Intervention method and device for apnea event

Abstract

The invention discloses an intervention method and device for an apnea event, and the method comprises the steps: obtaining a pressure signal collected by a pressure detection unit, predicting the sleeping posture and position information of a target object based on the pressure signal, and enabling the position information to comprise a chest position and an abdomen position; based on the chest position and the abdomen position, a chest respiration signal and an abdomen respiration signal are extracted respectively; determining whether the target object has an apnea event or not based on the chest breathing signal and the abdomen breathing signal; if the target object has the apnea event, determining an intervention scheme corresponding to the apnea event based on the sleeping posture and the position information, and adjusting the air pressure of the corresponding air bag based on the intervention scheme. The influence of different sleep positions and sleeping postures of the user on the accuracy of the measurement result is effectively reduced, the influence of apnea on the user is reduced on the premise of not influencing the sleep of the user as much as possible, accurate apnea judgment and personalized intervention are realized, and the monitoring cost is reduced.

Description

Technical Field

[0001] This application relates to the technical field of smart home, and particularly to an intervention method, device, computer device and storage medium for apnea events. Background Art

[0002] Sleep apnea encompasses various types such as obstructive sleep apnea (OSA) and central sleep apnea (CSA). Its hazards are extensive and profound, and are closely related to factors such as apnea duration, frequency, and degree of hypoxia, seriously threatening the health of multiple organ systems.

[0003] Currently, polysomnography (PSG), as the gold standard device for sleep apnea measurement at the laboratory level, can accurately calculate the apnea-hypopnea index (AHI), respiratory disturbance index (RDI), and oxygen desaturation index (ODI) by collecting multi-dimensional information such as electroencephalogram (EEG), electrooculogram (EOG), electromyogram (EMG), oronasal airflow, thoracoabdominal respiratory movement, and blood oxygen, and can accurately distinguish central and obstructive events and evaluate the sleep structure. However, PSG has problems such as complex equipment, the need for in-hospital monitoring, high single cost, and poor comfort, restricting its application scenarios mainly to laboratories and hospital sleep departments. Although portable grading diagnostic devices reduce the collection of some physiological signals and can be used at home, their comfort and convenience for long-term use are still insufficient, hindering their large-scale promotion. In recent years, emerging monitoring methods such as millimeter-wave radar, piezoelectric film mattresses, and smart watches have advantages in user body sensation and the feasibility of long-term monitoring, but are easily interfered by factors such as the environment, device wearing, and movement state, resulting in poor monitoring accuracy and the inability to intervene in a timely manner. Summary of the Invention

[0004] Based on this, it is necessary to provide an intervention method and device for apnea events to solve at least one of the above-mentioned problems in view of the above technical problems.

[0005] In a first aspect, an intervention method for apnea events is provided, which is applied to a respiratory intervention device. A plurality of pressure detection units and airbags are provided on the respiratory intervention device. The method includes: Obtain the pressure signal collected by the pressure detection unit, and based on the pressure signal, predict the sleeping posture and position information of the target object, where the position information includes the chest position and the abdominal position; Based on the chest position and the abdominal position, extract the chest breathing signal and the abdominal breathing signal respectively; Based on the chest breathing signal and the abdominal breathing signal, determine whether the target object has a apnea event; If the target object has a apnea event, based on the sleeping posture and the position information, determine an intervention plan corresponding to the apnea event, so as to adjust the air pressure of the corresponding airbag based on the intervention plan.

[0006] In a possible implementation manner, the determining whether the target object has a apnea event based on the chest breathing signal and the abdominal breathing signal includes: Based on the chest breathing signal and the abdominal breathing signal, determine the chest and abdomen signal features, the chest and abdomen breathing correlation features, and the chest and abdomen breathing time dimension features; Based on the determined chest and abdomen signal features, the chest and abdomen breathing correlation features, and the chest and abdomen breathing signal time dimension features, determine whether the target object has a apnea event.

[0007] In a possible implementation manner, after adjusting the air pressure of the corresponding airbag based on the intervention plan, further include: Detect whether the chest and abdomen breathing correlation features change; If the chest and abdomen breathing correlation features change, determine whether the change of the chest and abdomen breathing correlation features meets the preset intervention conditions; If it does not meet the preset intervention conditions, increase the adjustment range of the corresponding airbag; If it meets the preset intervention conditions, restore the current air pressure of the corresponding airbag to the air pressure before the intervention.

[0008] In a possible implementation manner, the determining the chest and abdomen signal features, the chest and abdomen breathing correlation features, and the chest and abdomen breathing time dimension features based on the chest breathing signal and the abdominal breathing signal includes: Respectively extract the chest signal features corresponding to the chest breathing signal and the abdominal signal features corresponding to the abdominal breathing signal, where the chest signal features include chest signal phase information and chest signal amplitude information, and the abdominal signal features include abdominal signal phase information and abdominal signal amplitude information; Based on the chest signal phase information, the abdominal signal phase information, the chest signal amplitude information, and the abdominal signal amplitude information, determine the chest and abdomen breathing correlation features; Determine the time - dimension features of the chest - abdomen respiratory signal based on the chest signal phase information, abdominal signal phase information, chest signal amplitude information, and abdominal signal amplitude information.

[0009] In a possible implementation, the steps of separately extracting the chest signal features corresponding to the chest respiratory signal and the abdominal signal features corresponding to the abdominal respiratory signal include: Parse the chest respiratory signal and the abdominal respiratory signal respectively, and obtain the chest signal phase information and the abdominal signal phase information based on the parsed chest respiratory signal and abdominal respiratory signal respectively; Based on the effective peak points and effective valley points of the chest respiratory signal and the abdominal respiratory signal, determine the chest respiratory amplitude information and the abdominal respiratory signal amplitude information respectively; Calculate the respiratory cycle variability based on the effective peak points of the chest respiratory signal and the abdominal respiratory signal.

[0010] In a possible implementation, the steps of separately parsing the chest respiratory signal and the abdominal respiratory signal, and obtaining the chest signal phase information and the abdominal signal phase information based on the parsed chest respiratory signal and abdominal respiratory signal respectively include: Perform band - pass filtering on the chest respiratory signal and the abdominal respiratory signal respectively; Perform Hilbert transform on the filtered chest respiratory signal and abdominal respiratory signal to obtain the transformed chest respiratory signal and abdominal respiratory signal; Based on the parsed chest respiratory signal and abdominal respiratory signal, obtain the chest respiratory signal phase and the abdominal respiratory signal phase; Perform phase unwrapping on the chest respiratory signal phase and the abdominal respiratory signal phase to obtain the chest signal phase information and the abdominal signal phase information.

[0011] In a possible implementation, the position information includes the body inclination of the target object. The steps of determining an intervention plan corresponding to the apnea event based on the sleeping position and the position information include: Based on the position information, determine the contact state between the target object and each airbag; Classify each airbag based on the body inclination of the target object to obtain the corresponding type of each airbag; Based on the sleeping position, contact state, and the corresponding type of each airbag, determine an intervention plan corresponding to the apnea event.

[0012] In a possible implementation, the steps of determining an intervention plan corresponding to the apnea event based on the sleeping position, contact state, and the corresponding type of each airbag include: If the sleeping position is supine or prone, based on the contact state, determine whether there are multiple contact airbags of the same type that have a contact relationship with the target object; If there are multiple contact airbags of the same type, determine the distribution mode corresponding to the multiple contact airbags of the same type, so as to correspondingly adjust the air pressure of the contact airbags based on the distribution mode; If there are no multiple contact airbags of the same type, correspondingly adjust the air pressure of airbags of different types respectively.

[0013] In a possible implementation manner, the extracting the chest respiration signal and the abdominal respiration signal respectively based on the chest position and the abdominal position includes: Based on the chest position and the abdominal position, determine the chest area and the abdominal area respectively, where the chest area and the abdominal area are respectively composed of corresponding pressure detection units to form a first pressure matrix and a second pressure matrix; Determine the first pressure mean value corresponding to the first pressure matrix, and use the first pressure mean value as the chest respiration signal; Determine the second pressure mean value corresponding to the second pressure matrix, and use the second pressure mean value as the abdominal respiration signal.

