CN113261951B

CN113261951B - Sleeping posture identification method and device based on piezoelectric ceramic sensor

Info

Publication number: CN113261951B
Application number: CN202110474897.1A
Authority: CN
Inventors: 高伟东; 胡迪坤
Original assignee: Beijing University of Posts and Telecommunications
Current assignee: Beijing University of Posts and Telecommunications
Priority date: 2021-04-29
Filing date: 2021-04-29
Publication date: 2022-03-29
Anticipated expiration: 2041-04-29
Also published as: CN113261951A

Abstract

The invention provides a sleeping posture identification method and a device based on a piezoelectric ceramic sensor, which comprises the following steps: under the condition that a user is in a stable sleep state, acquiring a mixed cardiac shock signal of a chest and abdomen area of the user based on a piezoelectric ceramic sensor system; determining a cardiorespiratory activity distribution characteristic of the user based on the mixed heart attack signal; inputting the cardiopulmonary activity distribution characteristics and the preset environment vector characteristics into the trained sleep posture recognition classification network model to obtain a sleep posture recognition result of the user. According to the method, under the condition that a user is in a stable sleep state, the multi-channel piezoelectric ceramic sensor is used for collecting mixed cardiac shock signals of the chest and abdomen area of the user and processing the signals to obtain the heart and lung activity distribution characteristics of the user, so that the heart and lung activity distribution characteristics and the preset environment vector characteristics are input into a trained sleep posture recognition classification network model to obtain a sleep posture recognition result of the user, non-invasive real-time user sleep posture monitoring is achieved, and the method has strong universality and environment anti-interference capacity.

Description

Sleeping posture identification method and device based on piezoelectric ceramic sensor

Technical Field

The invention relates to the technical field of sleep monitoring, in particular to a sleeping posture identification method and device based on a piezoelectric ceramic sensor.

Background

With the improvement of living standard, more and more people begin to pay attention to their sleep condition.

For sleep monitoring, a common way for monitoring the sleep posture of a user and the state of related data at present is a sleep multi-guide graph (PSG) and a Cyclic Alternating Pattern (CAP) in sleep, and the sleep posture condition of the user all night is judged by a gyroscope auxiliary video monitoring system. According to the method, the physiological parameters of a user are obtained through contact type measurement, and the user can obtain a complete report only by wearing more than ten electrodes and wearing a tension sensor for testing all night. The whole testing process is complicated, and the user is easy to generate the first night effect. Meanwhile, the invasive test also brings physiological and psychological burdens to the user, and directly influences the normal sleeping posture data state of the user.

Therefore, how to better realize the sleep posture recognition has become a research focus of interest in the industry.

Disclosure of Invention

The invention provides a sleeping posture identification method and device based on a piezoelectric ceramic sensor, which are used for better realizing sleeping posture identification.

The invention provides a sleeping posture identification method based on a piezoelectric ceramic sensor, which comprises the following steps:

under the condition that a user is in a stable sleep state, acquiring a mixed cardiac shock signal of a chest and abdomen area of the user based on a piezoelectric ceramic sensor system, wherein the piezoelectric ceramic sensor system comprises a plurality of piezoelectric ceramic sensors;

determining a cardiorespiratory activity distribution characteristic of the user based on the blended ballistocardiographic signal;

inputting the cardiopulmonary activity distribution characteristics and preset environment vector characteristics into a trained sleeping posture recognition classification network model to obtain a sleeping posture recognition result of the user;

the trained sleeping posture identification and classification network model is obtained by training according to cardiopulmonary activity distribution characteristics carrying sleeping posture labels and environment vector characteristic samples.

According to the sleeping posture identification method based on the piezoelectric ceramic sensor provided by the invention, under the condition that a user is in a stable sleeping state, a mixed heart impact signal of the chest and abdomen area of the user is acquired based on a piezoelectric ceramic sensor system, and the method specifically comprises the following steps:

acquiring first pressure information applied to the piezoelectric ceramic sensor system by the chest and abdomen area of a user;

determining first voltage information corresponding to the first pressure information according to a mapping relation between pressure and output voltage;

judging the sleep state of the user according to the variable quantity of the first voltage information, and determining second voltage information of the user in a stable sleep state;

and determining the mixed heart impact signal of the user in the stable sleep state according to the variable quantity of the second voltage information.

According to the sleep posture identification method based on the piezoelectric ceramic sensor, provided by the invention, the heart-lung activity distribution characteristics of a user are determined based on the mixed heart impact signal, and the method specifically comprises the following steps:

calculating cardiorespiratory activity intensity characteristics of the user based on the mixed heart impact signal;

and determining the cardiorespiratory activity distribution characteristics of the user according to the cardiorespiratory activity intensity characteristics of the user and the parameter eliminating processing parameters in the piezoelectric ceramic sensor system.

According to the sleeping posture identification method based on the piezoelectric ceramic sensor, provided by the invention, based on the mixed heart impact signal, the heart-lung activity intensity characteristics of a user are calculated, and the method specifically comprises the following steps:

extracting information of the mixed heart impact signal to obtain the voltage information of the cardiopulmonary activity signal of the user;

calculating the stress amplitude information of the cardiopulmonary activity of the user based on signal differentiation and Maclaurin formula approximate estimation according to the voltage information of the cardiopulmonary activity signal of the user;

and performing feature extraction of preset signal length on the cardiorespiratory activity stress amplitude information of the user to obtain cardiorespiratory activity intensity features of the user.

According to the sleep posture identification method based on the piezoelectric ceramic sensor, provided by the invention, the cardiopulmonary activity distribution characteristics of the user are determined according to the cardiopulmonary activity intensity characteristics of the user and the parameter elimination processing parameters in the piezoelectric ceramic sensor system, and the method specifically comprises the following steps:

calculating the cardio-pulmonary activity amplitude information acquired by each sensor according to the cardio-pulmonary activity intensity characteristics of the user and the parameter eliminating processing parameters in the piezoelectric ceramic sensor system;

wherein, the parameter eliminating processing parameter is obtained by calculating the ratio of the parameters of the rest sensors to the reference parameter;

the reference parameters are set by selecting the sensor parameters with the minimum parameters and normal operation in the piezoelectric ceramic sensor system;

and determining the cardio-pulmonary activity distribution characteristics of the user according to the cardio-pulmonary activity amplitude information acquired by the sensors.

According to the sleeping posture identification method based on the piezoelectric ceramic sensor, before the cardiopulmonary activity distribution characteristics and the preset environment vector characteristics are input into the trained sleeping posture identification classification network model, the method further comprises the following steps:

obtaining a plurality of cardio-pulmonary activity distribution characteristic and environment vector characteristic samples carrying sleeping posture labels; taking each cardiopulmonary activity distribution characteristic and environment vector characteristic sample carrying a sleeping posture label as a group of training samples to obtain a plurality of groups of training samples, and training the sleeping posture identification classification network model by using the plurality of groups of training samples.

According to the sleep posture identification method based on the piezoelectric ceramic sensor, provided by the invention, the sleep posture identification classification network model is trained by utilizing a plurality of groups of training samples, and the method specifically comprises the following steps:

for any group of training samples, inputting the training samples into the sleep posture recognition classification network model, and outputting the prediction probability corresponding to the training samples;

calculating a loss value according to the prediction probability corresponding to the training sample and the sleeping posture label in the training sample by using a preset loss function;

and if the loss value is smaller than a preset threshold value, finishing the training of the sleep posture identification and classification network model.

