Cochannel signal processing system

• US 6,018,317 A
• Filed: 11/22/1996
• Issued: 01/25/2000
• Est. Priority Date: 06/02/1995
• Status: Expired due to Term

• Current Assignee: Northrop Grumman Systems Corporation (Northrop Grumman Corporation)
• Original Assignee: TRW Inc. (Northrop Grumman Corporation)
• Inventors: Dogan, Mithat Can, Stearns, Stephen Deane
• Primary Examiner(s): Blum, Theodore M.
• Application Number:
  US08/755,775
• Time in Patent Office: 1,159 Days
• Field of Search: 342/378, 342/417, 342/451, 342/373
• US Class Current: 342/378
• CPC Class Codes: G01S 3/74 Multi-channel systems specially adapted for direction-finding, i.e. having a single antenna system capable of giving simultaneous indications of the directions of different signals systems in which the directions of different signals are determined sequentially and displayed simultaneously G01S3/04, G01S3/14
• CPC Class Codes: H04B 7/0857 using maximum ratio combining techniques, e.g. signal-to- interference ratio [SIR], received signal strenght indication [RSS]
• CPC Class Codes: H04B 7/086 using weights depending on external parameters, e.g. direction of arrival [DOA], predetermined weights or beamforming
• CPC Class Codes: H04L 25/03993 Noise whitening

Abstract

A method and apparatus for processing cochannel signals received at a sensor array in a cumulant-based signal processing and separation engine to obtain a desired set of output signals or parameters. For use in a signal recovery system, the output signals are recovered and separated versions of the originally transmitted cochannel signals. An important feature that distinguishes the cumulant-based system from other signal separation and recovery systems is that it generates an estimated generalized steering vector associated with each signal source, and representative of all received coherent signal components attributable to the source. This feature enables the invention to perform well in multipath conditions, by combining all coherent multipath components from the same source. In a receiver/transmitter system, the estimated generalized steering vectors associated with each source are used to generate transmit beamformer weight vectors that permit cochannel transmission to multiple user stations. The basic cumulant-based processing and separation engine can also be used in a variety of applications, such as high density recording, complex phase angle equalization, receiving systems with enhanced effective dynamic range, and signal separation in the presence of strong interference. Various embodiments and extensions of the basic cumulant-based system are disclosed.

Claims

1. A signal processing system for deriving at least one output quantity of interest from received cochannel input signals, the system comprising:

a signal receiving system, including means for generating a set of conditioned receiver signals from received signals of any modulation or type;

an estimated generalized steering vector (EGSV) generator, for computing an EGSV that results in optimization of a utility function that depends on fourth or higher even-order statistical cumulants derived from the received signals, the EGSV being indicative of a combination of signals received at the signal receiving system from a signal source; and

a supplemental computation module, for deriving at least one output quantities of interest from the conditioned receiver signals and the EGSV.

2. A signal processing system as defined in claim 1, wherein:

the EGSV generator uses an iterative technique to generate the EGSV and includes means for generating successively more accurate EGSVs based on iterative optimization of the utility function.

3. A signal processing system as defined in claim 2, wherein the means for generating successively more accurate EGSVs includes:

a linear combiner, for repeatedly computing a single channel combined signal from the conditioned receiver signals and an EGSV;

means for supplying an initial EGSV to the linear combiner, to produce the initial output of a single channel combined signal;

an EGSV computation module, for computing successive values of the EGSV from successive values of the single channel combined signal received from the linear combiner and the conditioned receiver signals; and

means for feeding the successive values of the EGSV back to the linear combiner for successive iteration cycles; and

means for terminating iterative operation upon convergence of the EGSV to a sufficiently accurate value.

4. A signal processing system as defined in claim 2, wherein the means for generating successively more accurate EGSVs includes:

a cross-cumulant matrix computation module, for generating a matrix of cross-cumulants of all combinations of the conditioned receiver signals;

a structured quadratic form computation module, for computing successive cumulant strength functions derived from successive EGSVs and the cross-cumulant matrix;

means for supplying an initial EGSV to the structured quadratic form computation module, to produce the initial output of a cumulant strength function;

an ESGV computation module, for generating successive EGSVs from successive cumulant strength functions received from the structured form computation module; and

means for feeding the successive values of the EGSV back to the structured quadratic form computation module for successive iteration cycles; and

means for terminating iterative operation upon convergence of the EGSV to a sufficiently accurate value.

