Beamforming techniques for MIMO communication systems

1. In a multiple input multiple output (MIMO) communication system including a source device that transmits, via a number N t of transmit antennas, a beamformed signal to a sink device, a beamforming method for the source device, the method comprising:

employing a processor executing computer readable instructions to perform the following acts:

determining a set of beamforming weights to be applied to a number N ss of signal streams to generate a number N t of transmit streams, wherein the number N ss is at least two, wherein the act of determining includes:

computing a singular value decomposition of a channel matrix H associated with a communication channel from the source device to the sink device;

defining a preliminary weight matrix based on the singular value decomposition; and

phase-rotating each column of the preliminary weight matrix such that a phase of a reference element in each column is equal to a target phase, thereby generating the set of beamforming weights;

applying the set of beamforming weights to the number N ss of signal streams to generate the number N t of transmit streams; and

transmitting the number N t of transmit streams via the number N t of transmit antennas.

2. The method of claim 1 , wherein the source device transmits signals to the sink device using a plurality of subcarriers, the method further comprising:

performing the act of determining the set of beamforming weights independently for each of the plurality of subcarriers.

3. The method of claim 2 , wherein for each subcarrier n in the plurality of subcarriers, the act of computing the singular value decomposition of the channel matrix H(n) associated with the subcarrier n includes:

factoring the channel matrix H(n) into a product H(n)=U(n)D(n)V H (n), wherein U(n) and V(n) are unitary matrices, V H (n) is a Hermetian transpose of matrix V(n) and D(n) is a matrix in which singular values of the channel matrix H(n) appear on a main diagonal thereof.

4. The method of claim 3 , wherein defining the preliminary weight matrix includes:

defining, as the preliminary weight matrix, a matrix that includes at least a subset of columns of the matrix V H (n).

5. The method of claim 2 , wherein the act of phase-rotating each column of the preliminary weight matrix is performed for respective weight matrices associated with each one of the subcarriers, and wherein elements in corresponding locations in the respective weight matrices are used as the reference elements for each weight matrix.

6. The method of claim 1 , further comprising:

prior to applying the set of beamforming weights, modifying the set of beamforming weights so as to normalize a power consumption per transmit antenna, wherein the modified set of beamforming weights, and not the set of beamforming weights is applied during the act of applying.

7. The method of claim 1 , further comprising:

prior to applying the set of beamforming weights, modifying the set of beamforming weights so as to normalize a power consumption per signal stream, wherein the modified set of beamforming weights, and not the set of beamforming weights, is applied during the act of applying.

8. The method of claim 1 , further comprising:

prior to applying the set of beamforming weights, modifying the set of beamforming weights so as to truncate an impulse response length associated with the set of beamforming weights, wherein the modified set of beamforming weights, and not the set of beamforming weights, is applied during the act of applying.

9. In a multiple input, multiple output (MIMO) communication system including a source device that transmits, via a number N t of transmit antennas, a beamformed signal to a sink device, a beamforming method for the source device, the method comprising:

employing a processor executing computer readable instructions to perform the following acts:

determining a beamforming weight matrix W to be applied to a number N ss of signal streams to generate a number N t of transmit streams, wherein the number N ss is at least two, wherein the act of determining includes:

computing an initial beamforming weight matrix using a minimum mean square error technique; and

if the number N ss of signal streams is less than a number N r of receive antennas:

optimizing the initial beamforming weight matrix by selecting a number N ss of columns in the initial beamforming weight matrix for which a minimum power is required to equalize a channel; and

determining the optimized beamforming weight matrix to be the beamforming weight matrix W;

if the number N ss of signal streams is not less than a number N r of receive antennas:

determining the initial beamforming weight matrix to be the beamforming weight matrix W;

applying the beamforming weight matrix to the number N ss of signal streams to generate the number N t of transmit streams; and

transmitting the number N t of transmit streams via the number N t of transmit antennas.

10. The method of claim 9 , wherein the act of determining the beamforming weight matrix W includes:

computing a matrix W=H H (αI+HH H ) −1 , wherein H is a channel matrix characterizing a channel between the source device and the sink device, H H is a Hermetian transpose of the channel matrix H, I is an identity matrix, and α is a complex scale factor chosen to prevent division by zero during the matrix inversion operation.

11. The method of claim 9 , wherein the source device transmits signals to the sink device using a plurality of subcarriers, the method further comprising:

computing a beamforming matrix W(n) for each of the subcarriers n.

12. The method of claim 11 , wherein selecting the number N ss of columns in the beamforming weight matrix W(n) for each subcarrier includes:

determining a power value P C (n) for each column of each beamforming weight matrix W(n);

averaging respective power values P C (n) of corresponding columns in the beamforming weight matrices W(n) over all subcarriers n; and

selecting, in each beamforming weight matrix W(n), the number N ss of columns with lowest average power values.

13. The method of claim 9 , further comprising:

prior to applying the beamforming weight matrix W, modifying the beamforming weight matrix W so as to normalize a power consumption per transmit antenna, wherein the modified beamforming weight matrix, and not the beamforming weight matrix W, is applied during the act of applying.

14. The method of claim 9 , further comprising:

prior to applying the beamforming weight matrix W, modifying the beamforming weight matrix W so as to normalize a power consumption per signal stream, wherein the modified beamforming weight matrix, and not the beamforming weight matrix W, is applied during the act of applying.

15. The method of claim 9 , further comprising:

prior to applying the beamforming weight matrix W, modifying the beamforming weight matrix W so as to truncate an impulse response length associated with the beamforming weight matrix W, wherein the modified beamforming weight matrix, and not the beamforming weight matrix W, is applied during the act of applying.