The invention also provides a sleeping posture recognition device based on the piezoelectric ceramic sensor, which comprises:

the mixed cardiac shock signal acquisition module is used for acquiring a mixed cardiac shock signal of a chest and abdomen area of a user based on a piezoelectric ceramic sensor system under the condition that the user is in a stable sleep state, wherein the piezoelectric ceramic sensor system comprises a plurality of piezoelectric ceramic sensors;

the heart-lung activity distribution characteristic generating module is used for determining the heart-lung activity distribution characteristic of the user based on the mixed heart impact signal;

the sleeping posture recognition result generation module is used for inputting the cardiopulmonary activity distribution characteristics and preset environment vector characteristics into a trained sleeping posture recognition classification network model to obtain a sleeping posture recognition result of the user;

The invention also provides an electronic device, which comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein the processor executes the program to realize the steps of the piezoelectric ceramic sensor-based sleeping posture identification method.

The present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when being executed by a processor, realizes the steps of the piezoelectric ceramic sensor-based sleeping posture identifying method as any one of the above.

According to the sleeping posture identification method and device based on the piezoelectric ceramic sensor, under the condition that a user is in a stable sleeping state, the multi-channel piezoelectric ceramic sensor is used for collecting mixed heart impact signals of the chest and abdomen area of the user, the heart and lung activity distribution characteristics of the user are obtained through information processing, the heart and lung activity distribution characteristics and the preset environment vector characteristics of the user are input into a trained sleeping posture identification classification network model, the sleeping posture identification result of the user is obtained, non-invasive real-time user sleeping posture monitoring is achieved, and the sleeping posture identification method and device based on the piezoelectric ceramic sensor have strong universality and environment anti-interference capability.

Drawings

In order to more clearly illustrate the technical solutions of the present invention or the prior art, the drawings needed for the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.

FIG. 1 is a schematic flow chart of a sleeping posture identification method based on a piezoelectric ceramic sensor provided by the invention;

FIG. 2 is a schematic flow chart illustrating steps of a sleep posture identification method based on a piezoelectric ceramic sensor according to an embodiment of the present invention;

FIG. 3 is a schematic diagram comparing output voltage curves of a user and an equal weight object on a piezoceramic sensor system in an embodiment of the present invention;

FIG. 4 is a schematic diagram of a respiratory frequency band signal and an original mixed signal in a mixed ballistocardiogram signal according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of the cardiac shock signal and the envelope of the heartbeat cycle thereof in the hybrid cardiac shock signal provided by the embodiment of the present invention;

FIG. 6 is a schematic diagram of a piezoceramic sensor voltage amplification circuit provided by an embodiment of the present invention;

FIG. 7 is a schematic diagram of an equivalent input circuit of a piezoceramic sensor voltage amplification circuit provided by an embodiment of the present invention;

FIG. 8 is a system framework diagram of the sleeping posture recognition method based on piezoelectric ceramic sensor provided by the invention;

FIG. 9 is a schematic structural diagram of a sleeping posture recognition device based on a piezoelectric ceramic sensor provided by the invention;

fig. 10 is a schematic structural diagram of an electronic device provided by the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Fig. 1 is a schematic flow chart of a sleeping posture recognition method based on a piezoelectric ceramic sensor, as shown in fig. 1, including:

and step S1, acquiring a mixed cardiac shock signal of the chest and abdomen area of the user based on a piezoelectric ceramic sensor system under the condition that the user is in a stable sleep state, wherein the piezoelectric ceramic sensor system comprises a plurality of piezoelectric ceramic sensors.

Alternatively, the sleep plateau described in the present invention refers to a state where the user has no significant limb movement and the entire body is in a state of restitution.

The mixed heart attack (BCG) signal of the chest-abdomen area is obtained by acquiring the signal through a piezoelectric ceramic sensor system below the chest-abdomen area of a user when the user is in a stable sleep state. The mixed cardiac shock signal is mainly voltage variation generated by heart and lung activities, and contains a large amount of physiological information including respiratory signals, cardiac shock signals, pulse signals and the like.

It should be noted that, in the conventional monitoring method, a dense sensor matrix is used to acquire the absolute pressure of the whole body, the number of sensors used in the method is large, so that the cost is high, and meanwhile, the multi-path sensors greatly increase the cost of data processing and the response time of an algorithm. In addition, the absolute pressure is easily interfered by external force, and the pressure distribution of the mattress can be influenced by the differences of the body types, the body structures and even the sleeping posture habits of different people. If the sleeping posture is recognized through absolute pressure distribution, a large amount of training and testing needs to be carried out aiming at complex scenes of various crowds, so that the experiment cost is further increased, and the trained model is difficult to ensure to have better adaptability and accuracy. Secondly, the use rate of absolute pressure data in a sleep monitoring system is low, and other modal data are needed for monitoring other data states such as sleep stages, heart rate, respiratory rate and the like of a human body.

In the embodiment of the invention, the piezoelectric ceramic sensor system is not a dense sensor matrix system and comprises a plurality of piezoelectric ceramic sensors, wherein in the piezoelectric ceramic sensor system, when the number of paths of the piezoelectric ceramic sensor acquisition channels is 10-30, better technical effects can be achieved.

And step S2, determining the heart-lung activity distribution characteristics of the user based on the mixed heart impact signal.

Optionally, the distribution characteristics of the cardiopulmonary activity of the user described in the present invention refer to the distribution characteristics of the amplitude of the heartbeat motion and the distribution characteristics of the amplitude of the respiratory motion of the user.

Further, the acquired mixed heart impact signal is subjected to information extraction and processing, so that the distribution characteristics of the respiratory motion amplitude and the distribution characteristics of the heartbeat motion amplitude can be obtained, and the distribution characteristics of the heart-lung activity of the user can be determined.

And step S3, inputting the cardiopulmonary activity distribution characteristics and preset environment vector characteristics into a trained sleep posture recognition classification network model to obtain a sleep posture recognition result of the user.

Optionally, the trained sleep posture recognition and classification network model described in the present invention is obtained by training according to a model training sample, and is used for performing user sleep posture recognition on the input cardiopulmonary activity distribution characteristics and outputting a recognition result.

The model training sample is composed of a plurality of groups of cardiopulmonary activity distribution characteristics and environment vector characteristic samples which carry sleeping posture labels, and is used for improving the accuracy of various sleeping posture identifications and the environment adaptability of the model.

In the embodiment of the application, in order to adapt to environmental influence and consider richer scene characteristics, personalized environment vector features are added, including the strength ratio of the heartbeat motion amplitude to the respiratory motion amplitude, the total number of the sensors occupied by the user when the user lies down in a standard manner, the total number of the sensors occupied by the user when the user lies down in a side manner, current data state judgment and previous data state judgment.

The sleeping position labels described by the invention are supine, left side lying, right side lying and prone positions, are predetermined according to the cardiopulmonary activity distribution characteristic samples and are in one-to-one correspondence with the cardiopulmonary activity distribution characteristic samples. That is, each cardiopulmonary activity distribution characteristic and environmental vector characteristic sample in the training sample is preset to carry a sleeping posture label corresponding to the cardiopulmonary activity distribution characteristic and the environmental vector characteristic sample.

According to the method provided by the embodiment of the invention, under the condition that the user is in a stable sleep state, based on a piezoelectric ceramic sensor system, the multi-channel piezoelectric ceramic sensor is used for acquiring the mixed cardiac shock signals of the chest and abdomen area of the user, and the signal processing is carried out to obtain the cardio-pulmonary activity distribution characteristics of the user, so that the cardio-pulmonary activity distribution characteristics and the preset environment vector characteristics of the user are input into the trained sleep posture recognition classification network model to obtain the sleep posture recognition result of the user, the non-invasive real-time sleep posture monitoring of the user is realized, and the method has strong universality and environment anti-interference capability.

Optionally, under the condition that the user is in a sleep steady state, acquiring a mixed cardiac shock signal of the chest and abdomen area of the user based on the piezoelectric ceramic sensor system, specifically:

Optionally, the first pressure information described in the present invention refers to the pressure information generated by the user on the piezo-ceramic sensor system in the chest and abdomen area before the user enters sleep, where the pressure information mainly comes from the user's own gravity, the acting force generated by the user's respiratory motion, and the acting force generated by the user's heartbeat motion.

