Method and apparatus for reconstructing digitized distorted signals

• US 8,228,218 B2
• Filed: 10/29/2008
• Issued: 07/24/2012
• Est. Priority Date: 10/30/2007
• Status: Expired due to Fees

• Current Assignee: TECHNION RESEARCH AND DEVELOPMENT FOUNDATION, LTD. (Technion-Israel Institute Of Technology)
• Original Assignee: TECHNION RESEARCH AND DEVELOPMENT FOUNDATION, LTD. (Technion-Israel Institute Of Technology)
• Inventors: Eldar, Yonina, Dvorkind, Tsvi Gregory
• Primary Examiner(s): WILLIAMS, HOWARD L
• Application Number:
  US12/739,745
• Publication Number:
  US 20100259428A1
• Time in Patent Office: 1,364 Days
• Field of Search: 341/110, 341/118, 341/120, 702/86, 702189-196
• US Class Current: 341/118
• CPC Class Codes: H03M 1/0614 of harmonic distortion H03M1/0617 takes precedence
• CPC Class Codes: H03M 1/12 Analogue/digital converters H03M1/001 – H03M1/10 take precedence

Abstract

A method of reproducing an original analog signal modified by a distortion and sampled. The distortion may be linear or nonlinear, and the samples may be either ideal or non-ideal. The method determines a stationary point of a cost function based on an error vector. The method iteratively computes approximated analog signals until the approximated analog signals converge to the original analog signal. The method may utilize Fr?chet derivatives and Moore-Penrose pseudo inverse transformations in order to iteratively compute the original analog signal. An apparatus performing said method is also disclosed.

Claims

1. A method for recovering an original analog signal from a sampling of the original analog signal, the method comprising:

obtaining the samples by applying a first distortion to the original analog signal forming a distorted original analog signal and then sampling the distorted original signal by a first digitizing process forming a representation of digitized samples of the original signal;

calculating a signal representing an initial approximation of the original analog signal;

reconstructing the original analog signal by;

(a) applying a second distortion on the calculated signal forming a distorted calculated signal;

wherein said second distortion is substantially equivalent to the first distortion;

(b) sampling the distorted calculated signal by a second digitizing process;

wherein said second digitizing process is substantially equivalent to the first digitizing process forming a representation of digitized samples of the distorted calculated signal;

(c) comparing the representations by forming an error representation of the difference between the representations of the digitized samples of the original signal and the digitized samples of the calculated signal;

if the representations are substantially equivalent then the calculated signal is considered to be substantially equivalent to the original signal and the calculated signal is output;

otherwise the error representation is used to calculate a refined approximation of the original signal and the reconstructing of the original signal is performed iteratively using the refined calculated signal.

2. The method of claim 1, wherein reconstructing the original analog signal is performed in a digital domain.

3. The method of claim 1, wherein reconstructing the original analog signal is performed in an analog domain.

4. The method of claim 1, wherein calculating a signal representing an initial approximation of the original analog signal comprises generating the calculated signal as a function of known characteristics of the original analog signal.

5. The method of claim 4, wherein the original analog signal is an electrical or electromagnetic signal.

6. The method of claim 1, wherein the distortion is a nonlinear distortion.

7. The method of claim 6, wherein the first digitizing process is a non-ideal sampling process.

8. The method of claim 7, wherein the original analog signal represents audio data.

9. The method of claim 7, wherein the original analog signal represents visual or image data.

10. The method of claim 6, wherein the original analog signal represents temperature, air pressure or chemical process.

11. The method of claim 1, wherein said step of reconstructing the original analog signal is performed by determining a stationary point in an error function based on the error representation.

12. The method of claim 1, wherein said calculated signal is also output upon a sufficient convergence of the calculated analog signal.

13. The method of claim 12, wherein the first digitizing process is a non-ideal sampling process.

14. The method of claim 11, wherein said determination of the stationary point utilizes an inductive equation;

the inductive equation is a_n+1=a_n−

α

_ne(a_n).

15. The method of claim 14, wherein said determination of the stationary point utilizes a derivative.

16. The method of claim 15, wherein said determination of the stationary point utilizes an inductive equation;

the inductive equation is a_n+1=a_n−

α

_n(S*M′

_nA)^†

e(a_n).

17. An apparatus for reconstructing an original analog signal, comprising:

an input/output (I/O) module for receiving a representation of digitized samples of a distorted version of the original analog signal distorted by a first distortion;

an initial signal generator for generating an initial calculated signal of an approximation of the original analog signal;

a distorting module for applying a second distortion on the calculated signal forming a distorted calculated signal;

wherein said second distortion is substantially equivalent to the first distortion;

a sampler for sampling the distorted calculated signal forming a representation of digitized samples of the calculated signal;

wherein said sampler performs a sampling process that is substantially equivalent to a sampling process that obtained the digitized samples of the distorted version of the original analog signal; and

a processing module for

comparing the representations by forming an error representation of the difference between the representation of the digitized samples of the original signal and the digitized samples of the calculated signal;

if the representations are substantially equivalent then the calculated signal is considered to be substantially equivalent to the original signal and the calculated signal is output from the I/O module by the processing module;

otherwise the error representation is used to calculate a refined approximation of the original signal and the reconstructing of the original signal is performed iteratively using the refined calculated signal.

18. The apparatus of claim 17 further comprising an error estimation module that compares the representations for the processing module.

19. The apparatus of claim 17, wherein the first distortion is a nonlinear distortion.

20. The apparatus of claim 17, wherein the sampler is a non-ideal sampler.

21. The apparatus of claim 17 further comprising a digital-to-analog converter to create the calculated signal.

22. The apparatus of claim 17, wherein at least one of said distorting module, said sampler and said processing module operate in a digital domain.

23. The apparatus of claim 17, wherein said distorting module utilizes a module that caused the first distortion.

24. The apparatus of claim 17 wherein said sampler utilizes sampler module that performed the sampling process that obtained the samples of the distorted version of the original analog signal.