[0014] In a second aspect, a device for intervening in apnea events is provided, which is applied to a respiratory intervention device. The respiratory intervention device is provided with a plurality of pressure detection units and airbags. The device includes: A prediction unit, configured to obtain the pressure signal collected by the pressure detection unit, and based on the pressure signal, predict the sleeping position and position information of the target object, where the position information includes the chest position and the abdominal position; A chest and abdomen signal extraction unit, configured to respectively extract a chest respiration signal and an abdominal respiration signal based on the chest position and the abdominal position; An apnea event determination unit, configured to determine whether the target object has an apnea event based on the chest respiration signal and the abdominal respiration signal; An apnea intervention unit, configured to, if the target object has an apnea event, determine an intervention plan corresponding to the apnea event based on the sleeping position and the position information, so as to adjust the air pressure of the corresponding airbag based on the intervention plan.

[0015] In a third aspect, a computer device is provided, including a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor. When the processor executes the computer-readable instructions, the above method for intervening in apnea events is implemented.

[0016] Fourthly, a readable storage medium is provided. When the computer-readable instructions are executed by one or more processors, the one or more processors are caused to execute the intervention method for the apnea event as described above.

[0017] For the above intervention method and device for the apnea event, the implementation of the method includes: obtaining the pressure signal collected by the pressure detection unit, and predicting the sleeping posture and position information of the target object based on the pressure signal, where the position information includes the chest position and the abdominal position; extracting the chest respiration signal and the abdominal respiration signal respectively based on the chest position and the abdominal position; determining whether the target object has an apnea event based on the chest respiration signal and the abdominal respiration signal; if the target object has an apnea event, determining an intervention plan corresponding to the apnea event based on the sleeping posture and the position information, so as to adjust the air pressure of the corresponding airbag based on the intervention plan. In the embodiment of the present application, the pressure detection unit is used to collect the pressure signal in real time, and the advanced algorithm is used to deeply analyze the pressure data to accurately lock the sleeping posture, chest and abdominal positions of the target object. This unique design effectively avoids the interference of the movement of the user's sleeping position and the frequent change of the sleeping posture on the monitoring result. Compared with the traditional monitoring method, the stability and accuracy of data collection are significantly improved. On this basis, the system can accurately identify different types of apnea events through the intelligent extraction and comparative analysis of the chest and abdominal respiration signals. No matter which type of sleep apnea, it can achieve fast and accurate judgment. When an apnea event is detected, the system will automatically match the preset personalized intervention plan according to the user's real-time sleeping posture, and effectively relieve the apnea symptoms by accurately adjusting the air pressure of the corresponding airbag, while minimizing the interference to the user's sleep and reducing the health risks such as cardiovascular diseases and cognitive disorders caused by apnea. This technical solution not only realizes a comfortable and convenient home monitoring experience, getting rid of the limitations of the traditional polysomnography (PSG) which requires hospitalization, has complex equipment and high costs, but also breaks through the technical bottlenecks of poor comfort of portable devices and insufficient accuracy of emerging monitoring technologies, and provides a more efficient and feasible solution for the monitoring and treatment of sleep apnea in an innovative way with low cost, high precision and humanization. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0019] Figure 1It is a schematic diagram of an implementation environment of the pressure detection unit and the airbag deployment method of the breathing intervention device in an embodiment of the present application; Figure 2 It is a schematic flowchart of a method for intervening in apnea events in an embodiment of the present application; Figure 3 It is a schematic diagram of the model structure of a method for a prediction model in an embodiment of the present application; Figure 4 It is a schematic diagram of the imaging result when the target object is in the supine position in an embodiment of the present application; Figure 5 It is a schematic diagram of the movement of the chest position frame and the abdominal position frame when the chest and abdomen positions of the target object change in an embodiment of the present application; Figure 6 It is a signal schematic diagram of the method for selecting the wave peaks and wave valleys of the breathing signal in an embodiment of the present application; Figure 7 It is a schematic diagram of an implementation scenario of the classification method of each airbag when the target object is in the supine position in an embodiment of the present application; Figure 8 It is a schematic diagram of the structure of an intervention device for apnea events in an embodiment of the present application; Figure 9 It is a schematic diagram of a computer device in an embodiment of the present application. Detailed implementation manners

[0020] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0021] In one embodiment, as Figure 1 , shown in FIG. 2, a method for intervening in apnea events is provided, which is applied to a breathing intervention device. A plurality of pressure detection units and airbags are provided on the breathing intervention device, and the method includes the following steps: In step S110, obtain the pressure signal collected by the pressure detection unit, and based on the pressure signal, predict the sleeping posture and position information of the target object, where the position information includes the chest position and the abdominal position; It should be noted that the breathing intervention device can be intelligent furniture such as an intelligent bed, an intelligent mattress, a sofa, etc. for users to sit, lie and rest. Multiple pressure sensing units arranged in an array can be set thereon, such as rectangular array, circular array, hexagonal array, etc. Such an arrangement can evenly sense the pressure changes at different positions on the surface of the breathing intervention device. For example, for a mattress, no matter whether the target object, such as a human body, is lying supine, on the side or prone, it can more accurately capture the pressure distribution of each part of the body on the mattress.

[0022] Among them, the pressure sensing unit is a component that can sense pressure changes and convert them into electrical signals or other detectable signals, such as a pressure sensor.

[0023] It should be noted that multiple airbags are also provided on the breathing intervention device, and the multiple airbags can be distributed in the breathing intervention device. The airbag can be connected to a barometric pressure sensor for monitoring the air pressure in the airbag. When the user is at any position on the intelligent monitoring device, the change in the air pressure in the airbag can be monitored. It can be understood that the pressure sensing unit can be arranged above the airbag, and multiple pressure sensing units can be deployed in one airbag deployment area, that is, one airbag can correspond to a pressure sub-matrix composed of multiple pressure sensing units. Such a design can achieve refined monitoring of the pressure in each airbag area. When a certain airbag undergoes a pressure change, the corresponding pressure sub-matrix can quickly and accurately sense this change and convert it into data such as electrical signals for transmission and analysis. For example, in an intelligent mattress, when a certain part of the user presses on the area corresponding to the airbag, the pressure sub-matrix corresponding to this airbag area can accurately measure information such as the magnitude and change trend of the pressure. By analyzing these data, the force condition of the user in this area can be inferred, thereby providing data support for analyzing the user's sleep posture, micro-movement of the body during breathing, etc.

[0024] Such as Figure 1As shown in the figure, a schematic diagram of the deployment method of the pressure sensing unit and the airbag of a breathing intervention device is provided. Among them, S11 is a pressure matrix, which can be 64*32, that is, there are 32 pressure sensors in each row, with a total of 64 rows, and the sampling accuracy can be 0.01 KPa. S12 - S17 can be airbags, which can be connected to a barometric pressure sensor. The range of the barometric pressure sensor can be 0 - 40 KPa, and the sampling accuracy can be 0.5 Pa. Taking the breathing intervention device as a mattress and the target object as a human body, the airbag can be deployed according to the supine state of the human body. Among them, S12 can be the position of the airbag on the left user's shoulder, S13 is the position of the airbag on the left user's waist, S14 is the position of the airbag on the left user's hip, S15 is the position of the airbag on the right user's shoulder, S16 is the position of the airbag on the right user's waist, and S17 is the position of the airbag on the right user's hip. Deploying S12 - S17 in the above manner can effectively sense the changes in the airbag air pressure in the chest and abdomen of the human body, so as to better measure the breathing depth. It should be noted that when the breathing intervention device is different and the target object is different, the deployment method of the airbag and the pressure matrix is different, and it can be deployed correspondingly according to actual needs. It should be noted that each airbag is connected to an air pump through a pipeline and can be inflated and deflated independently to perform apnea intervention.

