COMMUNICATIONS CHANNEL ESTIMATION

• US 20110038445A1
• Filed: 08/13/2009
• Published: 02/17/2011
• Est. Priority Date: 08/13/2009
• Status: Active Grant

• Current Assignee: Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: Wang, Jibing, Zhou, Yan
• Granted Patent: US 8,599,969 B2
• Time in Patent Office: Days
• US Class Current: 375/346
• CPC Class Codes: H04L 25/0204 of multiple channels
• CPC Class Codes: H04L 25/022 of frequency response
• CPC Class Codes: H04L 25/0228 with direct estimation from sounding signals
• CPC Class Codes: H04L 25/0248 Eigen-space methods

Abstract

Techniques for estimating the response of a communications channel. In an aspect, a channel autocorrelation matrix is derived based on, e.g., certain assumptions about the time delay profile, Doppler profile. An eigenvector decomposition is performed on the autocorrelation matrix, and a set of dominant eigenvectors is determined. A received pilot signal vector is projected onto a projection matrix derived from the dominant eigenvectors to generate a set of eigen-coefficients, which is then used to reconstruct an estimated channel vector. In an exemplary embodiment, the eigenvector-based approach is combined with a simplified minimum mean-square error based approach to generate an estimated channel vector when the received signal-to-noise ratio is below a threshold.

Claims

1. A method for estimating a channel vector for a communications channel, the method comprising:

generating an eigen-based weight matrix based on a dominant eigenvector matrix derived from an autocorrelation matrix of a channel vector, the weight matrix comprising the product of a data selection matrix, the dominant eigenvector matrix, and the pseudo-inverse of a projection matrix, the projection matrix comprising the product of a transmit pilot matrix, a pilot selection matrix, and the dominant eigenvector matrix; and

multiplying the eigen-based weight matrix with a received pilot vector to generate an estimated data channel vector.

2. The method of claim 1, the generating the eigen-based weight matrix comprising retrieving the eigen-based weight matrix from a memory, the eigen-based weight matrix in the memory being computed off-line.

3. The method of claim 2, the step of multiplying being performed while receiving a signal over the communications channel from a transmitter.

4. The method of claim 1, the dominant eigenvector matrix comprising dominant eigenvectors selected from the eigenvectors of the autocorrelation matrix, the dominant eigenvectors determined as those having corresponding eigenvalues greater than a constant times the mean of all eigenvalues.

5. The method of claim 1, further comprising:

estimating a signal-to-noise ratio (SNR) of the communications channel;

only performing the step of multiplying the eigen-based weight matrix with a received pilot vector if the SNR estimate is above a threshold.

6. The method of claim 5, further comprising, if the SNR estimate is below the threshold:

generating a simplified MMSE-based weight matrix, the weight matrix comprising the result of a data-pilot cross-correlation matrix divided by a noise estimate; and

multiplying the simplified MMSE-based weight matrix with the received pilot vector to generate the estimated data channel vector.

7. The method of claim 6, the generating the simplified MMSE-based weight matrix comprising using a fixed SNR estimate.

8. The method of claim 6, the generating the simplified MMSE-based weight matrix comprising retrieving the simplified MMSE-based weight matrix from a memory.

9. An apparatus for estimating a channel vector for a communications channel, the apparatus comprising:

a weight generation module configured to generate an eigen-based weight matrix, the weight matrix comprising the product of a data selection matrix, an eigenvector basis matrix, and the pseudo-inverse of a projection matrix, the projection matrix comprising the product of a pilot selection matrix and the eigenvector basis matrix, the eigenvector basis matrix comprising the dominant eigenvectors of an autocorrelation matrix of the channel vector; and

a module configured to multiply the output of the weight generation module with a received signal vector to generate an estimated data channel vector.

10. The apparatus of claim 9, the received signal vector being a received pilot vector.

11. The apparatus of claim 9, the dominant eigenvectors comprising eigenvectors having corresponding eigenvalues greater than a threshold.

12. The apparatus of claim 9, further comprising a memory storing the eigen-based weight matrix, the weight generation module configured to generate the eigen-based weight matrix by reading the matrix from the memory.

13. The apparatus of claim 9, the weight generation module further configured to generate a simplified minimum mean-square error (MMSE)-based weight matrix, the apparatus further comprising:

an SNR range estimation module configured to select the simplified MMSE-based weight matrix as the output of the weight generation matrix when the SNR is below a threshold, and to select the eigen-based weight matrix as the output of the weight generation matrix when the SNR is above the threshold.

14. An apparatus for estimating a channel vector for a communications channel, the apparatus comprising:

means for generating an eigen-based weight matrix based on a dominant eigenvector matrix derived from an autocorrelation matrix of a channel vector; and

means for multiplying the eigen-based weight matrix with a received pilot vector to generate an estimated data channel vector.

15. The apparatus of claim 14, further comprising means for generating a simplified minimum mean-square error (MMSE)-based weight matrix when an estimated SNR of the channel is less than a threshold.

16. A computer program product storing code for causing a computer to estimate a channel vector for a communications channel, the code comprising:

code for causing a computer to generate an eigen-based weight matrix based on a dominant eigenvector matrix derived from an autocorrelation matrix of a channel vector, the weight matrix comprising the product of a data selection matrix, the dominant eigenvector matrix, and the pseudo-inverse of a projection matrix, the projection matrix comprising the product of a transmit pilot matrix, a pilot selection matrix, and the dominant eigenvector matrix;

code for causing a computer to multiply the eigen-based weight matrix with a received pilot vector to generate an estimated data channel vector.

17. The computer program product of claim 16, the code for causing a computer to generate the eigen-based weight matrix comprising:

code for causing a computer to retrieve the eigen-based weight matrix from a memory, the eigen-based weight matrix stored in memory being computed off-line.

18. The computer program product of claim 16, the code for causing a computer to multiply comprising code for causing a computer to multiply while receiving a signal over the communications channel from a transmitter.

19. The computer program product of claim 16, the dominant eigenvector matrix comprising dominant eigenvectors selected from the eigenvectors of the autocorrelation matrix, the dominant eigenvectors determined as those having corresponding eigenvalues greater than a constant times the mean of all eigenvalues.

20. The computer program product of claim 16, the code further comprising:

code for causing a computer to estimate a signal-to-noise ratio (SNR) of the communications channel; and

code for causing a computer to only perform the step of multiplying the eigen-based weight matrix with a received pilot vector if the SNR estimate is above a threshold.

21. The computer program product of claim 20, the code further comprising:

code for causing a computer to, if the SNR estimate is below the threshold, generate a simplified MMSE-based weight matrix, the weight matrix comprising the result of a data-pilot cross-correlation matrix divided by a noise estimate; and

code for causing a computer to multiply the simplified MMSE-based weight matrix with the received pilot vector to generate the estimated data channel vector.