Recognition of a useful signal in a measurement signal

• US 6,631,281 B1
• Filed: 05/14/2002
• Issued: 10/07/2003
• Est. Priority Date: 06/10/1999
• Status: Expired due to Term

• Current Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Original Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Inventors: Kästle, Siegfried
• Primary Examiner(s): WINAKUR, ERIC FRANK
• Application Number:
  US10/018,067
• Time in Patent Office: 511 Days
• Field of Search: 600/300, 600/309, 600/310, 600/322, 600/323, 600/330, 600/336, 600/481, 600/508, 600/529, 702/19, 708/403, 708/404
• US Class Current: 600/336
• CPC Class Codes: G06F 17/148 Wavelet transforms

Abstract

Recognition of a useful signal in a measurement signal by: transforming the measurement signal for a given time slot into the frequency range; identifying frequency peaks in the transformed measurement signal; assigning identified frequency peaks to temporal progressions of identified frequency peaks of one or more preceding time slots to the extent the identified frequency peaks are already present; assigning the temporal progressions to one or more families which are comprised of a fundamental wave and/or of one or more harmonic waves; selecting a family as that which should represent the useful signal, and; selecting a frequency peak of the current time slot from the selected family as that which should represent the measured value of the useful signal in this time slot. The signal filtering is preferably used for medical measurement signals, preferably in the area of pulsoximetry and for measuring blood pressure or determining the heart rate.

Claims

1. A method of recognizing a useful signal in a physical measurement signal, the method comprising the steps of

(a) transforming the measurement signal for a given time slot into the frequency range, by means of a Fourier transform,

(b) identifying frequency peaks in the transformed measurement signal,

(c) assigning identified frequency peaks to temporal progressions of identified frequency peaks of one or more preceding time slots to when identified frequency peaks are already present,

(d) assigning the temporal progressions to one or more families which are comprised of a fundamental wave and/or of one or more attendant harmonic waves,

(e) selecting a family as that which should represent the useful signal, and

(f) selecting a frequency peak of the given time slot from the selected family as that which should represent the measured value of the useful signal in this time slot.

2. The method as claimed in claim 1, wherein, in step (c), the assignment is effected by

(c1) initialization, at which the first obtained set of frequency peaks is taken to establish a set of progressions,

(c2) continuous link-up of fitting frequency peaks, a frequency peak being considered to be fitting when only small deviations regarding predetermined criteria occur in the last link of the progression, and wherein preferably a gap remains if no new frequency peak can be assigned to an available frequency peak, and the progression is either ended or replaced by a new progression if the gap becomes too long.

3. The method as claimed in claim 1, wherein, in step (d), the assignment is effected by examining the extent that given characteristic features exist between the progressions, which characteristic features jointly indicate that the progressions belong to the same useful signal, by some combination of a harmonic frequency relation, an expected amplitude decrease of the harmonic wave range or a proportionally equal trend development of the frequencies and/or amplitudes.

4. The method as claimed in claim 1, wherein, in step (e), the selection is effected by determining a greatest probability for the representation of the useful signal by the selected family.

5. The method as claimed in claim 1, wherein, in step (e), the selection is effected by combining predetermined criteria, including the existence of fundamental wave and/or harmonic wave(s), average fit accuracy of the progressions, number of valid frequency peaks in a progression, uniformity of a progression or quality of relation between fundamental wave and harmonic wave (s).

6. The method as claimed in claim 1, wherein, in step (e), the selection is effected by means of a plausibility test of the family with respect to previous output values, with the most plausible family being selected.

7. The method as claimed in claim 1, wherein, in step (f), the measured value of the useful signal is determined from the selected frequency peak.

8. The method as claimed in claim 1, wherein, in step (f), the selection is made by determining a greatest probability for the representation of the measured value of the useful signal by the selected frequency peak, by combining predetermined criteria by means of a fuzzy logic, the criteria being those that are set for a plausibility of the current measured value with respect to previous measured values and/or expected or useful values.

9. The method as claimed in claim 1, wherein, after step (f), a plausibility check is performed to check whether the selected frequency peak also actually corresponds to an expected measured value of the useful signal and/or whether a measured value derived from the selected frequency peak is to lead to an output, or whether no measured value at all should be output for this time slot, by comparing the measured value with previous measured values and/or with expected or useful values.

10. The method as claimed in claim 1 wherein the measurement signal is a physiological measurement signal and the method further includes the step of determining oxygen content of the blood or heart rate.

11. A computer program product stored on a storage medium which is readable by means of a computer, the computer program comprising a code for performing the steps as claimed in claim 1, when the program is run on a computer.

12. A device for recognizing a useful signal in a physical measurement signal, the device comprising:

means for transforming the measurement signal for a current time slot into the frequency range by a Fourier transform,

means for identifying frequency peaks in the transformed measurement signal,

means for assigning identified frequency peaks to temporal progressions of identified frequency peaks of one or more preceding time slots when identified frequency peaks are already present,

means for assigning the temporal progressions to one or more families which are comprised of a fundamental wave and/or of one or more attendant harmonic waves,

means for selecting a family as that which should represent the useful signal, and

means for selecting a frequency peak of the current time slot from the selected family as that which should represent the measured value of the useful signal in this time slot.

13. A device for recognizing a useful signal in a physical measurement signal as claimed in claim 12, wherein the measurement signal is a pulsoximetry signal and the device further includes means for determining oxygen content of the blood from the selected frequency peak.