5. A signal processing system as defined in claim 1, wherein the EGSV generator uses a direct computational technique to generate the EGSV, and includes:

a cross-cumulant matrix computation module, for generating a matrix of cross-cumulants of all combinations of the conditioned receiver signals; and

an EGSV computation module for computing the EGSV directly from the cross-cumulant matrix by solving a fourth degree polynomial equation.

6. A signal processing system as defined in one of claims 1-5, wherein:

the means for generating the set of conditioned signals includes an eigendecomposition module for generating an estimate of the number of signal sources, a transformation matrix for conditioning the receiver signals, and an eigenstructure derived from the receiver signals; and

the EGSV generator employs signals output by the eigendecomposition module to compute EGSVs.

7. A signal processing system as defined in one of claims 1-5, wherein:

the means for generating the set of conditioned signals includes a covariance matrix computation module and a matrix decomposition module, for generating a matrix; and

the system further includes a beamformer, for generating a recovered signal from the receiver signals, the EGSV and the matrix obtained from the matrix decomposition module.

8. A signal processing system as defined in one of claims 1-5, wherein:

the means for generating the set of conditioned signals includes an eigendecomposition module for generating and estimate of the number of signal sources, a transformation matrix for conditioning the receiver signals, and an eigenstructure derived from the receiver signals;

the EGSV generator employs signals output by the eigendecomposition module to compute EGSVs

the system further comprises a multiple port signal recovery unit, including means for matching current EGSVs with EGSVs from a prior data block to impose waveform continuity from block to block.

9. A signal processing system as defined in one of claims 2-4, wherein:

signals are processed in successive blocks of data; and

the initial of EGSVs for each new processing block are computed by a means for combining a prior block EGSV and a cumulant vector derived from the utility function.

10. A signal processing system as defined in claim 9, wherein:

the means for combining takes the sum of the prior block EGSV multiplied by a first factor, and the cumulant vector multiplied by a second factor, wherein the first and second factors are selected to provide an initial EGSV that anticipates and compensates for movement of a signal source.

11. A signal processing system as defined in claim 3, wherein:

the system functions to separate a plurality (P) of received cochannel signals;

there is a plurality (P) of EGSV generators, including P EGSV computation modules and P linear combiners, for generating an equal plurality (P) of EGSVs associated with P signal sources; and

the supplemental computation module functions to recover P separate received signals from the P generalized steering vectors and the conditioned receiver signals.

12. A signal processing system as defined in claim 11, wherein:

the supplemental computation module includes a recovery beamformer weight vector computation module, for generating from all of the EGSVs a plurality (P) of receive weight vectors, and a plurality (P) of recovery beamformers, each coupled to receive one of the P receive weight vectors and the conditioned receiver signals, for generating a plurality (P) of recovered signals.

13. A signal processing system as defined in claim 4, wherein:

the system functions to separate a plurality (P) of received cochannel signals;

there is a plurality (P) of EGSV generators, including P EGSV computation modules and P structured quadratic form computation modules, for generating an equal plurality (P) of EGSVs associated with P signal sources; and

the supplemental computation module functions to recover P separate received signals from the P generalized steering vectors and the conditioned receiver signals.

14. A signal processing system as defined in claim 13, wherein:

the supplemental computation module includes a recovery beamformer weight vector computation module, for generating from all of the EGSVs a plurality (P) of receive weight vectors, and a plurality (P) of recovery beamformers, each coupled to receive one of the P receive weight vectors and the conditioned receiver signals, for generating a plurality (P) of recovered signals.

15. A signal processing system as defined in claim 5, wherein:

the system functions to separate a two received cochannel signals;

the ESGV computation module generates two EGSVs from the cross-cumulant matrix data; and

the supplemental computation module functions to recover two separate received signals from the two generalized steering vectors and the conditioned receiver signals.

