KR20100008239A

KR20100008239A - Eliminating method of motion artifact from ppg signal

Info

Publication number: KR20100008239A
Application number: KR1020080068702A
Authority: KR
Inventors: 박경철; 윤갑열; 이건기; 이주원; 이한욱; 정원근
Original assignee: (주)에스엔티
Priority date: 2008-07-15
Filing date: 2008-07-15
Publication date: 2010-01-25

Abstract

PURPOSE: A motion artifact removal method of PPG signals is provided to precisely detect PPG(Photo-plethysmography) signals by effectively filtering motion artifact from the PPG signals although the amplitude of motion artifact is high. CONSTITUTION: Periodically, PPG input signals in which waveform repeats are inputted on a real time basis. The whole inputted PPG input signals are successively separated according to each period and a plurality of division input signals are produced. Plural division inputs are accurately separated from the PPG input signals. Output signals composed of samples in which motion artifact is filtered are produced. The PPG output signals in which motion artifact is filtered are outputted on a real time basis by connecting the output signals according to the pass of the time.

Description

Eliminating method of motion artifact from PPG signal

The present invention relates to a method for removing noise from a PPG signal, and more particularly, to a method of removing noise from a PPG signal that can effectively filter out the noise from the PPG signal even if the amplitude of the mixed noise in the PPG signal is large. .

Currently, technology development for the construction of ubiquitous health care environment is in full swing, and medical equipment is also being miniaturized to be carried in the hands or worn as clothes.

On the other hand, PPG signal (Photo-plethysmography signal) is an optical signal for detecting the pulse wave generated by the pulse transmitted by the heart beat to the peripheral nerve, it can estimate the current state of blood vessels and the amount of blood. The PPG signal, unlike the conventional ECG signal (Electrocardiogram signal) which requires the installation of a complex detection device for signal detection, has a predetermined number of LEDs that can be easily attached to a single contact surface of the body. As it can be detected through a simple pulse wave detector consisting of a photodetector, it has been spotlighted as one of the vital bio signals to be monitored in the ubiquitous health care environment.

As a technique for measuring such a PPG signal, Korean Patent Laid-Open Publication No. 10-2008-0009030 "Pulsometer to wear on the wrist and related methods" and the like have been devised.

Here, as the PPG signal is an optical signal, various motion artifacts due to respiration, body movement, and the like can be easily mixed, so that accurate detection of the PPG signal is required above all.

However, the PPG signal generally has a frequency band of 0.5 to 4 Hz, and the body dynamic noise generally has a frequency band of 0.04 to 1.6 Hz, so that the PPG signal overlaps with the frequency band of the dynamic noise. When the filter is used, a part of the PPG signal is also removed along with the dynamic noise, causing distortion of the PPG signal.

In response to this, a technique for removing dynamic noise by applying an adaptive filer, a wavelet, and a filter bank to a PPG signal has been devised. In the case of removing A, when the amplitude of the dynamic noise mixed with the PPG signal is large, there is a problem in that the dynamic noise is not properly removed.

Accordingly, the present invention improves such a problem of the prior art, and after matching the number of samples through decimation or interpolation of the separated PPG signal for each period, and then each time period PPG A new form that can smoothly remove the noise even if the amplitude of the noise mixed in the PPG signal is large by filtering the noise by using a signal average method of extracting and averaging the same sequence of samples from the signal. An object of the present invention is to provide a method of removing noise of a PPG signal.

According to a feature of the present invention for achieving the above object, the present invention, in the method for removing the dynamic noise according to the body motion from the body's PPG signal detected from the pulse wave detector installed in the skin tissue of the body, A divisional input signal generation step of receiving the repeated PPG signal in real time and generating a plurality of divisional input signals by sequentially separating all the input PPG signals for each period during a predetermined time interval including a current time; Each split input signal generated in the split input signal generating step is subjected to the same sample using any one of a signal processing method selected from decimation and interpolation, which are multirate digital signal processing. A split input signal sampling step of sampling with a number; Generates an output signal that generates an output signal consisting of samples with dynamic noise filtering by sequentially applying a signal average method of extracting and averaging the same sequence of samples from each divided input signal to the samples forming the divided input signals. Including the step, through the step of generating the divided input signal, the step of sampling the input signal, the output signal generating step by connecting the output signal sequentially generated over time so that the PPG signal filtered by the noise is output in real time Characterized in that.

In the method of removing the dynamic noise of the PPG signal according to the present invention, the split input signal generating step sequentially performs decimation and interpolation signal processing at a predetermined ratio with respect to the entire PPG input signal during a predetermined time interval including the current time. Apply to generate a PPG periodic signal with a simplified waveform, detect the maximum and minimum values corresponding to the floor and valley of the PPG periodic signal sequentially, calculate the median value of the neighboring maximum and minimum values, and then And a time point corresponding to the detection point is allocated from the PPG input signal as a separation point for generating the plurality of divided input signals.

