JPH02172443A - Photoelectric pulse wave type oxymeter - Google Patents

Abstract

PURPOSE:To improve precision of a measuring value itself, to stabilize the display value of oxygen saturation, and to improve reliability by a method wherein a noise component contained in a photoelectric pulse wave signal before the degree of oxygen saturation of arterial blood is reduced. CONSTITUTION:With the source switch of an oxymeter turned ON, initialization is effected. A number of pulses detecting part 5 detects the number pulses from a pulse wave signal outputted from a logarithm converting part 4, and outputs it to a control part 6. Reliability of the number of pulses determined by the number of pulses detecting part 5 is evaluated by the control part 6, the selection frequency of a B.P.F. part 7 is decided and controlled by means of the number of pulses, and only the main frequency component of a pulse wave signal outputted from the logarithm converting part 4 passes the B.P.F. part 7, and noise of other frequency region is removed by the B.P.F. part 7. From reliability of number of pulses data, the width of the pass-band of the B.P.F. part 7 is decided. A signal on which filtering process is applied by the B.P.F. part 7 is inputted to a computing part 8 to compute an SaO2 value of the degree of oxygen saturation. A computing result is outputted to a display part 9 for display.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application This invention involves receiving light of at least two wavelengths that has passed through a living body, and obtaining a pulsating photoplethysmogram signal due to the influence of absorption by arterial blood within the living body. The present invention relates to a photoplethysmogram oximeter that measures the oxygen saturation of arterial blood based on this photoplethysmogram signal.

BACKGROUND ART Conventionally, in such oximeters, a method is widely known in which measurements are taken every second, and the displayed value is obtained by averaging the measured values over five seconds. However, this method has the disadvantage that if the measured values every second vary due to the respiration of the living body, the movement of the living body, etc., the oxygen saturation display value will also vary due to the influence. In particular, when such an oximeter is applied to measure the arterial oxygen saturation of a newborn baby, the living body moves rapidly, and the oxygen saturation display value therefore fluctuates greatly. Furthermore, if the moving average time is lengthened in order to reduce the dispersion of the displayed values, the response time will be long and it will be impossible to follow sudden changes.

In addition, in order to improve such fluctuations in displayed values, it has been proposed to perform various statistical processes on measured values.

In JP-A-59-160446, a confidence limit is established in advance, and if the data received by the processing system is within the confidence limit, the confidence limit is gradually narrowed, and if the data is outside the confidence limit, the confidence limit is gradually narrowed. The limits of confidence will be expanded. The final display value is then determined only from data that are within the confidence limits. In other words, if the data is very stable, the confidence bounds will be narrow;
If the data changes locally, the data is rejected (excluded), and if the data changes continuously, the confidence limits are expanded to follow the changes in the data, but the measurements taken every second If the dam changes to a small Shiram dam, the confidence limit remains wide, and the dispersion of the final display value cannot be suppressed.

In US Pat. No. 4,407,290, a weighted average of each measured value is taken as the final display value.

The weight is determined by the reciprocal of the difference between each measured value and the latest displayed value. With this method, if the measured value fluctuates locally, the weight of that data becomes lighter and its influence can be avoided. However, if local fluctuations cannot be completely removed and the measured values vary, the final displayed value will also vary.

In addition, in Special Publication No. 62-500843, the patient's heartbeat is detected from an electrocardiogram, etc., the photoplethysmogram signal from the oximeter is correlated with the heartbeat, and the photoplethysmogram signal is determined depending on whether the two are synchronized. Determine reliability. Pulse wave signals that do not meet preset criteria are rejected and oxygen saturation readings are calculated based only on accepted pulse wave signals.

Problems to be Solved by the Invention However, these statistical processes are related to calculation methods that are repeatedly performed to obtain oxygen saturation display values. Therefore, the temporal change in the final oxygen saturation display value becomes smoother, and the displayed value can be grasped in perspective without being confused by unnecessary small fluctuations, but it does not improve the accuracy of each measurement value itself, so there is a limit. was there.

An object of the present invention is to provide a photoplethysmogram oximeter that reduces variations in oxygen saturation display values by improving the accuracy of each measurement value itself.

Means for Solving the Problems In order to achieve the above objects, the photoplethysmogram oximeter according to the present invention includes means for detecting pulse and means for frequency filtering the photoplethysmogram signal. and a control means for changing the pass frequency band of the frequency filtering means based on the pulse detection means, and determining the oxygen saturation of arterial blood based on the frequency filtered photoplethysmogram signal. It is something.