16. A signal processing system as defined in claim 15, wherein:

the supplemental computation module includes a recovery beamformer weight vector computation module, for generating from both of the EGSVs two receive weight vectors, and two recovery beamformers, each coupled to receive one of the receive weight vectors and the conditioned receiver signals, for generating two recovered signals.

17. A signal processing system as defined in one of claims 1-10, wherein:

the system functions to derive a direction of arrival (DOA) of a received signal;

the supplemental computation module includes a memory for storing sensor array calibration data, and means for deriving the DOA of a received signal from its associated generalized steering vector and the stored sensor array manifold data.

18. A signal processing system as defined in claim 17, wherein:

the sensor array manifold data includes a table associating multiple DOA values with corresponding steering vectors; and

the means for deriving the DOA includes means for performing a reverse table lookup function to obtain an approximated DOA value from a steering vector supplied by the generalized steering vector generator.

19. A signal processing system as defined in claim 18, wherein:

the means for deriving the DOA further includes means for interpolating between two DOA values to obtain a more accurate result.

20. A signal processing system as defined in one of claims 1-10, wherein the supplemental computation module includes:

a signal recovery module for generating received signal beamformer weights from the conditioned receiver signals and the EGSV; and

for recovering the received signal therefrom; and

a transmitter, for generating transmit signal beamformer weights from the received signal beamformer weights, and for transmitting signals containing information in a direction determined by the transmit signal beamformer weights.

21. A signal processing system as defined in claim 12 or 14, wherein the supplemental computation module further includes:

a transmit weight vector computation module, for generating P transmit beamforming weight vectors from receive weight vectors generated by the recovery beamformer weight vector computation module; and

P linear combiners, for combining each of the P information signals to be transmitted with its associated transmit weight vector, to obtain a weighted transmit beam for each of the information signals to be transmitted, and then combining corresponding components of the weighted transmit beams, for coupling to a transmit array.

22. A signal processing system as defined in one of claims 11-16, wherein:

the signal receiving system includes a plurality of waveguide sensors for receiving signals transmitted onto a waveguide in different modes, wherein the modes are subject to scrambling in the waveguide; and

the supplemental computation module separates and recovers the signals and mitigates the effect of mode mixing in the waveguide.

23. A signal processing system as defined in claim 22, wherein the waveguide is a microwave waveguide.

24. A signal processing system as defined in claim 22, wherein the waveguide is an optical fiber.

25. A signal processing system as defined in claim 22, wherein the waveguide is a coaxial cable.

26. A signal processing system as defined in claim 22, wherein the waveguide includes at least one twisted pair of conductors.

27. A signal processing system as defined in one of claims 11-16, wherein:

the signals received by the signal receiving system are in-phase and quadrature components of two-dimensional communication signal, which has been subject to phase rotation during propagation; and

the recovered signals generated automatically from the supplemental computation module are in-phase and quadrature components of a two-dimensional communication signal that has been corrected for phase rotation, wherein the system functions as a complex phase equalizer.

28. A signal processing system as defined in one of claims 1-5, wherein:

the signals received by the signal receiving system have been subject to distortion by analog processing and analog-to-digital conversion in a radio receiving system; and

the output quantities include a recovered signal having significantly less distortion than the received signals, whereby the receiving system has improved dynamic range as a result of the use of the signal processing system.

29. A signal processing system as defined in one of claims 11-16, wherein:

the signals received by the signal receiving system include signals from a relatively weak desired source and much stronger signals from at least one interfering source;

wherein the recovered and separated signals include those from the relatively weak desired source, free of interference, and those from the stronger interfering source, which can be discarded.

30. A method for processing cochannel signals received at a sensor array, the method comprising the steps of:

conditioning a set of signals received at a sensor array;

generating an estimated generalized steering vector (EGSV) that results in optimization of a utility function that depends on fourth or higher even-order statistical cumulants derived from the received signals, the EGSV being indicative of a combination of signals received at the sensors from a signal source; and

performing supplemental computation to derive at least one output quantity of interest from the conditioned receiver signals and the EGSV.

