Manifold assisted channel estimation and demodulation for CDMA systems in fast fading environments

• US 6,501,747 B1
• Filed: 05/12/1999
• Issued: 12/31/2002
• Est. Priority Date: 08/20/1998
• Status: Active Grant

• Current Assignee: Hanger Solutions, LLC (IP Investments Group LLC)
• Original Assignee: Metawave Communications Corp.
• Inventors: Abbott, William L., Friedlander, Benjamin, Scherzer, Shimon B.
• Primary Examiner(s): Chin, Vivian
• Assistant Examiner(s): Lee, John J
• Application Number:
  US09/310,567
• Time in Patent Office: 1,329 Days
• Field of Search: 370/320, 370/334, 370/335, 370/339, 370/342, 370/441, 370/479, 370/442, 375/142, 375/144, 375/148, 375/150, 375/140, 375/145, 375/146, 375/147, 375/130, 375/152
• US Class Current: 370/342
• CPC Class Codes: H04B 7/0615 of weighted versions of same signal
• CPC Class Codes: H04B 7/0617 for beam forming
• CPC Class Codes: H04B 7/086 using weights depending on external parameters, e.g. direction of arrival [DOA], predetermined weights or beamforming

Abstract

Known antenna array manifold information is exploited to provide for fast and accurate channel estimation and demodulation on both the forward and reverse links, thereby increasing capacity in PCS and cellular CDMA networks that use adaptive antenna arrays. On the reverse link, an “extended” array manifold is used to assist the demodulator in maintaining a dynamic estimate of the spatial signature to use for beamforming and coherent demodulation. On the forward link, channel estimation is performed in the handset to provide a robust solution for beamforming.

Claims

1. A method of signal processing for a wireless communication system, comprising:

receiving a set of de-spread vectors y_i, wherein i=1 to N and N is the number of possible received symbols from a wireless transmission;

spatially correlating each of the de-spread vectors y_i with an array manifold matrix C to produce a correlation matrix P, wherein each vector y_i has M components, M being equal to the number of antenna output elements;

determining angle of arrival (AOA) information of each vector y_i from matrix P;

determining angular distribution information from a collection of AOA information;

creating a subspace matrix from the matrix C and the AOA and angular distribution information;

projecting each vector y_i into the subspace matrix to produce a set of N vectors z_i;

finding a vector z_max from the one of vectors z_i having the maximum energy;

averaging vector z_max to get a vector z_avg; and

multiplying vector z_avg with the subspace matrix to produce a beamforming coefficient vector w.

2. The method of claim 1, wherein the determining comprises finding the maximum element of the matrix P.

3. The method of claim 1, wherein the determining comprises finding the maximum element of the matrix resulting from averaging a set of P matrices.

4. The method of claim 1, further comprising, after the multiplying, determining the inner product of each vector y_i with the vector w.

5. The method of claim 1, wherein the array manifold matrix C is an M×

K matrix, K being equal to the number of angles in the array manifold.

6. The method of claim 5, wherein the elements of the i^th column of the correlation matrix P represent a level of correlation between vector y_i and the array manifold for each of the K angles.

7. The method of claim 1, wherein creating the subspace matrix comprises a singular value decomposition of a portion of the array manifold matrix C.

8. The method of claim 1, wherein the subspace matrix is an approximation of the actual subspace.

9. The method of claim 8, wherein the approximation is obtained by selecting columns of the actual subspace containing dominant singular values.

10. The method of claim 1, wherein the averaging is over a period approximately an order of magnitude shorter than the Doppler spread of the received signals.

11. The method of claim 1, wherein the system is an IS-95 based system.

12. The method of claim 11, wherein N is equal to 64.

13. The method of claim 11, wherein the determining comprises selecting the row and column of the matrix P containing the overall maximum value of the matrix p.

