Cancelling non-linear power amplifier induced distortion from a received signal by moving incorrectly estimated constellation points

• US 8,355,466 B2
• Filed: 09/25/2009
• Issued: 01/15/2013
• Est. Priority Date: 09/25/2009
• Status: Active Grant

• Current Assignee: General Dynamics C4 Systems Incorporated (General Dynamics Corporation)
• Original Assignee: General Dynamics C4 Systems Incorporated (General Dynamics Corporation)
• Inventors: Kleider, John, Smith, Anthony
• Primary Examiner(s): Ahn, Sam K
• Assistant Examiner(s): Wong, Linda
• Application Number:
  US12/567,505
• Publication Number:
  US 20110074506A1
• Time in Patent Office: 1,208 Days
• Field of Search: 375/219, 375/222, 375/261, 375/268, 375/273, 375/279, 375/280, 375/281, 375/283, 375/295, 375/297, 375/296, 375/298, 375/302, 375/308, 375/316, 375/318, 375/324, 375/329, 375/340, 375/345, 375/240, 37524026-24029, 375/259, 375/260, 375/284, 375/285, 375/278, 375/317, 375/346, 375/347, 375/358, 4551142-1143, 455/91, 455/113, 455/115.1, 455/115.3, 455/127.2, 455/127.1, 455/127.5, 455/130, 455/136, 455/139, 455/182.2, 455/182.3, 455/183.1, 455/183.2, 455/192.3, 455/194.2, 455/219, 455/232.1, 455/234.1, 455/239.1, 455/240.1, 455/247.1, 455/250.1, 455/192.2, 455/39, 455/500, 455/501, 455/63.1, 455/103
• US Class Current: 375/316
• CPC Class Codes: H03F 1/3294 Acting on the real and imaginary components of the input signal
• CPC Class Codes: H04L 27/3863 Compensation for quadrature error in the received signal

Abstract

A method and system are provided for reducing power amplifier induced distortion. Power amplifier induced distortion is iteratively estimated and cancelled. When the difference between the current estimated power amplifier distortion and the previous estimated power amplifier distortion is less than a convergence threshold, particular M-QAM constellation points that are still in error are determined A M-QAM constellation point correction routine is provided that can move the incorrectly estimated M-QAM constellation points that are in error towards their expected quadrants by generating updated M-QAM constellation points. The remaining estimated non-linear power amplifier induced distortion in the received signal can then be estimated and canceled.

Claims

1. A method for reducing power amplifier induced distortion, comprising:

receiving a received signal corresponding to a channel affected version of a transmitted signal, wherein the transmitted signal comprises transmit data that was distorted by a transmit power amplifier prior to transmission over the channel;

iteratively estimating distortion induced by the transmit power amplifier prior to transmission in the received signal via an estimation routine to generate estimated power amplifier distortion;

iteratively cancelling the estimated power amplifier distortion present in the received signal via a cancellation routine executed after each iteration of the estimation routine;

comparing, after each iteration of the cancellation routine, a current estimated power amplifier distortion to a previous estimated power amplifier distortion to determine whether difference between the current estimated power amplifier distortion and the previous estimated power amplifier distortion is less than a convergence threshold;

temporarily stopping the cancellation routine when the difference between the current estimated power amplifier distortion and the previous estimated power amplifier distortion is less than the convergence threshold;

determining M-QAM constellation points corresponding to estimates of the transmit data, and determining particular ones of the M-QAM constellation points that are incorrectly estimated and in error; and

executing, for each of the M-QAM constellation points, a M-QAM constellation point correction routine designed to move the incorrectly estimated M-QAM constellation points that are in error towards expected quadrants of those incorrectly estimated M-QAM constellation points.