31. A method as defined in claim 30, wherein:

the step of generating an EGSV includes providing an initial EGSV and then iteratively generating successive EGSVs until an acceptable convergence is attained.

32. A method as defined in claim 31, wherein generating an EGSV includes:

computing in a linear combiner successive values of a single channel combined signal derived from an input EGSV and the received signals;

supplying an initial EGSV to the linear combiner;

computing successive EGSVs from the received signals and the successive values of the single channel combined signal;

feeding the successive EGSVs back to the linear combiner for further iteration; and

terminating further iteration when the EGSV has satisfactorily converged.

33. A method as defined in claim 31, wherein generating an EGSV includes:

computing a matrix of cross-cumulants of the received signals;

computing in a structured quadratic form computation module successive values of a cumulant strength function derived from an input EGSV and the cross-cumulants of the received signals;

supplying an initial EGSV to the structured quadratic form computation module;

computing successive EGSVs from the successive values of the cumulant strength;

feeding the successive EGSVs back to the structured quadratic form computation module for further iteration; and

terminating further iteration when the EGSV has satisfactorily converged.

34. A method as defined in claim 30, wherein generating an EGSV includes:

generating a matrix of cross-cumulants of all combinations of the received signals; and

computing the EGSV directly from the cross-cumulants by solving a fourth degree polynomial equation.

35. A method as defined in one of claims 30-34, wherein:

conditioning the received signals includes generating by eigendecomposition an estimate of the number of signal sources, a transformation matrix for conditioning the receiver signals, and an eigenstructure derived from the receiver signals; and

the step of generating EGSVs employs signals generated in the foregoing step of generating by eigendecomposition.

36. A method as defined in one of claims 30-34, wherein:

conditioning the received signals includes generating a covariance matrix and generating therefrom another matrix; and

the method further comprises beamforming to generate a recovered signal from the receiver signals, the EGSV and the other matrix obtained from the covariance matrix.

37. A method as defined in one of claims 30-34, wherein:

conditioning the received signals includes generating by eigendecomposition an estimate of the number of signal sources, a transformation matrix for conditioning the receiver signals, and an eigenstructure derived from the receiver signals;

the step of generating EGSVs employs signals generated in the foregoing step of generating by eigendecomposition

the method further comprises the step of matching current EGSVs with EGSVs from a prior data block to impose waveform continuity from block to block.

38. A method as defined in one of claims 31-33, wherein:

signals are processed in successive blocks of data; and

the method further comprises a step of computing an initial EGSV for each new processing block by combining a prior block EGSV and a cumulant vector derived from the utility function.

39. A method as defined in claim 38, wherein:

the combining step includes taking the sum of the prior block EGSV multiplied by a first factor, and the cumulant vector multiplied by a second factor, wherein the first and second factors are selected to provide an initial EGSV that anticipates and compensates for movement of a signal source.

40. A method as defined in claim 32, wherein:

the method separates a plurality (P) of received cochannel signals;

the step of generating an EGSV is performed in a plurality (P) of EGSV generators, including P EGSV computation modules and P linear combiners, for generating an equal plurality (P) of EGSVs associated with P signal sources; and

the step of performing supplemental computation includes recovering P separate received signals from the P generalized steering vectors and the conditioned receiver signals.

41. A method as defined in claim 40, wherein:

the step of performing supplemental computation further includes generating, in a recovery beamformer weight vector computation module, from all of the EGSVs a plurality (P) of receive weight vectors, and generating a plurality (P) of recovered signals in a plurality (P) of recovery beamformers, each coupled to receive one of the P receive weight vectors and the conditioned receiver signals.

42. A method as defined in claim 33, wherein:

the system functions to separate a plurality (P) of received cochannel signals;

the step of generating EGSVs includes generating a plurality (P) of EGSVs associated with P signal sources; and

the step of performing supplemental computation includes recovering P separate received signals from the P generalized steering vectors and the conditioned receiver signals.