Effect: With the above means, the photoplethysmogram signal carrying information on oxygen saturation passes through the filtering means, and noise components in a frequency band different from the pulse frequency are cut by the filtering means.

Therefore, if the oxygen saturation is determined from the charged pulse wave signal after filtering, the displayed value, which conventionally fluctuated due to the influence of other noise components, becomes stable.

Note that the pass frequency band in the filtering means is
Since it changes according to the pulse rate, necessary information will not be cut out by filtering.

Embodiment FIG. 1 is a block diagram showing the overall configuration of a photoplethysmographic oximeter according to a first embodiment of the present invention. (1) is a light emitting part that emits two lights of different wavelengths, for example, infrared light IR and red light, and is a light emitting diode.
is provided.

The light emitted from the light emitting part (1) is transmitted to the measurement site (2) of the living body.
) and enters the light receiving section (3). The light receiving section (3) outputs an electrical signal according to the intensity of the received light. (4)
is a logarithmic conversion unit which logarithmically converts the output signal of the light receiving unit (3) to produce a photoplethysmogram signal (hereinafter referred to as a pulse wave signal).

). Then, this pulse wave signal is sent to a pulse rate detection unit (5) and a band pass filter (hereinafter referred to as B, P, and F).

shall be. ) output to section (7).

The pulse rate detection unit (5) detects the pulse rate based on the pulse wave signal output from the logarithmic conversion unit (4) by a method described later. In addition, in this embodiment, when detecting the pulse rate,
In order to calculate the oxygen saturation Sadz value of arterial blood, the pulse wave signal is used as the first method, but as a different method,
For example, the pulse rate may be detected based on an electrocardiographic waveform signal obtained by an independent electrocardiograph. Further, instead of this pulse rate detection section (5), a pulse cycle detection section that detects the pulse period from the pulse wave signal may be provided.

The control unit (6) evaluates the reliability of the pulse rate data obtained by the pulse rate detection unit (5), and determines the cut off frequency of the B, P, li, unit (7) based on the pulse rate and its reliability. Or control the Q value. The B, P, and F sections (7) are composed of analog circuits described later, and under the control of the control section (6), remove noise from the pulse wave signal output from the logarithmic conversion section (4). do. It should be noted that the pulse wave signal from the logarithmic conversion section (4) can be sequentially A/D converted into a digital signal, and then filtered by computer software to obtain B, P, 1. The same effect can be obtained. In addition, if the frequency components of noise included in the pulse wave signal are only low frequencies or only high frequencies, Is, l', F,
Section (7) is a low pass filter (hereinafter referred to as L, P, F) or a bypass filter (hereinafter referred to as I-1, P
, F, etc.).

The above operations from the light emitting section (1) to the B, P, l·', and section (7) are performed in parallel with respect to two lights of different wavelengths, that is, infrared light IR and red light lζ. (8)
is an arithmetic unit, and B, P, 1. Based on the two signals regarding infrared light IR and red light k that have been filtered by section (7), a psao2 value is calculated by a known method. This calculation result is displayed on the display section (9).

Next, regarding the operation of the embodiment shown in section ia, the pulse rate detection section (5), the control section (6), B, P, F, section (
7) will be mainly explained using the flowchart of FIG.

First, in step #1, the power switch of the oximeter is turned on, and initialization is performed in step #2.

In step #3, the pulse rate detection unit (5) converts the logarithmic conversion unit (4
) The pulse rate is detected from the output pulse wave signal and output to the control unit (6). A specific method for detecting the pulse rate in the pulse rate detection unit (5) is to binarize the pulse wave signal output from the logarithmic conversion unit (4), and then calculate the period of a predetermined number of the binarized pulse wave signals. It is possible to calculate the pulse rate by calculating the reciprocal of the period.

The reliability of the pulse rate determined by the pulse rate detection unit (5) is evaluated by the control unit (6) in step #4. Generally, signal fluctuations due to one body movement are larger than signal fluctuations based on changes in pulse rate due to changes in biological conditions. In other words, if there is a large variation in pulse rate data within a predetermined period in the past, the variation is due to body movement, etc.
It can be determined that the reliability of the pulse rate data is low. Conversely, if the variation in pulse rate data within the past predetermined period is small.

It can be determined that the reliability of the pulse rate data is high.

Therefore, the reliability of the pulse rate data is determined based on the magnitude of this variation. Note that if the pulse rate data has been stable for a predetermined period in the past and then suddenly begins to fluctuate, it can be determined that noise has clearly occurred due to body movement or the like. In this case, the pulse rate data at that time is deleted to obtain the pulse rate.