The mapping relation between the pressure and the output voltage described by the invention refers to that the signal acquisition module of the piezoelectric ceramic sensor system acquires external pressure information, and the external pressure information is processed by a digital circuit to obtain corresponding output digital voltage information, so that the mapping relation between the external pressure and the system output digital voltage is obtained.

The first voltage information described in the present invention refers to pressure generated by the user's own gravity, pressure generated by the user's respiratory motion, and pressure generated by the user's heartbeat motion in the first pressure information, and the voltage corresponding to the user's own gravity, the user's respiratory motion signal voltage, and the user's heartbeat motion signal voltage are obtained based on the mapping relationship between the pressure and the output voltage.

Further, according to the mapping relation between the pressure and the output voltage, first voltage information corresponding to the first pressure information can be determined.

The second voltage information described in the present invention refers to voltage change information generated by respiratory stress and heartbeat stress, which is output by the system because the gravity stress of the user is almost unchanged or slowly changed when the user is in a stable sleep state, and the output voltage of the system is hardly affected.

The mixed heart impact signal of the user in the stable sleep state refers to the mixed heart impact signal mainly based on the heart-lung movement acquired when the user does not have obvious limb movement. According to the voltage change, the energy entropy, the approximate entropy and other information output by the system in the steady state, whether the signal is the mixed heart impact signal in the steady state can be judged.

According to the method provided by the embodiment of the invention, the pressure information applied to the piezoelectric ceramic sensor system by the chest and abdomen area of the user is collected, the corresponding voltage information is obtained based on the mapping relation between the pressure and the output voltage, whether the user is in the stable sleep state or not is determined according to the variation of the voltage information, and then the mixed cardiac shock signal of the user in the stable sleep state is determined according to the variation of the voltage information in the stable state.

Optionally, determining a cardiorespiratory activity distribution characteristic of the user based on the mixed heart attack signal, specifically:

Optionally, the intensity characteristics of the cardiopulmonary activity of the user described in the present invention comprise an intensity characteristic of respiratory motion of the user and an intensity characteristic of heartbeat motion of the user.

The parameter elimination processing parameters described in the invention refer to processing parameters determined through calculation in order to eliminate the difference between the intrinsic parameters of each sensor. Through the processing mode, the technical problem that the output voltage of the system is not consistent due to difference among the sensors can be solved.

Further, according to the mixed heart impact signal of the user in the stable sleep state, the heart and lung activity characteristics are extracted through signal analysis and processing, and the heart and lung activity intensity characteristics of the user are obtained.

And performing optimization calculation on the cardiorespiratory activity intensity characteristics of the user according to the cardiorespiratory activity intensity characteristics of the user and the parameter eliminating processing parameters in the piezoelectric ceramic sensor system to determine the cardiorespiratory activity distribution characteristics of the user.

According to the method provided by the embodiment of the invention, the cardiorespiratory activity intensity characteristics of the user are obtained by performing signal analysis and characteristic extraction on the mixed heart impact signals of the user in a stable sleep state, and then the cardiorespiratory activity distribution characteristics of the user are accurately calculated on the basis of eliminating parameter differences among sensors.

Optionally, based on the mixed cardiac shock signal, calculating a cardiorespiratory activity intensity characteristic of the user, specifically:

Optionally, the information extraction described in the present invention refers to separating and extracting the signal of the respiratory frequency band and the signal of the heartbeat frequency band from the mixed heart impact signal.

The voltage information of the cardiopulmonary activity signal of the user described by the invention refers to a signal of a respiratory frequency band and a signal of a heartbeat frequency band, and the relationship between the respiratory motion signal voltage and the respiratory motion stress amplitude and frequency and the relationship between the heartbeat motion signal voltage and the heartbeat motion stress amplitude and frequency are obtained based on the pressure and output voltage mapping relationship.

The cardiorespiratory activity stress amplitude information of the user comprises respiratory motion stress amplitude of the user and heartbeat motion stress amplitude of the user.

In the embodiment of the invention, the feature extraction of the preset signal length refers to feature extraction of fixed signal length for the heart-lung activity stress amplitude information of the user aiming at the difference between the breathing and heartbeat cycles of different user individuals in different scenes.

Further, aiming at the stress amplitude information of the cardiopulmonary activity of the user, the feature extraction formula with fixed signal length is optimized by utilizing the real-time heart rate and the respiration rate of the signal, and the cardiopulmonary activity intensity feature of the user can be obtained.

In an embodiment of the invention, the intensity of cardiopulmonary activity characteristic of the user comprises a respiratory motion intensity characteristic of the user and a heartbeat motion intensity characteristic of the user.

According to the method provided by the embodiment of the invention, the mixed heart impact signal is separated, the signal of the respiratory frequency band and the signal of the heartbeat frequency band are extracted, and the feature extraction formula of the signal real-time heart rate and respiratory rate on the preset signal length is optimized, so that more accurate heart-lung activity intensity features are obtained.

Optionally, determining the cardiorespiratory activity distribution characteristics of the user according to the cardiorespiratory activity intensity characteristics of the user and the parameter canceling processing parameters in the piezoelectric ceramic sensor system, specifically:

the reference parameters are set by selecting the sensor with the minimum parameter in the piezoelectric ceramic sensor system and working normally;

Optionally, the cardiopulmonary activity amplitude information described herein includes an amplitude of respiratory motion and an amplitude of heartbeat motion.

The cardiorespiratory activity distribution characteristics of the user described by the invention refer to real-time distribution characteristics of the cardiorespiratory motion amplitude of the user obtained by monitoring through various sensors, and comprise amplitude distribution characteristics of respiratory motion of the user and amplitude distribution characteristics of heartbeat motion of the user.

In the embodiment of the invention, the sensor parameter with the minimum parameter and normal operation in the piezoelectric ceramic sensor system is selected as the reference parameter, and the ratio of the parameters of the other sensors to the reference parameter, namely the parameter elimination processing parameter, is calculated, so that the amplitude of respiratory motion and the amplitude of heartbeat motion acquired by different sensor channels at the same moment can be obtained, and the amplitude distribution characteristic of the respiratory motion of a user and the amplitude distribution characteristic of the heartbeat motion of the user are further determined.

According to the method provided by the embodiment of the invention, the influence of the difference among the parameters of the sensors is eliminated through the comparison of the vibration tests of different sensors, and further the heart-lung activity distribution characteristics of the user are determined more accurately according to the heart-lung activity intensity characteristics of the user.

Optionally, before inputting the cardiopulmonary activity distribution characteristics and preset environment vector characteristics into the trained sleep posture recognition classification network model, the method further comprises:

obtaining a plurality of cardio-pulmonary activity distribution characteristic and environment vector characteristic samples carrying sleeping posture labels;

and taking each cardiopulmonary activity distribution characteristic sample carrying the sleeping posture label as a group of training samples to obtain a plurality of groups of training samples, and training the sleeping posture identification and classification network model by using the plurality of groups of training samples.

Optionally, before inputting the cardiopulmonary activity distribution characteristics and the preset environment vector characteristics into the sleep posture recognition and classification network model, the sleep posture recognition and classification network model needs to be trained, and the specific training process is as follows:

the sleeping posture labels carried by the cardiopulmonary activity distribution characteristic and environment vector characteristic sample and the cardiopulmonary activity distribution characteristic sample are used as a group of training samples, namely, each cardiopulmonary activity distribution characteristic and environment vector characteristic sample with the sleeping posture label is used as a group of training samples, and therefore a plurality of groups of training samples can be obtained.

In the embodiment of the invention, the cardiopulmonary activity distribution characteristics and the environmental vector characteristic samples correspond to the sleeping posture labels carried by the cardiopulmonary activity distribution characteristics samples in a one-to-one mode.