[0025] Specifically, after generating a pressure signal through the pressure matrix, the validity of the pressure signal can be detected first. If the pressure signal is valid, the pressure signal can be input into a pre-trained prediction model to predict the sleeping position, chest position, and abdominal position of the target object. For example Figure 3As shown in the figure, the prediction model may include an input layer S21, a first convolutional layer S22, a region extraction layer S23, a second convolutional layer S24, a first fully connected layer S25, a second fully connected layer S26, and an output layer S27. The collected pressure signal is input into the input layer S21, and its tensor size is 64*32*1. After being input into the first convolutional layer S22 for convolution processing, a tensor with a size of 64*32*16 can be obtained. In this convolutional layer, K3 represents a convolutional kernel size of 3*3, s1 represents a stride of 1, p1 represents padding-based maxpooling, and c16 represents 16 convolutional kernels. The tensor with a size of 64*32*16 obtained after being processed by the first convolutional layer S22 is input into the region extraction layer S23 for region extraction. The region extraction method is to directly obtain it through a fixed-size window and a single-step sliding window. The fixed window sizes include 8*4, 8*8, 8*16, 8*32, 16*4, 16*8, 16*16, 16*32, 32*4, 32*8, 32*16, 32*32, 64*4, 64*8, 64*16, 64*32. The single-step sliding window means that each time the window is updated, it only moves one data point horizontally or vertically. The tensor extracted from the region is subjected to a 1*1 convolution operation, i.e., flattening, by the second convolutional layer S24 to obtain a one-dimensional array of size 16RC, where R is the fixed window row size and C is the fixed window column size. Taking an 8*4 fixed window as an example, R is 8 and C is 4, and a one-dimensional array of size 512 is obtained after flattening. Then, the first fully connected layer S25 performs a fully connected process on the flattened one-dimensional feature array after region extraction to obtain a one-dimensional array of size 1000. The second fully connected layer S26 performs a fully connected process on the one-dimensional array of size 1000 again to obtain a one-dimensional array of size 500. Finally, the output layer S27 can output a one-dimensional vector with a length of 17. It may include 4 kinds of sleeping posture probabilities and 13 kinds of position information. Among them, the sleeping postures may include supine, prone, side-lying, and other sleeping postures. It can be obtained through the activation function softmax. In addition, the 13 kinds of position information may include the abscissa of the user center , the ordinate of the user center , the height of the user box , the width of the user box , the abscissa of the chest center , the ordinate of the chest center , the width of the chest box , the height of the chest box , the abscissa of the abdomen center , the ordinate of the abdomen center , the width of the abdomen box , the height of the abdomen box , the body tilt angle It is obtained by linearly transforming and weighting the previous fully connected layer. The coordinate system of the horizontal and vertical coordinates takes the vertex in the upper left corner of the pressure matrix as the coordinate origin, with the horizontal rightward as the positive direction of the horizontal axis and the vertical downward as the positive direction of the vertical axis. The body tilt angle refers to the angle formed by the straight line connecting the feet to the head of a person and the vertically upward direction.

[0026] It should be noted that for different fixed window sizes, the S25 - S27 model parameters are not shared, that is, the corresponding parameters are different. For each extraction area, there is a one - dimensional output with a length of 17. That is to say, for the original input of 64*32*1, there will be a total of 10080 (140*72, 59 + 49 + 33 + 1, 29 + 25 + 17 + 1) one - dimensional outputs with a length of 17. Taking the case of a user lying on one side supine, the imaging result is as Figure 4 shown. The white dotted line box of S31 is the position box of the user. 、 is its central coordinate point. is its height. is its width; the white dotted line box of S32 is the position where the user's chest is located. 、 is its central coordinate point. is its height. is its width; the white dotted line box of S33 is the position where the user's abdomen is located. 、 is its central coordinate point. is its height. is its width.

[0027] Optionally, the validity detection of the pressure signal refers to determining whether the pressure corresponds to the pressure signal of the target object on the respiratory intervention device, as well as detecting its stability and signal quality, so as to eliminate the pressure signals generated when the target object is not on the respiratory intervention device, the pressure signals in an unstable state, and the pressure signals with poor signal quality, so as to avoid the interference of the signals on the final measurement result and affect the measurement accuracy.

[0028] Among them, when determining whether the pressure signal is the pressure signal collected by the target object on the respiratory intervention device, the sum value of all obtained pressure signals can be calculated. If the sum value is greater than the preset value TH1; and the number of pressure values greater than the preset value TH2 in the pressure signal is greater than TH3; and the preset window time length, such as the sum of the pressure signals within 5s; and perform band - pass filtering processing on it, such as [0.1, 2]Hz. If the sum of the absolute values of the pressure signals after band - pass filtering is greater than the preset value TH4, and all the above three conditions are met simultaneously, it is considered that the target object exists on the respiratory intervention device, otherwise it is considered that the target object does not exist on the respiratory intervention device.

[0029] It should be noted that TH1 is strongly related to the sensor characteristics. It can collect the data of the matrix pressure sensor on the breathing intervention device when the target object exists, sum the matrix pressure sensor signals at the current moment, and calculate the lower quartile of the signal sum in all scenarios. TH1 is obtained by multiplying the lower quartile by a certain coefficient, such as 0.5. TH2 can be an empirical value. For example, an object of a fixed size (such as 5 cm * 5 cm) and a fixed weight (such as 250 g) can be placed at different positions on different mattresses, and the maximum value of the pressure matrix in the placement area is statistically calculated, and the mean value of this maximum value is used as TH2. TH3 can be an empirical value. First, determine the coverage area of a single sensor. For example, the entire pressure sensor has a coverage area of 150 cm * 160 cm and 32 * 32 sensor points, so the coverage area of a single sensor is 4.68 cm * 5 cm. Then, according to the lower quartile of the contact area of the detected object lying on the breathing intervention device obtained statistically, multiply it by a certain coefficient, such as 0.5. Finally, divide the obtained area by the coverage area of a single sensor and round down the value to obtain the corresponding TH3. TH4 is strongly related to the sensor characteristics. It can collect the data of the matrix pressure sensor when different populations are distributed on the bed and there is no obvious body movement. Sum the matrix pressure sensor signals at the current moment, then perform a band-pass filter on the signal sum in the range of [0.1, 2] Hz, calculate the sum of the absolute values of the filtered signals within the window time. The window time can be 5 s, and calculate the lower quartile of the sum of the absolute values of all window signals. Then, TH4 is obtained by multiplying the lower quartile by a certain coefficient, such as 0.3.

[0030] Among them, when determining whether the pressure signal corresponds to the pressure signal generated under a stable state, the pressure signal can be summed to obtain the pressure signal sum value. Then, after performing a first-order difference operation on it, take its absolute value. If the absolute value is not greater than the preset value TH5; and the absolute value of the difference between the number of pressure values greater than the preset value TH2 in the pressure signal and the number greater than TH2 at the previous moment is not greater than TH6, it is considered stable, otherwise it is considered unstable.

[0031] It should be noted that TH5 is strongly related to the sensor characteristics. It can collect the data of the matrix pressure sensor when different populations are distributed on the bed and there is no obvious body movement. Sum the matrix pressure sensor signals at the current moment, and then calculate the upper quartile of the first-order difference of the sum of the absolute values of the signals. Then, TH5 is obtained by multiplying the upper quartile by a certain coefficient, such as 1.5. TH6 can be an empirical value. This threshold is used to evaluate whether a large movement occurs, that is, whether there is a large change in the contact area of the pressure matrix at adjacent moments. It is obtained by rounding up the allowable change area threshold divided by the coverage area of a single sensor. For example, it is expected that the change in the effective contact surface (points greater than TH2) is less than 25 cm * 25 cm, and the coverage area of a single sensor is 4.68 cm * 5 cm, and this threshold is 27.

[0032] When determining whether the signal quality of the pressure signal meets the preset signal quality condition, the passband signal f1 within the window time length of 5 s in the frequency range of [0.1, 2] Hz in the pressure matrix and the high-frequency component f2 with a cut-off frequency of 2 Hz in the pressure matrix can be extracted, and the ratio of the sum of the absolute values of f1 to the absolute value of f2 is calculated. If the ratio is less than the preset value TH7, the signal quality is considered too low; otherwise, the signal quality is considered to meet the requirements.

[0033] It should be noted that TH7 can be an empirical value used to evaluate the signal-to-noise ratio and is related to the sensor, such as 1.5.

