WO2017143581A1

WO2017143581A1 - Fetal monitoring band

Info

Publication number: WO2017143581A1
Application number: PCT/CN2016/074635
Authority: WO
Inventors: 尤佳; 陶肖明; 李钦; 郭振华; 陈乔珠
Original assignee: 香港纺织及成衣研发中心有限公司
Priority date: 2016-02-23
Filing date: 2016-02-26
Publication date: 2017-08-31
Also published as: CN107095646A

Abstract

A fetal monitoring band, comprising a carrier that can be worn on a pregnant woman's belly and a sensing pad provided on the carrier for sensing fetal movement information of a fetus in the pregnant woman's belly. The sensing pad comprises a sensor for sensing a fetal status signal and a control card for performing noise-removal and signal-enhancement on the fetal status signal to obtain a fetal movement signal. The fetal monitoring band adopts a new signal analysis algorithm to realize an accurate recognition of the fetal movement signal and further realize an accurate evaluation of fetal health.

Description

Fetal monitoring belt

Technical field

The invention relates to the field of fetal monitoring, and in particular to a fetal monitoring belt.

Background technique

Fetal monitoring is to determine the fetal state by identifying signals such as fetal movements, and plays an important role in determining the health of the fetus during pregnancy. Fetal movement is the oldest and most common means of assessing fetal health; the reduction or disappearance of fetal movement is a warning sign of fetal injury or death.

The mother can feel the fetal movement, and the mother's counting of the fetal movement is easy to implement. However, the counting process is subjective and not suitable for long-term monitoring. In this way, various sensors and devices have emerged for objective monitoring of fetal movement, enabling electronic fetal monitoring (EFM) to be widely used in practice.

At present, the conventional Doppler fetal monitor is generally used to determine the fetal heart rate by ultrasound, and the fetal movement is detected during the antenatal care of the pregnant woman. However, the prior art has some drawbacks:

1) Obstetricians and gynaecologists need to use fetal monitoring equipment in a clinical setting, which limits fetal monitoring equipment to technical support, mobility, cost, time, etc.

2) Due to the effects of radiation, operating procedures, costs, etc., pregnant women cannot frequently use fetal monitoring equipment at home.

3) At present, there is no fetal monitoring device that can meet all the requirements of safety, reliability, waterproof, non-radiative, convenient, comfortable, automatic detection and data interaction sharing.

Summary of the invention

The present invention is directed to the above existing problems, and provides a wearable and washable fetal monitoring belt which can automatically detect abnormal fetal movement, thereby reducing the incidence of fetal death, perinatal mortality and risk of pregnant women. It is safe, real-time, convenient, reliable, flexible and efficient.

The technical solution proposed by the present invention is as follows:

The invention provides a fetal monitoring belt comprising a carrier wearable on a pregnant woman's abdomen and a sensor pad arranged on the carrier for sensing a fetal movement signal of a fetus in the belly of the pregnant woman; the sensor pad comprises a sensor for sensing a fetal state signal And for denoising and signal enhancement of fetal status signals, thereby Get the control card for the fetal movement signal.

In the above fetal monitoring device of the present invention, the process of denoising and signal enhancing the fetal state signal by the control card includes the following steps:

S1, normalizing the pressure amplitude of the fetal state signal;

S2, using a fast Fourier transform algorithm to find a first frequency having a maximum amplitude in the fetal state signal and a second frequency having a second largest amplitude; and using a bandpass filter based on a second derivative of the Gaussian function to enhance the first a first input signal of a frequency and a second input signal of a second frequency; and then determining whether the first input signal and the second input signal have periodicity, respectively, and if one of the input signals has periodicity, having a period in the fetal state signal The sexual input signal is removed to obtain a fetal movement signal;

S3. Performing a scale product on the fetal movement signal to enhance the signal of the fetal movement signal; wherein the scale product of the fetal movement signal is defined as the product of the response output of the fetal movement signal filtered by two Gabor filters having different frequencies respectively.

In the above fetal monitoring device of the present invention, the sensing pad includes a sensor array for mapping a set of signals sensed by the sensor array to a fetal state signal using a Sigmoid function.

In the above fetal monitoring belt of the present invention, the sensor is a fabric pressure sensor, and the fabric pressure sensor comprises an elastic fabric base and a conductive layer; the elastic fabric base is used for generating a pressure change accompanying the undulation of the pregnant woman's abdomen, thereby changing the conductive layer. Conductive properties to allow the sensor to sense fetal status signals.

