Method, computer program, and system for automated real-time signal analysis for detection, quantification, and prediction of signal changes

• US 20030187621A1
• Filed: 04/01/2003
• Published: 10/02/2003
• Est. Priority Date: 04/03/2000
• Status: Active Grant

• Current Assignee: Flint Hills Scientific, LLC
• Original Assignee: Flint Hills Scientific, LLC
• Inventors: Frei, Mark G., Osorio, Ivan, Bhavaraju, Naresh C., Nikitin, Alexei V., Davidchack, Ruslan
• Granted Patent: US 6,904,390 B2
• Time in Patent Office: Days
• US Class Current: 702/189
• CPC Class Codes: G06F 17/10 Complex mathematical operations function generation by table look-up G06F1/03; evaluation of elementary functions by calculation G06F7/544
• CPC Class Codes: G06F 17/18 for evaluating statistical data , e.g. average values, frequency distributions, probability functions, regression analysis forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04
• CPC Class Codes: G06F 2218/12 Classification; Matching
• CPC Class Codes: G16B 40/00 ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
• CPC Class Codes: G16B 40/10 Signal processing, e.g. from mass spectrometry [MS] or from PCR
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 50/50 for simulation or modelling of medical disorders

Abstract

A method, computer program, and system for real-time signal analysis providing characterization of temporally-evolving densities and distributions of signal features of arbitrary-type signals in a moving time window by tracking output of order statistic filters (also known as percentile, quantile, or rank-order filters). Given a raw input signal of arbitrary type, origin, or scale, the present invention enables automated quantification and detection of changes in the distribution of any set of quantifiable features of that signal as they occur in time. Furthermore, the present invention'"'"'s ability to rapidly and accurately detect changes in certain features of an input signal can also enable prediction in cases where the detected changes associated with an increased likelihood of future signal changes.

Claims

1. A method of detecting a change in at least one feature of an input signal, the method comprising the steps of:

(a) receiving the input signal;

(b) determining at least one time-weighted cumulative distribution function of the input signal;

(c) comparing the time-weighted cumulative distribution function to a reference function in order to detect the change in the feature of the input signal; and

(d) communicating detection of the change in the feature.

2. The method as set forth in claim 1, wherein the input signal is of arbitrary type, origin, scale, and degree of complexity.

3. The method as set forth in claim 1, wherein the input signal is selected from the group consisting of:

continuous-time signals, discrete-time signals, analog signals, digital signals, scalar signals, multi-dimensional signals, deterministic signals, stochastic signals, stationary signals, non-stationary signals, linear signals, and nonlinear signals.

4. The method as set forth in claim 1, wherein the input signal is selected from the group consisting of:

biological signals, financial signals, physical signals, communication signals, mechanical signals, chemical signals, sequences of measured data, and lists of measured data.

5. The method as set forth in claim 1, wherein step (a) includes preprocessing the input signal to derive a feature signal, wherein the feature signal is thereafter used in place of the input signal.

6. The method as set forth in claim 5, wherein the feature signal is multi-dimensional, with each dimension being representative of a level of a particular feature of the input signal.

7. The method as set forth in claim 6, wherein the particular feature of the input signal is selected from the group consisting of:

derivatives, integrals, moments, wavespeed, signal power in a time window, signal power in a particular frequency band, frequency domain features, time domain features, measures from nonlinear dynamics, measures of rythmicity, wave shape, and amplitude.

8. The method as set forth in claim 1, wherein the time-weighted cumulative distribution function is determined from at least one function selected from the group consisting of:

time-weighted density functions, time-weighted feature density functions, time-dependent feature density functions, modulated threshold density functions, time-weighted modulated threshold density functions, time-dependent modulated threshold density functions, time-weighted amplitude density functions, time-dependent amplitude density functions, time-weighted counting density functions, time-dependent counting density functions, time-weighted threshold crossing density functions, time-dependent threshold crossing density functions, time-weighted acceleration density functions, time-dependent acceleration density functions.

