Low complexity precoding matrix selection

• US 20080304464A1
• Filed: 06/05/2007
• Published: 12/11/2008
• Est. Priority Date: 06/05/2007
• Status: Active Grant

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Texas Instruments, Inc.
• Inventors: Mujica, Fernando Alberto, Borkar, Milind Anil
• Granted Patent: US 7,769,098 B2
• Time in Patent Office: Days
• US Class Current: 370/342
• CPC Class Codes: H04B 7/0434 using multiple eigenmodes
• CPC Class Codes: H04B 7/0478 Special codebook structures directed to feedback optimisation
• CPC Class Codes: H04B 7/0639 Using selective indices, e.g. of a codebook, e.g. pre-distortion matrix index [PMI] or for beam selection
• CPC Class Codes: H04B 7/0848 Joint weighting
• CPC Class Codes: H04L 1/0045 Arrangements at the receiver end
• CPC Class Codes: H04L 1/0625 Transmitter arrangements
• CPC Class Codes: H04L 25/0204 of multiple channels
• CPC Class Codes: H04L 25/0246 with factorisation
• CPC Class Codes: H04L 25/0248 Eigen-space methods
• CPC Class Codes: H04L 5/0007 the frequencies being orthogonal, e.g. OFDM(A), DMT
• CPC Class Codes: H04L 5/0039 Frequency-contiguous, i.e. with no allocation of frequencies for one user or terminal between the frequencies allocated to another
• CPC Class Codes: H04L 5/0041 Frequency-non-contiguous

Abstract

A method of determining indices for matrix codewords in a matrix codeword codebook. The matrix codewords are adapted for communicating information between a transmitter and a receiver. The method includes retrieving from temporary storage, an eigenmode representation for a communications channel, where the eigenmode representation is based upon on a received signal precoded by a first matrix codeword. The method also includes performing a test on multiple vector codewords to identify a first vector codeword among the multiple vector codewords, where the test includes determining a relationship between the first vector codeword and the representation of an eigenmode. The first vector codeword is associated with a first vector codeword index that identifies the first vector codeword. The method also includes generating a matrix codeword index associated with a second matrix codeword in the matrix codeword codebook. The matrix codeword index is based upon the first vector codeword index, and the order of the first vector codeword is different from the order of the second matrix codeword.

Claims

1. A method of determining indices for matrix codewords in a matrix codeword codebook, the matrix codewords adapted for communicating information between a transmitter and a receiver, the method comprising:

retrieving from temporary storage an eigenmode representation for a communications channel, wherein the representation of the eigenmode is based upon on a received signal precoded by a first matrix codeword;

performing a test on multiple vector codewords to identify a first vector codeword among the multiple vector codewords, wherein the test includes determining a relationship between the first vector codeword and the eigenmode representation, the first vector codeword is associated with a first vector codeword index that identifies the first vector codeword; and

generating a matrix codeword index associated with a second matrix codeword in the matrix codeword codebook, wherein the matrix codeword index is based upon the first vector codeword index, and the order of the first vector codeword is different from the order of the second matrix codeword.

2. The method of claim 1, further comprising receiving the signal precoded by the first matrix codeword.

3. The method of claim 1, further comprising:

sending a representation of the matrix codeword index to the transmitter; and

receiving a second signal precoded by the second matrix codeword.

4. The method of claim 1, wherein the eigenmode representation for the communications channel includes a weakest eigenmode, and determining a relationship between the first vector codeword and the eigenmode representation includes determining correlation between the weakest eigenmode and the first vector codeword.

5. The method of claim 4, wherein the correlation between the weakest eigenmode and the first vector codeword is stronger than correlations between the weakest eigenmode and each of the vector codewords, other than the first vector codeword, among the multiple vector codewords.

6. The method of claim 1, wherein the eigenmode representation for the communications channel includes a strongest eigenmode, and determining a relationship between the first vector codeword and the eigenmode representation includes determining a correlation between the strongest eigenmode and the first vector codeword.

7. The method of claim 6, wherein the correlation between the strongest eigenmode and the first vector codeword is greater than correlations between the strongest eigenmode and each of the vector codewords, other than the first vector codeword, among the multiple vector codewords.

8. The method of claim 1, further comprising:

generating a vector by taking the product of a Householder operation on a vector codeword and the eigenmode representation; and

performing a second test on a second multiple vector codewords to identify a second vector codeword among the second multiple vector codewords, wherein the second test includes determining a relationship between the second vector codeword and the generated vector, and the second vector codeword is associated with a second vector codeword index that identifies the second vector codeword.

