Inter-symbol interference cancellation for wireless multiple access

• US 20070110196A1
• Filed: 06/29/2006
• Published: 05/17/2007
• Est. Priority Date: 11/15/2005
• Status: Active Grant

• Current Assignee: III Holdings 1, LLC
• Original Assignee: Rambus Inc.
• Inventors: McCloud, Michael, Guess, Tommy
• Granted Patent: US 7,733,941 B2
• Time in Patent Office: Days
• US Class Current: 375/348
• CPC Class Codes: H04B 2201/70702 Intercell-related aspects
• CPC Class Codes: H04L 2025/03375 Passband transmission
• CPC Class Codes: H04L 2025/0349 as a feedback filter
• CPC Class Codes: H04L 2025/03605 Block algorithms
• CPC Class Codes: H04L 25/03057 with a recursive structure H04L25/03031 takes precedence

Abstract

An interference-canceling receiver is configured for cancelling inter-symbol interference due to both inter- and intra-channel interference in coded, multiple-access, spread spectrum transmissions that propagate through frequency-selective communication channels. The receiver mitigates the effects of post-cursor inter-symbol interference using feedback of previously estimated symbols, and mitigates pre-cursor inter-symbol interference using an interference canceller, such as an iterative interference canceller.

Claims

1. An interference canceller for cancelling inter-symbol interference (ISI), intra-cell interference, and inter-cell interference in a multiple-access communication signal, comprising:

a post-cursor ISI synthesis module configured for synthesizing an estimated post-cursor ISI signal from post-cursor symbol estimates;

a post-cursor ISI cancellation module and a resolving module configured for processing the estimated post-cursor ISI signal and the received signal to produce a first interference-canceled signal;

an interference cancellation module configured for cancelling pre-cursor inter-symbol interference, intra-cell interference, and inter-cell interference from the first interference-cancelled signal for producing interference-cancelled symbol estimates and the post-cursor symbol estimates; and

a feedback loop configured for coupling the post-cursor symbol estimates to the post-cursor ISI synthesis module.

2. The interference canceller recited in claim 1, whereby the post-cursor ISI synthesis module is further configured to receive as input a plurality of symbol estimates in a predetermined number of post-cursor symbol intervals, the post-cursor ISI synthesis module comprising:

a weighting module configured for weighting the plurality of symbol estimates according to symbol-estimate merits to produce weighted symbol estimates;

a code-waveform modulator configured for modulating the weighted symbol estimates onto corresponding code waveforms for producing modulated code waveforms and summing the modulated code waveforms to generate an estimated transmit signal; and

a channel emulator configured for imparting multipath delays and gains to the estimated transmit signal.

3. The interference canceller recited in claim 2, whereby the interference cancellation module is configured to subtract a synthesized post-cursor inter-symbol interference signal from the received signal to cancel post-cursor inter-symbol interference in the received signal.

4. The interference canceller recited in claim 2, whereby each of the symbol-estimate merits comprises a function of average ratio of signal power to interference-plus-noise power of a corresponding symbol estimate.

5. The interference canceller recited in claim 4, whereby the average ratio of signal power to interference-plus-noise power is estimated from a pilot signal.

6. The interference canceller recited in claim 4, whereby the average ratio of signal power to interference-plus-noise power is estimated using techniques of statistical signal processing.

7. The interference canceller recited in claim 2, whereby the function of the average ratio of signal power to interference-plus-noise power is characterized by

$\pi$

= max ⁢

{ C , 1 1 + 1 / SINR } ,

where π

is a symbol weight, max { } is a function for selecting a maximum value from a set of quantities within brackets { }, SINR denotes an average ratio of signal power to interference-plus-noise power (SINR) of the symbol, and C is a non-negative real constant for enforcing a minimum symbol weight.

8. The interference canceller recited in claim 2, whereby the weighting module is further configured to assign all symbols a common weight.

9. The interference canceller recited in claim 8, whereby the weighting module is further configured to choose the weight to minimize mean square-error between the received signal and a synthesized signal.

10. The interference canceller recited in claim 9, whereby the weighting module is further configured to choose the weight according to

π

_n=(θ

_n^post)^Hθ

_n/∥

θ

_n^post∥

²,

where π

_n is the weight, θ

_n is a vector having multi-symbol elements produced from Rake processing with combining, θ

_n^post is a vector having multi-symbol outputs after despreading, H denotes a complex-conjugate transpose, and ∥

•

∥

² denotes a sum of an argument'"'"'s elements'"'"' magnitude squares.