Based on the pulse rate obtained as described above and the reliability of the data, the control unit (6) selects B and P in step #5.

Controls the characteristics of the F0 section (7). First, depending on the pulse rate υB
, 1m, F', the selection frequency of section (7) is determined and controlled. This allows only the main frequency components of the pulse wave signal output from the logarithmic conversion section (4) to pass through the B, P, F, section (7), and noise in the other frequency regions is passed through the B, P, F, section (7).
This is to remove it in step 7). Next, the reliability of the pulse rate data and the width of the passband of the sections B, P, and F (7) are determined. In other words, if the reliability of the pulse rate data is low,
P.F.

The passband width of the section (7) is widened, and when the reliability of the pulse rate data is high, the passband width of the B, P, F, sections (7) is narrowed. This is despite the low reliability of pulse rate data for B, P, and F. This is because if the passband width of section (7) is narrowed, the pulse wave signal itself may also be attenuated. For example, if pulse rate data that had been stable for a predetermined period in the past suddenly starts to fluctuate, it can be determined that noise has clearly occurred due to body movement, etc.
The passband of section (7) is gradually widened. After that, when the pulse rate data starts to stabilize again, the passbands of the B, P, and F sections (7) are gradually narrowed to the original width.

The signals filtered by the B, P, F sections (7) are input to the calculation section (8), and the oxygen saturation 5a02 value is calculated in step #6. Specifically, B,
P, I-, two lights input from section (7), infrared light I
The two signals related to R and red light are sent to the calculation unit (8
), and a final arterial blood saturation value of 5aOz is calculated based on the digital signal obtained thereby. This calculation result is outputted and displayed on the display section (9) in step #7, and the flow returns to step #3.

Here, an example of the configuration of portions B, P, and F (7) shown in FIG. 1 is shown in FIG. 3. In FIG. 3, the pulse wave signal related to the infrared light IR output from the logarithmic conversion unit (4) is transmitted to a plurality of B, P, F, each having a different selection frequency via an analog switch SL, 52.53, that is, B , P.F.

1, B, l", F, 2 and B, P, F, 3. For the red light, also switch S.
4,55° Via S6, B, P, F, 4, B, P
, F, 5 and B, P, F, 6, the pulse wave signal is input. At this time, B, P, F, l and B, P, F
, 4, B, P, IX, 2 and B, P, F, 5.

B, l'', 1=', 3 and B, P, F, 5 each have the same selection frequency, and the pulse wave signals for the infrared light IR and red light have the same selection frequency Ll, P, F, is input.

On the other hand, the control unit (6) selects the most appropriate selected frequency according to the pulse rate data from the pulse rate detection unit (5).
, l., and controls the analog switch so that the pulse wave signal is input to the selected B, P, and F.

At the same time, the passband widths of the B, P, and F sections are controlled through signal deterioration based on the reliability of the pulse rate data.

For example, immediately after turning on the power switch of the oximeter or when fluctuations in pulse rate data are observed, by controlling the Q value of each B, P, F, part (7). Widen the overall passband width.

The effects of the filtering process in sections B, P, and F (7) shown in FIG. 3 will be explained using FIGS. 4 to 7.

FIG. 4 shows temporal divisions based on the filter beat number output from the logarithmic conversion unit (4) and the reliability of the data, and the boundaries are indicated by broken lines. Before filtering, the pulse wave signal contains a lot of noise, so based on this, 5a02
When @ is calculated, as shown by the solid line in Figure 5, 5aOz
The values fluctuate widely and have low reliability.

Therefore, depending on the pulse rate and the reliability of the data, ls, P,
After controlling the selected frequency and Q value of the F section (7), the pulse wave signal is subjected to filtering processing.

For example, since the pulse rate is moderate in section a1 in FIG. Since the pulse wave signal is large, the selected frequencies are B, P, and F.
, l Since the pulse rate is low in 1 division a3, the selected frequencies are input to B, P, F, and 3, which have low selection frequencies. B, P, F at the same time
, the Q value of section (7) is also controlled according to the reliability of the pulse rate data and filtering processing is performed. The signal after the filtering process has reduced noise as indicated by the solid line in FIG. Similar processing is also performed on the pulse wave signal related to red light K, and the SaO2 value is calculated based on these two processed signals. As shown by the solid line in FIG. 7, the fluctuation is small and the reliability is high.

In this way, even when noise occurs due to body movement, B, l), l= depending on the fluctuation of the pulse rate data at that time.

