ON-DEMAND HEART RATE ESTIMATION BASED ON OPTICAL MEASUREMENTS
First Claim
1. A computer-implemented method for determining a heart rate from a heartbeat signal present in a first signal generated by a heartbeat sensor, the method comprising:
- computing an autocorrelation of data samples of the first signal over a first time period;
processing a second signal to evaluate motion of the heartbeat sensor during the first time period, the second signal indicative of a motion of the heartbeat sensor with respect to one or more directions;
determining whether the data samples of the first signal over the first time period pass a first set of confidence checks, the first set of confidence checks comprising at leastdetermining whether the motion of the heartbeat sensor during the first time period satisfies a motion-related criteria;
upon determination that the data samples of the first signal over the first time period pass the first set of confidence checks, computing a first estimate of the heart rate from the autocorrelation of the data samples of the first signal over the first time period;
computing a Discrete Fourier Transform (DFT) for data samples of the first signal over a second time period;
computing a second estimate of the heart rate from the DFT; and
determining the heart rate based on the first estimate and the second estimate.
1 Assignment
0 Petitions
Accused Products
Abstract
Activity monitors and smart watches utilizing optical measurements are becoming widely popular, and users expect to get an increasingly accurate estimate of their heart rate (HR) from these devices. These devices are equipped with a light source and an optical sensor which enable estimation of HR using a technique called photoplethysmography (PPG). One of the main challenges of HR estimation using PPG is the coupling of motion into the optical PPG signal when the user is moving randomly or exercising. The present disclosure describes a computationally feasible and fast HR estimation algorithm to be executed at instances of little or no motion. Resulting HR readings may be useful on their own, or be provided to systems that monitor HR continuously to prevent the problem of such systems being locked on an incorrect HR for long periods of time. Implementing techniques described herein leads to more accurate HR measurements.
-
Citations
20 Claims
-
1. A computer-implemented method for determining a heart rate from a heartbeat signal present in a first signal generated by a heartbeat sensor, the method comprising:
-
computing an autocorrelation of data samples of the first signal over a first time period; processing a second signal to evaluate motion of the heartbeat sensor during the first time period, the second signal indicative of a motion of the heartbeat sensor with respect to one or more directions; determining whether the data samples of the first signal over the first time period pass a first set of confidence checks, the first set of confidence checks comprising at least determining whether the motion of the heartbeat sensor during the first time period satisfies a motion-related criteria; upon determination that the data samples of the first signal over the first time period pass the first set of confidence checks, computing a first estimate of the heart rate from the autocorrelation of the data samples of the first signal over the first time period; computing a Discrete Fourier Transform (DFT) for data samples of the first signal over a second time period; computing a second estimate of the heart rate from the DFT; and determining the heart rate based on the first estimate and the second estimate. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
-
-
18. An apparatus for determining a heart rate from a heartbeat signal present in a first signal generated by a heartbeat sensor, the apparatus comprising:
-
at least one memory configured to store computer executable instructions, and at least one processor coupled to the at least one memory and configured, when executing the instructions, to; compute an autocorrelation function (AF) of data samples of the first signal over a first time period; compute an squared difference function (SDF) of the data samples of the first signal over the first time period; compute a combination function representative of the computed autocorrelation function and the computed squared difference function; process a second signal to evaluate motion of the heartbeat sensor during the first time period, the second signal indicative of a motion of the heartbeat sensor with respect to one or more directions; determine whether the data samples of the first signal over the first time period pass a first set of confidence checks, the first set of confidence checks comprising at least determining whether the motion of the heartbeat sensor during the first time period satisfies a motion-related criteria; upon determination that the data samples of the first signal over the first time period pass the first set of confidence checks, compute a first estimate of the heart rate from the combination function; compute a Discrete Fourier Transform (DFT) for data samples of the first signal over a second time period; compute a second estimate of the heart rate from the DFT; and determine the heart rate based on the first estimate and the second estimate. - View Dependent Claims (19)
-
-
20. A computer-implemented method for determining a heart rate from a heartbeat signal present in a first signal generated by a heartbeat sensor, the method comprising:
-
processing a second signal to evaluate motion of the heartbeat sensor during a first time period, the second signal indicative of a motion of the heartbeat sensor with respect to one or more directions; determining whether data samples of the first signal over the first time period pass a first set of confidence checks, the first set of confidence checks comprising at least determining whether the motion of the heartbeat sensor during the first time period satisfies a motion-related criteria; upon determination that the data samples of the first signal over the first time period pass the first set of confidence checks, computing a first estimate of the heart rate from the data samples of the first signal over the first time period using a first heart rate estimation algorithm, computing a second estimate of the heart rate from the data samples of the first signal over the first time period using a second heart rate estimation algorithm, and determining the heart rate based on the first estimate and the second estimate.
-
Specification