Method and apparatus for downlink multiuser MIMO transmission in a wireless network

1. A computer implemented method comprising:

at a multi-antenna base station that is communicating with K multi-antenna relay stations where K is an integer greater than 1, obtaining a column vector v k (1) of a beamforming matrix associated with each of said K multi-antenna relay stations;

assembling an intermediate matrix W SVD using said column vectors;

using zero-forcing on said intermediate matrix W SVD to calculate a transmit beamforming matrix W for use by said multi-antenna base station to transmit data to said K relay stations, wherein the transmit beaming matrix W includes using zero-forcing to calculate:

W 0 =W SVD ( W SVD H W SVD ) −1 , and

W ZF-SVD =W 0 /trace( W 0 H W 0 ),

wherein H is a conjugate transpose of the intermediate matrix W SVD and trace is the trace of (W 0 H W 0 ); and

transmitting data from said multi-antenna base station to said K relay stations using said transmit beamforming matrix W, wherein the relay stations are stationary and have stationary propagation channels.

2. The method of claim 1 wherein:

assembling an intermediate matrix W SVD using said column vectors includes assembling a matrix W SVD =[v 1 (1) , . . . , v K (1) ].

3. The method of claim 1 wherein:

using zero-forcing on said intermediate matrix W SVD to calculate said transmit beamforming matrix W includes using zero-forcing to calculate:

W=W SVD ( W SVD H W SVD ) −1 .

4. The method of claim 1 wherein:

obtaining a column vector v k (1) of a beamforming matrix associated with each relay station within said K multi-antenna relay stations includes receiving a column vector v 1 (1) from a first relay station, wherein said first relay station generated a channel matrix for a channel between said base station and said first relay station, performed a singular value decomposition (SVD) on the channel matrix to generate a beamforming matrix, and extracted said column vector v 1 (1) from said beamforming matrix before transmitting said column vector v 1 (1) to said base station.

5. The method of claim 1 wherein:

obtaining a column vector v k (1) of a beamforming matrix associated with each relay station within said K multi-antenna relay stations includes receiving at said base station, from a first relay station, a channel matrix for a channel between said base station and said first relay station, performing a singular value decomposition (SVD) on said channel matrix at said base station to generate a beamforming matrix, and extracting a column vector v 1 (1) from said beamforming matrix.

6. An apparatus comprising:

a beamformer to precode data to be transmitted from multiple antennas to K remote relay stations using a beamforming matrix W, where K is an integer greater than 1;

a data collection unit to collect a largest column vector v k (1) of a beamforming matrix associated with each of said K remote relay stations;

an intermediate matrix assembler to assemble an intermediate matrix W SVD using said largest column vectors v k (1) ; and

a zero-forcing filter to perform zero-forcing on said intermediate matrix W SVD to generate a transmit beamforming matrix W for use by the beamformer, wherein the zero-forcing filter performs zero-forcing on the intermediate matrix W SVD using:

W 0 =W SVD ( W SVD ( W SVD H W SVD ) −1 , and

W ZF-SVD =W 0 /trace( W 0 H W 0 ),

wherein H is a conjugate transpose of the intermediate matrix W SVD and trace is the trace of (W 0 H W 0 ).

7. The apparatus of claim 6 , wherein said intermediate matrix assembler assembles said intermediate matrix as:

W SVD =[V 1 (1) , . . . , v K (1) ].

8. The apparatus of claim 6 , wherein said zero-forcing filter performs zero-forcing on said intermediate matrix W SVD as follows:

W=W SVD ( W SVD H W SVD ) −1 .

9. The apparatus of claim 6 , wherein:

said data collection unit is configured to receive said largest column vectors v k (1) directly from said K remote relay stations.

10. The apparatus of claim 6 , wherein:

said data collection unit is configured to receive a channel estimate from each of said K remote relay stations, each channel estimate being for a channel between said apparatus and a corresponding remote relay station, wherein said data collection unit is configured to use a singular value decomposition (SVD) unit to perform a SVD on each received channel estimate to generate a beamforming matrix and extract a largest column vector v k (1) of said beamforming matrix.

11. A non-transitory computer readable storage medium having computer executable instructions stored thereon that, when executed by a computing platform, operate to:

obtain a column vector v k (1) of a beamforming matrix associated with each of K multi-antenna relay stations, where K is an integer greater than 1;

assemble an intermediate matrix W SVD using said column vectors;

use zero-forcing on said intermediate matrix W SVD to calculate a transmit beamforming matrix W for use by said multi-antenna base station to transmit data to said K multi-antenna wireless, wherein the transmit beaming matrix W includes using zero-forcing to calculate:

W 0 =W SVD ( W SVD H W SVD ) −1 , and

W ZF-SVD =W 0 /trace( W 0 H W 0 ),

wherein H is a conjugate transpose of the intermediate matrix W SVD and trace is the trace of (W 0 H W 0 ); and

transmit data to said K multi-antenna relay stations using said transmit beamforming matrix W, wherein the relay stations are stationary and have propagation channels.

12. The non-transitory storage medium of claim 11 , wherein:

said operation to assemble an intermediate matrix W svp using said column vectors includes an operation to assemble a matrix W SVD =[V 1 (1) , . . . , V K (1) ].

13. The non-transitory storage medium of claim 11 , wherein:

said operation to use zero-forcing on said intermediate matrix W SVD to calculate said transmit beamforming matrix W includes an operation to use zero-forcing to calculate:

W=W SVD (( W SVD H W SVD ) −1 .

14. The non-transitory storage medium of claim 11 , wherein:

said operation to obtain a column vector v k (1) of a beamforming matrix associated with each of K multi-antenna relay stations includes an operation to receive a column vector v 1 (1) from a first relay station of said K multi-antenna relay stations, wherein said first relay station had generated a channel matrix for a wireless channel associated with said first relay station, performed a singular value decomposition (SVD) on said channel matrix to generate a beamforming matrix, and extracted said column vector v 1 (1) from said beamforming matrix before transmitting said column vector v 1 (1) .

15. The non-transitory storage medium of claim 11 , wherein:

said operation to obtain a column vector v k (1) of a beamforming matrix associated with each of K multi-antenna relay stations includes an operation to receive, from a first relay station of said K multi-antenna relay stations, a channel matrix for a channel associated with said first relay station, an operation to perform a singular value decomposition (SVD) on said channel matrix to generate a beamforming matrix, and an operation to extract a column vector v 1 (1) from said beamforming matrix.