CN114767064B - Child sleep monitoring method, system and electronic device - Google Patents

Abstract

The invention discloses a child sleep monitoring method, which belongs to the field of medical signal processing, and comprises the steps of installing three-axis acceleration sensors, collecting information, processing respiratory signals of each three-axis acceleration sensor, processing heart rate signals of each three-axis acceleration sensor, comprehensively judging sleep conditions and the like, wherein the sensors are arranged at four positions of two upper arms, a chest and an abdomen, and when sleeping posture changes to press one sensor, other sensors can work continuously to prevent false alarms; by arranging the sensors on the clothes corresponding to the chest and the abdomen, the respiratory problem caused by sleep blockage can be monitored; early warning is carried out on life-threatening events, and general negative events are reminded and recorded; has good resolving power to sleeping gesture and low false alarm. The invention also relates to a child sleep monitoring system and a device for implementing the child sleep monitoring method.

Description

Children's sleep monitoring method, system and electronic device

Technical field

The present invention relates to the field of medical monitoring, and in particular to children's sleep monitoring methods, systems and electronic devices.

Background technique

Sleep is very important for physical and mental health, especially for children during their growth and development period. But there are currently huge problems with children's sleep monitoring. Because children's wrists are slender, wearable watches are easy to slip off and leak light; children's heart power is small, their heart rate is fast, and their sleeping posture is unstable. Mattress-type sleep monitoring systems are prone to miss signals, and children often sleep with adults, so children's heartbeat signals Being submerged and disturbed by adults; the camera/radar sleep monitoring system is also affected by sleeping posture, and its accuracy is limited when covered with thick quilts.

Contents of the invention

In order to overcome the shortcomings of the existing technology, one of the purposes of the present invention is to provide a children's sleep monitoring method that is not affected by sleeping posture and environment and has high measurement accuracy.

In order to overcome the shortcomings of the existing technology, the second object of the present invention is to provide a children's sleep monitoring system that is not affected by sleeping posture and environment and has high measurement accuracy.

In order to overcome the shortcomings of the prior art, the third object of the present invention is to provide a children's sleep monitoring device that is not affected by sleeping posture and environment and has high measurement accuracy.

One of the purposes of the present invention is achieved by adopting the following technical solutions:

A child sleep monitoring method includes the following steps:

Install three-axis acceleration sensors: Install four three-axis acceleration sensors on the outside of the clothing on the upper arms, chest and abdomen;

Collect information: Each of the three-axis acceleration sensors collects respiration and heart rate signals:

Process the respiratory signal of each three-axis acceleration sensor: filter the respiratory signal of each three-axis acceleration sensor, perform Fourier transform on the filtered signal, and when there is a respiratory wave in the transformed signal , obtain the respiratory waveform through Meyer wavelet transform and record it. When the transformed signal does not contain respiratory waves, it is judged to be in a state of stress, respiratory disorder or body movement through the strength and comparison of the overall signal;

Process the heart rate signal of each three-axis acceleration sensor: filter the heart rate of each three-axis acceleration sensor, and perform Fourier transform on the filtered signal. When there is a heart rate cycle in the transformed signal, Extract the periodic heart rate. When the converted signal does not contain a heart rate period, the state of stress is judged by the strength of the overall signal or the periodicity of the heartbeat is calculated by distinguishing the energy density;

Comprehensive judgment of sleep status: Discard the compressed breathing signals and heart rate signals of the four three-axis acceleration sensors, conduct a comprehensive analysis of the remaining signals, determine whether the child is in a state of irregular body movement, respiratory obstruction, or a normal state, and analyze the results respectively. Provides early warning for irregular body movements and respiratory obstruction, and records normal status.

Furthermore, in the step of comprehensively judging sleep conditions, respiratory obstruction is determined by using the three-axis acceleration sensors of the chest and abdomen. When the thorax expands and the abdomen contracts, it is respiratory obstruction.

Further, in the step of processing the heart rate signal of each of the three-axis acceleration sensors, whether there is a heart rate cycle is determined by whether there is a significant peak value in the converted signal.