2. A method according to claim 1, wherein the step of executing, for each of the M-QAM constellation points, a M-QAM constellation point correction routine, comprises the steps of:

receiving a frequency domain input that represents estimates of the transmit data as M-QAM constellation points, wherein the frequency domain input comprises correctly estimated M-QAM constellation points and incorrectly estimated M-QAM constellation points that are in error, and separating the frequency domain input into real and imaginary components to generate a real component of each M-QAM constellation point and an imaginary component of each M-QAM constellation point;

generating a border set based on an input specifying a constellation alphabet set of size M, and separating the border set into a real component that represents real boundary points of M-QAM regions and an imaginary component that represents imaginary boundary points of the M-QAM regions;

generating a threshold from a threshold lookup table (TLUT) based on channel characteristics, a signal-to-noise ratio (SNR) measurement, and distortion characteristics;

computing a minimum distance between the real component of each M-QAM constellation point and a nearest boundary in a real direction based on the real boundary points of the M-QAM regions, and a minimum distance between the imaginary component of each M-QAM constellation point and a nearest boundary in an imaginary direction based on the imaginary boundary points of the M-QAM regions; and

determining whether the real components of each M-QAM constellation point are in error, and determining whether the imaginary components of each M-QAM constellation point are in error; and

generating at least one of an updated real component and an updated imaginary component.

3. A method according to claim 2, wherein the step of determining whether the real components of each M-QAM constellation point are in error and determining whether the imaginary components of each M-QAM constellation point are in error, comprises:

determining whether the real components of each M-QAM constellation point are in error by determining whether an absolute value of the minimum distance between the real component of each M-QAM constellation point and the nearest boundary in the real direction is less than the threshold; and

determining whether the imaginary components of each M-QAM constellation point are in error by determining whether an absolute value of the minimum distance between the imaginary component of each M-QAM constellation point and the nearest boundary in the imaginary direction is less than the threshold.

4. A method according to claim 3, wherein the step of generating at least one of an updated real component and an updated imaginary component, comprises:

generating, when the absolute value of the minimum distance is less than the threshold, an updated real component for that incorrectly estimated M-QAM constellation point that is in error; and

generating, when the absolute value of the minimum distance is less than the threshold, an updated imaginary component for that incorrectly estimated M-QAM constellation point that is in error.

5. A method according to claim 4, further comprising the steps of:

multiplying each of the updated imaginary components by an imaginary multiplier; and

generating, for each incorrectly estimated M-QAM constellation point that is in error, an updated M-QAM constellation point based on at least one of the an updated real component and the updated imaginary component.

6. A method according to claim 2, wherein the updated real component for that incorrectly estimated M-QAM constellation point that is in error is:

equal to the difference between the real component and the minimum distance between the real component and the nearest boundary in the real direction such that the real component of that incorrectly estimated M-QAM constellation point that is in error is moved to an adjacent M-QAM region in the real direction.

7. A method according to claim 2, wherein the updated imaginary component for that incorrectly estimated M-QAM constellation point that is in error is:

equal to the difference between the imaginary component and the minimum distance between the imaginary component and the nearest boundary in the imaginary direction such that the imaginary component of that incorrectly estimated M-QAM constellation point that is in error is moved to an adjacent M-QAM region in the imaginary direction.

8. A method according to claim 1, wherein the M-QAM constellation points that are incorrectly estimated and in error correspond to incorrect symbols of the transmit data estimate that are still in error after the cancellation routine ends.

9. A method according to claim 8, wherein the M-QAM constellation points that are incorrectly estimated and in error are located outside an expected quadrant of a correct M-QAM location.

10. A method according to claim 9, wherein the M-QAM constellation points that are incorrectly estimated and in error lie along a boundary with an adjacent quadrant that is adjacent to the expected quadrant.

11. A method according to claim 1, further comprising:

estimating, after executing the M-QAM constellation point correction routine, remaining non-linear power amplifier induced distortion in the received signal via the estimation routine to generate remaining estimated power amplifier distortion; and

cancelling the remaining estimated power amplifier distortion present in the received signal via the cancellation routine.

12. A method according to claim 2, wherein the real boundary points of the M-QAM regions and the imaginary boundary points of the M-QAM regions are sequences containing border crossings in the real and imaginary directions, respectively.

