Transmitter beamforming for self-interference cancellation

1. A method for self-interference cancellation, comprising:

in an acquisition phase, iteratively acquiring received power measurements for different receive chains for transmissions sent via a plurality of beam directions;

estimating a self-interference subspace eigenvector based, at least in part, on the received power measurements;

based on the estimated self-interference subspace eigenvector, performing a local search of a nullspace to identify a filtering eigenvector, wherein performing the local search of the nullspace comprises one of:

iteratively estimating a gradient in the received power measurements for each of the plurality of beam directions, identifying a maximum gradient increase, and performing a line search in a direction opposite a direction of the maximum gradient increase, or

iteratively estimating a gradient in the received power measurements for each of the plurality of beam directions, identifying a minimum point in an inverted Hessian matrix estimated from the received power measurements, and performing a line search in a direction of the identified minimum point; and

applying the filtering eigenvector to one or more transmission chains to cancel self-interference on one or more receive chains.

2. The method of claim 1 , wherein the acquisition phase is performed while in a transmit-only mode.

3. The method of claim 2 , wherein the filtering eigenvector is applied while in a full-duplex operation mode.

4. The method of claim 1 , further comprising:

comparing a received power measurement to a saturation threshold, wherein estimating the self-interference subspace eigenvector is based on identifying a beam direction having a received power exceeding the saturation threshold.

5. A system for self-interference cancellation, comprising:

a processor configured to:

in an acquisition phase, iteratively acquire received power measurements for different receive chains for transmissions sent via a plurality of beam directions;

estimate a self-interference subspace eigenvector based, at least in part, on the received power measurements;

based on the estimated self-interference subspace eigenvector, perform a local search of a nullspace to identify a filtering eigenvector, wherein performing the local search of the nullspace comprises one of:

iteratively estimating a gradient in the received power measurements for each of the plurality of beam directions, identifying a maximum gradient increase, and performing a line search in a direction opposite a direction of the maximum gradient increase, or

iteratively estimating a gradient in the received power measurements for each of the plurality of beam directions, identifying a minimum point in an inverted Hessian matrix estimated from the received power measurements, and performing a line search in a direction of the identified minimum point; and

apply the filtering eigenvector to one or more transmission chains to cancel self-interference on one or more receive chains, and

a memory coupled to the processor.

6. The system of claim 5 , wherein the acquisition phase is performed while in a transmit-only mode.

7. The system of claim 6 , wherein the filtering eigenvector is applied while in a full-duplex operation mode.

8. The system of claim 5 , wherein the processor is further configured to:

compare a received power measurement to a saturation threshold, wherein estimating the self-interference subspace eigenvector is based on identifying a beam direction having a received power exceeding the saturation threshold.

9. A system for self-interference cancellation, comprising:

means for, in an acquisition phase, iteratively acquiring received power measurements for different receive chains for transmissions sent via a plurality of beam directions;

means for estimating a self-interference subspace eigenvector based, at least in part, on the received power measurements;

means for performing, based on the estimated self-interference subspace eigenvector, a local search of a nullspace to identify a filtering eigenvector, wherein the means for performing the local search of the nullspace comprises one of:

means for iteratively estimating a gradient in the received power measurements for each of the plurality of beam directions, means for identifying a maximum gradient increase, and means for performing a line search in a direction opposite a direction of the maximum gradient increase, or

means for iteratively estimating a gradient in the received power measurements for each of the plurality of beam directions, means for identifying a minimum point in an inverted Hessian matrix estimated from the received power measurements, and means for performing a line search in a direction of the identified minimum point; and

means for applying the filtering eigenvector to one or more transmission chains to cancel self-interference on one or more receive chains.

10. The system of claim 9 , wherein the acquisition phase is performed while in a transmit-only mode.

11. The system of claim 10 , wherein the filtering eigenvector is applied while in a full-duplex operation mode.

12. The system of claim 9 , wherein the means for performing the local search of a nullspace comprises:

means for iteratively estimating a gradient in the received power measurements for each of the plurality of beam directions;

means for identifying a minimum point in an inverted Hessian matrix estimated from the received power measurements; and

means for performing a line search in a direction of the identified minimum point.

13. A non-transitory computer-readable medium for self-interference cancellation comprising instructions that when executed cause a wireless device to:

in an acquisition phase, iteratively acquire received power measurements for different receive chains for transmissions sent via a plurality of beam directions;

estimate a self-interference subspace eigenvector based, at least in part, on the received power measurements;

based on the estimated self-interference subspace eigenvector, perform a local search of a nullspace to identify a filtering eigenvector, wherein performing the local search of the nullspace comprises one of:

iteratively estimating a gradient in the received power measurements for each of the plurality of beam directions, identifying a maximum gradient increase, and performing a line search in a direction opposite a direction of the maximum gradient increase, or

iteratively estimating a gradient in the received power measurements for each of the plurality of beam directions, identifying a minimum point in an inverted Hessian matrix estimated from the received power measurements, and performing a line search in a direction of the identified minimum point; and

apply the filtering eigenvector to one or more transmission chains to cancel self-interference on one or more receive chains.

14. The non-transitory computer-readable medium of claim 13 , wherein the acquisition phase is performed while in a transmit-only mode.

15. The non-transitory computer-readable medium of claim 14 , wherein the filtering eigenvector is applied while in a full-duplex operation mode.

16. The non-transitory computer-readable medium of claim 13 , further comprising instructions for:

comparing a received power measurement to a saturation threshold, wherein estimating the self-interference subspace eigenvector is based on identifying a beam direction having a received power exceeding the saturation threshold.