Then, after obtaining a plurality of groups of training samples, sequentially inputting the plurality of groups of training samples into the sleeping posture identification and classification network model, namely, the cardiopulmonary activity distribution characteristics, the environment vector characteristic samples and the corresponding sleeping posture labels in each group of training samples, simultaneously inputting the sleeping posture identification and classification network model, adjusting parameters of the sleeping posture identification and classification network model according to each output result of the sleeping posture identification and classification network model, and finally completing the training process of the sleeping posture identification and classification network model.

According to the method provided by the embodiment of the invention, the sleeping posture labels carried by the cardiopulmonary activity distribution characteristic and environment vector characteristic samples and the cardiopulmonary activity distribution characteristic samples are used as a group of training samples, and the plurality of groups of training samples are utilized to carry out model training on the sleeping posture recognition classification network model.

Optionally, the training of the sleep posture recognition and classification network model by using multiple groups of training samples specifically includes:

Optionally, in an embodiment of the present invention, the preset loss function refers to a loss function preset in the sleep posture recognition and classification network model, and is used for model evaluation; the preset threshold refers to a preset threshold of the model, and is used for obtaining a minimum loss value and completing model training

After a plurality of groups of training samples are obtained, for any group of training samples, the distribution characteristics of the cardiopulmonary activity in the training samples, the environmental vector characteristic samples and the sleeping posture labels carried correspondingly are input into a sleeping posture recognition classification network model at the same time, and the prediction probability corresponding to the training samples is output, wherein the prediction probability refers to the prediction probability corresponding to different sleeping posture recognition results of the training samples.

On the basis, a preset loss function is used for calculating a loss value according to the prediction probability corresponding to the training sample and the sleeping posture label in the training sample. The preset loss function may be a cross entropy loss function. In other embodiments, the sleep posture label representation and the preset loss function may be set according to actual requirements, and are not specifically limited herein.

Further, after the loss value is obtained through calculation, the training process is finished, the preset sleep posture recognition classification network model parameters can be updated through an error back propagation algorithm, and then the next training is carried out. In the training process, if the loss value obtained by calculation aiming at a certain group of training samples is smaller than a preset threshold value, the training of the sleep posture recognition classification network model is completed.

According to the method provided by the embodiment of the invention, the sleep posture identification and classification network model is trained, and the loss value of the sleep posture identification and classification network model is controlled within the preset range, so that the accuracy of the sleep posture identification result output by the sleep posture identification and classification network model is improved.

Fig. 2 is a schematic flow chart of steps of a sleep posture identification method based on a piezoelectric ceramic sensor according to an embodiment of the present invention, and as shown in fig. 2, the method includes the following steps: the method comprises the steps of analyzing the stress and the working state of a sensor, calculating an equivalent output circuit, outputting a voltage and stress mapping formula, judging the sleep stable state, extracting the intensity characteristic of the heart and lung activity in the stable state, optimizing the heart and lung activity characteristic, extracting the environmental characteristic, eliminating the influence of the super parameters among the sensors, and identifying the real-time distribution characteristic of the heart and lung motion and the sleep posture in the stable state.

Optionally, in an embodiment of the present invention, the step of analyzing the stress and the working state of the sensor is specifically as follows:

in the piezoelectric ceramic sensor system, the piezoelectric ceramic sensor has a piezoelectric effect, and the piezoelectric ceramic sensor system mainly utilizes the positive piezoelectric effect, namely when the piezoelectric ceramic sensor deforms along the action of external force in a certain direction, electric polarization occurs inside the piezoelectric ceramic sensor system, and meanwhile, charges with opposite signs can be generated on the surface of the piezoelectric ceramic sensor. When the external force is removed, the piezoelectric ceramic sensor is restored to an uncharged state, and the electric charge quantity generated by the piezoelectric ceramic sensor is in direct proportion to the magnitude of the external force within a certain range.

The external force that piezoceramics sensor received all can take place deformation in 3 dimensions directions in the cubical space to produce stress. According to Hooke's law of unity, i.e. the internal stress is proportional to the amount of strain. Since the compliance coefficient and the stiffness coefficient are different at three bit planes, 3 bit planes need to be discussed separately.

The stress state of any point is decomposed into 3 bit planes, and the force of each bit plane can be decomposed into 3 system coordinates of x-axis, y-axis and z-axis, so that the stress state of any point consists of 9 stress components T_xx，T_xy，T_xz，T_yx，T_yy，T_yz，T_zx，T_zy，T_zzAnd (6) determining. According to the reciprocal law of shear stress, T_yz＝T_zy，T_xy＝T_yx，T_zx＝T_xzTherefore, 9 stress components are reduced to 6. Wherein, T_xx，T_yy，T_zzRespectively, normal stress, denoted by T₁，T₂，T₃Represents; t is_yz，T_zx，T_xyRespectively representing shear stress, by T₄，T₅，T₆And (4) showing.

Electric displacement of piezoelectric ceramic sensor

Is of the formula

Wherein epsilon₀Is dielectric constant in vacuum, and the piezoelectric ceramic has no polarization intensity

The electric field strength in each direction.

Piezoelectric coefficient is the intensity of polarization

And external stress

The ratio of (A) to (B):

wherein the content of the first and second substances,

represents the polarization intensity with a component P at 3 planes₁、P₂、P₃；[d]Which represents the piezoelectric coefficient of the piezoelectric element,

the external stress is indicated.

Due to the fact that

Is 6 stress components, so the piezoelectric coefficient [ d ]]Is a 3 x 6 matrix of coefficients, the first subscript of the piezoelectric coefficients indicating the direction of polarization, i.e. the direction of the electric field, and the second subscript indicating the direction of the resolving force.

Analyzing the short circuit of the piezoelectric ceramic sensor system in the vertical direction of the Z axis:

the subscript of the piezoelectric ceramic sensor in the vertical direction of the Z axis is set to be 3, and when the polarized electrodes are short-circuited in the vertical direction, the electric field strength is 0, namely:

E₃＝0；

considering only the direction of the polarizing electric field, i.e.

According to the formula of piezoelectric coefficient, at this time

The directional polarization is produced by the positively stressed piezoelectric effect, i.e.

In the piezoelectric coefficient matrix, d₃₁＝d₃₃，

And (3) performing open circuit analysis on the piezoelectric ceramic sensor in the vertical direction:

when the circuit is broken between the electrodes polarized in the vertical direction, the charge between the plates cannot be displaced, i.e.

D₃＝0；

Considering only the direction of the polarizing electric field, i.e.

According to the formula of piezoelectric coefficients, at this time,

the directional polarization is generated by the positive stress piezoelectric effect.

Is provided with

The piezoelectric coefficient is also the same, and the relationship between the electric field intensity and the stress at this time is

Fig. 3 is a schematic diagram comparing output voltage curves of a user and an equal-weight object on a piezoelectric ceramic sensor system in an embodiment of the present invention, and as shown in fig. 3, an abscissa represents the number of sampling points, each sampling point is 10ms, that is, the abscissa also represents sampling time; the ordinate represents the digital voltage output by the analog-to-digital conversion circuit. It should be noted that, in the present application, the ordinate is similar to the output voltage value of the oscilloscope, theoretically, the working voltage of the analog-to-digital converter is-3.3V to 3.3V, when the numerical scale on the ordinate is 4096, the corresponding working voltage is 3.3V, and when the numerical scale on the ordinate is 2048, the corresponding working voltage is 0V; when no voltage is output, 2048 is output.

As shown in fig. 3 (a), the output voltage variation curve of the equal weight object placing process is shown. The leakage resistance of the sensor and the input resistance of the amplifier in the piezoelectric ceramic are large, the electric displacement is weak in the working state, the pressure change moment can be regarded as circuit break under the ideal condition, and the electric field intensity on the surface of the piezoelectric ceramic sensor is close to that of the piezoelectric ceramic sensor

As the stress increases, the voltage rises rapidly. However, the actual equivalent resistance is not ideal and infinite, and the actual circuit is not open circuit and has electric potential shift, so that the two poles of the piezoelectric ceramic are balanced after a period of time, and the electric field and the voltage return to zero.