[0034] In step S120, based on the chest position and the abdominal position, the chest respiration signal and the abdominal respiration signal are extracted respectively. It should be noted that before extracting the chest respiration signal and the abdominal respiration signal, it is necessary to first determine whether the sleeping posture of the target object has changed or whether the chest and abdominal positions have changed. Among them, a change in the sleeping posture means that the sleeping posture at the current moment is different from that at the previous moment. For example, if it changes from supine to lateral, it can be considered that the sleeping posture has changed. The sleeping posture at the current moment refers to the sleeping posture with the largest proportion within a preset time window TH8, such as 5 s. A change in the chest and abdominal positions means that if the difference between the coverage areas of the chest and abdominal position frames at the current moment and the previous moment and the area of the chest and abdominal position frame at the previous moment is greater than the preset threshold TH9, such as 0.25, it is considered that the position has changed significantly; otherwise, it is considered that the position change is not obvious. The area difference is Figure 5 the shaded area in, and the chest and abdominal position frame refers to the union of the chest position frame and the abdominal position frame.

[0035] If it is detected that the sleeping posture has changed or the chest and abdominal positions have changed significantly, the cache variables related to the respiration event detection are initialized. The cache variables can include the chest respiration signal cache array, the chest respiration signal phase cache array, the chest respiration signal amplitude cache array, the respiratory cycle variability (RCV) cache array, the abdominal respiration signal cache array, the abdominal respiration signal phase cache array, the abdominal respiration signal amplitude cache array, the abdominal respiration cycle variability RCV, and the chest-to-abdominal movement ratio cache array. This means that some variables storing data related to the respiration event detection are restored to the initial state, which can ensure that in the new state, the respiration event detection system can more accurately monitor and analyze respiration events based on accurate initial data.

[0036] If it is detected that the sleeping posture has not changed or there are no obvious changes in the chest and abdomen positions, the chest breathing signal and the abdominal breathing signal can be extracted. Taking the chest breathing signal as an example, the extraction process can be as follows: Based on the predicted chest position, the chest area is determined, the pressure detection units corresponding to the chest area are determined, and the pressure signals collected by all the pressure detection units corresponding to the chest area are summed to obtain the pressure mean value, which can be used as the chest breathing signal. Similarly, the abdominal breathing signal can also be obtained based on the above method.

[0037] In step S130, based on the chest breathing signal and the abdominal breathing signal, it is determined whether the target object has a apnea event; Optionally, based on the chest breathing signal and the abdominal breathing signal, relevant variables for detecting breathing events can be obtained. These relevant variables include chest and abdomen signal characteristics, chest and abdomen signal correlation characteristics, and chest and abdomen signal time dimension characteristics. Then, the chest and abdomen signal characteristics, chest and abdomen signal correlation characteristics, and chest and abdomen signal time dimension characteristics can be combined to form a feature array, and this feature array is input into a preset model for prediction to obtain the apnea event detection result. The apnea event detection result includes the apnea event type and the probability of each apnea event type. If the probability is greater than the preset threshold, it indicates that an apnea event has occurred, and the apnea event type can also be determined, such as obstructive sleep apnea (OSA), central sleep apnea (CSA), or mixed apnea, etc. It should be noted that the preset model can be a decision tree model or other classification models. Or, it is also possible to determine whether an apnea event has occurred by combining feature detection and logical judgment.

[0038] In step S140, if the target object has an apnea event, based on the sleeping posture and the position information, an intervention plan corresponding to the apnea event is determined to adjust the air pressure of the corresponding airbag based on the intervention plan.

[0039] Optionally, when a apnea event of the target object is detected, a corresponding intervention plan can be selected based on the current sleeping posture of the target object to adjust the airbag pressure, so as to change the current sleeping posture of the target object, and then intervene in the apnea event, realizing the intervention of apnea events during sleep. On the premise of minimizing the impact on the target object's sleep, the impact of apnea on the target object is reduced. It should be noted that different apnea events can correspond to different intervention plans, and different intervention plans can include the air pressure adjustment ranges for different airbags. If no apnea event of the target object is detected, step S120 can be returned, and the chest respiration signal and the abdominal respiration signal can be extracted again based on the pressure signal collected at the current moment, and then the judgment of the apnea event can be carried out to realize the continuous cyclic judgment of the apnea event.

[0040] In the embodiment of the present application, the pressure detection unit is used to collect pressure signals in real time, and advanced algorithms are used to deeply analyze the pressure data to accurately lock the sleeping posture, chest and abdominal positions of the target object. This unique design effectively avoids the interference of the user's sleep position movement and frequent changes in sleeping postures on the monitoring results. Compared with traditional monitoring methods, it significantly improves the stability and accuracy of data collection. On this basis, the system can accurately identify different types of apnea events through the intelligent extraction and comparative analysis of chest and abdominal respiration signals. No matter which type of sleep apnea, it can achieve fast and accurate judgment. When an apnea event is detected, the system will automatically match a pre-set personalized intervention plan according to the user's real-time sleeping posture, and effectively relieve the apnea symptoms by precisely adjusting the air pressure of the corresponding airbag, while minimizing the interference to the user's sleep, reducing health risks such as cardiovascular diseases and cognitive impairments caused by apnea. This technical solution not only realizes a comfortable and convenient home monitoring experience, getting rid of the limitations of traditional polysomnography (PSG) which requires hospitalization, complex equipment and high costs, but also breaks through the technical bottlenecks of poor comfort of portable devices and insufficient accuracy of emerging monitoring technologies, providing a more efficient and feasible solution for the monitoring and treatment of sleep apnea in an innovative way with low cost, high precision and user-friendliness.

[0041] In an embodiment of the present application, the extracting the chest respiration signal and the abdominal respiration signal based on the chest position and the abdominal position respectively includes: Based on the chest position and the abdominal position, the chest area and the abdominal area are determined respectively, wherein the chest area and the abdominal area are composed of the corresponding pressure detection units to form a first pressure matrix and a second pressure matrix respectively; Determine the first pressure mean value corresponding to the first pressure matrix, and use the first pressure mean value as the chest respiration signal; Determine the second average pressure corresponding to the second pressure matrix, and use the second average pressure as the abdominal breathing signal.

[0042] Optionally, there may be multiple pressure detection units, and they may be arranged in an array. Taking the device as a mattress as an example, when a user lies on the mattress, both the chest and abdominal regions correspond to pressure matrices composed of multiple pressure detection units. Taking the chest position as an example, the chest position output by the prediction model may include the central position coordinates of the chest position box, the height of the box, and the width of the box. Based on the central position coordinates, the height of the box, and the width of the box, the area where the chest is located can be calculated. Then, find all the pressure detection units in the pressure matrix corresponding to the area where the chest is located, obtain the pressure values collected by each pressure detection unit, and take the average value, and use this average value as the chest breathing signal. Similarly, the abdominal breathing signal can also be obtained in the above manner.

[0043] In an embodiment of the present application, based on the chest breathing signal and the abdominal breathing signal, determining whether the target object has an apnea event includes: Based on the chest breathing signal and the abdominal breathing signal, determine the chest and abdomen signal characteristics, the chest and abdomen breathing correlation characteristics, and the chest and abdomen breathing time dimension characteristics; Based on the determined chest and abdomen signal characteristics, the chest and abdomen breathing correlation characteristics, and the chest and abdomen breathing signal time dimension characteristics, determine whether the target object has an apnea event.

[0044] Optionally, based on the chest breathing signal and the abdominal breathing signal, relevant variables for detecting breathing events can be obtained. These relevant variables include chest and abdomen signal characteristics, chest and abdomen signal correlation characteristics, and chest and abdomen signal time dimension characteristics. Then, the chest and abdomen signal characteristics, the chest and abdomen signal correlation characteristics, and the chest and abdomen signal time dimension characteristics can be combined to form a feature array, and the feature array is input into a preset model for prediction to obtain the apnea event detection result. The apnea event detection result includes the apnea event type and the probability of each apnea event type. If the probability is greater than the preset threshold, it indicates that an apnea event has occurred, and the apnea event type can also be determined, such as obstructive sleep apnea (OSA), central sleep apnea (CSA), or mixed sleep apnea, etc. It should be noted that the preset model can be a decision tree model or other classification models.

