Correlation shaping matched filter receiver

• US 7,751,469 B2
• Filed: 02/20/2001
• Issued: 07/06/2010
• Est. Priority Date: 02/20/2001
• Status: Expired due to Fees

• Current Assignee: Massachusetts Institute of Technology
• Original Assignee: Massachusetts Institute of Technology
• Inventors: Oppenheim, Alan V., Eldar, Yonina C.
• Primary Examiner(s): Burd; Kevin M
• Application Number:
  US09/788,890
• Publication Number:
  US 20020146066A1
• Time in Patent Office: 3,423 Days
• Field of Search: 375/144, 375/152, 375/147, 375/150, 375/316, 375/324, 375/325, 375/130
• US Class Current: 375/152
• CPC Class Codes: H04B 1/7093 Matched filter type
• CPC Class Codes: H04B 1/71052 using decorrelation matrix
• CPC Class Codes: H04L 25/03299 with noise-whitening circuitry

Abstract

The present invention is directed toward an apparatus and method for receiving signals, demodulating the signals and shaping the correlation of the output of a correlation demodulator. The present invention can be used irrespective of whether the received signals have noise components that are Gaussian or non-Gaussian. Moreover, the present invention can be utilized when a predetermined set of received signals is linearly independent or linearly dependent.

Claims

1. A receiver comprising:

a correlation demodulator, the correlation demodulator receiving one or more predetermined signals and noise, the demodulator having plural correlators, each correlator cross correlating the one or more predetermined signals and noise with a signal and outputting a correlation signal, and the demodulator providing a set of correlation signals a, the correlation signals a having a first covariance matrix; and

a correlation shaper, the correlation shaper receiving the set of correlation signals a and transforming the correlation signals a to a second set of signals b, equal in number to the set of correlation signals a, the second set of signals b having a second covariance matrix and the second covariance matrix is proportional to an identity matrix on at least a subspace of the second covariance matrix.

2. The receiver of claim 1 wherein the correlation demodulator is a matched filter demodulator.

3. The receiver of claim 1, wherein the transformation is a whitening transformation determined substantially by minimizing the mean squared error of the difference between the second set of signals and the set of correlation signals.

4. The receiver of claim 1, wherein the transformation is determined substantially by minimizing the mean squared error of the difference between the second set of signals and the set of correlation signals.

5. The receiver of claim 1, wherein the transformation is a subspace whitening transformation performed on the set of correlation signals.

6. The receiver of claim 5, wherein the subspace whitening transformation is determined substantially by minimizing the mean squared error of the difference between the second set of signals and the set of correlation signals.

7. The receiver of claim 1, wherein the transformation is performed on a subspace.

8. A method for processing signals comprising the steps of:

receiving a signal, the signal having at least one member of a predetermined set of signals that has undergone some distortion;

demodulating the received signal in a correlation demodulator adapted to correlate the received signal in plural correlators and provide a set of correlation signals, the correlation signals having a first covariance matrix; and

transforming the set of correlation signals with a correlation shaper, the correlation shaper adapted to provide a second set of signals, equal in number to the set of correlation signals, the second set of signals having a second covariance matrix and the second covariance matrix is proportional to an identity matrix on at least a subspace of the second covariance matrix.

9. The method of claim 8, wherein the transforming comprises a whitening transformation and includes substantially minimizing the mean squared error of the difference between the second set of signals and the set of correlation signals.

10. The method of claim 8, wherein the transforming includes substantially minimizing the mean squared error of the difference between the second set of signals and the set of correlation signals.

11. The method of claim 8, wherein the transforming comprises a subspace whitening transformation of the set of correlation signals.

12. The method of claim 11, wherein the transforming further comprises substantially minimizing a mean squared error of the difference between the second set of signals and the set of correlation signals.