43. A method as defined in claim 42, wherein:

the step of performing supplemental computation further includes generating from of the EGSVs, in a recovery beamformer weight vector computation module, a plurality (P) of receive weight vectors, and generating a plurality (P) of recovered signals in a plurality (P) of recovery beamformers, each coupled to receive one of the P receive weight vectors and the conditioned receiver signals.

44. A method as defined in claim 34, wherein:

the system functions to separate two received cochannel signals;

the step of generating EGSVs generates two EGSVs from the cross-cumulant matrix data; and

the step of performing supplemental computation includes recovering two separate received signals from the two generalized steering vectors and the conditioned receiver signals.

45. A method as defined in claim 44, wherein:

the step of performing supplemental computation includes generating from both of the EGSVs two weight vectors, and generating two recovered signals in two recovery beamformers, each coupled to receive one of the weight vectors and the conditioned receiver signals.

46. A method as defined in one of claims 30-39, wherein:

the method functions to derive a direction of arrival (DOA) of a received signal;

the step of performing supplemental computation module storing sensor array manifold data in a memory, deriving the DOA of a received signal from its associated generalized steering vector and the stored sensor array manifold data.

47. A method as defined in claim 46, wherein:

the sensor array manifold data includes a table associating multiple DOA values with corresponding steering vectors; and

the step of deriving the DOA includes performing a reverse table lookup function to obtain an approximated DOA value from a steering vector supplied by the generalized steering vector generator.

48. A method as defined in claim 47, wherein:

the step of deriving the DOA further includes interpolating between two DOA values to obtain a more accurate result.

49. A method as defined in one of claims 30-39, wherein the step of performing supplemental computation includes:

generating received signal beamformer weights from the conditioned receiver signals and the EGSV;

recovering the received signal therefrom;

generating transmit signal beamformer weights from the received signal beamformer weights; and

transmitting signals containing information in a direction determined by the transmit signal beamformer weights.

50. A method as defined in claim 41 or 43, wherein the step of performing supplemental computation further includes:

generating P transmit beamforming weight vectors from receive weight vectors generated by the recovery beamformer weight vector computation module; and

P linear combiners, for combining each of the P information signals to be transmitted with its associated transmit weight vector, to obtain a weighted transmit beam for each of the information signals to be transmitted, and then combining corresponding components of the weighted transmit beams, for coupling to a transmit array.

51. A method as defined in one of claims 40-45, wherein:

the method further comprises receiving signals from a plurality of waveguide sensors positioned to detect signals transmitted onto a waveguide in different modes, wherein the modes are subject to scrambling in the waveguide; and

the step of performing supplemental computation includes separating and recovering the signals, while mitigating the effect of mode mixing in the waveguide.

52. A method as defined in claim 51, wherein the waveguide is a microwave waveguide.

53. A method as defined in claim 51, wherein the waveguide is an optical fiber.

54. A method as defined in claim 51, wherein the waveguide is a coaxial cable.

55. A method as defined in claim 51, wherein the waveguide includes at least one twisted pair of conductors.

56. A method as defined in one of claims 40-45, wherein:

the signals received by the signal receiving system are in-phase and quadrature components of two-dimensional communication signal, which has been subject to phase rotation during propagation; and

the step of recovering signals automatically generating in-phase and quadrature components of a two-dimensional communication signal that has been corrected for phase rotation, wherein the method functions as a complex phase equalizer.

57. A method as defined in one of claims 30-34, wherein:

the received signals have been subject to distortion by analog processing and analog-to-digital conversion in a radio receiving system; and

the output quantities include a recovered signal having significantly less distortion than the received signals, whereby the receiving system has improved dynamic range as a result of the use of the signal processing method.

58. A method as defined in one of claims 40-45, wherein:

the received signals include signals from a relatively weak desired source and much stronger signals from at least one interfering source;

wherein the step of recovering the signals includes recovering a signal from the relatively weak desired source, free of interference, and discarding signals from the stronger interfering source.