Further, in the step of processing the heart rate signal of each of the three-axis acceleration sensors, the periodicity of the heartbeat calculation based on the energy density is specifically: Then the heart rate is attempted to be obtained through information entropy. The calculation formula of entropy is: where I _e is the signal strength at a certain moment, p _i is the probability of occurrence of I _e , and H _s is the entropy value in a specific time interval. Calculate the entropy value within a fixed time through the sliding window, draw the entropy value fluctuation curve, and compare it Periodicity assessment (frequency domain transformation and finding significant peaks). If obvious periodicity can be found, the entropy method is used to obtain the heart rate.

Further, in the step of comprehensively judging sleep conditions, recording the normal state is specifically as follows: the final respiratory signal is obtained by weighting the data of four three-axis acceleration sensors.

Further, in the step of comprehensively judging sleep conditions, recording the normal state is specifically as follows: the final heart rate signal is obtained by weighting the data of the four three-axis acceleration sensors after Kalman filtering.

Furthermore, in the step of comprehensively judging the sleep situation, when the child is in a state of respiratory obstruction, when the respiratory obstruction is frequent but short-lasting, the child is reminded to change the position; when the respiratory obstruction continues to be unrelieved and the heart rate weakens, an early warning is provided.

Further, during the filtering process of the respiratory signal of each of the three-axis acceleration sensors, the retained signal is between 0.13-0.66Hz, that is, 7.8-39.6bpm; during the filtering of the heart rate signal of each of the three-axis acceleration sensors In the process, the first step separates the 4-11Hz signal, and the second step retains the signal at 0.8-3.5Hz, that is, 48-210bpm.

The second object of the present invention is achieved by adopting the following technical solutions:

A children's sleep monitoring system, which is used to implement any of the above children's sleep monitoring methods.

The third object of the present invention is achieved by adopting the following technical solutions:

A child sleep monitoring setup including

Four three-axis acceleration sensors, the four three-axis acceleration sensors are respectively installed on the outside of the children's upper arms, chest and abdomen;

processor;

Memory, the memory is communicatively connected to the processor;

The memory stores data collected by the four three-axis acceleration sensors and instructions that can be executed by the processor. The instructions are executed by the processor to implement any of the above children's sleep monitoring methods.

Compared with the existing technology, the children's sleep monitoring method of the present invention places sensors at four locations on the two upper arms, chest and abdomen. When the sleeping position changes and presses one sensor, the other sensors can continue to work and prevent false alarms; by placing sensors on the chest and abdomen Sensors are installed on the corresponding clothing, which can monitor breathing problems caused by sleep obstruction; have better discrimination ability for apnea, shortness of breath, and respiratory obstruction; provide early warning for life-threatening events, and remind and record general negative events; It has good ability to distinguish sleeping postures (left and right side, prone, etc.), and has low false alarms (compared to radar waves and other means); it has good measurement capabilities for heart rate and pulse intensity, and can self-correct measurements based on sleeping postures; Through the analysis of body movement, breathing, and heart rate, data related to sleep quality are obtained. Since children are light in weight, the technical mode of using a pressure sensor will cause the signal-to-noise ratio to be too low, but the acceleration sensor measurement will not be affected.

Description of the drawings

Figure 1 is a flow chart of the children's sleep monitoring method of the present invention;

Figure 2 is a flow chart for processing the respiratory signal of each three-axis acceleration sensor;

Figure 3 is a flow chart for processing the heart rate signal of each of the three-axis acceleration sensors;

Figure 4 is a flow chart for comprehensively judging sleep status;

Figure 5 is a schematic diagram of the implementation of the children's sleep monitoring method of the present invention;

Figure 6 is a three-dimensional view of the patch;

Figure 7 shows the original data of the respiratory signal;

Figure 8 shows the respiratory waveform after the respiratory signal has been transformed by Meyer wavelet.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

It should be noted that when a component is referred to as being "fixed to" another component, it can be directly on the other component or another intermediate component may be present through which it is fixed. When a component is said to be "connected" to another component, it can be directly connected to the other component or there may be another intermediate component present at the same time. When a component is said to be "disposed on" another component, it can be directly located on the other component or another intervening component may be present. The terms "vertical," "horizontal," "left," "right" and similar expressions are used herein for illustrative purposes only.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which the invention belongs. The terminology used herein in the description of the invention is for the purpose of describing specific embodiments only and is not intended to limit the invention. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.