11. The interference canceller recited in claim 1, whereby the interference cancellation module is further configured to process an input of symbol estimates of all sources for an on-cursor symbol interval for producing an output comprising updated, interference-cancelled symbol estimates of all sources for the on-cursor symbol interval.

12. The interference canceller recited in claim 11, whereby the input is further configured to include temporary symbol estimates of all sources for a predetermined number of pre-cursor symbol intervals.

13. The interference canceller recited in claim 11, whereby the output is further configured to include temporary symbol estimates of all sources for a predetermined number of pre-cursor symbol intervals.

14. The interference canceller recited in claim 1, whereby the interference cancellation module comprises,

a symbol-estimation module configured for generating initial symbol estimates;

a sequence of interference cancellation units, the sequence configured to receive the initial symbol estimates, each of the sequence of interference cancellation units configured to weight input symbol estimates with soft weighting, apply stabilizing step sizes to a difference signal, and perform mixed decisions to produce updated, interference-cancelled symbol estimates; and

a partitioning module configured to output final updated, interference-cancelled symbol estimates for an on-cursor symbol interval.

15. The interference canceller recited in claim 14, whereby the partitioning module is further configured to output temporary symbol estimates corresponding to a predetermined number of post-cursor symbol intervals.

16. The interference canceller recited in claim 14, whereby the symbol-estimation module is further configured to employ temporary symbol estimates for a predetermined number of post-cursor symbol intervals for generating the initial symbol estimates.

17. The interference canceller recited in claim 1, whereby the resolving module is configured for resolving the received signal and the estimated post-cursor ISI signal for producing a resolved received signal and a resolved estimated post-cursor ISI signal, and the post-cursor ISI cancellation module is configured to subtract the resolved estimated post-cursor ISI signal from the resolved received signal to produce the first interference-canceled signal.

18. The interference canceller recited in claim 17, whereby a processor configured to perform a matrix multiplication comprises the post-cursor ISI synthesis module and the resolving module, the processor being configured to perform the matrix multiplication on the post-cursor symbol estimates to produce the resolved estimated post-cursor ISI signal.

19. The interference canceller recited in claim 1, whereby the post-cursor ISI cancellation module is further configured to subtract the estimated post-cursor ISI signal from the received signal for producing an interference-cancelled signal, and the resolving module is further configured for resolving the interference-canceled signal onto a signal basis for a plurality of symbol sources to produce the first interference-canceled signal.

20. The interference canceller recited in claim 1, whereby the interference cancellation module is further configured for performing iterative pre-cursor ISI cancellation at the symbol level by employing a one-step matrix update equation, comprising an explicit matrix representation of a received-signal correlation matrix over multiple symbol intervals, an implementation matrix over multiple symbol intervals, a soft-weighting matrix, a scalar stabilizing step size, and a mixed decision symbol estimator.

21. The interference canceller recited in claim 20, whereby the matrix update equation is expressed by

{circumflex over (b)}_n^[i+1]=Ψ

(μ

_n^[i](q_n−

R_nn^[i]{circumflex over (b)}_n^[i])+F_nn^[i]{circumflex over (b)}_n^[i]),

wherein {circumflex over (b)}_n^[i] is a column vector containing input symbol decisions of sources of pre-cursor symbols for symbol intervals after an it iteration of interference cancellation, q_n is a vector of a resolved received signal over a plurality of post-cursor symbol intervals, μ

_n^[i] is a scalar stabilizing step, _n^[i] is a weighting matrix used for scaling the input symbol decisions, R_n is the received-signal correlation matrix over the plurality of post-cursor symbol intervals, F_n is an implementation matrix, Ψ

is a mixed-decision operator that returns mixed symbol decisions to produce the updated symbol decisions {circumflex over (b)}_n^[i+1].

22. A method whereby inter-symbol interference (ISI), intra-cell interference, and inter-cell interference in a multiple-access communication signal are cancelled, the method comprising:

providing for synthesizing an estimated post-cursor ISI signal from post-cursor symbol estimates;

providing for processing the estimated post-cursor ISI signal and the received signal to produce a first interference-canceled signal in which post-cursor ISI has been removed from the received signal;

providing for performing interference cancellation for cancelling pre-cursor inter-symbol interference, intra-cell interference, and inter-cell interference from the first interference-cancelled signal for producing interference-cancelled symbol estimates and the post-cursor symbol estimates; and

providing for feeding back the post-cursor symbol estimates to the step of providing for synthesizing an estimated post-cursor ISI signal.