Since the selection frequency and passband width of the section (7) are controlled, the B, P, F sections (7) always perform appropriate filtering processing on the pulse wave signal output from the logarithmic conversion section (4). be able to. Therefore, the 5a02 value calculated based on the signal after the filtering process is accurate with less error than 1.

In addition, B, P, l, which is different from the configuration shown in FIG.
-', an example of the configuration of section (7) is shown in FIG. Here, a plurality of glues, each having a different cut off frequency, P, F, and )1. Although P and F' are provided,
L, P, F, 1, P, F.

4, L, P, F, 2, P, F, 5, L, P
, F, 3, P, F, 6° and %H, P, F, 1 and H,
P, F, 4, H, P, F, 2 and H, P, F, 5゜H, 1'', 3 and H, P, F, 5 are the same - cu
t off frequency, and the pulse wave signals of infrared light IK and red light are the same (a pair of beams with the t off frequency, P and F
, and a pair of H, P, F.

The control unit (6) determines P, F, and )1. P
,F.

analog switches S7 to S5 so that the pulse wave signal is input.
18.

In this case, by controlling the combination of P, F and H, P, F, it is possible to control both the selection frequency and the passband width of the B, P, F section (7) as a whole. For example, if you want to increase the selected frequency, P, F, and H, P
, F, both of which have a high cut off frequency, and conversely, if you want to lower the selection frequency, select L, P, F.
, and HoP, F, both of which have low cut off frequencies are selected. Also, if you want to widen the passband width, use a high cut off frequency or use P, F.

Select H, P, F, which has a low cut off frequency, and conversely, if you want to narrow the passband width, select a low cut off frequency.
off frequency or high cut of P, F, etc.
i, P, F, with f frequencies.

Furthermore, the same effect can be obtained even if the order of L, P, F and I-1, P, F is reversed.

FIG. 9 is a block diagram showing the overall configuration of a photoplethysmographic oximeter according to a second embodiment of the present invention in which the functions of the b, p, F, section (7) are realized by computer software.

The configuration from the light emitting section (1) to the control section (6) is the same as that shown in FIG. The pulse wave signal output from the logarithmic conversion section (4) is sequentially A/D converted by the A/D conversion section (10), and the converted signal is input to the filtering processing section (11), where it undergoes filtering processing. will be administered. At this time, the control section (6) controls the filtering processing section (11) according to the pulse rate data from the pulse rate detection section (5). The signal from which noise other than the main frequency component has been removed by the filtering processing section (11) is input to the calculation section (8). Based on this, the arterial blood oxygen saturation Sa 02 value is calculated, and the calculation result is displayed on the display section (9).

A weighted moving average method can be used as one method of filtering processing in the filtering processing section (11). In this weighted moving average method, the A/D converter (1
A plurality of digital signals (hereinafter referred to as −x
(i). ) to calculate a smoothed value using a weighting function as shown in equation (■).

Calculating the smoothed value X (i) of the signal x(
Regarding the frequency i), the same effect as L, l), F can be obtained by removing the high frequency component.

Conversely, the value Y (y) obtained by subtracting the smoothed value obtained by the weighted moving average method of formula (■) from the original signal x(i) is: Values where N: number of smoothing points, N=2M+1°M=natural number ω(j): weighting function W: constant j for normalization: j=-M, ..., M, -1, 0, 1,・
...It is M.

x(i-M) using the weighted moving average method as shown in equation (2).

x(i −M+1 ), −x(i), −x(i×M−1
), X(1+M).

Calculating the value of Y (i) using the weighted moving average method as in equation
, you will get the same effect as .

Therefore, by using the weighted moving average method using equations (2) and (2), a band-pass filtering effect can be obtained similarly when L, P, F, and HoP, F are combined.

In addition, by controlling the weighting function ω(j) and the number of smoothing points N in equations ■ and
f frequency can be controlled. In other words, even when noise occurs due to body movement, etc., the selection frequency and passband width of the filtering processing section (11) are controlled according to the pulse rate data at that time and its reliability, so the filtering processing section (11) 11) can always perform appropriate filtering processing on the pulse wave signal from the logarithmic conversion section (4).

Next, the filtering processing operation of the embodiment shown in FIG. 9 will be explained using the flowchart in FIG. 10. Steps #11 to #14 in FIG.
This is the same as steps #1 to #4 in the figure.

In step #IIt', the control unit (6) controls appropriate high-frequency and low-frequency cues according to the evaluation of the pulse rate and the reliability of the data.
The weighting function ω('') and the number of smoothing points N are determined so as to obtain the t off frequency, and are input to the filtering processing unit (11) in step #16.