Figures 1 to 4 are flow charts of the children's sleep monitoring method of the present invention. The children's sleep monitoring method includes the following steps:

Install three-axis acceleration sensors: Install four three-axis acceleration sensors on the outside of the clothing on the upper arms, chest and abdomen;

Collect information: Each three-axis acceleration sensor collects respiration and heart rate signals:

Process the respiratory signal of each three-axis acceleration sensor: filter the respiratory signal of each three-axis acceleration sensor, perform Fourier transform on the filtered signal, and when there is a respiratory wave in the transformed signal, pass Meyer wavelet transform is used to obtain the respiratory waveform and record it (see attached Figure 7-8, which is a comparison of the signal before and after Meyer wavelet transform). When there is no respiratory wave in the transformed signal, it is judged to be in a stressed state by the strength of the overall signal. , in a respiratory disorder or in a state of physical movement;

Process the heart rate signal of each three-axis acceleration sensor: filter the heart rate of each three-axis acceleration sensor, and perform Fourier transform on the filtered signal. When there is a heart rate cycle in the transformed signal, Extract the periodic heart rate. When the converted signal does not contain a heart rate period, the state of stress is judged by the strength of the overall signal or the periodicity of the heartbeat is calculated by distinguishing the energy density;

Comprehensive judgment of sleep status: Discard the compressed breathing signals and heart rate signals of the four three-axis acceleration sensors, conduct a comprehensive analysis of the remaining signals, determine whether the child is in a state of irregular body movement, respiratory obstruction, or a normal state, and analyze the results respectively. Provides early warning for irregular body movements and respiratory obstruction, and records normal status.

The steps for installing the three-axis acceleration sensor are shown in Figure 5. The specific steps are: the three-axis acceleration sensor adopts a patch method (as shown in Figure 6). The x-axis specifically refers to the direction perpendicular to the dial surface perpendicular to the arm, and the y-axis specifically refers to The direction of the patch is parallel to the arm, and the z-axis specifically refers to the direction perpendicular to the surface of the patch. Each patch contains a three-axis acceleration sensor, which can sense the orientation and vibration of the patch. The sampling rate is 50Hz, the measurement range is ±1g, and it is set to 14-bit accuracy.

Four three-axis acceleration sensors are installed on the outside of the clothing on the upper arms, chest and abdomen respectively. In this application, four three-axis acceleration sensors are installed on the outside of the clothes on the upper arms, chest and abdomen respectively. Compared with the wrist, the patch is more suitable for the chest and upper arms. If the patch is placed on the wrist, such as a smart watch, the respiratory data measurement will become worse. The optimal measurement point for breathing is the chest area. In some previous technologies, pressure sensors (such as PVDF, pressure films) were placed on the chest. However, because the pressure sensors need to be close to the skin, the tight fit is not comfortable and affects sleep. The acceleration sensor is not subject to this restriction, and the clothing does not need to be tight. The sensor is placed on the chest, upper arms, and abdomen, and does not restrict the movement of the thorax during breathing. Currently, the acceleration sensor hardware used in smart watches and bracelets can technically achieve high accuracy. However, since daily activities require consideration of strenuous exercise scenarios and data storage capacity, most products are designed with a measurement range as high as ±16g, so the final resolution The rate is only 1mg-16mg, it cannot accurately measure heartbeat intensity, and the signal-to-noise ratio is low in non-close situations. The present invention uses patches only for sleep, and can limit the measurement range to ±1g. It still uses universal 14-bit data collection, and can measure the lowest acceleration of 0.1mg, which can reduce restrictions on the texture of pajamas and reduce the cost of the entire system ( Only patches are required, no special underwear material is required).