14. The method of claim 11, wherein the determining comprises:

selecting the maximum-valued element from each row of the matrix P to form a K-element {tilde over (p)} vector;

averaging the {tilde over (p)} vector; and

selecting the maximum value of the averaged {tilde over (p)} vector and corresponding position.

15. The method of claim 1, wherein the system is a W-CDMA or CDMA2000 proposed system.

16. The method of claim 15, wherein the y_i vectors are a combination of de-spread pilot vectors y_p and de-spread data vectors y_d.

17. The method of claim 15, wherein N is equal to 2.

18. The method of claim 16, wherein the combination comprises a real part and an imaginary part.

19. The method of claim 16, wherein the determining comprises finding the maximum element of the matrix P.

20. The method of claim 16, wherein the determining comprises finding the maximum element of the matrix resulting from averaging a set of P matrices.

21. The method of claim 16, wherein the determining comprises averaging over all the elements of the matrix P and finding the resulting maximum element and corresponding position.

22. A method of signal processing for a wireless communication system, comprising:

receiving, in a mobile station, a de-spread pilot signal vector y_fp, wherein the vector y_fp has M components, M being equal to the number of antenna output elements in the base station;

spatially correlating the vector y_fp with an array manifold matrix C to produce a correlation vector p_f;

determining angle of arrival (AOA) information for the vector y_fp from vector p_f;

determining angular distribution information from a collection of AOA information;

creating a subspace matrix from the matrix C and the AOA and angular distribution information;

projecting the vector y_fp into the subspace matrix to produce a filtered version of the vector y_fp;

calculating beamforming coefficients x_i, i=1 to J, from the matrix C and the AOA and angular distribution information; and

determining the inner product, x_i^Ty_fp, of the vector y_fp with each beamforming coefficient vector x_i.

23. The method of claim 22, further comprising, after determining the inner product, multiplying the conjugate of each inner product result with an associated one of J de-spread data signal values.

24. The method of claim 22, wherein pilot signals sent from each antenna element in the base station are each spread with a different extended Walsh code.

25. A beamforming system for wireless communication, comprising:

a primary spatial correlator (PSC), wherein the PSC receives de-spread data vectors y_i (i=1 to N) and signals representing an array manifold matrix C and transmits signals representing spatial information of the vectors y_i;

each vector y_i having M components, M being equal to the number of antenna output elements and N being equal to the number of possible symbols from a wireless transmission;

a digital signal processor (DSP) coupled to the PSC, wherein the DSP determines additional spatial information and determines a subspace containing most of the energy from the vectors y_i; and

an extended manifold processor (EMP) coupled to the DSP and the vectors y_i, wherein the EMP projects signals representing each vector y_i into the subspace to generate projection coefficient vectors z_i and outputs signals representing a beamforming coefficient vector w.

26. The system of claim 25, wherein the spatial information is angle of arrival (AOA) information.

27. The system of claim 25, wherein the additional spatial information is the angular distribution information resulting from a collection of AOA information.

28. The system of claim 25, wherein the PSC performs a spatial correlation between each of the vectors y_i and the matrix C to produce signals representing a correlation matrix P.

29. The system of claim 25, wherein the subspace matrix results from a singular value decomposition of a portion of the array manifold matrix C.

30. The system of claim 25, wherein the subspace matrix is an approximation of the actual subspace.

31. The system of claim 30, wherein the approximation consists of columns of the actual subspace which contain dominant singular values.

32. The system of claim 25, wherein the EMP comprises selection circuitry for selecting a vector z_max having the maximum energy of the projection coefficient vectors z_i.

33. The system of claim 32, wherein the EMP further comprises averaging circuitry for averaging vectors z_max to generate an average projection coefficient vector z_avg that when multiplied by the subspace matrix produces an estimate of the spatial signature.

34. The system of claim 33, wherein the EMP further comprises a multiplier for multiplying the vector z_avg with the subspace matrix to produce the vector w.