13. A method of processing a received signal comprising:

demodulating the received signal with plural correlators, the received signal being a combination of one or more predetermined signals s_k and distortion;

cross-correlating the received signal in each of plural correlators with one of a set of predetermined linearly independent signals s_k to produce one of a set of correlation output signals a_k, wherein the produced set of correlation output signals has a positive-definite covariance matrix; and

transforming the set of correlation output signals to a set of transformed output signals b_k according to a whitening transformation W such that

b=Wa

where a is a vector representation of the correlation output signals a_k and b is a vector representation of the transformed output signals b_k and W is a matrix defined by

W=(S*S)<sup>−

1/2</sup>

where the j,k^th element of S*S is given substantially by the inner product of the predetermined signals S_j,S_k.

14. A multi-signal receiver comprising:

a demodulator having plural correlators, each one of plural correlators cross-correlating a received signal with one of a set of predetermined linearly independent signals S_k to produce one of a set of correlation output signals a_k, wherein

the received signal is a combination of one or more predetermined signals S_k and distortion, and

the produced set of correlation output signals has a positive-definite covariance matrix; and

a correlation shaper, the correlation shaper operatively transforming the correlation output signals to a set of transformed output signals b_k according to a whitening transformation W such that

b=Wa

where a is a vector representation of the correlation output signals a_k and b is a vector representation of the transformed output signals b_k and W is a matrix defined by

W=(S*S)<sup>−

1/2</sup>

where the j, k^th element of S*S is given substantially by the inner product of the predetermined signals S_j,S_k.

15. A method of processing a received signal comprising:

demodulating the received signal with plural correlators, the received signal being a combination of one or more predetermined signals S_k and distortion;

cross-correlating the received signal in each of plural correlators with one of a set of predetermined linearly dependent signals S_k to produce one of a set of correlation output signals a_k; and

transforming the set of correlation output signals to a set of transformed output signals b_k according to a subspace whitening transformation W such that

b=Wa

where a is a vector representation of the correlation output signals a_k and b is a vector representation of the transformed output signals b_k and W is a matrix defined by

W=((S*S)^1/2 )†

where the j, k^th element of S*S is given substantially by the inner product of the predetermined signals S_j,S_k and ( )†

denotes the Moore-Penrose pseudo-inverse of the matrix.

16. A multi-signal receiver comprising:

a demodulator having plural correlators, each one of plural correlators cross-correlating a received signal with one of a set of predetermined linearly dependent signals S_k to produce one of a set of correlation output signals a_k, wherein

the received signal is a combination of one or more predetermined signals

S_k and distortion; and

a correlation shaper, the correlation shaper operatively transforming the correlation output signals to a set of transformed output signals b_k according to a subspace whitening transformation W such that

b=Wa

where a is a vector representation of the correlation output signals a_k and b is a vector representation of the transformed output signals b_k and W is a matrix defined by

W=((S*S)^1/2)†

where the j,k^th element of S*S is given substantially by the inner product of the predetermined signals S_j,S_k and ( )†

denotes the Moore-Penrose pseudo-inverse of the matrix.

17. A method of processing a received signal comprising:

demodulating the received signal with plural correlators, the received signal being a combination of one or more predetermined linearly independent signals S_k and distortion;

cross-correlating the received signal in each of plural correlators with one of a set of orthogonal signals q_j given substantially by

q_j=Σ

_k(W*)_jk S_k

where W* is the adjoint of a matrix W, matrix W defined by

W=(S*S)<sup>−

1/2</sup>

where the j,k^th element of S*S is given substantially by the inner product of the predetermined signals S_j,S_k.

18. A method of processing a received signal comprising:

demodulating the received signal with plural correlators, the received signal being a combination of one or more predetermined linearly dependent signals S_k and distortion;

cross-correlating the received signal in each of plural correlators with one of a set of projected orthogonal signals q_j given substantially by

q_j=Σ

_k(W*)_jk S_k

where W* is the adjoint of a matrix W, matrix W defined by

W=((S*S)^1/2)†

where the j,k^th element of S*S is given substantially by the inner product of the predetermined signals S_j,S_k and ( )†

denotes the Moore-Penrose pseudo-inverse of the matrix.