Methods and apparatus for channel estimation and precoding with incomplete channel observation and channel state information feedback

1. A method implemented by a node, the method comprising:

estimating first downlink (DL) channel information of first DL channels in accordance with reference signals received on first uplink (UL) channels, the first UL channels being a subset of available UL channels, and the first DL channels corresponding to the first UL channels;

receiving channel feedback for a set of DL channels;

estimating second DL channel information in accordance with the first DL channel information and a projection of the channel feedback; and

communicating in accordance with the second DL channel information.

2. The method of claim 1 , the second DL channel information comprising information derived from the first DL channel information and an orthogonal projection of the channel feedback.

3. The method of claim 1 , the estimating the second DL channel information comprising:

projecting, by the node, channel information derived from the channel feedback onto an orthogonal subspace of the first DL channels, thereby producing the projection;

selecting, by the node, one or more dominant eigenvectors of the projection; and

combining, by the node, the first DL channel information and the one or more dominant eigenvectors to form the second DL channel information.

4. The method of claim 1 , the communicating in accordance with the second DL channel information comprising:

precoding, by the node, data in accordance with the second DL channel information.

5. The method of claim 1 , further comprising:

processing, by the node, the channel feedback after the receiving the channel feedback, thereby producing processed channel feedback.

6. The method of claim 5 , the estimating the second DL channel information comprising:

projecting, by the node, channel information derived from the processed channel feedback onto an orthogonal subspace of the first DL channels, thereby producing the projection;

selecting, by the node, one or more dominant eigenvectors of the projection; and

combining, by the node, the first DL channel information and the one or more dominant eigenvectors to form the second DL channel information.

7. The method of claim 5 , the processing the channel feedback comprising:

filtering or denoising the channel feedback with channel statistics.

8. The method of claim 7 , the filtering or the denoising the channel feedback comprising at least one of:

minimum mean square error (MMSE) filtering the channel feedback, projecting of dominant subspaces of the channel feedback, maximum fairness (MF) or maximum ratio combining (MRC) based filtering of the channel feedback, or quadratic programming of the channel feedback.

9. The method of claim 5 , the processing the channel feedback comprising:

determining, by the node, a conditional covariance for the received channel feedback; and

filtering or denoising, by the node, the channel feedback in accordance with the conditional covariance.

10. The method of claim 9 , the filtering or the denoising the channel feedback comprising at least one of:

MMSE filtering the channel feedback, projecting of dominant subspaces of the channel feedback, MF or MRC based filtering of the channel feedback, or quadratic programming of the channel feedback.

11. A method implemented by a node, the method comprising:

estimating first downlink (DL) channel information of first DL channels in accordance with reference signals received on first uplink (UL) channels, the first UL channels being a subset of available UL channels, and the first DL channels being associated with the first UL channels;

receiving channel feedback for a set of DL channels;

eigen decomposing the first DL channel information to generate an eigen decomposition of the first DL channel information;

selecting one or more dominant first eigenvectors of the eigen decomposition as a partial precoder;

projecting channel information derived from the channel feedback onto an orthogonal subspace of the first DL channels;

selecting one or more dominant second eigenvectors of the channel information after the projecting of the channel information as second DL channel information of second DL channels, the second DL channels being unassociated with the first UL channels; and

combining the partial precoder and the one or more dominant second eigenvectors to form a precoder.

12. The method of claim 11 , further comprising:

processing the channel feedback after the receiving the channel feedback.

13. The method of claim 12 , the processing the channel feedback comprising:

filtering or denoising the channel feedback with channel statistics.

14. The method of claim 13 , further comprising:

determining a conditional covariance for the channel feedback, the filtering or the denoising the channel feedback being in accordance with the conditional covariance.

15. The method of claim 13 , the filtering or the denoising the channel feedback comprising at least one of:

minimum mean square error (MMSE) filtering the channel feedback, projecting of dominant subspaces of the channel feedback, maximum fairness (MF) or maximum ratio combining (MRC) based filtering of the channel feedback, or quadratic programming of the channel feedback.

16. A node comprising:

a non-transitory memory storage comprising instructions; and

one or more processors in communication with the non-transitory memory storage, wherein the one or more processors execute the instructions to cause the node to perform operations including:

estimating first downlink (DL) channel information of first DL channels in accordance with reference signals received on first uplink (UL) channels, the first UL channels being a subset of UL channels, and the first DL channels corresponding to the first UL channels;

receiving channel feedback for a set of DL channels;

estimating second DL channel information in accordance with the first DL channel information and a projection of the channel feedback; and

communicating in accordance with the second DL channel information.

17. The node of claim 16 , the node to the estimating the second DL channel information comprising:

projecting channel information derived from the channel feedback onto an orthogonal subspace of the first DL channels, thereby producing the projection;

selecting one or more dominant eigenvectors of the projection; and

combining the first DL channel information and the one or more dominant eigenvectors to form the second DL channel information.

18. The node of claim 16 , the communicating in accordance with the second DL channel information comprising:

precoding data in accordance with the second DL channel information.

19. The node of claim 16 , the operations further comprising:

processing the channel feedback after the receiving the channel feedback, thereby producing processed channel feedback.

20. The node of claim 19 , the estimating the second DL channel information comprising:

projecting channel information derived from the processed channel feedback onto an orthogonal subspace of the first DL channels, thereby producing the projection;

selecting one or more dominant eigenvectors of the projection; and

combining the first DL channel information and the one or more dominant eigenvectors to form the second DL channel information.

21. The node of claim 20 , the processing the channel feedback comprising:

filtering or denoising the channel feedback with channel statistics.

22. The node of claim 19 , the processing the channel feedback comprising:

determining a conditional covariance for the channel feedback; and

filtering or denoising the channel feedback in accordance with the conditional covariance.