Method and apparatus for identifying recurring signal patterns

• US 4,211,237 A
• Filed: 04/14/1978
• Issued: 07/08/1980
• Est. Priority Date: 04/14/1977
• Status: Expired due to Term

• Current Assignee: Biotronik Mess-Und Therapiegeraete GmbH & Company Ingenieurbuero Berlin (Biotronik SE & Company KG)
• Original Assignee: Biotronik Mess-Und Therapiegeraete GmbH & Company Ingenieurbuero Berlin (Biotronik SE & Company KG)
• Inventors: Nagel, Joachim
• Primary Examiner(s): Kamm, William E.
• Application Number:
  US05/896,771
• Time in Patent Office: 816 Days
• Field of Search: 128/2.06 B, 128/2.06 A, 128/2.06 R, 128/2.06 F, 128/696, 128/698, 128/702, 128/703, 128/704, 128/706, 128/708, 128/901, 128/902, 364/415, 364/417, 364/728, 364/819
• US Class Current: 600/511
• CPC Class Codes: A61B 5/344 Foetal cardiography
• CPC Class Codes: A61B 5/35 by template matching
• CPC Class Codes: A61B 5/4362 Assessing foetal parameters

Abstract

A periodically occurring signal pattern that is part of a signal mixture including interference components, in particular QRS complexes of fetal heart signals in an abdominally derived fetomaternal electrocardiogram, is detected by storing the amplitude-time waveform pattern of a signal that has initially been recognized according to a first criterion, preferably an amplitude criterion, in a memory and then using this pattern for identifying subsequent signals according to a second criterion which is the degree of coincidence of the signal in question with the amplitude shape of the stored pattern.

Claims

1. A method for identifying a periodically recurring signal pattern contained in a signal mixture which also contains interference components, comprising:

detecting the amplitude pattern of a segment of the signal mixture according to a first criterion;

storing the detected pattern in a memory, where the stored pattern constitutes a sample;

continuously matching the amplitude pattern of the stored sample to subsequently occurring signal mixture segments having a similar amplitude pattern; and

determining the degree of coincidence between the amplitude pattern of subsequently occurring signal segments and that of the sample in order to identify such subsequently occurring segments according to a second criterion, which is different from the first criterion.

2. A method as defined in claim 1 wherein the signal pattern to be detected is a fetal heart QRS complex and the signal mixture is a fetomaternal electrocardiogram.

3. Method as defined in claim 1 wherein said step of matching comprises averaging the corresponding amplitude values of the stored sample and of a subsequently occurring segment, and said step of storing includes thereafter storing the averaged amplitude shape as the sample.

4. Method as defined in claim 3 wherein said step of averaging is effected for a given period of time from signals which have been detected exclusively according to the first criterion before said step of determining.

5. Method as defined in claim 1 further comprising removing interference components from the signal mixture before step of detecting.

6. Method as defined in claim 5 wherein said step of removing interference components is carried out by filtering out predetermined interference components.

7. Method as defined in claim 5 comprising, before said step of detecting, preliminarily detecting a first occurrence of a periodically recurring interference component signal according to a third criterion, preliminarily storing the amplitude pattern of such signal in a second memory, subsequently detecting the occurrence of a further identical interference signal in the signal mixture, and, as soon as the occurrence of such a further identical interference signal, subtracting this further signal from the amplitude pattern stored in the second memory in identical phase with respect to its amplitude.

8. Method as defined in claim 7 wherein said step of subsequently detecting is carried out according to a second criterion by determining the degree of coincidence of the signal mixture with the amplitude shape of the interference signal that is stored in the second memory.

9. Method as defined in claim 7 wherein the interference signal is subject to long term amplitude variations, and said step of preliminarily storing in a second memory includes continuously adjusting the amplitude pattern stored in the second memory to the current signal shape of the interference signal.

10. Method as defined in claim 7 wherein said step of preliminarily storing in a second memory comprises averaging the corresponding amplitude values of the stored interference signal and of the currently detected interference signal and storing the averaged amplitude shape, thereby creating an improved matching of the stored amplitude pattern of the interference signal with the pattern of later interference signals that have been detected.

11. Method as defined in claim 10 wherein said step of subtracting is effected only if averages of the amplitude pattern of the interference signal have been obtained for a predetermined period of time.

12. Method as defined in claim 7 wherein the third criterion is that at least one of the amplitude pattern and autocorrelation function of the signal mixture exceeds a selected threshold value, and comprising adjusting the magnitude of the selected threshold value to the average amplitude of the interference components contained in the signal mixture.

13. Method as defined in claim 12 wherein said step of adjusting comprises providing a scaling factor for the amplitude of the interference components, which is a function of the ratio of the cross correlation function of the pattern stored in the second memory and the signal mixture to the autocorrelation function of the pattern stored in the second memory.

14. Method as defined in claim 12 wherein said step of adjusting comprises providing a scaling factor for the amplitude of the interference components, which is a function of the ratio of the cross power of the pattern stored in the second memory and the signal mixture to the power of the pattern stored in the second memory.

15. Method as defined in claim 1 wherein said step of determining includes adapting the amplitude level of the sample to the amplitude level of the signal mixture.

16. Method as defined in claim 15 wherein said step of adapting is carried out with respect to the maximum values of the amplitudes of the sample and signal mixture.