In the above fetal monitoring device of the present invention, the sensor is an acceleration sensor or an acoustic sensor.

The fetal monitoring belt provided by the invention overcomes the defects of the prior art in fetal monitoring, and can be applied not only in a clinical environment but also in general health care services. The fetal monitoring belt of the invention also adopts a novel signal analysis algorithm to realize accurate identification of the fetal movement signal, thereby realizing accurate assessment of the fetal health state.

DRAWINGS

The present invention will be further described below in conjunction with the accompanying drawings and embodiments, in which:

Figure 1 is a diagram showing the variation of the imaginary part value of a Gabor filter at a frequency;

2 is a diagram showing changes in the imaginary part value of a Gabor filter at another frequency and using the same Gaussian factor standard deviation σ as the Gabor filter of FIG. 1;

Figure 3 shows the same high frequency at the same frequency and using the same Gabor filter as Figure 1. Schematic diagram of the change of the imaginary part value of the Gabor filter when the standard deviation σ is sigma;

FIG. 4 shows a calculation effect diagram of a scale product for improving a signal-to-noise ratio (SNR);

FIG. 5 is a flowchart showing the operation of the adaptive bandwidth filtering for denoising proposed by the present invention; FIG.

Figure 6 is a circuit diagram of an acceleration sensor of the control card of the present invention;

Figure 7 is a circuit diagram of a resistive touch screen of the control card of the present invention;

Figure 8 is a circuit diagram of an LCD screen of the control card of the present invention;

Figure 9 is a circuit diagram of a control transistor of the control card of the present invention;

Figure 10 is a circuit diagram of the memory of the control card of the present invention;

11 is a circuit diagram of a clock circuit of a control card of the present invention;

Figure 12 is a circuit diagram of the bell of the control card of the present invention;

Figure 13 is a circuit diagram of a reset switch of the control card of the present invention;

14 is a schematic structural view of a sensor array in an inductive pad;

15 is a circuit diagram of a sensor array in an inductive pad;

Figure 16 is a schematic diagram of an ideal situation of a fetal state signal;

17 is a schematic diagram of acquiring a fetal state signal according to the present invention;

18 is a flow diagram of using a Sigmoid function to map a set of signals sensed by a sensor array to a fetal state signal;

19 is a mapping diagram of a Sigmoid function according to the present invention;

Figure 20 shows the fetal status signal in the simulation test;

Figure 21 shows the normalized fetal state signal in the simulation test;

Figure 22 shows the fetal movement signal obtained after denoising the fetal state signal in the simulation test;

Figure 23 shows a data distribution plot for a preliminary clinical trial.

detailed description

The fetal monitoring device of the present invention comprises a carrier wearable on the abdomen of the pregnant woman and a sensor pad disposed on the carrier for sensing a fetal movement signal of the fetus in the belly of the pregnant woman; the sensor pad includes a sensor for sensing a fetal state signal, and A control card for denoising and signal enhancement of the fetal state signal to obtain a fetal movement signal.

Here, the sensor is sensitive to pressure and/or motion and/or sound, so that it can sense the fetal movement signal. Number and other interfering signals (mainly fetal respiratory signals, etc.), and these signals are all fetal status signals. The control card is used to remove interfering signals in the fetal status signal while preserving the fetal movement signal. In this way, the user can determine the health status of the fetus based on the accurate fetal movement signal.

Specifically, the sensor may be a fabric pressure sensor comprising an elastic fabric substrate and a conductive layer; the elastic fabric substrate is capable of generating a pressure change accompanying the undulation of the pregnant woman's abdomen, thereby changing the conductive property of the conductive layer to enable the sensor A fetal status signal is sensed. In other embodiments, the sensor can also be an acceleration sensor or a sound sensor.

The fetal movement signal has a wide bandwidth, so the denoising process of the fetal state signal is very important for the accuracy of the obtained fetal movement signal. Thus, the present invention proposes an adaptive bandpass filtering algorithm based on a control card to achieve denoising and signal enhancement of fetal state signals.

The Gabor filter is an excellent bandpass filter for signal enhancement. Considering the function of the Gabor filter, the present invention proposes an adaptive bandpass filter obtained by extending a 1DGabor filter with an adaptive interval. algorithm. The conventional 1DGabor filter is defined as:

g(t)=g _e (t)+ig _o (t)

among them,

Here, g(t) is a function value of the Gabor filter; σ is a standard deviation of a Gaussian factor of the Gabor filter; t is time; f ₀ is a frequency.