As shown in fig. 3 (b), the output voltage variation curve of the process of picking up the equal-weight object is shown. When the object leaves, the stress of the sensor is reduced, the voltage drops rapidly, and the electric field and the voltage are balanced after a period of time.

As shown in (c) of fig. 3, an output voltage variation curve during the user's lying down is shown.

As shown in fig. 3 (d), the output voltage variation curve is the user rising.

At the moment of lying down and rising, although the voltage change is consistent with the voltage change of the equal-weight object during placement and picking up, after a period of time, the voltage change is a periodic sine-like signal, namely a mixed heart shock signal in medicine. According to the medical meaning of the mixed heart-shocking signal, the signal is the voltage change generated by the heart-lung activity, the amplitude of the normal breathing motion is far larger than that generated by the heartbeat, and in the open circuit state, the stress is in direct proportion to the electric field intensity, namely the surface voltage. Therefore, the user can be deduced to be in a sleep stable state, and the user has sine-like acting force generated by respiratory motion besides relatively stable gravity. Moreover, the stress variation generated by the breathing movement is much smaller than the self gravity of the user, so that the pressure is consistent with the pressure when the user gets up and lies down and the object with the same weight is taken up and placed. After a period of time, the charge displacement caused by gravity is completed, namely, when the short circuit state is returned, the pressure change of breathing motion at the moment leads the output voltage change to generate sine-like piezoelectric change on the surface of the piezoelectric ceramic sensor.

In an embodiment of the present application, the mixed ballistocardiogram signal contains a ballistocardiogram signal in addition to the respiration signal.

Fig. 4 is a schematic diagram of a respiratory frequency band signal and an original mixed signal in a mixed cardiac shock signal according to an embodiment of the present invention, as shown in fig. 4, a thick solid line is the respiratory frequency band signal, and a thin solid line is the mixed cardiac shock signal, where an abscissa represents the number of sampling points, each sampling point is 10ms, that is, the abscissa also represents sampling time; the ordinate represents the digital voltage output by the analog-to-digital conversion circuit.

Fig. 5 is a schematic diagram of a heartbeat signal and an envelope of a heartbeat cycle thereof in a mixed heartbeat signal provided by an embodiment of the present invention, as shown in fig. 5, a thin solid line represents the heartbeat signal, and a thick solid line represents the envelope of the heartbeat cycle of the heartbeat signal, wherein an abscissa represents the number of sampling points, each sampling point is 10ms, that is, the abscissa also represents sampling time; the ordinate represents the digital voltage output by the analog-to-digital conversion circuit.

The ballistocardiogram signal can be separated from the respiration signal according to the difference between the frequency of the respiration wave (0.1-0.5Hz) and the frequency of the ballistocardiogram signal (5-20 Hz). Like the respiration signal, the ballistocardiogram signal is produced by pressure changes during the heartbeat, which is a periodic signal, although not sinusoidal, one cycle being a complete heartbeat in medical definition.

Further, establishing a stress model of the surface of the piezoelectric ceramic sensor:

wherein the content of the first and second substances,

representing the total stress experienced by the piezoceramic sensor surface,

representing the stress created by the user's weight,

representing the stress generated by the breathing movements of the user,

representing the stress caused by the user's heartbeat movement.

Although the three acting forces are mixed together in time domain to influence the output voltage, the respective frequency and voltage change period do not interfere with each other.

Further, in the embodiment of the present invention, the steps of calculating the equivalent output circuit are specifically as follows:

in the circuit, the piezoelectric ceramic is equivalent to a voltage source connected with a self capacitor C in series₁Leakage resistance R of parallel sensor₁Since the voltage generated by the sensor itself is weak, the surface voltage of the sensor needs to be amplified by an operational amplifier in the device.

FIG. 6 is a schematic diagram of a voltage amplifying circuit of a piezoceramic sensor according to an embodiment of the present invention, as shown in FIG. 6, C₂Representing the input capacitance of the operational amplifier, C₃Representing the equivalent capacitance of the line, R₁Indicating the leakage resistance, R, of the piezoceramic sensor₂Representing the input resistance of the operational amplifier.

FIG. 7 is a schematic diagram of an equivalent input circuit of a voltage amplifying circuit of a piezoceramic sensor provided by an embodiment of the present invention, as shown in FIG. 7, the equivalent input resistance is

An equivalent input capacitance of

C＝C₁+C₂；

Further, in the embodiment of the present invention, the steps of outputting the voltage and stress mapping formula are specifically as follows:

stress T of pressure sensor to vertical direction of sensor when user lies₃＞＞T₁,T₂，

According to U ═ Ed, d is the equipotential surface distance, i.e. the thickness of the piezoelectric ceramic, T₃Is a force in the vertical direction per unit area, i.e.

Wherein, F₃Is the force in the direction perpendicular to the Z axis, S is the surface area of the piezoelectric ceramic, and ε is the dielectric constant.

Assuming that the current user is in a sleep plateau state, namely

F₃≈F_N3+F_M3+F_H3；

Wherein, F_N3The gravitational stress is constant force, F in a stable state_N3The component equivalent short circuit generates a voltage of 0. F_M3The respiratory stress may be equivalent to a magnitude of F_mOf sinusoidal signal f_mIs the breathing rate.

F_M3≈F_msinωt；

ω＝2πf_m；

Sensor surface voltage U₃Respiratory motion component of

Input voltage of operational amplifier

Input voltage amplitude value of operational amplifier

Leakage resistance R in equivalent input resistance R₁Amplifier input resistance R₂Larger, i.e. ω R (C)₁+C₂+C₃)>>1；

Operational amplifier voltage U_3MinAfter amplification, the digital voltage U 'is acquired through low-pass filtering, voltage lifting, digital-to-analog conversion and the like in sequence'_3MoutOnly analogue voltage U_3MoutIs digitized, i.e. U'_3Mout≈U_3Mout. The low-pass filtering mainly filters hardware noise above 50Hz, and the amplitude value of the collected voltage is hardly influenced after the filtering.

Setting a voltage amplification factor B, raising the voltage to A, and acquiring a digital voltage U 'of respiratory motion at the moment'_3MoutIs composed of

The ballistocardiogram signal is a periodic signal, such as the thin solid line signal shown in fig. 5, and according to the medical nature, the envelope of the ballistocardiogram signal is a chirp-like signal, and the frequency is determined by the current heart rate. According to the heart rate change in a short time (30 seconds), the envelope signal of the heartbeat frequency band of the heart impact signal in the short time can be equivalent to a sinusoidal signal, such as the thick solid line signal shown in fig. 5, which is generally less than 0.3 Hz. The equivalent sinusoidal signal contains the peak amplitude and the periodic characteristic of the original heart impact signal, and the equivalent periodic motion can be used according to the condition that the output voltage of the sensor surface is in direct proportion to the stress of the generated voltage

The envelope signal of fig. 5 is generated, β being the current heart beat angular velocity. So, the equivalent heartbeat

Is the stress of heartbeat movement F_hThe periodic envelope of the light source(s),

amplitude F of heart-beat stress_hHas consistency.

Like respiratory motion, the relationship between the impact envelope voltage signal and the equivalent heartbeat stress can be obtained in the same way as

Further, in the embodiment of the present invention, the sleep stable state determination method includes the following steps:

according to the calculation process, the collected data is the mapping of the external stress on the digital voltage, but the invention only considers that under the ideal steady state, the state of the data is changed due to the difference of the force structure. After the section of data abnormity is eliminated, normal data can be divided into a non-stable state and a stable state, wherein the non-stable state comprises the states of getting up, lying down and body movement with obvious limb movement, and the state of leaving bed without load

The stress is analyzed according to the state, and the force in the vertical direction of the Z axis can be divided into gravity stress, respiratory motion stress and heartbeat motion stress, wherein the gravity stress is far larger than the respiratory motion stress, and the respiratory motion stress is far larger than the heartbeat motion stress.