23. The method recited in claim 22, whereby providing for synthesizing is further configured to receive as input a plurality of symbol estimates in a predetermined number of post-cursor symbol intervals, providing for synthesizing comprising:

providing for weighting the plurality of symbol estimates according to symbol-estimate merits to produce weighted symbol estimates;

providing for modulating the weighted symbol estimates onto corresponding code waveforms for producing modulated code waveforms and summing the modulated code waveforms to generate an estimated transmit signal; and

providing for performing channel emulation to impart multipath delays and gains to the estimated transmit signal.

24. The method recited in claim 23, whereby providing for performing interference cancellation further comprises subtracting a synthesized post-cursor inter-symbol interference signal from the received signal to cancel post-cursor inter-symbol interference in the received signal.

25. The method recited in claim 23, whereby each of the symbol-estimate merits comprises a function of average ratio of signal power to interference-plus-noise power of a corresponding symbol estimate.

26. The method recited in claim 25, whereby providing for performing interference cancellation further comprises estimating the average ratio of signal power to interference-plus-noise power from a pilot signal.

27. The method recited in claim 25, whereby providing for performing interference cancellation further comprises using techniques of statistical signal processing for estimating the average ratio of signal power to interference-plus-noise power.

28. The method recited in claim 23, whereby the function of the average ratio of signal power to interference-plus-noise power is characterized by

$\pi$

= max ⁢

{ C , 1 1 + 1 / SINR } ,

where π

is a symbol weight, max { } is a function for selecting a maximum value from a set of quantities within brackets { }, SINR denotes an average ratio of signal power to interference-plus-noise power (SINR) of the symbol, and C is a non-negative real constant for enforcing a minimum symbol weight.

29. The method recited in claim 23, whereby providing for weighting further comprises assigning all symbols a common weight.

30. The method recited in claim 29, whereby providing for weighting further comprises choosing the weight to minimize mean square-error between the received signal and at least one synthesized signal.

31. The method recited in claim 30, whereby providing for weighting further comprises choosing the weight as

π

_n=(θ

_n^post)^Hθ

_n/∥

θ

_n^post∥

²,

where π

_n is the weight, θ

_n is a vector having multi-symbol elements produced from Rake processing with combining, θ

_n^post is a vector having multi-symbol outputs after despreading, H denotes complex-conjugate transpose, and ∥

•

∥

² denotes a sum of an argument'"'"'s elements'"'"' magnitude squares.

32. The method recited in claim 22, whereby providing for performing interference cancellation further comprises processing an input of symbol estimates of all sources for an on-cursor symbol interval to produce an output comprising updated interference-cancelled symbol estimates of all sources for the on-cursor symbol interval.

33. The method recited in claim 32, whereby the input is further configured to include temporary symbol estimates of all sources for a predetermined number of pre-cursor symbol intervals.

34. The method recited in claim 32, whereby the output is further configured to include temporary symbol estimates of all sources for a predetermined number of pre-cursor symbol intervals.

35. The method recited in claim 22, whereby providing for performing interference cancellation comprises,

providing for generating initial symbol estimates;

providing for a sequence of interference cancellation steps, the sequence configured to receive the initial symbol estimates, each of the sequence of interference cancellation steps configured to weight input symbol estimates with soft weighting, apply stabilizing step sizes to a difference signal, and perform mixed decisions to produce updated, interference-cancelled symbol estimates; and

providing for outputting final updated, interference-cancelled symbol estimates for an on-cursor symbol interval.

36. The method recited in claim 35, whereby providing for outputting further comprises outputting temporary symbol estimates corresponding to a predetermined number of post-cursor symbol intervals.

37. The method recited in claim 35, whereby providing for generating initial symbol estimates is further configured to employ temporary symbol estimates for a predetermined number of post-cursor symbol intervals for generating the initial symbol estimates.

38. The method recited in claim 22, whereby providing for processing the estimated post-cursor ISI signal and the received signal comprises providing for resolving the received signal and the estimated post-cursor ISI signal to produce a resolved received signal and a resolved estimated post-cursor ISI signal, and providing for subtracting the resolved estimated post-cursor ISI signal from the resolved received signal to produce the first interference-canceled signal.