[0045] Alternatively, it is also possible to determine whether an apnea event has occurred by a combination of feature detection and logical judgment. Exemplarily, if the phase difference of the thoracic-abdominal respiration signal is greater than 120° and the phase difference is gradually increasing; the thoracic respiration amplitude increases by more than a preset threshold TH17, such as 2; the abdominal respiration amplitude increases by more than the preset threshold TH17; the amplitude ratio of the thoracic-abdominal respiration signal is greater than 2:1 and the duration exceeds 10 seconds; if the above conditions are met simultaneously, it is considered that an obstructive sleep apnea (OSA) event occurs currently. If the thoracic respiration amplitude suddenly drops below 10% of the baseline; the abdominal respiration amplitude suddenly drops below 10% of the baseline; the chest-abdominal motion synchrony index (CSI) is lower than 0.2; the thoracic / abdominal respiration amplitudes are both lower than 10% of the baseline and the duration exceeds 10 seconds; if the above conditions are met simultaneously, it is considered that a central sleep apnea (CSA) event occurs. If the thoracic-abdominal respiration amplitude suddenly drops below 10% of the baseline first and the duration does not exceed 10 seconds; subsequently, severe paradoxical movement occurs, that is, the phase difference of the thoracic-abdominal respiration signal is greater than 30° and the phase difference continues to increase and the duration exceeds 5 seconds; then it is considered that a mixed sleep apnea event occurs.

[0046] In an embodiment of the present application, determining the thoracic-abdominal signal characteristics, the thoracic-abdominal respiration correlation characteristics, and the thoracic-abdominal respiration time dimension characteristics based on the thoracic respiration signal and the abdominal respiration signal includes: Respectively extract the thoracic signal characteristics corresponding to the thoracic respiration signal and the abdominal signal characteristics corresponding to the abdominal respiration signal. Among them, the thoracic signal characteristics include thoracic signal phase information and thoracic signal amplitude information, and the abdominal signal characteristics include abdominal signal phase information and abdominal signal amplitude information; Based on the thoracic signal phase information, the abdominal signal phase information, the thoracic signal amplitude information, and the abdominal signal amplitude information, determine the thoracic-abdominal respiration correlation characteristics; Based on the thoracic signal phase information, the abdominal signal phase information, the thoracic signal amplitude information, and the abdominal signal amplitude information, determine the thoracic-abdominal respiration signal time dimension characteristics.

[0047] Among them, the thoracic-abdominal signal characteristics may include thoracic signal characteristics and abdominal signal characteristics. The thoracic signal characteristics may include thoracic respiration signal phase information, thoracic respiration signal amplitude information, and thoracic respiration cycle variability (RCV) information; the abdominal signal characteristics may include abdominal respiration signal phase information, abdominal respiration signal amplitude information, and abdominal respiration cycle variability (RCV) information.

[0048] Among them, the chest and abdomen respiration correlation features include the phase difference between the chest respiration phase and the abdomen respiration phase, the ratio of the chest respiration amplitude to the abdomen respiration amplitude, and the chest and abdomen movement synchronization index CSI. Based on the chest and abdomen signal features, the chest respiration signal phase and the abdomen respiration signal phase can be obtained. By subtracting them, the phase difference can be obtained. Based on the chest and abdomen signal features, the chest respiration amplitude and the abdomen respiration amplitude can be obtained. By dividing them, the ratio can be obtained. The chest and abdomen movement synchronization index CSI can be calculated by the following formula: ; wherein, is the standard deviation of the phase difference within the preset time threshold TH15 window, and the TH15 can be 10 seconds.

[0049] Among them, the time dimension features of the chest and abdomen respiration signals include the amplitude change feature of the chest respiration signal, the amplitude change feature of the abdomen respiration signal, the phase difference change feature of the chest and abdomen respiration signals, and the change feature of the ratio of the chest signal amplitude to the abdomen signal amplitude. The change feature includes the ratio of the mean value of the relevant feature from 3 s before the current moment to the current moment to the baseline value of the feature, and the difference between the feature at the current moment and the feature at the previous moment. It should be noted that the baseline value can be obtained by performing band-pass filtering on the feature, or by calculating the mean value from TH16 before the current moment, such as 6 seconds to the current moment.

[0050] In an embodiment of the present application, the separately extracting the chest signal features corresponding to the chest respiration signal and the abdomen signal features corresponding to the abdomen respiration signal includes: respectively parsing the chest respiration signal and the abdomen respiration signal, and respectively obtaining the chest signal phase information and the abdomen signal phase information based on the parsed chest respiration signal and abdomen respiration signal; based on the effective peak points and effective valley points of the chest respiration signal and the abdomen respiration signal, respectively determining the chest respiration amplitude information and the abdomen respiration signal amplitude information; calculating the respiratory cycle variability based on the effective peak points of the chest respiration signal and the abdomen respiration signal.

[0051] Optionally, after the chest respiration signal and the abdomen respiration signal are extracted, the chest respiration signal and the abdomen respiration signal can be respectively subjected to Hilbert transform to obtain the parsed chest respiration signal and abdomen respiration signal, and then the chest signal phase information and the abdomen signal phase information can be obtained. Search for the effective maximum value and the effective minimum value in the chest respiration signal and the abdomen respiration signal, and accordingly determine the effective peak point and the effective valley point. The respiration amplitude is the difference between the mean value of the effective peak points and the mean value of the effective valley points within the preset window time threshold TH10.

[0052] Respiratory Cycle Variability (RCV) refers to the degree of change or irregularity of the respiratory cycle over time. It reflects the differences between adjacent respiratory cycles. It can be calculated using the following formula: ; Wherein, is the standard deviation of the durations of consecutive respiratory cycles. The duration of consecutive respiratory cycles refers to the time interval between the above-mentioned valid adjacent peak points, is the average value of the durations of consecutive respiratory cycles. The consecutive breaths refer to detecting a preset threshold TH13, such as 50 respiratory cycles, or a TH14 window time length, such as 3 minutes. The respiratory cycles need to be subjected to abnormal value removal and resampling operations. The abnormal value removal can be performed using a window length of 5. If the duration of a respiratory cycle exceeds the mean of the duration of the respiratory cycle, it is removed, where is the standard deviation; the resampling can be cubic spline interpolation on the array of respiratory cycles after abnormal value removal, and the interpolated data can be 10 Hz.

[0053] In an embodiment of the present application, the chest respiratory signal and the abdominal respiratory signal are respectively analyzed, and chest signal phase information and abdominal signal phase information are respectively obtained based on the analyzed chest respiratory signal and abdominal respiratory signal, including: Performing band-pass filtering on the chest respiratory signal and the abdominal respiratory signal respectively; Performing Hilbert transform on the filtered chest respiratory signal and abdominal respiratory signal to obtain the transformed chest respiratory signal and abdominal respiratory signal; Based on the analyzed chest respiratory signal and abdominal respiratory signal, obtaining the chest respiratory signal phase and the abdominal respiratory signal phase; Performing phase unwrapping on the chest respiratory signal phase and the abdominal respiratory signal phase to obtain the chest signal phase information and the abdominal signal phase information.

[0054] Optionally, taking the chest respiratory signal as an example, band-pass filtering can be performed on it. For example, it can be implemented through an IIR-type Butterworth filter, and then a preset window time TH10 is used to perform Hilbert transform on the filtered chest respiratory signal to obtain the analyzed signal, as shown in the following formula: ; Wherein, t is the t-th sampling point, j represents an imaginary number, x(t) is the input signal, represents the Hilbert transform.

[0055] Performing phase calculation on the analyzed signal, specifically as shown in the following formula: ; Among them, arctan represents the arctangent calculation.

[0056] Then, phase unwrapping is performed on the obtained phase to eliminate the 2π jump of the arctangent function, as follows: Starting from the first sampling point, set the initial unwrapped phase ; For each subsequent point i, calculate the phase difference between adjacent points ; If , it indicates a downward jump, and compensate , that is ; If , it indicates an upward jump, and compensate , that is ; Otherwise, keep the original value.

[0057] Record the cumulative integer multiple k of the compensation to ensure the continuity of subsequent points.

[0058] Finally, perform phase smoothing. The smoothing method can be moving average or low-pass filtering to obtain the phase information of the chest signal. Similarly, the phase information of the abdominal signal can also be obtained in this way.

[0059] In an embodiment of the present application, determining the chest respiration amplitude information and the abdominal respiration signal amplitude information based on the effective peak points and effective valley points of the chest respiration signal and the abdominal respiration signal respectively includes: Detect the minimum points and maximum points of the chest respiration signal and the abdominal respiration signal respectively to obtain the effective maximum points and effective minimum points of the chest respiration signal, and the effective maximum points and effective minimum points of the abdominal respiration signal; Based on the effective maximum points and effective minimum points of the chest respiration signal, and the effective maximum points and effective minimum points of the abdominal respiration signal, determine the effective peak points and effective valley points of the chest respiration signal, and the effective peak points and effective valley points of the abdominal respiration signal respectively; Based on the effective peak points and effective valley points of the chest respiration signal, and the effective peak points and effective valley points of the abdominal respiration signal, obtain the chest respiration signal amplitude information and the abdominal respiration signal amplitude information respectively.