The getting-up process is the transition from a stable in-bed state to out-of-bed state, the gravity stress is rapidly reduced during getting-up, the change speed is far greater than the stress of heartbeat breathing movement, and the output voltage is rapidly reduced at the moment.

The bed leaving state is the no-load state of the mattress, at the moment, no stress is applied, and the output is 0.

The lying-down process is the transition from a bed-leaving state to a stable bed-in state, during the lying-down process, the gravity stress is rapidly increased, the change speed is far greater than the stress of heartbeat breathing movement, and at the moment, the output voltage is rapidly increased.

The body movement state is the state that the user takes place the motion in the lower limbs body of stationary state, and all are sleep stationary state before and after the body movement takes place, and gravity stress probably increases also can reduce during the body movement, and the change rate is far more than heartbeat respiratory motion stress, and output voltage can take place the rapid change this moment, but is sleep stationary state before and after the body movement takes place.

In the process of a steady state, the gravity stress changes almost or slowly, the output voltage is hardly influenced, only the output trend wave is influenced, and the respiratory cycle stress is far greater than the heartbeat cycle stress, so the output in the steady state is mainly the voltage change U 'generated by the respiratory stress'_3Mout。

Further, in the embodiment of the present invention, the steps of the intensity characteristics of the cardiopulmonary activity in the steady state are specifically as follows:

in the normal sleep period, the stable state occupies most of the time, the mixed heart impact signal can be well observed in the stable state, and the respiratory motion change signal and the heartbeat motion change signal are respectively extracted. For the states of bed leaving, rising, lying down and body movement, which are all unstable or data-free, the invention only distinguishes the states for monitoring the sleep state without further processing.

Through extraction and processing of the collected mixed cardiac shock signal, a respiratory motion frequency signal U 'can be obtained'_3MoutEnvelope signal U 'with heartbeat motion frequency'_3HoutAccording to the mapping relation between the pressure and the output voltage, a real-time signal and the lifting voltage can be obtained, and the relevant parts of the current signal, the phase ω t and the lifting voltage A can be eliminated through signal differentiation, namely

Wherein, Δ t is a minimum time interval, and the minimum time interval is a sampling time difference in the implementation process; ω represents the breathing angular velocity; f represents the respiration rate, which approaches 0 at 0.1Hz-0.5Hz, Δ t ω < 2 π.

Thus, it can be seen that:

sinω(t+Δt)-sinωt＝sinωt cosωΔt+sinωΔt cosωt-sinωt；

according to the mclaurin formula:

therefore, the following steps are carried out:

sinω(t+Δt)-sinωt≈sinωt(1)+(ωΔt)cosωt-sinωt

＝(ωΔt)cosωt；

wherein, U_3Mout(t) represents the digital signal acquired, Δ t represents the sampling interval of the selection signal, d₃₃Representing the piezoelectric coefficient, B representing the operational amplifier amplification coefficient, C representing the equivalent input internal capacitance, and omega representing the respiratory motionDynamic angular velocity, F_mRepresenting the respiratory motion stress magnitude.

Acquisition data U_3Mout(t) is a periodic signal, so that it is related to the current time, i.e. the instantaneous phase ω t, and the phase interference is reduced by the characteristic that the integral value is constant under the complete period of the sinusoidal signal. Setting the period of one breath as

Selecting N times of period NT > Deltat, and sampling interval of voltage output is Deltat

Wherein d is₃₃C, B, respectively, indicate the super-parameter of the piezo-ceramic sensor, and N indicates the number of cycles of the preset signal.

Thereby, the acquisition of the respiratory motion voltage U is established_3Mout(t) stress amplitude F of respiratory motion_mThe relationship (2) of (c).

Similarly, an envelope signal U of the heartbeat motion voltage can be obtained_3Hout(t) stress amplitude in relation to the heartbeat movement

The relationship of (a) is as follows:

wherein d is₃₃C, B respectively represent the super parameter of the piezoelectric ceramic sensor, beta represents the current heartbeat acceleration,

indicating the period of one breath and M indicating the preset number of heart beat treatment cycles.

Further, in the embodiment of the present invention, the step of optimizing the heart-lung activity feature and extracting the environmental feature specifically includes:

by mixing the frequency division and the signal processing of the heart impact signal, a respiratory motion voltage signal U can be obtained_3Mout(t) and a heartbeat voltage signal U_3Hout(t) a signal.

In the characteristic formula obtained according to the above steps, the length of the integral in the formula corresponding to the respiratory motion is NT₁The length of the integral in the formula corresponding to the heartbeat motion is MT₂The two are different in length, and different individuals have heartbeat periods T under different scenes₁And the respiratory cycle T₂With a large variance. Therefore, the invention presets a feature extraction formula with fixed signal length, and selects the optimal signal length for processing the heart rate and the respiration within 30 seconds to 3 minutes per frame. The longer the signal processing time, the larger N and M in each frame signal, the more stable the characteristics, the less influence on the phase, but the higher the possibility of the change degree of the heart rate and the respiratory rate in unit time. Therefore, the respiratory motion intensity characteristic and the heartbeat motion intensity characteristic are processed in 30 seconds to 1 minute by comprehensively considering each frame of signal, taking 30 seconds as an example:

having a breath cycle number within 30 seconds

Has a heartbeat cycle number within 30 seconds

Integral of absolute value of sinusoidal signal, per

Are all 1, so the pair 120f₁、120f₂Get the whole, i.e.

At this time, the signal with incomplete cycles can be estimated with an accuracy of 0.25 cycles and with small error, i.e.:

further, in the embodiment of the present invention, the step of eliminating the influence of the super-parameter between the sensors is specifically as follows:

according to the steps, the relation between the characteristic value and the output voltage under a single sensor can be obtained, d₃₃C, B is a hyper-parameter of the sensor, i.e. represents a parameter intrinsic to the sensor. Although the hyper-parameter of the same sensor is constant, when the motion force distribution of a plurality of sensors is calculated, the difference exists between different sensors, and even if the same batch of elements are adopted, the consistency of an output circuit cannot be ensured, so that the hyper-parameter between the sensors needs to be eliminated.

In the embodiment of the present invention, let d be assumedⁱ ₃₃，Cⁱ，BⁱThe upper mark of (1) is the serial number of the sensor, and the feature K of the respiratory motion intensity is outputⁱI.e. by

With a constant vibration source placed above each sensor, the output voltage, F, is measured several timesⁱ _m、ωⁱThe constant remains unchanged. The characteristic value of the nth sensor and the characteristic value of the jth sensor are used as a ratio to obtain the ratio between the super parameters of the sensors, namely the ratio

The sensor with the minimum parameter and normal operation is set as a unit reference

The other sensors obtain the ratio D of the excess parametersⁱThen, the different sensors acquire the respiratory movement intensity characteristic K at the same timeⁱWith amplitude of pressure F of respiratory movementⁱ _mThe relationship between:

where ω represents the breathing frequency. Amplitude of pressure F for respiratory motion between different sensorsⁱ _mUnlike this, the breathing frequency is uniform.

For the distribution of the intensity of the respiratory movement, i.e. by the characteristic KⁱCalculating respiratory motion amplitude F under different sensorsⁱ _mIn relation to each other, i.e.

Similarly, the heartbeat motion amplitude under different sensors can be calculated

The relationship between them will not be described in detail herein.

Further, in the embodiment of the present invention, the steps of real-time distribution of cardiopulmonary motion are specifically as follows:

through the steps, the collected voltage characteristic K can be obtainedⁱAnd the measured sensor characteristic DⁱRespectively calculating the intensity distribution of the heartbeat movement and the respiratory movement

Meanwhile, the sleep information such as Heart Rate, respiration, Heart Rate Variability (HRV) and the like at the moment can be obtained.

