System for reconstruction of non-uniformly sampled signals

• US 20070244669A1
• Filed: 04/12/2006
• Published: 10/18/2007
• Est. Priority Date: 04/12/2006
• Status: Active Grant

• Current Assignee: Infineon Technologies AG
• Original Assignee: Infineon Technologies AG
• Inventors: Draxelmayr, Dieter, Vogel, Christian
• Granted Patent: US 7,403,875 B2
• Time in Patent Office: Days
• US Class Current: 702/190
• CPC Class Codes: H03M 1/0836 of phase error, e.g. jitter
• CPC Class Codes: H03M 1/1215 using time-division multiplexing

Abstract

A signal processing system allows reconstruction of a non-uniformly sampled analog signal including sampling an analog signal at a sampling period of resulting in a quantized non-uniformly sampled signal. The non-uniformly sampled signal includes a sampled signal and an amplitude error between a signal sampled with the equidistant sample period and the non-uniformly sampled signal. A reconstructed amplitude error is determined through the time offset and the non-uniformly sampled signal, and is subtracted from the non-uniformly sampled signal. The signal processing system may be implemented in an electrocardiogram monitoring device or a mobile phone device.

Claims

1. A method that reconstructs a non-uniformly sampled analog signal, comprising:

sampling a signal, with an equidistant sample period, by at least one of an analog-digital converter or a time interleaved analog-digital converter;

determining a quantized non-uniformly sampled signal comprising the sampled signal with an amplitude error between the signal, and the non-uniformly sampled signal;

determining a reconstructed amplitude error based on a time offset and the non-uniformly sampled signal;

subtracting the reconstructed amplitude error from the non-uniformly sampled signal to form a reconstructed non-uniformly sampled signal; and

outputting the reconstructed non-uniformly sampled signal.

2. The method according to claim 1, further comprising delaying the non-uniformly sampled signal by a delay time needed to reconstruct the amplitude error prior to subtracting the reconstructed amplitude error from the non-uniformly sampled signal.

3. The method according to claim 2, where determining the reconstructed amplitude error comprises:

differentiating the non-uniformly sampled signal; and

multiplying the non-uniformly sampled signal with a weighting factor, where the weighting factor is determined from the time offset using one or more coefficients of approximating polynomials.

4. The method according to claim 3, where determining the reconstructed amplitude error comprises delaying, by a cascade time, the subtracting of the reconstructed amplitude error from the non-uniformly sampled signal.

5. The method according to claim 1, where determining the reconstructed amplitude error comprises transforming the non-uniformly sampled signal into a discrete-fourier-transformed signal using a plurality of samples, where the plurality of samples comprises a positive integer.

6. The method according to claim 5, where determining the reconstructed amplitude error comprises determining the reconstructed amplitude error through a plurality of cascaded reconstruction steps in the frequency domain.

7. The method according to claim 1, where sampling the analog signal comprises sampling every channel by a TI-ADC with an individual time offset based on a number of channels, determining weighting factors used for the reconstruction of the amplitude error according to a number of cascaded reconstruction steps, and storing the weighting factors.

8. A signal processing apparatus that reconstructs a non-uniformly sampled analog signal, sampled by at least one of an analog-digital converter or a time interleaved analog-digital converter, comprising:

an input operable to receive an analog signal;

an output configurable to provide a digital output signal;

an evaluator operable to reconstruct an amplitude error based on a time offset and the non-uniformly sampled signal;

a subtract unit operable to subtract the amplitude error from the non-uniformly sampled signal to form a reconstructed non-uniformly sampled signal; and

an output configurable to output the reconstructed non-uniformly sampled signal.

9. A signal processing apparatus according to claim 8, where the evaluator comprises a differentiator operable to differentiate the non-uniformly sampled signal and a multiplier operable to perform a subsequent multiplication of the non-uniformly sampled signal with a weighting factor, where the weighting factor is determined from the time offset using a plurality of coefficients of approximating polynomials.

10. A signal processing apparatus according to claim 9, where the evaluator is operable to execute a plurality of cascaded reconstruction steps, further comprising:

a delay unit operable to delay the non-uniformly sampled signal by a cascade time based on the cascaded reconstruction steps; and

where the subtract unit is operable to subtract the reconstructed amplitude error from the delayed non-uniformly sampled signal.

