Method for processing signal in communications system having plurality antennas

• US 6,876,693 B2
• Filed: 08/31/2001
• Issued: 04/05/2005
• Est. Priority Date: 09/02/2000
• Status: Expired due to Term

• Current Assignee: LG Electronics, Inc. (LG Corporation)
• Original Assignee: LG Electronics, Inc. (LG Corporation)
• Inventors: Sim, Dong Hi
• Primary Examiner(s): Ha, Dac V.
• Application Number:
  US09/942,999
• Publication Number:
  US 20020044616A1
• Time in Patent Office: 1,313 Days
• Field of Search: 375/144, 375/148, 375/267, 375/285, 375/346, 375/347, 375/349, 455/50.1, 455/52.3, 455/52.1, 455/63, 455/65, 455/67.3, 455/296, 455/303, 370/317, 370/320, 370/334, 370/335, 370/342, 370/441, 370/479, 342368-384, 342/457
• US Class Current: 375/144
• CPC Class Codes: H04B 7/0854 using error minimizing algorithms, e.g. minimum mean squared error [MMSE], "cross-correlation" or matrix inversion

Abstract

The present invention relates to an adaptive algorithm and device for the same, for calculating a weight vector to be applied to an adaptive array antenna system, suggesting an adaptive algorithm suitable for calculating an autocovariance matrix of signal vectors of signals received from antennas before despreading in a known channel code, and an autocovariance matrix of signal vectors of signals received from antennas after being despread in the known channel code. Particularly, the calculation is carried out with the autocovariance matrix of the signal vector before the despread divided into diagonal components and non-diagonal components.

Claims

1. A method for processing signal in a communication system having a plurality of antennas, comprising the steps of:

(a) extracting a first signal mostly including a specific signal, and a second signal mostly including a signal other than the specific signal, from a signal received at the plurality of antennas;

(b) calculating an autocovariance matrix of each of the first and second signals;

(c) dividing at least any one of the calculated autocovariance matrices into a diagonal component and a non-diagonal component, to separate the matrix into matrices of the components;

(d) calculating a weighted value for the specific signal by using the separated autocovariance matrices; and

,

(e) applying the weighted value to a transmission, or reception signal related to the specific signal, and forwarding a signal having the weighted value applied thereto.

2. A method as claimed in claim 1, wherein the communication system is a CDMA type radio communication system, and the first signal is a signal the received signal is despread in a particular code.

3. A method as claimed in claim 2, wherein the second signal is a signal before the received signal is despread in a particular code.

4. A method as claimed in claim 1, wherein the received signal includes a vector having a number of elements equal to, or less than a number of the antennas.

5. A method as claimed in claim 4, wherein the autocovariance matrix has a number of rows or columns the same with the number of elements of the vector of the received signal.

6. A method as claimed in claim 1, wherein the calculated weighted value is a vector having a number of element the same with the number of rows or columns of the autocovariance matrix.

7. A method as claimed in claim 1, wherein the step (d) includes the step of obtaining a weighted vector which leads a ratio of a product of the autocovariance of the first signal and the weighted value to a product of the autocovariance of the second signal and the weighted value to a maximum.

8. A method as claimed in claim 7, wherein the separated matrix is an autocovariance matrix of the first signal.

9. A method as claimed in claim 8, wherein the weight vector w can be calculated by w=(λ

Rxxw−

Ryy^Ow)(Ryy^D)<sup>−

1</sup>, where x denotes a second signal vector, y denotes a first signal vector, Rxx denotes the autocovariance matrix of the x, Ryy denotes the autocovariance matrix of the y, Ryy^D denotes a matrix of diagonal components of the Ryy, Ryy^O denotes a matrix of non-diagonal components of the Ryy, (Ryy^D)<sup>−

1</sup> denotes an inverse matrix of the Ryy^D, and λ

denotes a greatest proper value of Ryyw−

λ

Rxxw, which is a generalized Eigenvalue problem.