17. Method as defined in claim 1 wherein the first criterion is that at least one of the amplitude pattern and the autocorrelation function of the signal mixture exceeds a given threshold value.

18. Method as defined in claim 1 comprising varying the given threshold value according to the average amplitude of the signal mixture.

19. Method as defined in claim 17 comprising varying the given threshold value in accordance with the amplitude of the signal sample stored in the memory.

20. Method as defined in claim 17 comprising varying the given threshold value according to the amplitude of the autocorrelation function.

21. Method as defined in claim 17 comprising reducing the given threshold value in amplitude whenever the existing threshold value has not been exceeded for a certain period of time.

22. Method as defined in claim 17 wherein the second criterion is that the cross correlation function formed between the signal mixture and the amplitude pattern of the stored sample exceeds a second given threshold value, and comprising revising the amplitude of the second threshold value when a change is made from evaluation of the autocorrelation function to evaluation of the cross correlation function.

23. Method as defined in claim 1 wherein the first criterion is that at least one of the amplitude pattern and autocorrelation function of the signal mixture reaches a maximum.

24. Method as defined in claim 1 wherein the second criterion is that the cross correlation function formed between the signal mixture and the amplitude pattern of the stored sample exceeds a given threshold value.

25. Method as defined in claim 24 comprising varying the given threshold value according to the average amplitude of the signal mixture.

26. Method as defined in claim 24 comprising varying the given threshold value according to the amplitude of the cross correlation function.

27. Method as defined in claim 24 comprising reducing the given threshold value in amplitude whenever the existing threshold value has not been exceeded for a certain period of time.

28. Method as defined in claim 1 wherein the second criterion is that the cross correlation function formed between the signal mixture and the amplitude pattern of the stored sample reaches a maximum.

29. A method as defined in claim 1 wherein the signal pattern is subject to long term amplitude variations and said step of continuously matching comprises exponentially averaging the corresponding amplitude values of the sample and of the current signal, and thereafter storing the averaged amplitude shape as the sample.

30. A method as defined in claim 29 wherein said step of exponential averaging comprises employing an averaging constant whose value is a whole number power of two.

31. A method as defined in claim 30 wherein the value of the averaging constant is sixteen.

32. A method as defined in claim 1 wherein said step of determining comprises effecting a signal evaluation for individual amplitude values of the signal pattern separately by a repeated calculation.

33. A method as defined in claim 1 wherein the signal pattern is subject to long term amplitude variations, and said step of storing includes continuously averaging the sample and the current signal only when their amplitude maxima occur in time coincidence.

34. A method as defined in claim 1 wherein the signal pattern to be detected is a fetal heart signal, the signal mixture is an abdominally derived fetomaternal electrocardiogram, and comprising measuring the fetal heart frequency from the time spacing between the maxima of the degree of coincidence occurring during said step of determining.

35. A method as defined in claim 1 wherein the signal pattern to be detected is a fetal heart QRS complex subject to long term amplitude variations, the signal mixture is a fetomaternal electrocardiogram, and said step of continuously matching comprises averaging the corresponding amplitude values of the sample and of the subsequently occurring signal mixture segments for a given period of time from signals which have been detected exclusively according to the first criterion before the second criterion is used for the detection of signals, and thereafter storing the averaged amplitude shape of the sample.

36. A method as defined in claim 35 wherein the given period of time corresponds to the occurrence of a predetermined number of maternal QRS complexes.

37. A method as defined in claim 35 wherein the given period of time corresponds to the occurrence of a predetermined number of fetal QRS complexes.

38. A method as defined in claim 1 wherein the signal mixture is constituted by myographically sensed signals.

39. A method as defined in claim 38 comprising the step of rectifying the signal mixture before said step of storing.

40. Method as defined in claim 1 wherein the first criterion is that at least one of the amplitude pattern and the autocorrelation function of the signal mixture remains above a selected threshold during a given time interval and passes through a maximum.

41. Method as defined in claim 1 wherein the second criterion is that the magnitude of the cross correlation function formed between the signal mixture and the amplitude pattern of the stored sample remains above a selected threshold during a given time interval and passes through a maximum.

42. A method for identifying a periodically recurring first signal pattern contained in a signal mixture which also contains a periodically recurring second signal pattern having a larger average amplitude than the first pattern, comprising:

detecting the second signal pattern in the signal mixture according to a first criterion;

storing the detected second pattern;

subtracting the stored pattern from subsequent occurrences of the second pattern in the signal mixture to form a difference signal;

detecting the first signal pattern in the difference signal according to a second criterion; and

determining, according to a third criterion, the degree of coincidence between the amplitude pattern of subsequently occurring difference signal segments and the detected first signal pattern.

43. Apparatus for identifying a periodically recurring first signal pattern contained in a signal mixture which also contains a periodically recurring second signal pattern having a larger amplitude than the first pattern, comprising:

first detecting means for detecting the second signal pattern in the signal mixture according to a first criterion;

memory means connected to said first detecting means for storing the detected second pattern;

arithmetic means connected to said memory means and to said first detecting means for subtracting the pattern stored in said memory means from subsequent occurrences of the second pattern detected by said first detecting means to form a difference signal;

second detecting means connected to receive the difference signal and to detect the first signal pattern therein according to a second criterion; and

signal processing means connected to said second detecting means for determining, according to a third criterion, the degree of coincidence between the amplitude pattern of subsequently occurring difference signal patterns and the detected first signal pattern.