[0060] Optionally, as Figure 6As shown in the figure, a schematic diagram of detecting valid peaks and valid valleys of a breathing signal is presented. For the breathing signal, minimum points are detected. If a minimum point is detected, and the minimum value corresponding to the minimum point is less than the preset threshold TH11 - 1, and the time interval between this minimum point and the previous valid minimum point is greater than the preset threshold TH12, such as 2 seconds, then this minimum point is considered a valid minimum. At this time, the flag of this minimum point can be set to a preset value, such as 1, and then the cached valid minimum point and the preset threshold TH11 are updated. Similarly, for the breathing signal, maximum points can be detected. If a maximum point is detected, and the maximum value corresponding to the maximum point is greater than the preset threshold TH11 - 2, and the minimum point flag is 1, then this point is used as a candidate peak point, the maximum value cache array and the preset threshold TH11 - 2 are updated, and the minimum point flag is set to 0. If the minimum point flag is 0, and there is a candidate peak point, and this maximum point is greater than the candidate peak point, then the candidate peak point is updated, that is, this point is used as the candidate peak point. If a valid minimum is detected, the minimum flag is 0, and there is a candidate peak point, then this candidate peak point is used as a valid peak point, the candidate peak point is cleared, and the minimum flag is set to 1, and the minimum value cache array and TH11 - 1 are updated. When detecting valid valley points, the breathing signal can be inverted.

[0061] It should be noted that if no valid peak point is detected within the preset window time threshold TH10, such as 20 seconds, the breathing amplitude is considered 0.

[0062] It should be noted that TH11 - 1 is strongly related to the sensor characteristics. The initial TH11 - 1 is obtained through statistics. Matrix pressure sensor data of different populations lying in bed without obvious body movement can be collected, the current collected breathing signal can be extracted, the minimum value of the breathing signal can be extracted, and then the upper quartile of the minimum value can be statistically calculated. This TH11 - 1 can be the upper quartile * a certain coefficient, such as 0.5. The subsequent TH11 - 1 is updated from the minimum value cache array. TH11 - 2 is strongly related to the sensor characteristics. Matrix pressure sensor data of different populations lying in bed without obvious body movement can be collected, the current collected breathing signal can be extracted, the maximum value of the breathing signal can be extracted, and then the lower quartile of the maximum value can be statistically calculated. Then TH11 - 2 can be the lower quartile * a certain coefficient, such as 0.5. The subsequent TH11 - 2 is updated from the maximum value cache array.

[0063] In an embodiment of the present application, the position information includes the body inclination of the target object. Based on the sleeping position and the position information, determining an intervention plan corresponding to the apnea event includes: Based on the position information, determining the contact state between the target object and each airbag; Classify each of the airbags based on the body inclination of the target object to obtain the corresponding type of each airbag; Determine an intervention plan corresponding to the apnea event based on the sleeping posture, contact state, and the corresponding type of each airbag.

[0064] Optionally, if it is detected that the target object has an apnea event and the duration of the apnea event exceeds a preset threshold TH18, such as 10 seconds, the contact state between the target object and each contacted airbag can be obtained. This contact state can be determined by the abscissa of the user center , the ordinate of the user center , the height of the user frame , the width of the user frame of the user frame formed by the user frame and the positions of each preset airbag, and the body inclination angle of the user is obtained. Specifically, taking Figure 7 as an example, S11 - S17 are respectively the positions of the array pressure sensors and each airbag, the frame shown by S41 is the user frame, the dashed arrow shown by S42 is the vertically upward direction, the arrow shown by S43 is the direction from the user's feet to the head, and the angle between S42 and S43 is . If the overlapping area between the user frame and the position frame of a single airbag is greater than a preset threshold TH19, such as 0.1, it is considered that the airbag has been contacted by the user, and the contact state of the airbag is set to 1, otherwise the contact state is 0; if is in the range of 0° - 45° or 315° - 360°, it is considered that the user is sleeping longitudinally with the head upward, and the airbags from top to bottom are respectively set as Class 1, Class 2, and Class 3. Figure 7 The airbags S12 and S15 shown are of Class 1, S13 and S16 are of Class 2, and S14 and S17 are of Class 3; if is in the range of 45° - 135°, it is considered that the user is sleeping horizontally with the head to the left, and the airbags from left to right are set as Class 1 and Class 2. Figure 7 The airbags S12, S13, and S14 shown are of Class 1, and S15, S16, and S17 are of Class 2; if is in the range of 135° - 225°, it is considered that the user is sleeping longitudinally with the head downward. Figure 7 The airbags S14 and S17 shown are of Class 1, S13 and S16 are of Class 2, and S12 and S15 are of Class 3; if is in the range of 225° - 315°, it is considered that the user is sleeping horizontally with the head to the right. Figure 7 The airbags S15, S16, and S17 shown are of Class 1, and S12, S13, and S14 are of Class 2.

[0065] Then, form a feature array with the sleeping posture, the contact state of each airbag, and the type of each airbag, and use this feature array to determine the intervention plan corresponding to the apnea event.

[0066] In an embodiment of the present application, determining an intervention plan corresponding to the apnea event based on the sleeping posture, contact state, and corresponding types of each airbag includes: If the sleeping posture is supine or prone, based on the contact state, determine whether there are multiple contact airbags of the same type that have a contact relationship with the target object; If there are multiple contact airbags of the same type, determine the corresponding distribution mode of the multiple contact airbags of the same type, so as to correspondingly adjust the air pressure of the contact airbags based on the distribution mode; If there are not multiple contact airbags of the same type, correspondingly adjust the air pressure of airbags of different types respectively.

[0067] Optionally, based on the predicted sleeping posture of the target object, determine whether it is in a supine state or a prone state. If so, determine whether multiple airbags of the same type are in contact. The contact airbags of the same type refer to airbags that belong to the same category after classifying the airbags based on the body inclination of the above-mentioned target object, such as Class 1, Class 2, or Class 3. If there are multiple contact airbags of the same type, determine the corresponding distribution mode of the multiple contact airbags of the same type. For example, the distribution is horizontal or vertical. If the distribution is horizontal, the air pressure of the airbags on the right side contacted by all target objects can be increased, the air pressure of the airbags on the left side contacted by all target objects can be decreased, and a certain difference in air pressure on both sides is maintained. If multiple contacted airbags are vertically distributed, the air pressure of the airbags on the upper side contacted by all target objects is increased, and the air pressure of the airbags on the lower side contacted by all target objects is decreased, so that the user can change from supine to side-lying.

[0068] If it is detected that the target object is in a side-lying state, other sleeping postures, or there are not multiple contact airbags of the same type, the air pressure of airbags of each category can be correspondingly adjusted. For example, the air pressure of Class 1 airbags > the air pressure of Class 2 airbags > the air pressure of Class 3 airbags, so that the target object forms a C shape to adjust the breathing mode.

[0069] In an embodiment of the present application, after adjusting the air pressure of the corresponding airbag based on the intervention plan, it further includes: Detect whether the respiratory correlation characteristics of the chest and abdomen change; If the respiratory correlation characteristics of the chest and abdomen change, determine whether the change in the respiratory correlation characteristics of the chest and abdomen meets the preset intervention conditions; If the preset intervention conditions are not met, increase the adjustment amplitude of the corresponding airbag; If the preset intervention conditions are met, restore the current air pressure of the corresponding airbag to the air pressure before the intervention.

[0070] Among them, the chest and abdomen respiration correlation features include the phase difference between the chest respiration phase and the abdomen respiration phase, the ratio of the chest respiration amplitude to the abdomen respiration amplitude, and the chest and abdomen movement synchronization index CSI.