In step #17, the pulse wave signal x(i) sequentially A/D converted by the A/D converter (10) is
11). At this time, X (signal X other than itself)
(i + j) are also input. In step #18, the filtering processing unit (11) performs a filtering operation to remove the high frequency component shown in equation (2).
At step 19, a filtering operation is performed to remove the low frequency component shown in equation (2), and the band-pass filtering process by the filtering processing section (11) is completed.

Note that step #18. Even if the order of #19 is reversed, there is no problem in bandpass filtering processing.

In step #20, the calculation section (8) calculates the arterial oxygen saturation S based on the input signal from the filtering processing section (11).
Calculate the a02 value. The calculated value is displayed on the display section (9) in step #21, and the flow continues in step #1.
Return to 3.

In the above embodiment, the seeded moving average method is used as a method of filtering processing in the filtering processing section (11), but Fourier transformation may also be used. In this case, first, the control unit (6) determines a Fourier function according to the evaluation of the pulse rate and the reliability of the data, and inputs it to the filtering processing unit (11). The signal x(i) that has been A/D converted by the A/D converter (10) is then input to the filtering processor (11). The filtering processing unit performs Fourier transform on this signal Bandpass filtering processing is completed.

Effects of the Invention According to the present invention, by reducing noise components included in the photoplethysmogram signal before determining the oxygen saturation of arterial blood, variations in oxygen saturation calculation values are reduced, so each measurement value itself This improves the accuracy of oxygen saturation readings, making them more stable and more reliable.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the overall configuration of a photoplethysmographic oximeter according to a first embodiment of the present invention, FIG. 2 is a flowchart showing the operation of the photoplethysmographic oximeter according to the first embodiment, and FIG. FIG. 3 is a graph showing the pulse wave signal output from the logarithmic conversion section where i≠half, and FIG. 5 is a graph showing the signal after calculation ruttering processing based on the pulse wave signal shown in FIG. 4.
FIG. 7 is a diagram showing the acidity of arterial blood calculated based on the signals shown in FIG. 6, and FIG. 9 is a block 1 diagram showing the overall configuration of a photoplethysmographic oximeter according to a second embodiment of the present invention. Figure 10: Photoplethysmographic conversion unit according to the second embodiment, 5...Pulse rate detection unit, 6...Control unit, 7...B
, 1m, F, section, 8... calculation section, 9... display section, 1
0: A/D conversion unit; 11: filtering processing unit.

Claims (1)

[Claims] 1. Receive light of at least two wavelengths that has passed through the living body, obtain a photoplethysmogram signal that pulsates due to absorption by arterial blood in the living body, and detect the arterial blood based on this photoplethysmographic signal. A photoplethysmographic oximeter for measuring the oxygen saturation of a person, which includes a means for detecting pulse, a means for frequency filtering the photoplethysmographic signal, and a frequency filtering means for frequency filtering based on the pulse detecting means. 1. A photoplethysmogram oximeter, comprising a control means for changing a band, and determining the oxygen saturation of arterial blood based on the frequency-filtered photoplethysmogram signal. 2. The above-mentioned pulse detection means includes means for detecting the frequency of the pulse, and the control means includes means for shifting the pass frequency band of the filtering means based on the frequency of the pulse. The photoplethysmographic oximeter according to item 1. 3. The pulse detection means includes means for determining reliability of the detected pulse, and the control means includes means for changing the pass frequency bandwidth of the filtering means based on the determined reliability. A photoplethysmographic oximeter according to claim 1 or 2. 4. The photoplethysmogram oximeter according to any one of claims 1 to 3, wherein the pulse detection means detects the pulse based on a photoplethysmogram signal. 5. The photoplethysmogram according to any one of claims 1 to 3, wherein the pulse detection means detects the pulse based on a signal source different from the photoplethysmogram signal. oximeter. 6. The photoplethysmographic oximeter according to any one of claims 1 to 5, wherein the filtering means includes a bandpass filter. 7. The photoplethysmographic oximeter according to any one of claims 1 to 5, wherein the filtering means includes a high-pass filter. 8. The photoplethysmographic oximeter according to any one of claims 1 to 5, wherein the filtering means includes a low-pass filter. 9. Means for performing A/D conversion on the photoplethysmographic signal, and the filtering means converts the A/D
The photoplethysmogram oximeter according to any one of claims 1 to 5, characterized in that it includes a digital processing circuit that removes unnecessary frequency components contained in the converted photoplethysmogram signal. .