Heart rate estimation apparatus with state sequence optimization

• US 10,321,831 B2
• Filed: 08/16/2016
• Issued: 06/18/2019
• Est. Priority Date: 11/25/2015
• Status: Active Grant

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Texas Instruments, Inc.
• Inventors: Pande, Tarkesh, Magee, David Patrick, Narasimha, Rajan
• Primary Examiner(s): Bertram, Eric D.
• Application Number:
  US15/237,941
• Publication Number:
  US 20170143218A1
• Time in Patent Office: 1,036 Days
• CPC Class Codes: A61B 5/02416 using photoplethysmograph signals, e.g. generated by infrared radiation A61B5/14552 takes precedence
• CPC Class Codes: A61B 5/02427 Details of sensor
• CPC Class Codes: A61B 5/7225 Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation A61B5/30, A61B5/316 take precedence; transformers for specific applications H01F38/00; amplifiers H03F
• CPC Class Codes: A61B 5/7275 Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor
• CPC Class Codes: A61B 5/7278 Artificial waveform generation or derivation, e.g. synthesising signals from measured signals A61B5/327 takes precedence

Abstract

Disclosed examples include heart rate monitor systems and methods to estimate a patient heart rate or rate of another pulsed signal, in which rate hypotheses or states, are identified for a current time window according to digital sample values of the pulsed signal for a current sample time window, and a rate change value is computed for potential rate transitions between states of the current and previous time windows. Transition pair branch metric values are computed as a function of the rate change value and a frequency domain amplitude of the corresponding rate hypothesis for the current time window, and the pulsed signal rate estimate is determined according to a maximum path metric computed according to the branch metric value and a corresponding path metric value for the previous time window.

Claims

1. A heart rate monitor system to estimate a heart rate of a patient, the heart rate monitor system comprising:

a light source configured to direct a first light signal toward a patient;

an optical sensor configured to;

detect a second light signal from the patient, the second light signal representing transmission or reflection of the first light signal in the patient; and

generate an analog photoplethysmogram (PPG) signal representing the second light signal;

an analog to digital converter configured to generate a plurality of digital PPG sample values of the PPG signal in a time window; and

a processor circuit configured to;

for the time window;

obtain a frequency domain spectrum according to the digital PPG sample values; and

identify a plurality of heart rate hypothesis states having amplitudes in the frequency domain spectrum greater than a threshold value;

identify a plurality of potential heart rate transition pairs, each representing a respective heart rate transition between;

a corresponding one of the heart rate hypothesis states for the time window; and

one of a plurality of heart rate hypothesis states for a previous time window;

for each potential heart rate transition pair, compute;

a respective heart rate change value; and

a respective branch metric value as a function of the respective heart rate change value and a frequency domain amplitude of the potential heart rate transition pair'"'"'s corresponding one of the heart rate hypothesis states for the time window;

for each heart rate hypothesis state for the time window;

compute a plurality of candidate path metric values, each as a sum of a corresponding branch metric value and a path metric value for a corresponding one of the heart rate hypothesis states for the previous time window; and

select a largest one of the candidate path metric values to be the heart rate hypothesis state'"'"'s respective path metric value; and

determine a heart rate estimate value, representing the heart rate of the patient for the time window, as the heart rate hypothesis state whose respective path metric value is largest.

2. The heart rate monitor system of claim 1, wherein the function is a monotonically decreasing function of an absolute value of the respective heart rate change value.

3. The heart rate monitor system of claim 2, wherein the function is a monotonically decreasing exponential function of the absolute value of the respective heart rate change value.

4. The heart rate monitor system of claim 3, wherein the function is a monotonically decreasing exponential function of the form α

*e<sup>(−

β

X)</sup>, where X is a heart rate change between the previous time window and the time window, α

is a constant, and wherein the processor circuit is configured to adapt a bandwidth of a heart rate search space by selectively decreasing β

if the patient is transitioning between low and high activity states and by selectively increasing β

if the patient has a constant activity state.

5. The heart rate monitor system of claim 1, wherein the processor circuit is further configured to:

filter the digital PPG sample values to generate a plurality of filtered values corresponding to the time window;

perform motion compensation processing on the filtered values to generate a plurality of motion compensated values corresponding to the time window; and

compute frequency content of the motion compensated values using a fast Fourier transform algorithm to generate the frequency domain spectrum for the time window.

6. The heart rate monitor system of claim 1, wherein the processor circuit is further configured to:

if no heart rate hypothesis state for the time window is within a predetermined distance threshold value of a previous heart rate estimate for the previous time window;

determine the heart rate estimate value, representing the heart rate of the patient for the time window, as the previous heart rate estimate for the previous time window, and

propagate at least some of the heart rate hypothesis states from the previous time window as the heart rate hypothesis states for the time window.

7. The heart rate monitor system of claim 1, wherein the processor circuit is further configured to:

selectively discard potential heart rate transition pairs corresponding to heart rate change values that are greater than a predetermined distance threshold value; and

compute the respective branch metric value for each remaining potential heart rate transition pair.

8. The heart rate monitor system of claim 1, wherein the processor circuit is further configured to selectively discard one or more heart rate hypothesis states for the time window having a low path metric value.

9. The heart rate monitor system of claim 8, wherein the threshold value is a first threshold value, and the processor circuit is further configured to:

selectively discard one or more heart rate hypothesis states for the time window having a respective path metric value less than a second threshold value; and

selectively discard one or more heart rate hypothesis states for the time window having a respective path metric value less than a difference between a largest one of the respective path metric values and a third threshold value.

10. The heart rate monitor system of claim 1, wherein the processor circuit is further configured to selectively merge two or more heart rate hypothesis states for the time window that are within a predetermined distance threshold value of each other.