[0071] If the respiration phase difference gradually shrinks and is less than the preset threshold TH20, such as 120 degrees; the ratio of the chest / abdomen respiration amplitude is within the preset range; the chest and abdomen movement synchronization index is greater than the preset threshold TH21, such as 0.2; if the above requirements are not simultaneously met within the duration threshold TH22, if the airbag pressure is in an adjustable state, the adjustment amplitude of the intervention plan can be increased, and the adjustment amplitude refers to maintaining the pressure difference between different airbags. If the above requirements are simultaneously met within the duration threshold TH22, the airbag pressure distribution before the intervention can be restored, and the airbag pressure distribution before the intervention refers to the average value of the air pressures of each airbag within the preset window time TH23 before the respiratory abnormality event is recognized by S25.

[0072] In the embodiment of the present application, the pressure detection unit is used to collect pressure signals in real time, and advanced algorithms are used to deeply analyze the pressure data to accurately lock the sleeping position, chest and abdomen positions of the target object. This unique design effectively avoids the interference of the movement of the user's sleeping position and frequent changes in sleeping postures on the monitoring results, and significantly improves the stability and accuracy of data collection compared with traditional monitoring methods. On this basis, the system can accurately identify different types of apnea events through intelligent extraction and comparative analysis of the chest and abdomen respiration signals. No matter which type of sleep apnea occurs, it can achieve fast and accurate judgment. When an apnea event is detected, the system will automatically match a pre-set personalized intervention plan according to the user's real-time sleeping posture, and effectively relieve the apnea symptoms and reduce the health risks such as cardiovascular diseases and cognitive disorders caused by apnea by precisely adjusting the air pressure of the corresponding airbag while minimizing the interference to the user's sleep. This technical solution not only realizes a comfortable and convenient home monitoring experience, getting rid of the limitations of traditional polysomnography (PSG) that requires hospitalization, complex equipment and high costs, but also breaks through the technical bottlenecks of poor comfort of portable devices and insufficient accuracy of emerging monitoring technologies, and provides a more efficient and feasible solution for the monitoring and treatment of sleep apnea in an innovative way with low cost, high precision and user-friendliness.

[0073] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution, and the execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0074] In one embodiment, an intervention device for apnea events is provided, and the intervention device for apnea events corresponds one-to-one with the intervention method for apnea events in the above embodiment. As Figure 8As shown in the figure, the intervention device for apnea events includes a prediction unit 10, a chest and abdomen signal extraction unit 20, an apnea event determination unit 30, and an apnea intervention unit 40. The detailed description of each functional module is as follows: The prediction unit 10 is configured to obtain the pressure signal collected by the pressure detection unit, and based on the pressure signal, predict the sleeping posture and position information of the target object, where the position information includes the chest position and the abdomen position; The chest and abdomen signal extraction unit 20 is configured to respectively extract the chest respiration signal and the abdomen respiration signal based on the chest position and the abdomen position; The apnea event determination unit 30 is configured to determine whether the target object has an apnea event based on the chest respiration signal and the abdomen respiration signal; The apnea intervention unit 40 is configured to, if the target object has an apnea event, determine an intervention plan corresponding to the apnea event based on the sleeping posture and position information, so as to adjust the air pressure of the corresponding airbag based on the intervention plan.

[0075] In an embodiment of the present application, the apnea event determination unit 30 is further configured to: Determine the chest and abdomen signal characteristics, the chest and abdomen respiration correlation characteristics, and the chest and abdomen respiration time dimension characteristics based on the chest respiration signal and the abdomen respiration signal; Determine whether the target object has an apnea event based on the determined chest and abdomen signal characteristics, the chest and abdomen respiration correlation characteristics, and the chest and abdomen respiration signal time dimension characteristics.

[0076] In an embodiment of the present application, the device further includes an intervention plan adjustment unit, which is configured to: Detect whether the chest and abdomen respiration correlation characteristics change; If the chest and abdomen respiration correlation characteristics change, determine whether the change of the chest and abdomen respiration correlation characteristics meets the preset intervention conditions; If the preset intervention conditions are not met, increase the adjustment range of the corresponding airbag; If the preset intervention conditions are met, restore the current air pressure of the corresponding airbag to the air pressure before the intervention.

[0077] In an embodiment of the present application, the apnea event determination unit 30 is further configured to: Respectively extract the chest signal characteristics corresponding to the chest respiration signal and the abdomen signal characteristics corresponding to the abdomen respiration signal, where the chest signal characteristics include chest signal phase information and chest signal amplitude information, and the abdomen signal characteristics include abdomen signal phase information and abdomen signal amplitude information; Determine the chest and abdomen respiration correlation characteristics based on the chest signal phase information, abdomen signal phase information, chest signal amplitude information, and abdomen signal amplitude information; Determine the time dimension characteristics of the chest and abdomen respiration signals based on the chest signal phase information, abdomen signal phase information, chest signal amplitude information, and abdomen signal amplitude information.

[0078] In an embodiment of the present application, the apnea event determination unit 30 is further configured to: Analyze the chest respiration signal and the abdomen respiration signal respectively, and obtain the chest signal phase information and the abdomen signal phase information based on the analyzed chest respiration signal and abdomen respiration signal respectively; Determine the chest respiration amplitude information and the abdomen respiration signal amplitude information respectively based on the effective peak points and effective valley points of the chest respiration signal and the abdomen respiration signal; Calculate the respiratory cycle variability based on the effective peak points of the chest respiration signal and the abdomen respiration signal.

[0079] In an embodiment of the present application, the apnea event determination unit 30 is further configured to: Perform band-pass filtering on the chest respiration signal and the abdomen respiration signal respectively; Perform Hilbert transform on the filtered chest respiration signal and abdomen respiration signal to obtain the transformed chest respiration signal and abdomen respiration signal; Obtain the chest respiration signal phase and the abdomen respiration signal phase based on the analyzed chest respiration signal and abdomen respiration signal; Perform phase unwrapping on the chest respiration signal phase and the abdomen respiration signal phase to obtain the chest signal phase information and the abdomen signal phase information.

[0080] In an embodiment of the present application, the position information includes the body inclination of the target object, and the apnea intervention unit 40 is further configured to: Determine the contact state between the target object and each airbag based on the position information; Classify each airbag based on the body inclination of the target object to obtain the corresponding types of each airbag; Determine an intervention plan corresponding to the apnea event based on the sleeping posture, contact state, and the corresponding types of each airbag.

[0081] In an embodiment of the present application, the apnea intervention unit 40 is further configured to: If the sleeping posture is supine or prone, determine whether there are multiple contact airbags of the same type in contact with the target object based on the contact state; If there are multiple contact airbags of the same type, determine the corresponding distribution modes of the multiple contact airbags of the same type, so as to correspondingly adjust the air pressure of the contact airbags based on the distribution modes; If there are no multiple contact airbags of the same type, correspondingly adjust the air pressure of airbags of different types respectively.

[0082] In an embodiment of the present application, the chest and abdomen signal extraction unit 20 is further configured to: Based on the chest position and the abdomen position, respectively determine the area where the chest is located and the area where the abdomen is located, wherein the area where the chest is located and the area where the abdomen is located are respectively composed of corresponding pressure detection units to form a first pressure matrix and a second pressure matrix; Determine the first pressure mean value corresponding to the first pressure matrix, and use the first pressure mean value as the chest respiration signal; Determine the second pressure mean value corresponding to the second pressure matrix, and use the second pressure mean value as the abdomen respiration signal.

[0083] In the embodiment of the present application, the pressure detection unit is used to collect pressure signals in real time, and advanced algorithms are used to deeply analyze the pressure data to accurately lock the sleeping posture, chest and abdomen positions of the target object. This unique design effectively avoids the interference of the movement of the user's sleeping position and the frequent change of the sleeping posture on the monitoring results. Compared with the traditional monitoring method, it significantly improves the stability and accuracy of data collection. On this basis, the system can accurately identify different types of apnea events through the intelligent extraction and comparative analysis of the chest and abdomen respiration signals. No matter which type of sleep apnea occurs, it can achieve fast and accurate judgment. When an apnea event is detected, the system will automatically match a pre-set personalized intervention plan according to the user's real-time sleeping posture, and effectively relieve the apnea symptoms and reduce the health risks such as cardiovascular diseases and cognitive impairments caused by apnea by precisely adjusting the air pressure of the corresponding airbag while minimizing the interference to the user's sleep. This technical solution not only realizes a comfortable and convenient home monitoring experience, getting rid of the limitations of traditional polysomnography (PSG) that requires hospitalization, complex equipment and high cost, but also breaks through the technical bottlenecks of poor comfort of portable devices and insufficient accuracy of emerging monitoring technologies, providing a more efficient and feasible solution for the monitoring and treatment of sleep apnea in an innovative way with low cost, high precision and humanization.