39. The method recited in claim 38, whereby performing a matrix multiplication on the post-cursor symbol estimates to produce the resolved estimated post-cursor ISI signal further comprises providing for synthesizing and providing for resolving.

40. The method recited in claim 22, whereby providing for processing the estimated post-cursor ISI signal and the received signal further comprises subtracting the estimated post-cursor ISI signal from the received signal to produce an interference-cancelled signal, and providing for resolving further comprises resolving the interference-canceled signal onto a signal basis for a plurality of symbol sources to produce the first interference-canceled signal.

41. The method recited in claim 22, whereby providing for performing interference cancellation is further configured for performing iterative pre-cursor ISI cancellation at the symbol level by employing a one-step matrix update equation, comprising an explicit matrix representation of a received-signal correlation matrix over multiple symbol intervals, an implementation matrix over multiple symbol intervals, a soft-weighting matrix, a scalar stabilizing step size, and a mixed decision symbol estimator.

42. The method recited in claim 41, whereby the matrix update equation is expressed by

{circumflex over (b)}_n^[i+1]=Ψ

(μ

_n^[i](q_n−

R_nn^[i]{circumflex over (b)}_n^[i])+F_nn^[i]{circumflex over (b)}_n^[i]),

where {circumflex over (b)}_n^[i] is a column vector containing input symbol decisions of sources of pre-cursor symbols for symbol intervals after an i^th iteration of interference cancellation, q_n is a vector of a resolved received signal over a plurality of post-cursor symbol intervals, μ

_n^{[i] is a scalar stabilizing step,} _n^[i] is a weighting matrix used for scaling the input symbol decisions, R_n is the received-signal correlation matrix over the plurality of post-cursor symbol intervals, F_n is an implementation matrix, Ψ

is a mixed-decision operator that returns mixed symbol decisions to produce the updated symbol decisions {circumflex over (b)}_n^[i+1].

43. A system whereby inter-symbol interference (ISI), intra-cell interference, and inter-cell interference in a multiple-access communication signal are cancelled, the system comprising:

a post-cursor ISI synthesis means for synthesizing an estimated post-cursor ISI signal from post-cursor symbol estimates;

a post-cursor ISI cancellation means and a resolving means for processing the estimated post-cursor ISI signal and the received signal to produce a first interference-canceled signal;

an interference cancellation means for cancelling pre-cursor inter-symbol interference, intra-cell interference, and inter-cell interference from the first interference-cancelled signal for producing interference-cancelled symbol estimates and the post-cursor symbol estimates; and

a feedback means for coupling the post-cursor symbol estimates to the post-cursor ISI synthesis means.

44. The system recited in claim 43, whereby the post-cursor ISI synthesis means is configured to receive as input a plurality of symbol estimates in a predetermined number of post-cursor symbol intervals, the post-cursor ISI synthesis means comprising:

a weighting means for weighting the plurality of symbol estimates according to symbol-estimate merits to produce weighted symbol estimates;

a code-waveform modulation means for modulating the weighted symbol estimates onto corresponding code waveforms for producing modulated code waveforms and summing the modulated code waveforms to generate an estimated transmit signal; and

a channel-emulation means for imparting multipath delays and gains to the estimated transmit signal.

45. The system recited in claim 44, whereby the interference cancellation means is further configured to subtract a synthesized post-cursor inter-symbol interference signal from the received signal to cancel post-cursor inter-symbol interference in the received signal.

46. The system recited in claim 44, whereby each of the symbol-estimate merits comprises a function of average ratio of signal power to interference-plus-noise power of a corresponding symbol estimate.

47. The system recited in claim 46, whereby the average ratio of signal power to interference-plus-noise power is estimated from a pilot signal.

48. The system recited in claim 46, whereby the average ratio of signal power to interference-plus-noise power is estimated using techniques of statistical signal processing.

49. The system recited in claim 44, whereby the function of the average ratio of signal power to interference-plus-noise power is characterized by

$\pi$

= max ⁢

{ C , 1 1 + 1 / SINR } ,

where π

is a symbol weight, max { } is a function for selecting a maximum value from a set of quantities within brackets { }, SINR denotes an average ratio of signal power to interference-plus-noise power (SINR) of the symbol, and C is a non-negative real constant for enforcing a minimum symbol weight.