9. The method as set forth in claim 1, wherein the time-weighted cumulative distribution function is determined using a foreground moving time window of the input signal.

10. The method as set forth in claim 1, wherein the time-weighted cumulative distribution function depends upon at least one other auxiliary signal.

11. The method as set forth in claim 10, wherein the auxiliary signal may be determined from the input signal.

12. The method as set forth in claim 1, wherein the time-weighted cumulative distribution function utilizes exponential forgetting.

13. The method as set forth in claim 1, wherein the time-weighted cumulative distribution function is determined parametrically.

14. The method as set forth in claim 1, wherein the time-weighted cumulative distribution function is determined by first determining at least one parameter upon which the time-weighted cumulative distribution function depends and where the time-weighted cumulative distribution function is selected from the group consisting of:

Gaussian distributions, Chi-square distributions, Uniform distributions, Exponential distributions, Beta distributions, and Triangle distributions.

15. The method as set forth in claim 1, wherein the time-weighted cumulative distribution function is determined non-parametrically.

16. The method as set forth in claim 1, wherein the time-weighted cumulative distribution function is determined by interpolation from a plurality of percentile values of a second time-weighted cumulative distribution function.

17. The method as set forth in claim 1, wherein the time-weighted cumulative distribution function is determined by extrapolation from a plurality of percentile values of a second time-weighted cumulative distribution function.

18. The method as set forth in claim 1, wherein the reference function is a second time-weighted cumulative distribution function determined from a background moving time window of the input signal.

19. The method as set forth in claim 1, wherein step (c) further includes quantifying the change in the feature.

20. The method as set forth in claim 1, wherein step (c) is performed in real time.

21. The method as set forth in claim 1, further including step (e) using the detected change in the feature to predict a future change in the input signal.

22. The method as set forth in claim 1, further including step (e) using the detected change in the feature to control a process.

23. The method as set forth in claim 1, further including step (e) using the detected change in the feature to recognize the presence of a pattern in a particular feature of the input signal.

24. A system for detecting a change in at least one feature of an input signal, the system comprising circuitry operable to:

receive the input signal;

determine at least one time-weighted cumulative distribution function of the input signal;

compare the time-weighted cumulative distribution function to a reference function in order to detect the change in the feature of the input signal; and

communicate detection of the change in the feature.

25. The system as set forth in claim 24, wherein the input signal is of arbitrary type, origin, scale, and degree of complexity.

26. The system as set forth in claim 24, wherein the input signal is selected from the group consisting of:

continuous-time signals, discrete-time signals, analog signals, digital signals, scalar signals, multi-dimensional signals, deterministic signals, stochastic signals, stationary signals, non-stationary signals, linear signals, and nonlinear signals.

27. The system as set forth in claim 24, wherein the input signal is selected from the group consisting of:

biological signals, financial signals, physical signals, communication signals, mechanical signals, chemical signals, sequences of measured data, and lists of measured data.

28. The system as set forth in claim 24, wherein the circuitry is further operable to preprocess the input signal to derive a feature signal, wherein the feature signal may thereafter be used in place of the input signal.

29. The system as set forth in claim 28, wherein the feature signal is multi-dimensional, with each dimension being representative of a level of a particular feature of the input signal.

30. The system as set forth in claim 29, wherein the particular feature of the input signal is selected from the group consisting of:

derivatives, integrals, moments, wavespeed, signal power in a time window, signal power in a particular frequency band, frequency domain features, time domain features, measures from nonlinear dynamics, measures of rythmicity, wave shape, and amplitude.

31. The system as set forth in claim 24, wherein the circuitry is further operable to quantify the change in the feature.

32. The system as set forth in claim 24, wherein the detected change in the feature is used to predict a future change in the input signal.

33. The system as set forth in claim 24, wherein the detected change in the feature is used to control a process.

34. The system as set forth in claim 24, wherein the detected change in the feature is used to recognize the presence of a pattern in a particular feature of the input signal.

35. The system as set forth in claim 24, wherein the time-weighted cumulative distribution function is determined using exponential forgetting.