10. A method as claimed in claim 9, wherein the greatest proper value λ

of Ryyw−

λ

Rxxw, which is a generalized Eigenvalue problem, can be calculated by $\mathrm{\lambda <br><br>}=\frac{{\underset{\_}{w}}^H<br><br>R_{\mathrm{yy}}<br><br>\underset{\_}{w}}{{\underset{\_}{w}}^H<br><br>R_{\mathrm{xx}}<br><br>\underset{\_}{w}}$

with respect to ‘

H’

, the Hermitian operator.

11. A method as claimed in claim 8, wherein the weight vector w can be calculated by w(k+1)=λ

Rxxw(k) (Ryy^D)<sup>−

1</sup>−

Ryy^Ow(k)(Ryy^D)<sup>−

1</sup>, where x denotes a second signal vector, y denotes a first signal vector, Rxx denotes the autocovariance matrix of the x, Ryy denotes the autocovariance matrix of the y, Ryy^D denotes a matrix of diagonal components of the Ryy, Ryy^O denotes a matrix of non-diagonal components of the Ryy, (Ryy^D)<sup>−

1</sup> denotes an inverse matrix of the Ryy^D, and λ

denotes a greatest proper value of Ryyw−

λ

Rxxw, which is a generalized Eigenvalue problem.

12. A method as claimed in claim 7, wherein the separated matrix is the autocovariance matrix of the second signal.

13. A method as claimed in claim 12, wherein the weighted value w can be calculated by $\underset{\_}{w}=\frac{\{{R_{\mathrm{yy}}}^{}}{}$

(RxxD)-1-λ

⁢

 

⁢

RxxO⁡

(RxxD)-1}⁢

w_λ

,

where x denotes a second signal vector, y denotes a first signal vector, Rxx denotes the autocovariance matrix of the x, Ryy denotes the autocovariance matrix of the y, Rxx^D denotes a matrix of diagonal components of the Rxx, Rxx^O denotes a matrix of non-diagonal components of the Rxx, (Rxx^D)<sup>−

1</sup> denotes an inverse matrix of the Rxx^D, and λ

denotes a greatest proper value of Ryyw−

λ

Rxxw, which is a generalized Eigenvalue problem.

14. A method as claimed in claim 12, wherein the weighted value w can be renewed with respect to a snapshot index ‘

k’ and

a following snapshot index (k+1) by $\underset{\_}{w}<br><br>\left(k+1\right)=\frac{\left\{{R_{\mathrm{yy}}<br><br>\left(R_{\mathrm{xx}}^D\right)}^{-1}-\mathrm{\lambda <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>{R_{\mathrm{xx}}^O<br><br>\left(R_{\mathrm{xx}}^D\right)}^{-1}\right\}<br><br>\underset{\_}{w}<br><br>\left(k\right)}{\mathrm{\lambda <br><br>}},$

where x denotes a second signal vector, v denotes a first signal vector, Rxx denotes the autocovariance matrix of the x, Ryy denotes the autocovariance matrix of the y, Rxx^D denotes a matrix of diagonal components of the Rxx, Rxx^O denotes a matrix of non-diagonal components of the Rxx, (Rxx^D)<sup>−

1</sup> denotes an inverse matrix of the Rxx^D, and λ

denotes a greatest proper value of Ryyw−

λ

Rxxw, which is a generalized Eigenvalue problem.

15. A method as claimed in claim 12, wherein the greatest proper value λ

of Ryyw−

λ

Rxxw, which is a generalized Eigenvalue problem, can be calculated by $\mathrm{\lambda <br><br>}=\frac{{\underset{\_}{w}}^H<br><br>R_{\mathrm{yy}}<br><br>\underset{\_}{w}}{{\underset{\_}{w}}^H<br><br>R_{\mathrm{xx}}<br><br>\underset{\_}{w}}$

with respect to ‘

H’

, the Hermitian operator.