Method and apparatus for estimating physiological parameters using model-based adaptive filtering

Abstract

A method and apparatus for reducing the effects of noise on a system for measuring physiological parameters, such as, for example, a pulse oximeter. The method and apparatus of the invention take into account the physical limitations on various physiological parameters being monitored when weighting and averaging a series of measurements. Varying weights are assigned different measurements, measurements are rejected, and the averaging period is adjusted according to the reliability of the measurements. Similarly, calculated values derived from analyzing the measurements are also assigned varying weights and averaged over adjustable periods. More specifically, a general class of filters such as, for example, Kalman filters, is employed in processing the measurements and calculated values. The filters use mathematical models which describe how the physiological parameters change in time, and how these parameters relate to measurement in a noisy environment. The filters adaptively modify a set of averaging weights to optimally estimate the physiological parameters.

1 . A method, comprising:
 receiving streams of data from a sensor, wherein each stream of data corresponds to one of a plurality of wavelengths of electromagnetic energy detected by the sensor;   calculating a first pulse rate by tracking a fundamental frequency in a one of the streams of data using an adaptive comb filter and a processor;   calculating a frequency power spectrum of the one of the streams of data with the processor; and   calculating a second pulse rate based on the frequency power spectrum using the processor.   
     
     
         2 . The method of  claim 1 , comprising emitting electromagnetic energy at the plurality of wavelengths from one or more emitters of the sensor into a target;
 detecting the electromagnetic energy with a detector of the sensor after the electromagnetic energy has passed through the target and transmitting the streams of data to a monitor.   
     
     
         3 . The method of  claim 1 , comprising preprocessing the streams of data with the processor, including taking a natural logarithm of the streams of data and band pass filtering the streams of data to attenuate the streams of data below approximately 0.5 Hz and above approximately 10 Hz. 
     
     
         4 . The method of  claim 1 , comprising whitening the one of the streams of data with the processor prior to using the adaptive comb filter. 
     
     
         5 . The method of  claim 1 , wherein calculating the value of the second pulse rate based on the frequency power spectrum comprises determining which of a plurality of peaks in the frequency power spectrum corresponds to the fundamental frequency. 
     
     
         6 . The method of  claim 1 , comprising preprocessing the streams of data with the processor, including normalizing the streams of data. 
     
     
         7 . The method of  claim 1 , comprising finding the fundamental frequency by isolating noise in the one of the streams of data with the adaptive comb filter. 
     
     
         8 . The method of  claim 1 , comprising using the processor to calculate an average pulse rate period over a given number of data samples from the one of the streams of data, wherein the processor recognizes sequences of crests and troughs in the one of the streams of data and rejects data indicative of pulses that are potentially contaminated by motion artifact. 
     
     
         9 . The method of  claim 1 , comprising using the first pulse rate as a trigger for a cardiac gated averaging processor configured for averaging the plurality of wavelengths of electromagnetic energy detected by the sensor to generate a plurality of filtered waveforms. 
     
     
         10 . The method of  claim 1 , comprising using the processor and characteristics stored in a memory to match patterns in the one of the streams of data with the characteristics to determine a third pulse rate. 
     
     
         11 . The method of  claim 10 , comprising arbitrating between the first pulse rate and the third pulse rate based on confidence levels associated with the first and third pulse rates. 
     
     
         12 . A physiological monitoring system, comprising:
 sensor emitters configured to transmit a plurality of wavelengths of electromagnetic energy into tissue of a patient;   one or more sensor detectors configured to detect the electromagnetic energy transmitted through the tissue and provide streams of data corresponding the electromagnetic energy transmitted through the tissue;   a monitor configured to monitor a pulse rate of the patient using a one of the streams of data, wherein the monitor comprises a processor configured to:
 calculate a first pulse rate by tracking a fundamental frequency in the one of the streams of data using an adaptive comb filter; 
 calculate a frequency power spectrum of the one of the streams of data; and 
 use the frequency power spectrum in a pulse rate calculator to determine a second pulse rate or to verify the first pulse rate. 
   
     
     
         13 . The physiological monitoring system of  claim 12 , wherein the processor is configured to normalize the streams of data. 
     
     
         14 . The physiological monitoring system of  claim 12 , wherein the processor is configured to preprocess the streams of data by taking the natural logarithm of the streams of data and band pass filtering the streams of data to attenuate the streams of data below approximately 0.5 Hz and above approximately 10 Hz. 
     
     
         15 . The physiological monitoring system of  claim 12 , wherein the processor is configured to compare characteristics stored in a memory of the system to patterns in the one of the streams of data to determine a third pulse rate. 
     
     
         16 . The physiological monitoring system of  claim 12 , wherein the processor is configured to track the fundamental frequency by defining the adaptive comb filter to remove signal energy from the one of the streams of data corresponding to the fundamental frequency and harmonics thereof. 
     
     
         17 . The physiological monitoring system of  claim 16 , wherein the processor is configured to determine a particular harmonic frequency which minimizes noise energy at an output of the adaptive comb filter, the particular frequency corresponding with the fundamental frequency. 
     
     
         18 . A method, comprising:
 transmitting a plurality of wavelengths of electromagnetic energy from sensor emitters into tissue;   detecting the electromagnetic energy with a detector and producing time-domain data corresponding to the plurality of wavelengths of the electromagnetic energy, wherein the time-domain data includes signal portions corresponding to a plethysmographic waveform;   transforming one or more of the plurality of wavelengths of the time-domain data into a spectral domain data to obtain a frequency power spectrum;   classifying spectral lines in the frequency power spectrum into spectral values corresponding to a fundamental frequency and one or more harmonics of the fundamental frequency; and   estimating a pulse rate within the tissue through which the electromagnetic energy has passed from the spectral values by selectively applying one or more rules based on a confidence measure associated with the estimated pulse rate.   
     
     
         19 . The method of  claim 18 , wherein transforming comprises periodically calculating the frequency power spectrum for the time-domain data. 
     
     
         20 . The method of  claim 18 , comprising estimating the pulse rate by tracking the pulse rate in the time-domain data using an adaptive comb filter.