Signal component processor
View Patent ↗A signal processor generates basis functions and identifies at least one basis function component of a sensor signal in order to calculate a physiological measurement. The signal processor is advantageously applied to pulse oximetry so as to directly calculate oxygen saturation and so as to calculate a robust oxygen saturation measurement. In particular, a signal component transform may be calculated within a window around a derived pulse rate estimate. A signal component transform may also utilize sinusoidal basis functions, and an optimization of a signal component transform may occur at a particular frequency or a set of frequencies. A sinusoidal waveform or waveforms at that frequency or set of frequencies is generated to select associated red and infrared components of the sensor signal, and an oxygen saturation is calculated from a magnitude ratio of these components.
1. A signal processing method capable of improving a determination of a pulse rate of a monitored patient based on knowledge that plethysmograph signal energy may be concentrated around an actual pulse rate frequency of the monitored patient, the improved pulse rate being used to determine a more robust indication of oxygen saturation, the method comprising:
receiving a sensor signal indicative of an oxygen saturation and a pulse rate of a patient to whom a sensor is applied;
calculating an estimated pulse rate from the sensor signal;
defining a window of data within the sensor signal around the estimated pulse rate, the window of data including frequencies spaced at a resolution of interest;
determining within the window an indication of signal energy at each of the spaced frequencies;
detecting an extrema of the indications of signal energy;
defining a refined frequency corresponding to said extrema; and
outputting an indication of oxygen saturation calculated using said sensor signal at said refined frequency.
2. The signal processing method according to claim 1 wherein said detecting said extrema comprises:
transforming said sensor signal to a frequency spectrum at said spaced frequencies; and
detecting said an extrema of said spectrum indicative of said refined frequency.
3. The signal processing method according to claim 1 wherein said extrema comprises minima of said signal energy.
4. The signal processing method according to claim 1 , comprising iteratively defining increasingly narrower windows of data within said sensor signal around a previous iteration's refined frequency, each increasingly narrower window of data including frequencies spaced at increasingly greater resolutions of interest, thereby iteratively refining said refined frequency toward said actual pulse rate frequency.
5. A signal processor capable of improving a determination of a pulse rate of a monitored patient based on knowledge that plethysmograph signal energy may be concentrated around an actual pulse rate frequency of the monitored patient, the improved pulse rate being used to determine a more robust indication of oxygen saturation, the processor comprising:
an input configured to receive a sensor signal indicative of an oxygen saturation and a pulse rate of a patient to whom a sensor is applied;
a frequency calculator operably communicating with said input to be responsive to said sensor signal, said frequency calculator configured to calculate an estimated pulse rate from the sensor signal, to define a window of data within the sensor signal around the estimated pulse rate, the window of data including frequencies spaced at a resolution of interest, to determine within the window an indication of signal energy at each of the spaced frequencies, to detect an extrema of the indications of signal energy, to define a refined frequency corresponding to said extrema, and to output said refined frequency; and
a saturation calculator operably communicating with said input and said frequency calculator to be responsive to said refined frequency, said saturation calculator configured to output an indication of oxygen saturation calculated using said sensor signal at said refined frequency.
6. The processor of claim 5 wherein said frequency calculator detects said extrema by at least:
transforming said sensor signal to a frequency spectrum at said spaced frequencies; and
detecting said extrema of said spectrum indicative of said refined frequency.
7. The processor of claim 5 wherein said extrema comprises a minima of said signal energy.
8. The processor of claim 5 , wherein said frequency calculator iteratively defines increasingly narrower windows of data within said sensor signal around a previous iteration's refined frequency, each increasingly narrower window of data including frequencies spaced at increasingly greater resolutions of interest, thereby iteratively refining said refined frequency toward said actual pulse rate frequency.