Further, in the embodiment of the present invention, the sleep posture recognition step in the stable state specifically includes:

the data are distinguished and identified in the steps of getting up, getting out of bed, lying down, body movement and stable state, and further, the sleeping posture is identified in the stable sleeping state of the user.

In order to adapt to environmental influence and consider richer scene characteristics, the invention also adds personalized environmental vector characteristics, which comprise the strength ratio of the heartbeat motion amplitude to the respiratory motion amplitude, the total number of the sensors occupied by the user when the user lies down in a standard way, the total number of the sensors occupied by the user when the user lies down in a side way, the judgment of the current data state and the judgment of the data state at the previous moment.

In the embodiment of the invention, the adopted training sample is 4000 cases, wherein the feature matrix is the amplitude distribution feature 2 x 8 of respiratory motion, the feature matrix is the amplitude distribution feature 2 x 8 of heartbeat motion, the environment vector feature 1 x 5, and the sleeping posture labels are lying, lying on the left side, lying on the right side and lying on the prone side. The invention adopts a compression-and-Excitation network (SENet) classification network framework to classify, processes the feature map obtained after convolution to obtain a one-dimensional vector as the number of channels as the evaluation score of each channel, and then applies the score to the corresponding channel to obtain the result. Therefore, the mutual dependency relationship among the channels is explicitly modeled, the characteristic response of the channels is adaptively recalibrated, and the accuracy of model classification is improved.

Fig. 8 is a schematic diagram of a system framework of the sleep posture identification method based on the piezoelectric ceramic sensor, as shown in fig. 8, when a user is in a stable sleep state, the pressure applied to the piezoelectric ceramic sensor system by the chest and abdomen area of the user is collected through a signal collection module of the piezoelectric ceramic sensor system to obtain a corresponding output voltage, and a mixed cardiac shock signal of the chest and abdomen area of the user is obtained through an operational amplifier, voltage lifting, a low-pass filter and digital-to-analog conversion.

Further, the mapping relation between the output voltage and the stress is obtained through sensor state analysis, sensor stress analysis and output equivalent circuit calculation.

Further, the heart-lung movement signals can be obtained by analyzing and processing the mixed heart impact signals of the chest and abdomen area of the user, and the heart-lung movement distribution characteristics can be determined by eliminating the super-parameter test among all sensors and performing characteristic extraction on the heart-lung movement signals by utilizing the real-time heart rate and the respiration rate of the signals.

Further, according to the distribution characteristics of the cardiopulmonary activity of the user and the state environment vector characteristics, the sleeping posture of the user is identified based on a structure model of a compression and Excitation network (SEnet).

The method of the embodiment of the invention provides a distribution method for identifying the combined activities of the heart and the lung of a user in sleep by a multi-channel sensor device. The invention firstly adopts a piezoelectric ceramic sensor and an MSP430 chip to design a 16-path pressure acquisition mattress to obtain a mixed heart impact signal of the chest and abdomen part of a user during sleeping. According to the principle of an equivalent circuit, the piezoelectric ceramic sensor can be equivalent to a voltage source and a capacitor, the voltage value of the voltage source can be changed by the change of external pressure, and then a voltage signal passes through an operational amplifier circuit, a filter circuit, a lifting circuit and an analog-digital conversion circuit and is finally uploaded to a server. The data collected by the invention is also the function change of the surface voltage of the piezoelectric ceramics, and the operational relation between the surface voltage of the piezoelectric ceramics and the output voltage can be calculated through the equivalent circuit of the piezoelectric ceramics and the operational amplifier. Ideally, the leakage resistance R of the piezoelectric sensor₁And operational amplifier input resistance R₂Much larger than the circuit equivalent capacitance C, the equivalent resistance R can be regarded as an open circuit with infinite resistance. Therefore, almost no electric displacement occurs at the moment when the piezoelectric ceramic is subjected to a force, and only the voltage of the two electrodes is changed. However, in practical situations, the equivalent circuit is not a true open circuit, and a small amount of electric displacement occurs at the moment of stress. If the stress is kept unchanged along with the time, the charges of the two poles of the piezoelectric ceramics are finally balanced for a long time, and the internal electric field intensity and the voltage return to zero.

The invention changes the voltage into a voltage rapid change part and a voltage stable part by comparing the output digital voltage of the user in the lying, rising and stable states and the placing, taking and stable states of the equal weight object. For the part with the rapid voltage change, the voltage change conditions of a user and an object with the same weight are basically consistent, at the moment, an output circuit is equivalently opened, electric charge generated by rapid increase or decrease of the stress of the sensor hardly generates electric displacement, and the sensor generates large voltage change up and down. For the voltage plateau, the supporting force of the equal-weight object is kept constant at the moment, and the voltage of the plateau is almost consistent with that of the no-load state. Because the actual output equivalent resistance is not infinite, the circuit has weak electric displacement, the displacement reaches balance after a period of time, and the voltage of the equivalent short circuit of the output circuit returns to zero. The stable state of the user still has periodic pressure change, and according to the medical significance of the mixed heart impact signals, the mixed heart impact signals collected by the user in the stable state are mainly generated by respiratory movement and include heart-lung cycles of heartbeat and pulse vibration. The mixed heart impact signal is collected in a stable state and changes in a sine-like period, and the amplitude of the pressure of the heart-lung activity within a period of time is F by calculating and amplifying the voltage in direct proportion to the instantaneous stress_mThe variation period is consistent with the breathing frequency. The heart-lung activity distribution is then identified, i.e. the real-time distribution of the heart-lung movement amplitudes, F, monitored by the sensors is analyzed_m，

The larger the sensor is, the stronger the respiratory and heartbeat movement intensity monitored by the sensor is.

The invention designs a respiratory frequency self-adaptive feature extraction method by combining a calculus and a Merlan formula. When the model is in a stable state, the model meets the working conditions, and the voltage and the respiratory frequency output by the sensor are utilized to establish the respiratory amplitude F_mFeatures of linear dependenceAnd mapping, namely adjusting the sampling length and the phase difference according to the respiratory frequency of the section of signal to obtain the more accurate respiratory amplitude characteristic of the piezoelectric ceramic sensor. Through the comparison of the vibration tests of the stable vibration source on different sensors, the influence of the super parameters among the sensors is eliminated, and the real-time distribution of the respiratory motion amplitude during the sleep period is obtained. The average amplitude of the heartbeat can be calculated through the voltage signal of the heartbeat frequency band

The linear correlation characteristic mapping is that the heartbeat signal has periodicity but is not a standard sinusoidal signal, and the heartbeat average amplitude characteristic can be obtained by combining the change of the heartbeat rate

Thereby obtaining the real-time distribution of the heartbeat motion average amplitude during the sleep period.

The invention combines the medical characteristics of the piezoelectric effect, the equivalent circuit and the mixed heart impact signal with the Mecanolin limit thought to establish a mapping relation between the amplitude of the respiratory motion and the mean amplitude of the heartbeat motion and the output voltage signal. The signal features extracted by the invention have strong anti-interference performance, are easy to distinguish data states, solve the problem of difference of absolute pressures of different test scenes and different test groups, overcome the great demand of a conventional discrimination model on the sensor, reduce the data acquisition amount, optimize the distribution features by utilizing the multi-mode fusion of the respiration signals and the heartbeat signals and adjusting the period and the phase based on the heart rate and the respiration rate, standardize the signal length corresponding to the distribution features and obtain the joint features of two key physiological activities of the heartbeat and the respiration. Compared with the traditional absolute pressure distribution characteristic, the distribution characteristic of the invention has better ductility and has great advantages in sleep posture identification and data state combined monitoring.