The imaginary part g ₀ (t) of the function value g(t) of the Gabor filter can be used as a band-pass filter to enhance the signal of a specific bandwidth and smooth the signal by the frequency f ₀ , as shown in Figure 1-3 Show. Here, FIGS. 1-3 show imaginary part values at different frequencies and using the same Gaussian factor standard deviation σ.

Further, it is very important to select the appropriate bandwidth for band pass filtering. Unlike existing algorithms that employ predefined bandwidths, the present invention uses the fetal motion signal f(t) to be substituted into the scaled product of its matched filter to determine the optimal bandwidth. Specifically, the response output of the fetal motion signal f(t) filtered by the Gabor filter of the specific frequency i is defined as:

Where f(t) is a fetal movement signal;

Is the value of g(t) when the frequency f ₀ is equal to i.

Further, the scale product of the fetal movement signal f(t) is defined as the product of the response output of the fetal movement signal f(t) filtered by two Gabor filters having adjacent frequencies i and j, respectively:

P ^i,j (t)=R ⁱ (t)R ^j (t)

Where R ⁱ (t) is the response output of the fetal motion signal f(t) filtered by a Gabor filter with a frequency f ₀ equal to i; R ^j (t) is the fetal motion signal f(t) with a frequency f ₀ equal to j The output of the response after filtering by the Gabor filter.

Fig. 4 shows a calculation effect diagram for improving the scale product of the signal-to-noise ratio (SNR), wherein the meanings of the signals in the rows in Fig. 4 are as follows:

Line 1: input signal s for testing;

Line 2: a signal f after the noise signal is added to the input signal s;

Line 3-5: response output results R ₁ , R ₂ and R _{3 of} signal f filtered by Gabor filters of different frequencies;

Line 6: a maximum map of R ₁ , R ₂ and R ₃ ;

Line 7: scale product P _1, ₂ of R ₁ and R ₂ ;

Line 8: Scaled product of R ₂ and R ₃ P _2,3 .

After the calculation of the scale product, it can be seen that the noise signal is smoothed, and the characteristic peak of the input signal s still exists. This means that denoising of fetal status signals using a Gabor filter is feasible.

Here, the adaptive band pass filtering algorithm proposed by the present invention is implemented by a specific circuit which is embedded in the control card. FIG. 5 is a flowchart showing the operation of the adaptive filtering for denoising proposed by the present invention; FIG. 6 to FIG. 13 are circuit diagrams for implementing the algorithm; wherein the circuit structure includes FIG. The illustrated acceleration sensor, the resistive touch screen shown in FIG. 7, the LCD screen shown in FIG. 8, the control transistor shown in FIG. 9, the memory shown in FIG. 10, the clock circuit shown in FIG. 11, The bell shown in Fig. 12 and the reset switch shown in Fig. 13.

As shown in FIG. 5, the process of denoising and signal enhancing the fetal status signal by the control card includes the following steps:

S1, normalizing the pressure amplitude of the fetal state signal;

In this step, the maximum pressure of the fetal state signal sensed by the sensor pad is about 50g; The pressure is in a similar range. As a rule of thumb, the baseline can be set to 300 g and the amplitude of 350 g is used to normalize the fetal status signal.

S2, using a fast Fourier transform algorithm (FFT) to find a first frequency having a maximum amplitude and a second frequency having a second largest amplitude in the fetal state signal; and using a bandpass filter based on a second derivative of the Gaussian function A first input signal of a first frequency and a second input signal of a second frequency are enhanced; and then respectively determining whether the first input signal and the second input signal have periodicity, and if one of the input signals has periodicity, the fetal state signal is The periodic input signal is removed to obtain a fetal movement signal;

In this step, the bandwidth of the human body breathing is in the range of 0.2 Hz - 0.8 Hz, and the bandwidth of the fetal movement signal is in the range of about 0.5 Hz - 10 Hz. The human body has a periodicity, so that the periodic input signal in the fetal movement state signal is the human body breathing signal, and after the human body breathing signal is removed, the fetal movement signal is obtained.

S3. Performing a scale product on the fetal movement signal to denoise and enhance the fetal movement signal.

In this step, the smoothing of the fetal movement signal is also achieved by testing 10 adjacent sample points, and then the fetal product is scaled to achieve signal enhancement.

The applicant also performed a linearity test on the fetal movement signals induced by the two existing sensor pads; here, the data was collected from the subject. Table 1 lists the test results.