Changes in sleeping position will compress the patch, and the accuracy of respiration measurement will decrease at this time. Therefore, the patch needs to be placed in four positions to prevent false alarms. When the chest is compressed, the significance of the respiratory peak can be determined by analyzing the spectrum. When the amplitude of the respiratory wave is obviously suppressed, it jumps to the sensor of the upper arm. The data processing method of the upper arm is the same as that of the chest. In the sleeping state, it is extremely unlikely that the chest, abdomen, and arms on both sides will be pressed at the same time. Therefore, the patented invention can avoid respiratory data abnormalities and false warnings caused by sleeping postures. Another advantage of this design is that when the chest breathing wave is not periodic, the other three patches can be used to determine whether there is physical movement or a synchronized irregular breathing event.

A patch placed on the abdomen is mainly used to monitor breathing problems caused by sleep obstruction. The gold standard for respiratory measurement is oronasal airflow, generally accompanied by cyclic thoracic movements. However, some patients with respiratory obstruction try hard to breathe (chest rises and falls), but the airflow is very small, causing blood oxygen to decrease. One of the methods to distinguish respiratory obstruction is to measure the movement of the chest and abdomen. If the thorax and abdomen expand simultaneously, it is normal breathing. If the thorax expands and the abdomen contracts, it is respiratory obstruction. Because the abdominal patch is only used to identify respiratory obstruction, it does not need to be tightly attached to the skin and does not create any resistance to breathing. Abdominal measurement has another advantage. That is, there are large blood vessels in the abdomen and there is no rib obstruction. When the blood vessels pulse, there is also vibration perpendicular to the abdomen, which can complement the heart vibration transmitted by the body and reduce false alarms. The amplitude of vascular pulsation is larger when it is slightly compressed (refer to the upper arm sphygmomanometer, when the external pressure reaches the average blood pressure, the vascular pulsation amplitude is the largest), so even when prone, it is possible to distinguish whether there is a risk of cardiac arrest. Significantly reduces the false alarm rate for a single chest patch.

The specific steps for collecting information are: measure the breathing rate of each patch independently, because the sleep breathing component is concentrated at 0.13-0.66Hz, that is, 7.8-39.6bpm, and signals outside the frequency band are considered noise. Each patch independently measures heart rate. Children's heart rate is faster than that of adults, so the heart rate window is set to 0.8-3.5Hz, that is, 48-210bpm. It can be individually adjusted according to the user's age. Due to the elastic vibration of the body caused by each heart beat The signal is between 4-11Hz. Before calculating the heart rate, the external interference is eliminated by band-pass filtering of 4-11Hz, and then the second band-pass filtering is performed through the heart rate window (0.8-3.5Hz), which can effectively obtain a highly stable signal. heart rate signal

The specific steps for processing the respiratory signal of each three-axis acceleration sensor are as follows: during the filtering process, the retained signal is between 0.13-0.66Hz, that is, 7.8-39.6bpm; the Fourier transform window is a window of 30s. Whether there is a respiratory wave is judged by whether there is a significant peak value. The overall signal strength is judged by whether the maximum amplitude is less than 25mg.

The specific steps for processing the heart rate signal of each of the three-axis acceleration sensors are as follows: in the filtering process, the first step is to separate the 4-11Hz signal, eliminate external interference through 4-11Hz bandpass filtering, and the second step is to retain the signal At 0.8-3.5Hz, that is 48-210bpm. The window of Fourier transform is a window of 10s. Whether there is a heart rate cycle is judged by whether there is a significant peak. The overall signal strength is judged by whether the maximum amplitude is less than 2mg. The specific calculation periodicity of heartbeat calculation through energy density is: Then try to obtain the heart rate through information entropy. The calculation formula of entropy is:

Calculate the entropy value within a fixed period of time through the sliding window, draw the entropy value fluctuation curve, and evaluate its periodicity. If obvious periodicity can be found, the entropy method is used to obtain the heart rate.

