module electing to process an input received from an inverse discrete Fourier transform (IDFT) module when the LLR module is enabled to perform the two-symbol MLSD based log likelihood ratio processing.

6. The receiver of claim 5 , wherein the pairing and whitening module divides the symbols of a same codeword into multiple pairs and obtains a decision vector and an effective channel matrix for each pair.

7. The receiver of claim 5 , wherein the pairing and whitening module divides the symbols of two different codewords into multiple pairs and obtains a decision vector and an effective channel matrix for each pair.

8. The receivers of claim 6 , wherein LLRs corresponding to coded bits in each symbol pair are computed by the two-symbol MLSD responsive to the decision vector and the effective channel matrix of that pair.

9. The receiver of claim 6 , wherein the pairing and whitening module provides an output consisting of length-2 decision vectors {z m } that are expanded as z m =T m s m +{hacek over (n)} m , where s m denotes the vector consisting of the m th pair of symbols from the same codeword and T m denotes the effective channel matrix for the m th pair obtained after pairing and noise whitening with {hacek over (n)} m being the whitened noise vector having two uncorrelated elements.

10. The receiver of claim 7 , wherein for multiple received signal codewords, the pairing and whitening module provides an output consisting of length-2 decision vectors {z m } that are expanded as z m =T m s m +{hacek over (n)} m , where s m denotes the vector consisting of the m th pair of symbols from two different codewords and T m denotes the effective channel matrix for the m th pair obtained after pairing and noise whitening with {hacek over (n)} m being the whitened noise vector having two uncorrelated elements.

11. The receiver of claim 9 , wherein the pairing and whitening module provides an output consisting of length-2 decision vectors {z m } that are expanded as z m =Ts m +{hacek over (n)} m , where s m denotes the vector consisting of the m th pair of symbols from the same codeword and T denotes the effective channel matrix which is identical for all pairs with {hacek over (n)} m being the whitened noise vector having two uncorrelated elements.

12. The receiver of claim 10 , wherein the pairing and whitening module provides an output consisting of length-2 decision vectors {z m } that are expanded as z m =Ts m +{hacek over (n)} m , where s m denotes the vector consisting of the m th pair of symbols from two different codewords and T denotes the effective channel matrix which is identical for all pairs with {hacek over (n)} m being the whitened noise vector having two uncorrelated elements.

13. The receiver of claim 5 , wherein a pairing procedure is consistent with the following signal

relationships

r=F[h 1 † R 1 −1 h 1 , . . . , h M † R M −1 h M ] T

and compute {circumflex over (q)} as

q

^

=

arg

⁢

⁢

max

2

≤

k

≤

M

/

2

+

1

⁢

{



r

k



}

-

1

,

where h m denotes the frequency domain channel response vector on an m th tone, R m denotes a covariance matrix for the m th tone and F denotes a discrete Fourier transform matrix, {circumflex over (q)} being determined by subtracting 1 from the index of the element of the vector r that among all the elements of the vector r with indices in {2, . . . , M/2+1} has the maximum magnitude, and an m th symbol of the codeword being then paired with an (m+{circumflex over (q)}) th symbol of a same codeword.

14. A receiver for discrete Fourier transform-spread-orthogonal frequency division multiplexing (DFT-S-OFDM) based systems, comprising:

a prefilter for demodulating one of a received signal codeword and received signal codewords; and

a log-likelihood ratio LLR module responsive to the prefilter,

wherein the prefilter includes a pairing and whitening module that, based on channel estimates and a data rate, enables the LLR module to perform either a Serial-In-Serial-Out (SISO) based log likelihood ratio processing of an output from the pairing and whitening module or a two-symbol max-log soft output demodulator (two-symbol MLSD) based log likelihood ratio processing of the output from the pairing and whitening module, the prefilter including a per-tone equalizer for one of the received signal codeword and received signal codewords and an inverse discrete Fourier transform IDFT module responsive to the equalizer, the pairing and whitening module being responsive to the IDFT module.