59. A method for recovery and separation of multiple cochannel signals of any modulation or type received at an array of sensors, the method comprising the steps of:

receiving a plurality of cochannel signals from separate signal sources at an array of sensors;

preprocessing the received signals to provide preprocessed signals;

coupling the preprocessed signals to a plurality of signal extraction ports, each of which is in one of two states referred to as an active state and an inactive state;

in association with each signal extraction port in the active state, generating an estimated steering vector and a recovered signal corresponding to one of the signal sources, without regard for manifold data of the sensor array;

orthogonalizing the estimated steering vectors to ensure that each signal extraction port generates a recovered signal for a separate signal source; and

controlling the steps of orthogonalizing and generating recovered signals to ensure an orderly association of signal sources with signal extraction ports.

60. A method as defined in claim 59, wherein the step of preprocessing includes:

estimating the number of signal sources of which the signals are being received at the sensor array; and

transforming the received signals, which have a dimensionality based on the number sensors, to preprocessed signals, which have a dimensionality based on the estimated number of signal sources.

61. A method as defined in claim 60, wherein the step of preprocessing further includes:

performing an eigendecomposition of the received signals, to produce signal subspace and noise subspace eigenvalues and eigenvectors, and noise subspace eigenvalues and eigenvectors; and

computing a transformation matrix from eigenvalues and eigenvectors, for use in the transforming step.

62. A method as defined in claim 60, wherein the step of generating an estimated recovered signal in each active signal recovery port includes:

computing an even-order cumulant vector of fourth or higher order from the preprocessed signals and auxiliary signals related to the estimated recovered a signals; and

using the even-order cumulant vector to compute an estimated recovered signal, in an iterative process that rapidly converges on a solution for the recovered signal.

63. A method as defined in claim 61, wherein the step of generating an estimated recovered signal includes:

initially selecting a random vector to serve as an initial steering vector a_k ;

computing a weight value v_k from the steering vector a_k, by using the transformation matrix computed in the preprocessing step, wherein the weight vector has a dimensionality based on the estimated number of sources P_e and the weight value has a dimensionality based on the number sensors in the array;

computing an estimate of the recovered signal g_k (t) from the value v_k and the preprocessed signal y(t);

computing an auxiliary signal u_k (t) from the value v_k and the preprocessed input signal y(t);

computing a fourth-order cumulant vector b_k from the preprocessed signals y(t) and the auxiliary signal v_k ;

orthogonalizing the cumulant vector b_k in the orthogonalizer, to produce an orthogonalized cumulant vector d_k ;

deriving an updated steering vector a_k from the orthogonalized cumulant vector d_k ; and

repeating the steps of computing the value v_k, computing an estimate of the recovered signal, computing an auxiliary signal u_k (t), computing a fourth-order cumulant, orthogonalizing, and deriving an updated steering vector, until the recovered signal converges to an acceptably accurate solution.

64. A method as defined in claim 63, wherein the step of generating a recovered signal further includes:

calculating a capture strength c_k for the each active port.

65. A method as defined in claim 63, and further comprising the step of:

demodulating each of the recovered signals from the signal extraction ports.

66. A method as defined in claim 63, and further comprising the step of:

generating direction finding data, by converting the steering vector from each active signal extraction port to a corresponding angular position in three-dimensional space.

67. A method as defined in claim 63, wherein the step of computing a fourth-order cumulant vector uses as input variables the auxiliary signal v_k, two instances of the complex conjugate of the auxiliary signal u_k *(t) and the preprocessed signals y(t).

68. A method as defined in claim 64, wherein the controlling step includes:

detecting changes in the number of signal sources; and

, if the number of signal sources is changed;

identifying which signal extraction port was processing a lost source; and

identifying which signal extraction port will process a new source.

69. A receiver/transmitter system for receiving cochannel signals simultaneously from multiple remote units and transmitting cochannel signals to the remote units simultaneously, the system comprising:

a signal receiving system, including means for generating from signals received at a receive sensor array a set of conditioned receiver signals;

a plurality of estimated generalized steering vector (EGSV) generators, for computing for each transmitting remote unit an EGSV that results in optimization of a utility function that depends on fourth or higher even-order statistical cumulants derived from the received signals, each EGSV being indicative of a combination of signals received at the sensors from the remote unit;

a recovery beamformer weight vector computation module, for generating from all of the EGSVs a plurality of receive beamforming weight vectors;

a plurality of recovery beamformers, each coupled to receive one of the receive beamforming weight vectors and the conditioned receiver signals, for generating a plurality of recovered signals;

a transmit weight vector computation module, for generating transmit beamforming weight vectors from the receive beamforming weight vectors generated by the recovery beamformer weight vector computation module; and

a plurality of linear combiners, for combining each information signal to be transmitted with an associated transmit weight vector, to obtain a weighted transmit beam for each of the information signals to be transmitted, and then combining corresponding components of the weighted transmit beams, for coupling to a transmit array.