13. A system for reducing distortion in a received signal corresponding to a channel affected version of a transmitted signal that comprises transmit data that was distorted by a transmit power amplifier prior to transmission of the transmitted signal over the channel, the system comprising:

an estimation module designed to iteratively estimate power amplifier induced distortion, that was induced by the transmit power amplifier prior to transmission of the transmitted signal, in the received signal via an estimation routine to generate estimated power amplifier distortion;

a cancellation module designed to iteratively cancel the estimated power amplifier distortion present in the received signal via a cancellation routine executed after each iteration of the estimation routine;

a comparison module designed to compare, after each iteration of the cancellation routine, a current estimated power amplifier distortion to a previous estimated power amplifier distortion to determine whether difference between the current estimated power amplifier distortion and the previous estimated power amplifier distortion is less than a convergence threshold, wherein the cancellation routine is temporarily stopped when the difference between the current estimated power amplifier distortion and the previous estimated power amplifier distortion is less than the convergence threshold;

an error determination module designed to determine M-QAM constellation points corresponding to estimates of the transmit data, and to determine particular ones of the M-QAM constellation points that are incorrectly estimated and in error; and

a correction module designed to execute, for each of the M-QAM constellation points, a M-QAM constellation point correction routine designed to move the incorrectly estimated M-QAM constellation points that are in error towards expected quadrants of those incorrectly estimated M-QAM constellation points.

14. A system according to claim 13, wherein the M-QAM constellation point correction routine is designed to:

receive a frequency domain input that represents estimates of the transmit data as M-QAM constellation points, wherein the frequency domain input comprises correctly estimated M-QAM constellation points and incorrectly estimated M-QAM constellation points that are in error;

separate the frequency domain input into real and imaginary components to generate a real component of each M-QAM constellation point and an imaginary component of each M-QAM constellation point;

generate a border set based on an input specifying a constellation alphabet set of size M;

separate the border set into a real component that represents real boundary points of M-QAM regions and an imaginary component that represents imaginary boundary points of the M-QAM regions, wherein the real boundary points of the M-QAM regions and the imaginary boundary points of the M-QAM regions are sequences containing border crossings in the real and imaginary directions, respectively;

generate a threshold from a threshold lookup table (TLUT) based on channel characteristics, a signal-to-noise ratio (SNR) measurement, and distortion characteristics;

compute a minimum distance between the real component of each M-QAM constellation point and a nearest boundary in a real direction based on the real boundary points of the M-QAM regions, and a minimum distance between the imaginary component of each M-QAM constellation point and a nearest boundary in an imaginary direction based on the imaginary boundary points of the M-QAM regions; and

determine whether the real components of each M-QAM constellation point are in error, and determine whether the imaginary components of each M-QAM constellation point are in error; and

generate, for ones of the real components of each M-QAM constellation point or the imaginary components of each M-QAM constellation point are in error, at least one of an updated real component and an updated imaginary component.

15. A system according to claim 14, wherein the M-QAM constellation point correction routine is designed to:

determine whether the real components of each M-QAM constellation point are in error by determining whether an absolute value of the minimum distance between the real component of each M-QAM constellation point and the nearest boundary in the real direction is less than the threshold; and

determine whether the imaginary components of each M-QAM constellation point are in error by determining whether an absolute value of the minimum distance between the imaginary component of each M-QAM constellation point and the nearest boundary in the imaginary direction is less than the threshold.

16. A system according to claim 15, wherein the M-QAM constellation point correction routine is designed to:

generate, when the absolute value of the minimum distance is less than the threshold, an updated real component for that incorrectly estimated M-QAM constellation point that is in error; and

generate, when the absolute value of the minimum distance is less than the threshold, an updated imaginary component for that incorrectly estimated M-QAM constellation point that is in error.

17. A system according to claim 16, wherein the M-QAM constellation point correction routine is designed to:

multiply each of the updated imaginary components by an imaginary multiplier; and

generate, for each incorrectly estimated M-QAM constellation point that is in error, an updated M-QAM constellation point based on at least one of the updated real component and the updated imaginary component.

18. A system according to claim 14, wherein the updated real component for that incorrectly estimated M-QAM constellation point that is in error is equal to the difference between the real component and the minimum distance between the real component and the nearest boundary in the real direction such that the real component of that incorrectly estimated M-QAM constellation point that is in error is moved to an adjacent M-QAM region in the real direction.

19. A system according to claim 14, wherein the updated imaginary component for that incorrectly estimated M-QAM constellation point that is in error is:

equal to the difference between the imaginary component and the minimum distance between the imaginary component and the nearest boundary in the imaginary direction such that the imaginary component of that incorrectly estimated M-QAM constellation point that is in error is moved to an adjacent M-QAM region in the imaginary direction.