Table 1 Linearity test of two sensor pads

	100g～300g100g～300g	(300+)10g～(300+)50g(300+) 10g～(300+)50g
感应垫1的线性度误差Linearity error of sensor pad 1	8.3％8.3%	37.1％37.1%
感应垫2的线性度误差Linearity error of the sensor pad 2	17.5％17.5%	25.6％25.6%

It can be seen through experiments that the linearity error of the existing induction pad is large, which is caused by the layout of the sensor array and the supporting material in the sensing pad. Since the data we use to analyze is the change in the test data rather than the absolute value of the test data, such large errors have no significant effect on the test results. However, if the linearity error is reduced, the data will be more reliable.

In order to more reliably identify the fetal movement signal, Applicants used a Competitive Algorithm to determine the position of the sensor array node, which is most helpful in changing the force change between adjacent nodes. Applicants use the Sigmoid function to map the set of signals sensed by the sensor array to the tire The child state signal, as shown in FIG. 18, satisfies a linear relationship within a certain range. Figures 14-16 illustrate the structure of the sensor array in the inductive pad, the circuitry of the sensor array, and the ideal state of the fetal status signal, respectively. 17-19 illustrate the processing of data sensed by the sensor.

Further, the applicant has simulated the denoising effect of the fetal monitoring tape of the present invention, wherein fetal movement signals and other interference signals (human body motion, breathing, etc.) can be simulated by using a specific design water channel. Gaussian noise can be generated by using the maximum amplitude of the fetal motion. Figure 20 shows the fetal state signal in the simulation test; Figure 21 shows the normalized fetal state signal in the simulation test; Figure 22 shows the fetal movement signal obtained after the fetal state signal is denoised in the simulation test. The SNR in the simulation test is recorded in Table 2.

Table 2 SNR record

	胎儿状态信号Fetal status signal	胎动信号Fetal movement signal
SNR(dB)SNR (dB)	12.612.6	140.1140.1

Further, the applicant also conducted preliminary clinical trials at the Guangzhou Women&Children Medical Center; 6 groups of volunteers participated in the trial; the experiment included automatic fetal movement counting and voluntary acceptance of the fetal monitoring belt of the present invention. Manual counting of the tester. Figure 23 shows the data distribution of the preliminary clinical trial, in which Figure 23 shows nine signal channels in sequence; the fetal movement signal is captured and recorded in the seventh signal channel and the eighth signal channel, and is voluntarily subject to The marker of the sample of the tester is recorded in the ninth signal channel. The fetal movement monitoring of the present invention has an accuracy of 85.7%, and a competitive algorithm and an integrated learning algorithm can obtain better results.

It is to be understood that those skilled in the art will be able to make modifications and changes in accordance with the above description, and all such modifications and variations are intended to be included within the scope of the appended claims.

Claims

A fetal monitoring belt characterized by comprising a carrier wearable on a pregnant woman's abdomen and a sensor pad disposed on the carrier for sensing a fetal movement signal of a fetus in a pregnant woman's abdomen; the sensor pad includes a sensor for sensing a fetal state signal And a control card for denoising and signal enhancement of the fetal status signal to obtain a fetal movement signal.
The fetal monitoring device of claim 1 wherein the process of denoising and signal enhancing the fetal status signal by the control card comprises the steps of:

S1, normalizing the pressure amplitude of the fetal state signal;

S2, using a fast Fourier transform algorithm to find a first frequency having a maximum amplitude in the fetal state signal and a second frequency having a second largest amplitude; and using a bandpass filter based on a second derivative of the Gaussian function to enhance the first a first input signal of a frequency and a second input signal of a second frequency; and then determining whether the first input signal and the second input signal have periodicity, respectively, and if one of the input signals has periodicity, having a period in the fetal state signal The sexual input signal is removed to obtain a fetal movement signal;

S3. Performing a scale product on the fetal movement signal to enhance the signal of the fetal movement signal; wherein the scale product of the fetal movement signal is defined as the product of the response output of the fetal movement signal filtered by two Gabor filters having different frequencies respectively.
The fetal monitoring device of claim 1 wherein the sensing pad comprises a sensor array for mapping a set of signals sensed by the sensor array to a fetal status signal using a Sigmoid function.
The fetal monitoring belt according to claim 1, wherein the sensor is a fabric pressure sensor comprising an elastic fabric substrate and a conductive layer; and the elastic fabric substrate is adapted to generate pressure changes accompanying the undulation of the pregnant woman's abdomen , thereby changing the electrical conductivity of the conductive layer to cause the sensor to sense the fetal state signal.
The fetal monitoring belt according to claim 1, wherein the sensor is an acceleration sensor or an acoustic sensor.