70. A receiver/transmitter system as defined in claim 69, wherein:

signals received from at least one particular remote unit are propagated over multiple paths to the receive sensor array;

the ESGV associated with the signal received over multiple paths is representative of a composite of all coherent components of the signal arriving over the multiple paths, wherein multipath components of the same signal are automatically and dynamically combined but non-coherent cochannel signals are separated; and

the transmit weight vector used to transmit signals back to the particular remote unit results in transmit array directivity pattern that achieves transmission over generally the same multiple paths traversed by the received signals, using a weighted combination of the multiple paths;

whereby the receiver/transmitter system achieves diversity gain by virtue of its combination of multipath components.

71. A receiver/transmitter system as defined in claim 69, wherein the transmit weight vector computation module includes:

means for selecting from a plurality of fixed pre-formed transmit weight vectors, based on the receive beamforming weight vectors.

72. A receiver/transmitter system as defined in claim 69, wherein the transmit weight vector computation module includes:

means for generating transmit weight vectors adaptively to reflect as accurately as possible the characteristics of the receive beamforming weight vectors, whereby signals are transmitted to the respective remote units with minimum interference because a transmit weight vector associated with a particular remote unit is selected to be practically orthogonal to all of the estimated generalized steering vectors associated with transmissions to and from the other remote units.

73. A receiver/transmitter system as defined in claim 72, wherein:

the system includes receive and transmit sensor arrays that are identical in shape but are scaled in dimension in the same ratio as transmit and receive frequencies, respectively.

74. A method for using a receiver/transmitter system for receiving cochannel signals simultaneously from multiple remote units and transmitting cochannel signals to the remote units simultaneously, the method comprising the steps of:

receiving signals from a receive sensor array;

generating from the received signals a set of conditioned receiver signals;

computing for each transmitting remote unit an estimated generalized steering vector that results in optimization of a utility function that depends on fourth or higher even-order statistical cumulants derived from the received signals, the estimated generalized steering vector being indicative of a combination of signals received at the sensors from the remote unit;

generating from all of the generalized steering vectors a plurality of receive beamforming weight vectors;

generating from the receive beamforming weight vectors and the conditioned receiver signals a plurality of recovered signals corresponding to the signals received from the respective remote units;

generating transmit beamforming weight vectors from the receive beamforming weight vectors;

combining each information signal to be transmitted with an associated transmit weight vector, to obtain a weighted transmit beam for each of the information signals to be transmitted; and

combining corresponding components of the weighted transmit beams, for coupling to a transmit array.

75. A method as defined in claim 74, wherein:

signals received from at least one particular remote unit are propagated over multiple paths to the receive sensor array;

the estimated generalized steering vector associated with the signal received over multiple paths is representative of a composite of all coherent components of the signal arriving over the multiple paths, wherein multipath components of the same signal are automatically and dynamically combined but non-coherent cochannel signals are separated; and

the transmit weight vector used to transmit signals back to the particular remote unit results in transmit array directivity pattern that achieves transmission over generally the same multiple paths traversed by the received signals, using a weighted combination of the multiple paths;

whereby the receiver/transmitter system achieves diversity gain by virtue of its combination of multipath components.

76. A method defined in claim 74, wherein the step of generating transmit beamforming weight vectors includes:

selecting from a plurality of fixed pre-formed transmit weight vectors, based on the receive beamforming weight vectors.