20. A system according to claim 13, wherein the M-QAM constellation points that are incorrectly estimated and in error correspond to incorrect symbols of the transmit data estimate that are still in error after the cancellation routine ends, wherein the M-QAM constellation points that are incorrectly estimated and in error are located outside an expected quadrant of a correct M-QAM location, and wherein the M-QAM constellation points that are incorrectly estimated and in error lie along a boundary with an adjacent quadrant that is adjacent to the expected quadrant.

21. A system according to claim 13, wherein, after executing the M-QAM constellation point correction routine, the estimation module is designed to estimate the remaining non-linear power amplifier induced distortion in the received signal via the estimation routine to generate remaining estimated power amplifier distortion, and wherein the cancellation module is designed to cancel the remaining estimated power amplifier distortion present in the received signal via the cancellation routine.

22. A system for reducing power amplifier distortion in a received signal that is a channel affected version of a transmitted signal that comprises transmit data distorted by a transmit power amplifier prior to transmission, the system comprising:

a frequency domain input that represents estimates of the transmit data as M-QAM constellation points, wherein the frequency domain input comprises correctly estimated M-QAM constellation points and incorrectly estimated M-QAM constellation points that are in error;

a first separation module designed to separate the frequency domain input into real and imaginary components to generate a real component of each M-QAM constellation point and an imaginary component of each M-QAM constellation point;

a border set computation module designed to receive an input specifying a constellation alphabet set of size M, and designed to compute a border set;

a second separation module designed to separate the border set into a real component of the border set that represents real boundary points of the M-QAM regions and an imaginary component of the border set that represents imaginary boundary points of the M-QAM regions;

a threshold lookup table (TLUT) designed to receive inputs comprising channel characteristics, a signal-to-noise ratio (SNR) measurement, and distortion characteristics, and designed to generate a threshold (TH) based on the channel characteristics, the SNR measurement, and the distortion characteristics;

a boundary distance determination module designed to receive the real components of each M-QAM constellation point, the imaginary components of each M-QAM constellation point, the real boundary points of the M-QAM regions, and the imaginary boundary points of the M-QAM regions, and designed to compute a minimum distance between the real component of each M-QAM constellation point and a nearest boundary in a real direction based on the real boundary points of the M-QAM regions, and to compute a minimum distance between the imaginary component of each M-QAM constellation point and a nearest boundary in an imaginary direction based on the imaginary boundary points of the M-QAM regions;

a real sign flipping determination module designed to;

determine whether the real components of each M-QAM constellation point are in error by determining whether an absolute value of the minimum distance between the real component of each M-QAM constellation point and the nearest boundary in the real direction is less than the threshold; and

generate, when the absolute value of the minimum distance is less than the threshold, an updated real component for that incorrectly estimated M-QAM constellation point that is in error, wherein the updated real component for that incorrectly estimated M-QAM constellation point that is in error is equal to the difference between the real component and the minimum distance between the real component and the nearest boundary in the real direction such that the real component of that incorrectly estimated M-QAM constellation point that is in error is moved to an adjacent M-QAM region in the real direction;

an imaginary sign flipping determination module designed to;

determine whether the imaginary components of each M-QAM constellation point are in error by determining whether an absolute value of the minimum distance between the imaginary component of each M-QAM constellation point and the nearest boundary in the imaginary direction is less than the threshold; and

generate, when the absolute value of the minimum distance is less than the threshold, an updated imaginary component for that incorrectly estimated M-QAM constellation point that is in error, wherein the updated imaginary component for that incorrectly estimated M-QAM constellation point that is in error is equal to the difference between the imaginary component and the minimum distance between the imaginary component and the nearest boundary in the imaginary direction such that the imaginary component of that incorrectly estimated M-QAM constellation point that is in error is moved to an adjacent M-QAM region in the imaginary direction;

an imaginary multiplier module designed to multiply each of the updated imaginary components by an imaginary multiplier; and

a frequency domain combiner module designed to generate, for each incorrectly estimated M-QAM constellation point that is in error, an updated M-QAM constellation point based on at least one of the an updated real component and the updated imaginary component.