Method and apparatus for real-time estimation of physiological parameters

• US 6,575,905 B2
• Filed: 09/21/2001
• Issued: 06/10/2003
• Est. Priority Date: 09/22/2000
• Status: Expired due to Fees

• Current Assignee: Knobbe LIM & Buckingham
• Original Assignee: Knobbe LIM & Buckingham
• Inventors: Lim, Wah L., Knobbe, Edward J., Buckingham, Bruce A.
• Primary Examiner(s): Hindenburg, Max F.
• Assistant Examiner(s): Natnithithadha, Navin
• Application Number:
  US09/960,846
• Publication Number:
  US 20020099282A1
• Time in Patent Office: 627 Days
• Field of Search: 600/365, 600/347, 600/345, 600/316, 600/309, 128/923, 128/924, 128/925
• US Class Current: 600/365
• CPC Class Codes: A61B 2560/0242 adapted to measure environmental factors, e.g. temperature, pollution
• CPC Class Codes: A61B 5/0002 Remote monitoring of patients using telemetry, e.g. transmission of vital signals via a communication network A61B5/07 takes precedence; transmission systems for measured values G08C; transmission H04B; transmission of digital information H04L; wireless communication networks H04W
• CPC Class Codes: A61B 5/14532 for measuring glucose, e.g. by tissue impedance measurement
• CPC Class Codes: A61B 5/725 using specific filters therefor, e.g. Kalman or adaptive filters A61B5/316 take precedence
• CPC Class Codes: A61M 2230/201 Glucose concentration
• CPC Class Codes: A61M 5/142 Pressure infusion, e.g. using pumps
• CPC Class Codes: G16H 10/40 for data related to laboratory analysis, e.g. patient specimen analysis
• CPC Class Codes: G16H 20/13 delivered from dispensers
• CPC Class Codes: G16H 20/30 relating to physical therapies or activities, e.g. physiotherapy, acupressure or exercising
• CPC Class Codes: G16H 50/50 for simulation or modelling of medical disorders
• CPC Class Codes: Y02A 90/10 Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
• CPC Class Codes: Y10S 128/923 by comparison of patient data to other data
• CPC Class Codes: Y10S 128/924 using artificial intelligence

Abstract

A real-time glucose estimator uses a linearized Kalman filter to determine a best estimate of glucose level in real time. The real-time glucose estimator receives at least one measurement corresponding to glucose level. The measurement can be obtained with one or more sensors and is provided to the linearized Kalman filter in real time. The linearized Kalman filter has dynamic models and executes a recursive routine to determine the best estimate of glucose level based upon the measurement. Additional information can be provided to the linearized Kalman filter for initialization, configuration, and the like. Outputs of the linearized Kalman filter can be provided to a patient health monitor for display or for statistical testing to determine status of the real-time glucose estimator. The real-time glucose estimator can be implemented using a software algorithm.

Claims

1. A method for providing a best estimate of glucose level in real time comprising the acts of:

obtaining a measurement which is a function of glucose level, wherein noise associated with the measurement is within limits of a predefined measurement uncertainty;

supplying the measurement to an extended Kalman filter in real time, wherein the extended Kalman filter has a dynamic process model, a dynamic measurement model, a state vector with at least one element corresponding to glucose level, and an error covariance matrix of the state vector; and

determining the best estimate of glucose level in real time using the extended Kalman filter.

2. The method of claim 1, wherein the extended Kalman filter is implemented using a software algorithm.

3. The method of claim 1, wherein determination of the best estimate of glucose level in real time uses a recursive process comprising the acts of:

computing a current estimate of the state vector using a preceding best estimate of the state vector and the dynamic process model;

computing a current error covariance matrix of the state vector using a preceding error covariance matrix of the state vector, uncertainties associated with the dynamic process model, and the dynamic process model linearized about the current estimate of the state vector;

computing a Kalman gain using the current covariance matrix, uncertainties associated with the dynamic measurement model, and the dynamic measurement model linearized about the current estimate of the state vector;

computing a new error covariance matrix of the state vector using the current error covariance matrix of the state vector, the Kalman gain, and the dynamic measurement model linearized about the current estimate of the state vector; and

computing a new best estimate of the state vector using the current estimate of the state vector, the Kalman gain, the measurement, and the dynamic measurement model.

4. The method of claim 1, wherein development of the dynamic process model and the dynamic measurement model is an iterative process comprising the acts of:

defining estimation variables and uncertain parameters which become elements of the state vector;

defining nominal time propagation of the estimation variables and the uncertain parameters;

defining a nominal sensor measurement model;

defining relationships between the estimation variables, the uncertain parameters, and the measurement;

defining uncertainties associated with the estimation variables, the uncertain parameters, and the measurement; and

verifying a nominal dynamic process model and a nominal measurement model.