9. The method of claim 8, wherein the matrix codeword index is also based upon the second vector codeword index.

10. The method of claim 1, further comprising:

generating a matrix by taking the product of a Householder operation on a vector codeword and the eigenmode representation; and

performing a second test on a second multiple vector codewords to identify a second vector codeword among the second multiple vector codewords, wherein the second test includes determining a relationship between the second vector codeword and a column of the generated matrix, and the second vector codeword is associated with a second vector codeword index that identifies the second vector codeword.

11. The method of claim 10, wherein the matrix codeword index is also based upon the second vector codeword index.

12. An apparatus that determines indices for matrix codewords in a matrix codeword codebook, the matrix codewords adapted for communicating information between a transmitter and a receiver, the apparatus comprising:

an electronically accessible storage that is to store an eigenmode representation for a communications channel, wherein the representation of the eigenmode is based upon a received signal precoded by a first matrix codeword;

control logic coupled to the temporary storage and that is to perform a test on multiple vector codewords to identify a first vector codeword among the multiple vector codewords, wherein the test includes determining a relationship between the first vector codeword and the eigenmode representation, the first vector codeword is associated with a first vector codeword index that identifies the first vector codeword; and

wherein the control logic generates a matrix codeword index associated with a second matrix codeword in the matrix codeword codebook, wherein the matrix codeword index is based upon the first vector codeword index, and the order of the first vector codeword is different from the order of the second matrix codeword.

13. The apparatus of claim 12, further comprising a receiver coupled to the control logic and that is to receive the signal precoded by the first matrix codeword.

14. The apparatus of claim 13, further comprising a second transmitter coupled to the control logic and that is to send a representation of the matrix codeword index to the first transmitter, and wherein the receiver is to receive a second signal precoded by the second matrix codeword.

15. The apparatus of claim 13, wherein the eigenmode representation for the communications channel includes a weakest eigenmode, and the control logic is to determine a relationship between the first vector codeword and the eigenmode representation by determining correlation between the weakest eigenmode and the first vector codeword.

16. The apparatus of claim 13, wherein the eigenmode representation for the communications channel includes a strongest eigenmode, and the control logic is to determine a relationship between the first vector codeword and the eigenmode representation by determining a correlation between the strongest eigenmode and the first vector codeword.

17. The apparatus of claim 13, wherein the control logic is to generate a vector by taking the product of a Householder operation on a vector codeword and the eigenmode representation, and is to perform a second test on a second multiple vector codewords to identify a second vector codeword among the second multiple vector codewords, and wherein the second test includes determining a relationship between the second vector codeword and the generated vector, the second vector codeword is associated with a second vector codeword index that identifies the second vector codeword., and the matrix codeword index is also based upon the second vector codeword index.

18. The apparatus of claim 13, wherein the control logic is to generate a matrix by taking the product of a Householder operation on a vector codeword and the eigenmode representation, and perform a second test on a second multiple vector codewords to identify a second vector codeword among the second multiple vector codewords, and wherein the second test includes determining a relationship between the second vector codeword and a column of the generated matrix, the second vector codeword is associated with a second vector codeword index that identifies the second vector codeword, and the matrix codeword index is also based upon the second vector codeword index.

19. A method of determining indices for matrix codewords in a matrix codeword codebook, the matrix codewords adapted for communicating information between a transmitter and a receiver, the method comprising:

determining number of transmit antennas, N_t, being used to transmit to the receiver and the number of data streams, N_s, being transmitted to the receiver;

identifying a Householder operation and a vector codebook for generating a matrix codeword codebook, CB(N_t, N_s, L), wherein L is a representation of a number of bits in an index for identifying each matrix codeword in the matrix codeword codebook;

based upon the identified Householder operation and the vector codebook, performing a first test on multiple vector codewords associated with the vector codebook, wherein the result of the first test is a first vector code index that identifies a first vector codeword among the multiple vector codewords; and

based upon the result of the first test generating a matrix codeword index associated with a second matrix codeword in the matrix codeword codebook, wherein the order of the first vector codeword is different from the order of the second matrix codeword.

20. The method of claim 19, further comprising:

generating a matrix by taking the product of a Householder operation on a vector codeword and the eigenmode representation; and

performing a second test on a second multiple vector codewords to identify a second vector codeword among the second multiple vector codewords, wherein the second test includes determining a relationship between the second vector codeword and the generated matrix, the second vector codeword is associated with a second vector codeword index that identifies the second vector codeword, the matrix codeword index is also based upon the second vector codeword index, and the generated matrix has an order that is either equivalent to or greater than the order of the second vector codeword.