16. In a multiple input, multiple out (MIMO) communication system including a source device that transmits, via a number N t , of transmit antennas, a beamformed signal to a sink device, a beamforming method for the source device, the method comprising:

employing a processor executing computer readable instructions to perform the following acts:

determining a beamforming weight matrix W to apply to a number N ss of signal streams to generate a number N t of transmit streams, wherein the number N ss is at least two, wherein the act of determining includes:

computing a singular value decomposition of a channel matrix associated with a communication channel from the source device to the sink device, the singular value decomposition including a singular value matrix D and at least one unitary matrix U; and

determining the beamforming weight matrix W based on inverting a product matrix obtained from the unitary matrix U and the singular value matrix D;

applying the beamforming weight matrix W to the number N ss of signal streams to generate the number N t of transmit streams; and

transmitting the number N t of transmit streams via the number N t of transmit antennas.

17. The method of claim 16 , wherein the source device transmits signals to the sink device using a plurality of subcarriers, the method further comprising:

performing the acts of computing the singular value decomposition and determining the beamforming weight matrix W independently for each of the plurality of subcarriers.

18. The method of claim 17 , wherein for each subcarrier n in the plurality of subcarriers, the act of computing the singular value decomposition of the channel matrix H(n) associated with the subcarrier n includes:

factoring the channel matrix H(n) into a product H(n)=U(n)D(n)V H (n), wherein U(n) and V(n) are unitary matrices, V H (n) is a Hermetian transpose of matrix V(n) and D(n) is a matrix in which singular values of the channel matrix H(n) appear on a main diagonal thereof.

19. The method of claim 18 , wherein for each subcarrier n in the plurality of subcarriers, determining a beamforming weight matrix W(n) associated with the subcarrier n includes:

computing W(n)=V(n)(αI+D(n)) −1 , wherein I is an identity matrix and α is a complex scale factor chosen to prevent division by zero during the matrix inversion operation.

20. The method of claim 19 , wherein, if a number N r of receive antennas, the number N t of transmit antennas, and the number N ss of signal streams are not equal to one another, the matrices V(n) and D(n) are respectively represented by truncated versions of the matrices V(n) and D(n), wherein the truncated versions are equal to upper-left N ss ×N ss elements of V(n) and D(n), the matrices V(n) and D(n) being respectively represented by the truncated versions prior to the act of computing W(n)=V(n)(αI+D(n)) −1 .

21. The method of claim 16 , further comprising:

prior to applying the beamforming weight matrix W, modifying the beamforming weight matrix W so as to normalize a power consumption per transmit antenna, wherein the modified beamforming weight matrix, and not the beamforming weight matrix W, is applied during the act of applying.

22. The method of claim 16 , further comprising:

prior to applying the beamforming weight matrix W, modifying the beamforming weight matrix W so as to normalize a power consumption per signal stream, wherein the modified beamforming weight matrix, and not the beamforming weight matrix W, is applied during the act of applying.

23. The method of claim 16 , further comprising:

prior to applying the beamforming weight matrix W, modifying the beamforming weight matrix W so as to truncate an impulse response length associated with the beamforming weight matrix W, wherein the modified beamforming weight matrix, and not the beamforming weight matrix W, is applied during the act of applying.

24. In a multiple input, multiple output (MIMO) communication system including a source device that transmits, via a number N t of transmit antennas, a beamformed signal to a sink device, a beamforming method for the source device, the method comprising:

employing a processor executing computer readable instructions to perform the following acts:

determining a set of beamforming weights to apply to a number N ss of signal streams to generate a number N t of transmit streams, wherein the number N ss is at least two;

truncating an impulse response time associated with the set of beamforming weights;

applying the truncated-response set of beamforming weights to the number N ss of signal streams to generate the number N t of transmit streams; and

transmitting the number N t of transmit streams via the number N t of transmit antennas.

25. The method of claim 24 , wherein the source device transmits signals to the sink device using a plurality of subcarriers and wherein the act of determining a set of beamforming weights is performed for each subcarrier, thereby generating a plurality of beamforming weight matrices W(n).

26. The method of claim 25 , wherein the act of truncating includes:

converting the set of beamforming weights to a time domain representation, thereby producing a set of time-domain weights;

truncating the set of time-domain weights at a cutoff time; and

converting the truncated set of time-domain weights to a frequency domain representation, thereby producing the truncated-response set of beamforming weights.

27. In a multiple input, multiple output (MIMO) communication system including a source device that transmits, via a number N t of transmit antennas, a beamformed signal to a sink device, a beamforming method for the source device, the method comprising:

employing a processor executing computer readable instructions to perform the following acts:

determining a set of beamforming weights to apply to a number N ss of signal streams to generate a number N t of transmit streams, wherein the number N ss is at least two;

normalizing the set of beamforming weights, wherein normalizing the set of beamforming weights optimizes power required per signal stream;

applying the normalized set of beamforming weights to the number N ss of signal streams to generate the number N t of transmit streams; and

transmitting the number N t , of transmit streams via the number N t of transmit antennas.

28. In a multiple input, multiple output (MIMO) communication system including a source device that transmits, via a number N t of transmit antennas, a beamformed signal to a sink device, a beamforming method for the source device, the method comprising:

employing a processor executing computer readable instructions to perform the following acts:

determining a set of beamforming weights to apply to a number N ss of signal streams to generate a number N t of transmit streams, wherein the number N ss is at least two;

normalizing the set of beamforming weights to optimize power required per each of the number N t of transmit antennas;

rescaling the set of normalized beamforming weights;

applying the rescaled set of normalized beamforming weights to the number N ss of signal streams to generate the number N t of transmit streams; and

transmitting the number N t of transmit streams via the number N t of transmit antennas.