5. The method of claim 4, wherein a database of measurements is used to empirically verify the nominal dynamic process model and the nominal measurement model.

6. The method of claim 5, wherein the database has a plurality of sensor measurements with corresponding direct measurements.

7. A real-time glucose estimator comprising:

a plurality of measurement inputs, wherein at least one of the measurement inputs is configured to receive an input indicative of glucose level;

a plurality of control inputs; and

an extended Kalman filter algorithm configured to receive the plurality of measurement inputs and the plurality of control inputs to provide an optimal estimate of glucose level in real time.

8. The real-time glucose estimator of claim 7, wherein at least two independent glucose sensors are coupled to the measurement inputs respectively.

9. The real-time glucose estimator of claim 8, wherein at least two independent glucose sensors are of different types.

10. The real-time glucose estimator of claim 7, wherein at least one of the measurement inputs is configured to receive a direct measurement of glucose.

11. The real-time glucose estimator of claim 10, wherein the direct measurement is derived from a capillary blood glucose measurement.

12. The real-time glucose estimator of claim 7, wherein at least one of the control inputs is configured to receive configuration data.

13. The real-time glucose estimator of claim 7, wherein at least one of the control inputs is configured to receive environment data.

14. The real-time glucose estimator of claim 7, wherein the extended Kalman filter algorithm produces the optimal estimate of glucose level in real time using a recursive loop formed by a time update module and a measurement update module.

15. The real-time glucose estimator of claim 14, wherein the plurality of control inputs is processed in real time by the time update module.

16. The real-time glucose estimator of claim 14, wherein the plurality of measurement inputs is processed in real time by the measurement update module.

17. The real-time glucose estimator of claim 16, wherein measurement inputs of a substantially identical time interval are processed serially in one cycle of the measurement update module.

18. The real-time glucose estimator of claim 10, wherein the direct measurement is provided during initialization of the real-time glucose estimator and intermittently thereafter.

19. The real-time glucose estimator of claim 7, wherein at least one of the measurement inputs is provided periodically to the real-time glucose estimator.

20. The real-time glucose estimator of claim 7, wherein at least one of the measurement inputs is coupled to a non-invasive glucose sensor.

21. The real-time glucose estimator of claim 7 further comprising a patient health monitor configured to accept user inputs, to derive configuration data based on the user inputs, and to provide the configuration data in real time to the real-time glucose estimator.

22. The real-time glucose estimator of claim 7 further comprising a patient health monitor configured to display in real time the optimal estimate of glucose level and to provide an audible alarm when the optimal estimate of glucose level is outside a predetermined range.

23. The real-time glucose estimator of claim 10 further comprising a patient health monitor configured to display in real time a performance status for the real-time glucose estimator, wherein the performance status is based upon results of statistical testing performed on residuals of the extended Kalman filter.

24. The real-time glucose estimator of claim 23, wherein the patient health monitor outputs a request for the direct measurement of glucose when the performance status indicates the optimal estimate of glucose level is outside predetermined limits.

25. An estimator for monitoring a physiological parameter comprising:

a sensor which outputs a measurement as a function of the physiological parameter;

an electronic processor coupled to an output of the sensor, wherein the electronic processor executes an algorithm that implements an extended Kalman filter to estimate the physiological parameter in real time; and

an interface coupled to an output of the electronic processor to display the estimate of the physiological parameter in real time.

26. The estimator of claim 25, wherein the interface accepts user inputs and provides a control signal in real time to the electronic processor based on the user inputs.

27. The estimator of claim 26, wherein the user inputs describe environmental conditions of the sensor in real time.

28. The estimator of claim 26, wherein the user inputs describe activities of a patient using the estimator in real time.

29. The estimator of claim 26, wherein the user inputs describe real-time administration of medication to a patient using the estimator.

30. The estimator of claim 25, wherein the estimator is a portable device contained in a relatively small package and battery operated.

31. An estimator comprising:

means for obtaining a measurement related to a physiological parameter; and

means for processing the measurement in real time using a linearized Kalman filter algorithm to provide a real-time estimate of the physiological parameter.

32. The estimator of claim 31 further comprising means for displaying the real-time estimate of the physiological parameter in real time.

33. The estimator of claim 31 further comprising means for updating the linearized Kalman filter algorithm in real time with user inputs.

34. The estimator of claim 31 further comprising means for indicating performance status of the estimator in real time.