Fig. 9 is a schematic structural diagram of a sleeping posture identifying device based on a piezoelectric ceramic sensor, as shown in fig. 9, including:

the mixed cardiac shock signal acquisition module 910 is configured to acquire a mixed cardiac shock signal of a chest and abdomen region of a user based on a piezoelectric ceramic sensor system when the user is in a stable sleep state, where the piezoelectric ceramic sensor system includes a plurality of piezoelectric ceramic sensors;

a heart-lung activity distribution feature generating module 920, configured to determine a heart-lung activity distribution feature of the user based on the mixed heart attack signal;

a sleeping posture identification result generating module 930, configured to input the cardiopulmonary activity distribution characteristics and preset environment vector characteristics into a trained sleeping posture identification classification network model, so as to obtain a sleeping posture identification result of the user;

According to the sleeping posture recognition device based on the piezoelectric ceramic sensor, provided by the embodiment of the invention, under the condition that a user is in a stable sleeping state, a mixed cardiac shock signal of a chest and abdomen area of the user is obtained based on a piezoelectric ceramic sensor system; the method comprises the steps of obtaining cardiopulmonary activity distribution characteristics of a user based on extraction and processing of mixed cardiac shock signals, inputting the cardiopulmonary activity distribution characteristics and preset environment vector characteristics of the user into a trained sleep posture recognition classification network model, obtaining a sleep posture recognition result of the user, realizing non-invasive real-time sleep posture monitoring of the user, and having strong universality and environment anti-interference capability.

The sleeping posture identifying device based on the piezoelectric ceramic sensor can be used for executing the method embodiments, the principle and the technical effect are similar, and the details are not repeated here.

Fig. 10 is a schematic structural diagram of an electronic device provided in the present invention, and as shown in fig. 10, the electronic device may include: a processor (processor)1010, a communication Interface (Communications Interface)1020, a memory (memory)1030, and a communication bus 1040, wherein the processor 1010, the communication Interface 1020, and the memory 1030 communicate with each other via the communication bus 1040. Processor 1010 may invoke logic instructions in memory 1030 to perform the piezo ceramic sensor based sleep gesture recognition method, comprising: under the condition that a user is in a stable sleep state, acquiring a mixed cardiac shock signal of a chest and abdomen area of the user based on a piezoelectric ceramic sensor system, wherein the piezoelectric ceramic sensor system comprises a plurality of piezoelectric ceramic sensors; determining a cardiorespiratory activity distribution characteristic of the user based on the blended ballistocardiographic signal; inputting the cardiopulmonary activity distribution characteristics and preset environment vector characteristics into a trained sleeping posture recognition classification network model to obtain a sleeping posture recognition result of the user; the trained sleeping posture identification and classification network model is obtained by training according to cardiopulmonary activity distribution characteristics carrying sleeping posture labels and environment vector characteristic samples.

Furthermore, the logic instructions in the memory 1030 can be implemented in software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.

In another aspect, the present invention also provides a computer program product, the computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions, when the program instructions are executed by a computer, the computer being capable of executing the method for recognizing sleep posture based on piezoelectric ceramic sensor provided by the above methods, the method comprising: under the condition that a user is in a stable sleep state, acquiring a mixed cardiac shock signal of a chest and abdomen area of the user based on a piezoelectric ceramic sensor system, wherein the piezoelectric ceramic sensor system comprises a plurality of piezoelectric ceramic sensors; determining a cardiorespiratory activity distribution characteristic of the user based on the blended ballistocardiographic signal; inputting the cardiopulmonary activity distribution characteristics and preset environment vector characteristics into a trained sleeping posture recognition classification network model to obtain a sleeping posture recognition result of the user; the trained sleeping posture identification and classification network model is obtained by training according to cardiopulmonary activity distribution characteristics carrying sleeping posture labels and environment vector characteristic samples.

In yet another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, the computer program, when being executed by a processor, implementing the piezoceramic sensor-based sleep posture identification method provided by the above methods, the method comprising: under the condition that a user is in a stable sleep state, acquiring a mixed cardiac shock signal of a chest and abdomen area of the user based on a piezoelectric ceramic sensor system, wherein the piezoelectric ceramic sensor system comprises a plurality of piezoelectric ceramic sensors; determining a cardiorespiratory activity distribution characteristic of the user based on the blended ballistocardiographic signal; inputting the cardiopulmonary activity distribution characteristics and preset environment vector characteristics into a trained sleeping posture recognition classification network model to obtain a sleeping posture recognition result of the user; the trained sleeping posture identification and classification network model is obtained by training according to cardiopulmonary activity distribution characteristics carrying sleeping posture labels and environment vector characteristic samples.

The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.

Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.

Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims

1. A sleeping posture identification method based on a piezoelectric ceramic sensor is characterized by comprising the following steps:

under the condition that a user is in a stable sleep state, acquiring a mixed cardiac shock signal of a chest and abdomen area of the user based on a piezoelectric ceramic sensor system, wherein the piezoelectric ceramic sensor system comprises a plurality of piezoelectric ceramic sensors; the number of the piezoelectric ceramic sensor acquisition channels is 10-30;

wherein, the determining the distribution characteristics of the cardiopulmonary activity of the user based on the mixed heart impact signal specifically comprises:

determining the cardiorespiratory activity distribution characteristics of the user according to the cardiorespiratory activity intensity characteristics of the user and the parameter eliminating processing parameters in the piezoelectric ceramic sensor system;

wherein, the calculating the cardiorespiratory activity intensity characteristics of the user based on the mixed heart impact signal specifically comprises:

performing feature extraction of preset signal length on the cardiorespiratory activity stress amplitude information of the user to obtain cardiorespiratory activity intensity features of the user;

the sleep posture identification and classification network model is constructed based on a compression and excitation network classification network framework, and the trained sleep posture identification and classification network model is obtained by training according to cardiopulmonary activity distribution characteristics carrying sleep posture labels and environment vector characteristic samples; the preset environment vector characteristics comprise the strength ratio of the heartbeat motion amplitude to the respiratory motion amplitude, the total number of the sensors occupied by the user when the user lies down in a standard manner, the total number of the sensors occupied by the user when the user lies down on one side, the judgment of the current data state and the judgment of the data state at the previous moment.

2. The sleeping posture identification method based on the piezoceramic sensor according to claim 1, wherein under the condition that the user is in a stable sleep state, the mixed cardiac shock signal of the chest and abdomen area of the user is acquired based on the piezoceramic sensor system, and specifically comprises the following steps:

3. The sleeping posture identification method based on the piezoceramic sensor according to claim 1, wherein the cardiopulmonary activity distribution characteristics of the user are determined according to the cardiopulmonary activity intensity characteristics of the user and the parameter elimination processing parameters in the piezoceramic sensor system, and specifically are as follows:

4. The sleeping posture recognition method based on the piezoceramic sensor according to claim 1, wherein before inputting the cardiopulmonary activity distribution characteristics and the preset environment vector characteristics into the trained sleeping posture recognition classification network model, the method further comprises:

5. The sleeping posture identification method based on the piezoelectric ceramic sensor, according to claim 4, characterized in that the sleeping posture identification classification network model is trained by using a plurality of groups of training samples, specifically:

6. The utility model provides a sleep appearance recognition device based on piezoceramics sensor which characterized in that includes:

the mixed cardiac shock signal acquisition module is used for acquiring a mixed cardiac shock signal of a chest and abdomen area of a user based on a piezoelectric ceramic sensor system under the condition that the user is in a stable sleep state, wherein the piezoelectric ceramic sensor system comprises a plurality of piezoelectric ceramic sensors; the number of the piezoelectric ceramic sensor acquisition channels is 10-30;

7. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program implements the steps of the piezoceramic sensor-based sleep posture identification method according to any one of claims 1 to 5.

8. A non-transitory computer readable storage medium, on which a computer program is stored, wherein the computer program, when being executed by a processor, implements the steps of the piezoceramic sensor-based sleep posture identification method according to any one of claims 1 to 5.