According to the method for removing the dynamic noise of the PPG signal according to the present invention, even if the amplitude of the dynamic noise mixed in the PPG signal is large, it is possible to effectively filter the dynamic noise from the PPG signal, so that the PPG signal is accurately and precisely detected.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings, FIGS. 1 to 3, and like reference numerals denote like elements for performing the same functions in FIGS. 1 to 3. Meanwhile, in the drawings and the detailed description, the illustration and the description of the construction and the operation which can be easily understood by those skilled in the art from the general method of removing the noise of the PPG signal are briefly or omitted. In particular, in the drawings and detailed description of the drawings, detailed descriptions and illustrations of specific technical configurations and operations of elements not directly related to technical features of the present invention are omitted, and only the technical configurations related to the present invention are briefly shown or described. It was. In addition, although the size ratio between elements is somewhat different in the drawings of the drawings, or the size between the parts that are coupled to each other is expressed differently, the representation differences in these drawings can be easily understood by those skilled in the art. As they are parts, a separate description is omitted.

1 is a block diagram showing the steps of the method of removing the noise of the PPG signal according to the present invention, Figure 2 is a conceptual diagram showing the technical idea of the method of removing noise of the PPG signal according to an embodiment of the present invention, (A) of FIG. 3 is a time series graph of the PPG signal with no dynamic noise, and FIG. 3 (b) is a time series graph of the signal with the same noise mixed with the PPG signal of FIG. 3 (a). c) is a time series graph obtained by filtering the PPG signal mixed with the noise shown in FIG. 3 (b) with an adaptive filter, and FIG. 3 (d) shows the PPG signal mixed with the noise shown in FIG. 3 (b). Is a time series graph filtered using the method for removing noise according to the present invention.

The method for removing the dynamic noise of the PPG signal according to the present invention is a method for removing motion artifacts caused by movement of the body from a photo-plethysmograph (PPG) signal measuring pulse waves. Such a method of removing the noise of the PPG signal according to the present invention is schematically illustrated in FIG. 2, wherein a PPG signal having a signal characteristic in which similar waveforms are periodically repeated as shown in FIGS. 3A and 3B is illustrated. The signal average is separated by each period and the same number of samples is extracted from each sampled PPG signal by decimation or interpolation to match the number of samples. It is a technical feature to filter the noise by using the method. Accordingly, even if the amplitude of the dynamic noise mixed with the PPG signal is large, the dynamic noise is smoothly removed.

In the method of removing the noise of the PPG signal according to the present invention, as shown in FIG. 1, the PPG input signal 10 inputted in real time from the pulse wave detector installed in the skin tissue of the body is divided into a signal generation step, a split input signal sampling step, and an output signal. It is filtered through the generation stage so that the noise is removed.

The divisional input signal generation step is a step of generating a plurality of divisional input signals 12 by sequentially dividing the entire PPG input signal 10 inputted from a pulse wave detector to a present time by each period.

In this case, as the PPG signal has a signal characteristic in which similar waveforms are periodically repeated, the divided input signals 12 having similar waveforms may be generated by separating the respective waveforms as described above. The plurality of split input signals 12 generated as described above are used as data for a signal average to generate an output signal at the current time (cycle of one unit) from which the noise is removed.

As described above, in order to continuously generate the output signal at the current point in time (a unit of one unit), the time-series range of the PPG input signal 10 separated for each cycle is inputted during the predetermined time interval including the current point in time. The entire input signal 10 is targeted.

That is, the method of removing the noise of the PPG signal according to the present invention is to apply a signal processing method of a moving average filter.

Here, in the generating of the divided input signal according to the preferred embodiment of the present invention, the plurality of divided input signals 12 may be accurately separated at each period from the PPG input signal 10 input during the predetermined time interval including the current time point. The signal processing for this will be described in detail with reference to FIG. 2 as follows.

First, the PPG input signal 10 is minimized so as to have a simplified waveform characteristic as a periodic signal by performing decimation signal processing at a predetermined rate with respect to the input PPG input signal 10 during a predetermined time interval including the current time point. After sampling (sample 20a in FIG. 2) by the number of samples, the interpolation signal processing is performed at the same rate as the decimated ratio to have the same number of samples as the original PPG input signal 10. As a signal, a PPG periodic signal (reference numeral 20b of FIG. 2) having a simplified waveform is generated.

Next, the maximum and minimum values corresponding to the floor and the valley forming the waveform of the PPG periodic signal 20b generated as described above are sequentially detected, and after calculating the intermediate values of the neighboring maximum and minimum values, respectively, The PPG input signal 10 detects a time point corresponding to each intermediate value calculated in the PPG input signal 10 and assigns a time point corresponding to the intermediate value as a period separation point for generating a plurality of divided input signals. Separate each cycle.

Through the signal processing step described above, the plurality of split input signals 12 are accurately separated from the PPG input signal 10.