Comprehensive judgment of sleep conditions also includes: four-patch respiration and heart rate data integration steps. The four-patch respiration and heart rate data integration steps are specifically: there is a certain difference in the heart rate and respiration output by each sensor, and due to different data quality, respiration Different from the error in heart rate extraction, in order to minimize the measurement error, Kalman filtering and data quality weighting are used for integration.

By comparing with the gold standard, we can predict the data quality, that is, under certain noise and spectrum diffusion conditions, how much error will there be in a single measurement. The larger the error, the less credible the data is and the lower the weight. Since respiration can undergo sudden changes (such as suddenly decreasing to 0, holding your breath), the Kalman filter based on time series is not applicable, and the final measurement value of respiration is weighted by four sensors.

Since the heart rate amplitude is weaker and the change amplitude is limited, the Kalman filtering step can be added. In the time series, the heart rate data output by each sensor is a combination of the previous data and the change trend. The change trend generally conforms to the Gaussian distribution, that is, the heart rate is unlikely to undergo huge mutations. For example, if the previous data (t-1) is extremely reliable and the current data (t) is less reliable, you can use

HR(t-1)+k _t *(HR _m (t)-HR _m (t-1)) (2)

To predict the current heart rate, and compare it with HR _m (t) to reduce the measurement error, where HR (t-1) is the heart rate value of the previous measurement (heart rate), and HR _m (t) is the current independently measured heart rate value. , the subscript m represents measurement, meaning measurement. k _t is calculated as follows: Suppose P(t|t-1) represents the probability that the previous data is HR _m (t-1) and the current data is HR _m (t), r _t represents the reliability of the current HR _m measurement , then k _t =P(t|t-1)/[P(t|t-1)+r _t ]. r _t is calculated as the ratio of the spectral energy occupied by the peak with the highest amplitude after Fourier transform to the overall spectral energy. The final heart rate measurement data is weighted by four Kalman filtered heart rates.

Comprehensively judge the sleep apnea events that are not very harmful in sleep conditions, remind parents through mobile phones, slow them down through intervention measures such as turning over, and provide relevant data to parents to help determine whether medical intervention is needed.

This application uses a smart sleep patch to monitor children's sleep at low cost and without restraint. It is not disturbed by the environment and is comfortable to use. Have good ability to distinguish apnea, shortness of breath, and respiratory obstruction; provide early warning for life-threatening events, and remind and record general negative events; have good ability to distinguish sleeping positions (left and right side, prone, etc.), false It has low alarms (compared to radar waves and other means); it has good measurement capabilities for heart rate and pulse intensity, and can self-correct measurements based on sleeping posture; it can obtain data related to sleep quality through the analysis of body movement, breathing, and heart rate. Due to the light weight of children, the technical mode using the pressure sensor will result in a too low signal-to-noise ratio, but the acceleration sensor measurement will not be affected.

The present application also relates to a children's sleep monitoring system that implements a child's sleep monitoring method.

The present application also relates to a children's sleep monitoring device that implements a child's sleep monitoring method. The children's sleep monitoring device includes four patches, a processor and a memory. The memory is communicatively connected to the processor. The memory stores instructions that can be executed by the processor and stores breathing and heart rate information collected by the four patches. The instructions are executed by the processor. The above method for monitoring children’s sleep.

The above embodiments only express several embodiments of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the invention. It should be noted that, for those of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present invention, which are equivalent modifications to the above embodiments based on the essential technology of the present invention. and evolution, these all belong to the protection scope of the present invention.

Claims (10)

1. A method for monitoring sleep of a child, comprising the steps of:

and (3) installing a triaxial acceleration sensor: the method comprises the steps that 4 triaxial acceleration sensors are respectively arranged outside clothes of two upper arms, a chest and an abdomen;

collecting information: each of the three-axis acceleration sensors acquires respiration and heart rate signals:

processing the respiration signal of each triaxial acceleration sensor: filtering the respiratory signals of each triaxial acceleration sensor, carrying out Fourier transform on the filtered signals, acquiring respiratory waveforms through meyer wavelet transform and recording when respiratory waves exist in the transformed signals, and judging whether the signals are in a pressed state, a respiratory disorder or a body movement state through the strength and comparison of the total signals when the respiratory waves do not exist in the transformed signals;