50. The system recited in claim 44, whereby the weighting means is further configured to assign all symbols a common weight.

51. The system recited in claim 50, whereby the weighting means is further configured to choose the weight to minimize mean square-error between the received signal and at least one synthesized signal.

52. The system recited in claim 51, whereby the weighting means is further configured to choose the weight according to

π

_n=(θ

_n^post)^Hθ

_n/∥

θ

_n^post∥

²,

where π

_n is the weight, θ

_n is a vector having multi-symbol elements produced from Rake processing with combining, θ

_n^post is a vector having multi-symbol outputs after despreading, H denotes complex-conjugate transpose, and ∥

•

∥

² denotes a sum of an argument'"'"'s elements'"'"' magnitude squares.

53. The system recited in claim 43, whereby the interference cancellation means is further configured to process an input of symbol estimates of all sources for an on-cursor symbol interval for producing an output comprising updated, interference-cancelled symbol estimates of all sources for the on-cursor symbol interval.

54. The system recited in claim 53, whereby the input is further configured to include temporary symbol estimates of all sources for a predetermined number of pre-cursor symbol intervals.

55. The system recited in claim 53, whereby the output is further configured to include temporary symbol estimates of all sources for a predetermined number of pre-cursor symbol intervals.

56. The system recited in claim 43, whereby the interference cancellation means comprises,

a symbol-estimation means for generating initial symbol estimates;

a sequential interference cancellation means for receiving the initial symbol estimates, the sequential interference cancellation means configured to weight input symbol estimates with soft weighting, apply stabilizing step sizes to a difference signal, and perform mixed decisions to produce updated, interference-cancelled symbol estimates; and

a partitioning means configured for outputting final updated, interference-cancelled symbol estimates for an on-cursor symbol interval.

57. The system recited in claim 56, whereby the partitioning means is further configured to output temporary symbol estimates corresponding to a predetermined number of post-cursor symbol intervals.

58. The system recited in claim 56, whereby the symbol-estimation means is further configured to employ temporary symbol estimates for a predetermined number of post-cursor symbol intervals for generating the initial symbol estimates.

59. The system recited in claim 43, whereby the resolving means is configured for resolving the received signal and the estimated post-cursor ISI signal for producing a resolved received signal and a resolved estimated post-cursor ISI signal, and the post-cursor ISI cancellation means is configured to subtract the resolved estimated post-cursor ISI signal from the resolved received signal to produce the first interference-canceled signal.

60. The system recited in claim 59, whereby a processing means configured to perform a matrix multiplication comprises the post-cursor ISI synthesis means and the resolving means, the processing means configured to perform the matrix multiplication on the post-cursor symbol estimates to produce the resolved estimated post-cursor ISI signal.

61. The system recited in claim 43, whereby the post-cursor ISI cancellation means is configured to subtract the estimated post-cursor ISI signal from the received signal for producing an interference-cancelled signal, and the resolving means is configured for resolving the interference-canceled signal onto a signal basis for a plurality of symbol sources to produce the first interference-canceled signal.

62. The system recited in claim 43, whereby the interference cancellation means is further configured for performing iterative pre-cursor ISI cancellation at symbol level by employing a one-step matrix update equation, comprising an explicit matrix representation of a received-signal correlation matrix over multiple symbol intervals, an implementation matrix over multiple symbol intervals, a soft-weighting matrix, a scalar stabilizing step size, and a mixed decision symbol estimator.

63. The system recited in claim 62, whereby the matrix update equation is expressed by

{circumflex over (b)}_n^[i+1]=Ψ

(μ

_n^[i](q_n−

R_nn^[it9)+F_nn^[i]{circumflex over (b)}_n^[i]),

where {circumflex over (b)}_n^[it] is a column vector containing input symbol decisions of sources of pre-cursor symbols for symbol intervals after an i^th iteration of interference cancellation, q_n is a vector of a resolved received signal over a plurality of post-cursor symbol intervals, μ

_n^[i] is a scalar stabilizing step, _n^[i] is a weighting matrix used for scaling the input symbol decisions, R_n is the received-signal correlation matrix over the plurality of post-cursor symbol intervals, F_n is an implementation matrix, Ψ

is a mixed-decision operator that returns mixed symbol decisions to produce the updated symbol decisions {circumflex over (b)}_n^[i+1].