The split input signal sampling step is a step of sampling each split input signal 12 generated in the split input signal generation step with the same number of samples. To this end, in the divided input signal sampling step according to the present invention, multirate digital signal processing is performed on each of the divided input signals 12. Such far-rate digital signal processing is performed at a sampling frequency. By changing the, we distinguish between decimation for lowering the sampling frequency and interpolation for increasing the sampling frequency.

Here, in the split input signal sampling step according to the present invention, the divided input signal 12 having a higher sample number than the predetermined sample number is determined based on a predetermined number of samples, and then decimated, and the predetermined sample is used as a reference. The divided input signals 12 having a lower number of samples are interpolated so that each divided input signal 12 samples the same number of samples.

In this way, as each divided input signal 12 is sampled with the same number of samples, a signal average process for the samples of each divided input signal 12 is possible in a next step (output signal generation step). do.

In the output signal generating step, a signal average method is sequentially applied to the samples constituting each of the divided input signals 12 sampled with the same number of samples as described above, thereby generating an output signal composed of the samples from which the noise is filtered. It's a step.

That is, the samples are taken in the same order from each of the divided input signals 12 (for example, the samples corresponding to the sixth sample from each of the divided input signals 12) are summed, and then the average value is calculated. The output signal sampled from the samples generated by sequentially applying the entire sample of each split input signal 12 is generated.

The generated output signal is an average of samples of the plurality of divided input signals 12 including the divided input signals 12 in which the dynamic noise is not mixed as described above. And removed.

Here, the output signal generated in the output signal generating step is an output signal of the current time point (cycle of one unit) from which the noise is removed from the PPG input signal currently input from the pulse wave detector. The noise elimination method is such that the PPG input signals continuously input over time are sequentially generated through the split input signal generation step, the split input signal sampling step, and the output signal generation step. By connecting the generated output signal, the PPG output signal from which the noise is filtered is output in real time.

The dynamic noise removal efficiency according to the dynamic noise removal method of the PPG signal according to the present invention is filtered using the dynamic noise removal method according to the present invention, the PPG signal mixed with the dynamic noise shown in FIG. 3 (d) shows a time series graph, which is compared with FIG. 3 (c) showing a time series graph in which the PPG signal mixed with the noise shown in FIG. 3 (b) is filtered by an adaptive filter. As a result, the PPG signal using the conventional dynamic noise removing method using the adaptive filter is not properly removed, while the PPG signal applying the dynamic noise removing method of the PPG signal according to the present invention has the perfect dynamic noise. You can see it removed.

As described above, the method for removing the noise of the PPG signal according to the embodiment of the present invention has been shown in accordance with the above description and the drawings, but this is only an example and various modifications may be made without departing from the spirit of the present invention. It will be understood by those skilled in the art that variations and modifications are possible.

1 is a block diagram showing the steps of a method for removing noise in a PPG signal according to the present invention;

2 is a conceptual diagram illustrating a technical idea of a method for removing a noise of a PPG signal according to a preferred embodiment of the present invention;

Figure 3 (a) is a time series graph of the PPG signal is not mixed with noise;

3B is a time series graph of a signal in which dynamic noise is mixed with the PPG signal of FIG. 3A;

FIG. 3 (c) is a time series graph of filtering a PPG signal including dynamic noise shown in FIG. 3 (b) by an adaptive filter;

FIG. 3D is a time series graph of filtering the PPG signal including the dynamic noise shown in FIG.

* Description of the symbols for the main parts of the drawings *

10: PPG input signal 12: split input signal

20a, 20b: PPG periodic signal 30: PPG output signal

32: output signal

Claims

In the method for removing the dynamic noise according to the body movement from the PPG input signal of the body detected from the pulse wave detector installed in the skin tissue of the body,

Generates a split input signal that receives a PPG input signal in which a waveform is periodically repeated in real time, and generates a plurality of split input signals by sequentially separating all the input PPG input signals for each period sequentially for a predetermined time interval including a current time point. Steps;

Each split input signal generated in the split input signal generating step is subjected to the same sample using any one of a signal processing method selected from decimation and interpolation, which are multirate digital signal processing. A split input signal sampling step of sampling with a number;

Generates an output signal that generates an output signal consisting of samples with dynamic noise filtering by sequentially applying a signal average method of extracting and averaging the same sequence of samples from each divided input signal to the samples forming the divided input signals. Including the steps,

Through the split input signal generation step, the split input signal sampling step, and the output signal generation step, the output signals sequentially generated over time are connected to output the PPG output signal filtered with the noise in real time. Dynamic noise cancellation method of PPG signal.

The method of claim 1,

The divisional input signal generating step generates a PPG periodic signal having a simplified waveform by sequentially applying decimation and interpolation signal processing at a predetermined ratio with respect to the entire PPG input signal input during a predetermined time interval including a current time point. Detecting the maximum value and the minimum value corresponding to the floor and the valley of the PPG periodic signal sequentially, calculating the intermediate value between the neighboring maximum value and the minimum value, and detecting the time point corresponding to the intermediate value from the PPG input signal. A method of removing noise from a PPG signal, characterized by allocating split points for generating two split input signals.