processing heart rate signals of each triaxial acceleration sensor: filtering heart rate of each triaxial acceleration sensor, carrying out Fourier transform on the filtered signals, extracting periodic heart rate when heart rate period exists in the transformed signals, and judging whether the signals are in a pressed state or the heart rate is resolved through energy density to calculate the periodicity through strength and contrast of the overall signals when the heart rate period does not exist in the transformed signals;

comprehensively judging sleeping conditions: discarding the breathing signals and heart rate signals of the four triaxial acceleration sensors in the pressed state, comprehensively analyzing the residual signals, judging that the child is in an irregular body movement state, a breathing obstruction or a normal state, respectively carrying out early warning on the irregular body movement state and the breathing obstruction, and recording the normal state.

2. The method for monitoring sleep of a child according to claim 1, wherein: in the step of comprehensively judging the sleeping condition, the respiratory obstruction is judged by a triaxial acceleration sensor of the chest and the abdomen, and the respiratory obstruction is judged when the chest expands and the abdomen contracts.

3. The method for monitoring sleep of a child according to claim 1, wherein: and in the step of processing the heart rate signals of each triaxial acceleration sensor, judging whether a heart rate period exists or not according to whether a significant peak exists in the signals after frequency domain transformation.

4. A method of monitoring sleep in a child according to claim 3, wherein: in the step of processing the heart rate signal of each triaxial acceleration sensor, if the frequency domain transformation fails to find a significant peak value, heart rate fitting is tried through energy density, and periodicity is calculated through energy density resolution heart beat specifically as follows: attempting to obtain the heart rate by means of information entropy, wherein the calculation formula of the entropy is as follows:whereinI _e For the signal strength at a certain moment in time,p _i to appear to occurI _e Is a function of the probability of (1),H _s for the entropy value of a specific time interval, calculating the entropy value in a fixed time through a sliding window, drawing an entropy fluctuation curve, and evaluating the periodicity of the entropy fluctuation curve in a mode of searching for a significant peak by adopting frequency domain transformation, if the periodicity is obvious, obtaining the heart rate in an entropy mode.

5. The method for monitoring sleep of a child according to claim 1, wherein: in the step of comprehensively judging the sleeping condition, the recording of the normal state is specifically as follows: the final respiratory signal is obtained by data weighting of four triaxial acceleration sensors, wherein the data weight is a signal quality index and is calculated by the proportion of the frequency domain peak to the whole frequency domain energy.

6. The method for monitoring sleep of a child according to claim 1, wherein: in the step of comprehensively judging the sleeping condition, the recording of the normal state is specifically as follows: and finally weighting the data of the four triaxial acceleration sensors of the heart rate signal after Kalman filtering.

7. The method for monitoring sleep of a child according to claim 1, wherein: in the step of comprehensively judging the sleeping situation, when the child is in a respiratory obstruction state, reminding the child of changing the body position when respiratory obstruction is frequent but the duration time is short; early warning is given when respiratory obstruction continues unrelieved and heart rate decreases.

8. The method for monitoring sleep of a child according to claim 1, wherein: in the process of filtering the respiratory signals of each triaxial acceleration sensor, the reserved signals are in the range of 0.13-0.66Hz, namely 7.8-39.6bpm; in the process of filtering the heart rate signal of each triaxial acceleration sensor, the first step separates 4-11Hz signals, and the second step retains signals at 0.8-3.5Hz, namely 48-210bpm.

9. A child sleep monitoring system, characterized by: the child sleep monitoring system for implementing the child sleep monitoring method of any one of claims 1-8.

10. A child sleep monitoring arrangement, characterized by: comprising

Four triaxial acceleration sensors which are respectively arranged outside clothes of two upper arms, the chest and the abdomen of the child;

a processor;

a memory communicatively coupled to the processor;

the memory stores data collected by the four tri-axial acceleration sensors and instructions executable by the processor to implement the method of child sleep monitoring of any one of claims 1-8.