77. A method as defined in claim 74, wherein the step of generating transmit beamforming weight vectors includes:

generating transmit weight vectors adaptively to reflect as accurately as possible the characteristics of the receive beamforming weight vectors, whereby signals are transmitted to the respective remote units with minimum interference because a transmit weight vector associated with a particular remote unit is selected to be practically orthogonal to all of the generalized steering vectors associated with transmissions to and from the other remote units.

78. A method as defined in claim 77, and further comprising:

selecting receive and transmit sensor arrays that are identical in shape but are scaled in dimension in the same ratio as transmit and receive frequencies.

79. A two-way communication system using cochannel signals and diversity path multiple access (DPMA) for transmission in both directions, the system comprising:

at least one receiver/transmitter base station for communicating with a plurality of mobile devices having omnidirectional antennas for transmitting uplink signals at an assigned frequency and receiving downlink signals at another assigned frequency, wherein the receiver/transmitter base station includes

a receive antenna array,

a plurality of estimated generalized steering vector (EGSV) generators, for computing for each mobile device an EGSV that results in optimization of a utility function that depends on fourth or higher even-order statistical cumulants derived from the received signals, the EGSV being indicative of a combination of uplink signals received at the receive antenna array from the mobile device over possible multiple paths,

receiver processing means for generating from the EGSVs a recovered signal corresponding to each uplink signal from a mobile device, and a receive beamforming weight vector corresponding to the uplink signal,

a transmitter, including means for generating from each receive beamforming weight vector a corresponding transmit beamforming weight vector, and a modulator for modulating a downlink transmission signal with a desired information signal,

a transmit antenna array coupled to the transmitter and having a similar geometrical shape as the receive antenna array, wherein downlink transmission signals intended for a particular mobile device are propagated along generally the same multiple paths as the received uplink signals from the same mobile device;

wherein coherent uplink signals received over multiple paths from the same mobile device are automatically combined, providing a gain enhancement effect that allows weaker transmissions to be detected, and downlink signals transmitted over the same multiple paths also benefit from the gain enhancement effect and provide a stronger downlink signal to the mobile device.

80. A two-way communication system as defined in claim 79, wherein the receiver processing means includes:

a recovery beamformer weight vector computation module, for generating from all of the EGSVs the plurality of receive beamforming weight vectors; and

a plurality of recovery beamformers, each coupled to receive one of the receive beamforming weight vectors and conditioned receiver signals, for generating the plurality of recovered signals.

81. A two-way communication system as defined in claim 80, wherein the transmitter includes:

a transmit weight vector computation module, for generating transmit beamforming weight vectors from the receive beamforming weight vectors generated by the recovery beamformer weight vector computation module; and

a plurality of linear combiners, for combining each information signal to be transmitted with an associated transmit weight vector, to obtain a weighted transmit beam for each of the information signals to be transmitted, and then combining corresponding components of the weighted transmit beams, for coupling to the transmit antenna array.

82. A two-way communication system as defined in claim 79, wherein:

the uplink and downlink transmission frequencies are offset from each other to avoid interference between mobile devices; and

the transmit antenna array has dimensions scaled with respect to those of the receive antenna array in the inverse ratio of the uplink and downlink transmission frequencies.

83. A two-way communication system as defined in claim 81, wherein the transmit weight vector computation module includes:

means for selecting from a plurality of fixed pre-formed transmit weight vectors, based on the receive beamforming weight vectors.

84. A two-way communication system as defined in claim 81, wherein the transmit weight vector computation module includes:

means for generating transmit weight vectors adaptively to reflect as accurately as possible the characteristics of the receive beamforming weight vectors, whereby downlink signals are transmitted to the respective mobile devices with minimum interference because a transmit weight vector associated with a particular mobile device is selected to be practically orthogonal to all of the generalized steering vectors associated with transmissions to and from the other mobile devices.

85. A method of radio direction finding (DF) using a subarray of calibrated antennas, the method comprising the steps of:

receiving signals from multiple sources, at an antenna array of which only a small number of antenna elements are calibrated;

separating the signals using a cumulant recovery (CURE) system to generate the separated signals and estimates of their generalized steering vectors; and

processing the estimated generalized steering vectors and signals from the calibrated antenna elements, to obtain accurate signal directions for the multiple sources.