11. A signal processing apparatus according to claim 8, where the evaluator is operable to execute a reconstruction step in a frequency domain, comprising

a transforming unit operable to transform the non-uniformly sampled signal into a discrete-fourier-transformed signal based on a plurality of samples, where the plurality of sample comprises a positive integer;

a calculator operable to determine the reconstructed amplitude error; and

a subtract unit operable to subtract the reconstructed amplitude error from the discrete-fourier-transformed signal.

12. A signal processing apparatus according to claim 11, where the evaluator is operable to execute a plurality of cascaded reconstruction steps in the frequency domain.

13. An electrocardiogram monitoring device that reconstructs a non-uniformly sampled analog signal, sampled by at least one of an analog-digital converter or a time interleaved analog-digital converter, comprising:

an input operable to receive an analog signal;

an output configurable to provide a digital output signal;

an evaluator operable to reconstruct an amplitude error based on a time offset and the non-uniformly sampled signal;

a subtract unit operable to subtract the amplitude error from the non-uniformly sampled signal to form a reconstructed non-uniformly sampled signal; and

an output configurable to output the reconstructed non-uniformly sampled signal.

14. A mobile phone that reconstructs a non-uniformly sampled analog signal, sampled by at least one of an analog-digital converter or a time interleaved analog-digital converter, comprising:

an input operable to receive an analog signal;

an output configurable to provide a digital output signal;

an evaluator operable to reconstruct an amplitude error based on a time offset and the non-uniformly sampled signal; and

a subtract unit operable to subtract the amplitude error from the non-uniformly sampled signal to form a reconstructed non-uniformly sampled signal; and

an output configurable to output the reconstructed non-uniformly sampled signal.

15. A computer program product comprising:

a computer useable medium having computer readable code embodied in the medium, the computer readable code comprising;

computer readable code executable to sample a signal by a at least one of a analog-digital converter or a time interleaved analog-digital converter;

computer readable code executable to determine a quantized non-uniformly sampled signal comprising a sampled signal and an amplitude error between a signal sampled with an equidistant sample period and the non-uniformly sampled signal;

computer readable code executable to determine a reconstructed amplitude error based on the time offset and the non-uniformly sampled signal;

computer readable code executable to subtract the reconstructed amplitude error from the non-uniformly sampled signal to form a reconstructed non-uniformly sampled signal and;

computer readable code executable to output the reconstructed non-uniformly sampled signal.

16. A signal processing apparatus that reconstructs a non-uniformly sampled analog signal, sampled by at least one of an analog-digital converter or a time interleaved analog-digital converter, a sampling period comprises a period, comprising:

input means for receiving an analog signal;

output means for providing a digital output signal;

evaluation means for reconstructing an amplitude error based on the time offset and the non-uniformly sampled signal;

subtraction means for subtracting the amplitude error from the non-uniformly sampled signal to form a reconstructed non-uniformly sampled signal; and

means for outputting the reconstructed non-uniformly sampled signal.

17. A signal processing apparatus according to claim 16, where the evaluating means comprises differentiating means for differentiating the non-uniformly sampled signal and multiplying means for multiplying the non-uniformly sampled signal with a weighting factor, where the weighting factor is determined from the time offset using a plurality of coefficients of approximating polynomials.

18. A method that reconstructs a non-uniformly sampled analog signal, comprising:

determining a quantized non-uniformly sampled signal comprising a sampled signal and an amplitude error between a signal and the non-uniformly sampled signal, where the signal is sampled with an equidistant sample period;

determining a reconstructed amplitude error based on a time offset and the non-uniformly sampled signal; and

subtracting the reconstructed amplitude error from the non-uniformly sampled signal to form a reconstructed non-uniformly sampled signal.

19. The method according to claim 18, further comprising delaying the non-uniformly sampled signal by a delay time needed to reconstruct the amplitude error prior to subtracting the reconstructed amplitude error from the non-uniformly sampled signal.

20. The method according to claim 19, where determining the reconstructed amplitude error comprises:

differentiating the non-uniformly sampled signal; and

multiplying the non-uniformly sampled signal with a weighting factor, where the weighting factor is determined from the time offset using one or more coefficients of approximating polynomials.