[0084] For the specific limitations of the intervention device for apnea events, reference may be made to the limitations of the intervention method for apnea events in the foregoing text, which will not be elaborated here. Each module in the above intervention device for apnea events can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0085] In one embodiment, a computer device is provided. The computer device can be a terminal device, and its internal structural diagram can be as Figure 9 shown. The computer device includes a processor, a memory, and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a readable storage medium. The readable storage medium stores computer-readable instructions. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer-readable instructions are executed by the processor, an intervention method for apnea events is implemented. The readable storage medium provided in this embodiment includes a non-volatile readable storage medium and a volatile readable storage medium.

[0086] In an embodiment of the present application, a computer device is provided, including a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor. When the processor executes the computer-readable instructions, the steps of the intervention method for apnea events as described above are implemented.

[0087] In an embodiment of the application, a readable storage medium is provided. The readable storage medium stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the steps of the intervention method for apnea events as described above are implemented.

[0088] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through computer-readable instructions. The computer-readable instructions can be stored in a non-volatile readable storage medium or a volatile readable storage medium. When the computer-readable instructions are executed, they can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided in this application can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or an external cache. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0089] Those skilled in the art can clearly understand that for the convenience and brevity of description, only the above division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0090] The above embodiments are only used to illustrate the technical solutions of the present application, not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for intervening in apnea events, characterized in that, Applied to a respiratory intervention device, on which there are multiple pressure detection units and airbags, the method includes: Obtain the pressure signals collected by the pressure detection units, and based on the pressure signals, predict the sleeping position and position information of the target object, where the position information includes the chest position and the abdominal position; Based on the chest position and the abdominal position, extract the chest respiratory signal and the abdominal respiratory signal respectively; Based on the chest respiratory signal and the abdominal respiratory signal, determine whether the target object has a respiratory arrest event; If the target object has a respiratory arrest event, based on the sleeping position and the position information, determine an intervention plan corresponding to the respiratory arrest event, so as to adjust the air pressure of the corresponding airbag based on the intervention plan.

2. The intervention method for apnea events according to claim 1, characterized in that, The determining whether the target object has a respiratory arrest event based on the chest respiratory signal and the abdominal respiratory signal includes: Based on the chest respiratory signal and the abdominal respiratory signal, determine the chest and abdomen signal characteristics, the chest and abdomen respiratory correlation characteristics, and the chest and abdomen respiratory time dimension characteristics; Based on the determined chest and abdomen signal characteristics, the chest and abdomen respiratory correlation characteristics, and the chest and abdomen respiratory signal time dimension characteristics, determine whether the target object has a respiratory arrest event.

3. The intervention method for apnea events according to claim 2, wherein, The determining the chest and abdomen signal characteristics, the chest and abdomen respiratory correlation characteristics, and the chest and abdomen respiratory time dimension characteristics based on the chest respiratory signal and the abdominal respiratory signal includes: Extract the chest signal characteristics corresponding to the chest respiratory signal and the abdominal signal characteristics corresponding to the abdominal respiratory signal respectively, where the chest signal characteristics include the chest signal phase information and the chest signal amplitude information, and the abdominal signal characteristics include the abdominal signal phase information and the abdominal signal amplitude information; Based on the chest signal phase information, the abdominal signal phase information, the chest signal amplitude information, and the abdominal signal amplitude information, determine the chest and abdomen respiratory correlation characteristics; Based on the chest signal phase information, the abdominal signal phase information, the chest signal amplitude information, and the abdominal signal amplitude information, determine the chest and abdomen respiratory signal time dimension characteristics.

4. The method for intervening in apnea events according to claim 3, wherein, The extracting the chest signal characteristics corresponding to the chest respiratory signal and the abdominal signal characteristics corresponding to the abdominal respiratory signal respectively includes: Analyze the chest respiratory signal and the abdominal respiratory signal respectively, and obtain the chest signal phase information and the abdominal signal phase information respectively based on the analyzed chest respiratory signal and abdominal respiratory signal; Based on the effective peak points and effective valley points of the chest respiratory signal and the abdominal respiratory signal, determine the chest respiratory amplitude information and the abdominal respiratory signal amplitude information respectively; Based on the effective peak points of the chest respiratory signal and the abdominal respiratory signal, calculate the respiratory cycle variability.

5. The intervention method for apnea events according to claim 4, wherein, The analyzing the chest respiratory signal and the abdominal respiratory signal respectively, and obtaining the chest signal phase information and the abdominal signal phase information respectively based on the analyzed chest respiratory signal and abdominal respiratory signal includes: Perform band-pass filtering on the chest respiratory signal and the abdominal respiratory signal respectively; Perform Hilbert transform on the filtered chest breathing signal and abdominal breathing signal to obtain the transformed chest breathing signal and abdominal breathing signal; Based on the parsed chest breathing signal and abdominal breathing signal, obtain the chest breathing signal phase and abdominal breathing signal phase; Perform phase unwrapping on the chest breathing signal phase and abdominal breathing signal phase to obtain the chest signal phase information and abdominal signal phase information.

6. The intervention method for apnea events according to claim 1, characterized in that The position information includes the body inclination of the target object. Based on the sleeping posture and the position information, determining an intervention plan corresponding to the apnea event includes: Based on the position information, determine the contact state between the target object and each airbag; Based on the body inclination of the target object, classify each airbag to obtain the corresponding type of each airbag; Based on the sleeping posture, contact state, and the corresponding type of each airbag, determine an intervention plan corresponding to the apnea event.

7. The intervention method for apnea events according to claim 6, wherein, The determining an intervention plan corresponding to the apnea event based on the sleeping posture, contact state, and the corresponding type of each airbag includes: If the sleeping posture is supine or prone, based on the contact state, determine whether there are multiple contact airbags of the same type in contact with the target object; If there are multiple contact airbags of the same type, determine the corresponding distribution mode of the multiple contact airbags of the same type, and correspondingly adjust the air pressure of the contact airbags based on the distribution mode; If there are no multiple contact airbags of the same type, correspondingly adjust the air pressure of airbags of different types respectively.

8. The intervention method for apnea events according to any one of claims 1-7, characterized in that, After adjusting the air pressure of the corresponding airbag based on the intervention plan, it further includes: Detect whether the chest-abdominal breathing correlation feature changes; If the chest-abdominal breathing correlation feature changes, determine whether the change of the chest-abdominal breathing correlation feature meets the preset intervention condition; If it does not meet the preset intervention condition, increase the adjustment amplitude of the corresponding airbag; If it meets the preset intervention condition, restore the current air pressure of the corresponding airbag to the air pressure before the intervention.

9. The method for intervening in apnea events according to any one of claims 1-7, characterized in that, The extracting the chest breathing signal and abdominal breathing signal based on the chest position and abdominal position respectively includes: Based on the chest position and abdominal position, respectively determine the chest area and abdominal area, where the chest area and abdominal area are respectively composed of corresponding pressure detection units to form a first pressure matrix and a second pressure matrix; Determine the first pressure mean value corresponding to the first pressure matrix, and use the first pressure mean value as the chest breathing signal; Determine the second pressure mean value corresponding to the second pressure matrix, and use the second pressure mean value as the abdominal breathing signal.

10. An intervention device for apnea events, characterized in that, Applied to a respiratory intervention device, the respiratory intervention device is provided with a plurality of pressure detection units and airbags. The device includes: A prediction unit, configured to obtain the pressure signal collected by the pressure detection unit, and based on the pressure signal, predict the sleeping posture and position information of the target object, where the position information includes the chest position and abdominal position; A chest and abdomen signal extraction unit, configured to respectively extract a chest respiration signal and an abdominal respiration signal based on the chest position and the abdominal position; An apnea event determination unit, configured to determine whether an apnea event occurs to the target object based on the chest respiration signal and the abdominal respiration signal; An apnea intervention unit, configured to, if an apnea event occurs to the target object, determine an intervention plan corresponding to the apnea event based on the sleeping posture and the position information, so as to adjust the air pressure of the corresponding airbag based on the intervention plan.

Images