35. The system of claim 25, further comprising a beamformer coupled to the EMP and the signal vectors y_i, wherein the beamformer determines the inner product, w^Hy_i, of each vector y_i with the vector w.

36. The system of claim 25, wherein the system is an IS-95 system.

37. The system of claim 36, wherein N is equal to 64.

38. The system of claim 28, wherein the system is an IS-95 system.

39. The system of claim 38, wherein the PSC selects the row and column of the matrix P containing the overall maximum value of the matrix P.

40. The system of claim 38, wherein the PSC comprises:

first selection circuitry for selecting the maximum-valued element from each row of the matrix P to form a K-element {tilde over (p)} vector;

averaging circuitry for averaging the {tilde over (p)} vector; and

second selection circuitry for selecting the maximum value of the averaged {tilde over (p)} vector and corresponding position.

41. The system of claim 28, wherein the system is a W-CDMA or CDMA2000 proposed system.

42. The system of claim 41, further comprising M data estimators, each data estimator receiving a de-spread pilot vector y_p and de-spread data vector y_d to form combinations of y_p and y_d to use as vectors y_i.

43. The system of claim 42, wherein:

the PSC further comprises a receiver for receiving a set of vectors y_i from the data estimators and a transmitter for transmitting spatial information of the vectors y_i; and

the DSP is adapted to determine a subspace containing most of the energy from the vectors y_i.

44. The system of claim 41, wherein N is equal to 2.

45. The system of claim 41, further comprising a beamformer coupled to the EMP and the de-spread data vectors y_d, wherein the beamformer determines the inner product, w^Hy_d, of each vector y_d with the vector w.

46. A demodulation system for wireless communication comprising a plurality of manifold assisted demodulator (MAD) fingers, a first one of the plurality of MAD fingers comprising:

a first block for de-spreading pilot and data signals and generating a de-spread pilot vector y_p and a de-spread data vector y_d;

a first data estimator block that forms complex valued combination vectors, y_i, of the de-spread pilot and data signal vectors;

a first primary spatial correlator (PSC), wherein the PSC receives the de-spread data vectors y_i (i=1 to N) and an array manifold matrix C and transmits spatial information of the vectors y_i, each vector y_i having M components, M being equal to the number of antenna output elements and N being equal to the number of possible symbols from a wireless transmission;

a first extended manifold processor (EMP), wherein the first EMP receives the vectors y_i and projects each vector y_i into a subspace to generate projection coefficient vectors z_i and outputs a beamforming coefficient vector w; and

a first beamformer coupled to the first EMP and the data signal vectors y_d, wherein the first beamformer determines the inner product, w^Hy_d, of each vector y_d with the vector w.

47. The system of claim 46, wherein a second one of the plurality of MAD fingers consists of:

a second block for de-spreading the pilot and data signals and generating de-spread pilot vector y_p and de-spread data vector y_d;

a second data estimator block that forms the complex valued combination vectors, y_i, of the de-spread pilot and data signal vectors; and

a second primary spatial correlator (PSC), wherein the second PSC receives the de-spread data vectors y_i (i=1 to N) and the array manifold matrix C and transmits spatial information of the vectors y_i, each vector y_i having M components, M being equal to the number of antenna output elements and N being equal to the number of possible symbols from a wireless transmission.

48. The system of claim 47, wherein a third one of the plurality of MAD fingers consists of:

a third block for de-spreading the pilot and data signals and generating de-spread pilot vector y_p and de-spread data vector y_d;

a third data estimator block that forms the complex valued combination vectors, y_i, of the de-spread pilot and data signal vectors; and

a second extended manifold processor (EMP), wherein the second EMP receives the vectors y_i and projects each vector y_i into a subspace to generate projection coefficient vectors z_i and outputs a beamforming coefficient vector w; and

a second beamformer coupled to the second EMP and the data signal vectors y_d, wherein the first beamformer determines the inner product, w^Hy_d, of each vector y_d with the vector w.