Intelligent mouth-nose mask multimode physiological parameter monitoring system for closed-loop control of breathing machine

Abstract

The embodiment of the invention provides an intelligent mouth-nose mask multimode physiological parameter monitoring system for closed-loop control of a breathing machine, and relates to the technical field of medical monitoring. The device comprises a respirator connected with a preset respirator; the elastic pressure sensor is bound on the respirator; the electroencephalogram electrode is positioned on the elastic pressure sensor; the optical sensor and the body temperature sensor are located on the elastic pressure sensor; the air pressure sensor, the airflow sensor, the airflow temperature sensor and the vibration acoustic sensor are located on the respirator; and the electro-oculogram electrode is positioned on the elastic pressure sensor. The sensors used for collecting the physiological signals of the target object are arranged on the respirator or the elastic pressure sensor bound to the respirator, so that the target object only needs to wear the respirator connected with the respirator and the elastic pressure sensor bound to the respirator, and does not need to additionally wear other monitoring equipment; the burden of the target object caused by wearing the equipment can be effectively reduced.

Description

Technical Field

[0001] The present application relates to the field of medical monitoring technology, and in particular, to an intelligent oral and nasal mask multi-mode physiological parameter monitoring system for closed-loop control of a ventilator. Background Art

[0002] Ventilators are widely used auxiliary respiratory devices in the medical field, mainly used to improve patients' respiratory function, support lung function recovery and relieve symptoms of respiratory-related diseases. Its application range covers respiratory support for patients with chronic cardiopulmonary diseases, respiratory management for patients with sleep apnea syndrome (snoring), and treatment and monitoring of various respiratory-related diseases such as nocturnal hypoventilation and restrictive ventilation disorders. By providing stable airflow and air pressure, ventilators can significantly improve patients' ventilation quality and help patients maintain a more comfortable physiological state in various environments.

[0003] However, existing ventilator technology is usually adjusted based on preset parameters and lacks the ability to adapt to the patient's individual physiological state in real time. In fact, the effect of respiratory regulation is often closely related to a variety of physiological parameters, such as blood oxygen saturation, respiratory rate, blood pressure, body temperature, and brain electrical activity. However, the monitoring of these physiological parameters usually requires multiple independent devices, resulting in patients needing to wear multiple monitoring devices when comprehensively monitoring multiple parameters. This complex configuration is only feasible in highly specialized medical environments such as intensive care units. In daily life, this monitoring method is difficult to promote due to high equipment costs, complex configuration, and significant increase in patient discomfort. Summary of the invention

[0004] The purpose of this application is to solve at least one of the above technical defects. The technical solutions provided by the embodiments of this application are as follows: In a first aspect, an embodiment of the present application provides a physiological signal monitoring system, comprising: A breathing mask connected to a pre-set ventilator; An elastic pressure sensor attached to the breathing mask; EEG electrodes located on elastic pressure sensors; an optical sensor and a body temperature sensor located on the elastic pressure sensor; An air pressure sensor, an air flow sensor, an air flow temperature sensor, and a vibroacoustic sensor located on the breathing mask; an electrooculogram electrode located on the elastic pressure sensor; Among them, the elastic pressure sensor is used to collect the brain pressure distribution signal of the target object; the EEG electrode is used to collect the EEG signal of the target object; the optical sensor is used to collect the pulse wave signal of the target object; the body temperature sensor is used to collect the body temperature signal of the target object; the air pressure sensor is used to collect the respiratory air pressure signal of the target object; the airflow sensor is used to collect the respiratory airflow signal of the target object; the airflow temperature sensor is used to collect the respiratory airflow temperature signal of the target object; the vibroacoustic sensor is used to collect the respiratory frequency signal of the target object; and the electrooculogram electrode is used to collect the electrooculogram eye movement signal of the target object.

[0005] In a second aspect, an embodiment of the present application provides a ventilator adjustment method, comprising: The physiological signal monitoring system provided by each of the above embodiments collects various physiological signals of the target object in real time; wherein each physiological signal includes at least one of the following: brain pressure distribution signal, EEG signal, pulse wave signal, body temperature signal, respiratory pressure signal, respiratory airflow signal, respiratory airflow temperature signal, respiratory frequency signal and electrooculogram signal; Extract features of various physiological signals to obtain initial features of various physiological signals; Input each initial feature into the multimodal learning network model in real time, and output the real-time physiological features of the target object; Based on the difference value between the pre-acquired first standard physiological characteristic and the real-time physiological characteristic, the preset parameters of the ventilator are adjusted in real time until the difference value is less than a preset threshold value.

[0006] In a third aspect, an embodiment of the present application provides a ventilator adjustment device, comprising: A physiological signal acquisition module is used to acquire various physiological signals of the target object in real time through a physiological signal monitoring system; wherein each physiological signal includes at least one of the following: a pressure signal, an electroencephalogram signal, a pulse wave signal, a body temperature signal, a respiratory pressure signal, a respiratory airflow signal, a respiratory airflow temperature signal, a respiratory frequency signal, and an electrooculogram signal; A feature extraction module is used to extract features of various physiological signals to obtain initial features of various physiological signals; A real-time feature output module is used to input each initial feature into the multimodal learning network model in real time and output the real-time physiological features of the target object; The adjustment module is used to adjust the preset parameters of the ventilator in real time based on the difference value between the pre-acquired first standard physiological characteristics and the real-time physiological characteristics until the difference value is less than a preset threshold.

[0007] In a fourth aspect, an embodiment of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory; The processor executes the computer program to implement the method provided in the embodiment of the first aspect or any optional embodiment of the first aspect.

[0008] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method provided in the embodiment of the first aspect or any optional embodiment of the first aspect is implemented.

[0009] The beneficial effects of the technical solution provided by the embodiment of the present application are: The physiological signal monitoring system provided in the present application, by setting various sensors for collecting the physiological signals of the target object on a breathing mask or an elastic pressure sensor bound to the breathing mask, allows the target object to only wear the breathing mask connected to the ventilator and the elastic pressure sensor bound to the breathing mask, without the need to wear other additional monitoring equipment, thereby effectively reducing the burden on the target object caused by wearing equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in describing the embodiments of the present application are briefly introduced below.

[0011] Figure 1 A schematic diagram of a physiological signal monitoring system provided in an embodiment of the present application; Figure 2 A schematic diagram of a physiological signal monitoring system in an example of an embodiment of the present application; Figure 3 A schematic diagram of a physiological signal monitoring system in an example of an embodiment of the present application; Figure 4 A schematic diagram of a flow chart of a ventilator adjustment method provided in an embodiment of the present application; Figure 5 This is a schematic diagram of the overall process of a ventilator adjustment method in an example of an embodiment of the present application; Figure 6 A structural block diagram of a ventilator adjustment device provided in an embodiment of the present application; Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0012] The embodiments of the present application are described below in conjunction with the drawings in the present application. It should be understood that the implementation methods described below in conjunction with the drawings are exemplary descriptions for explaining the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions of the embodiments of the present application.

[0013] It will be understood by those skilled in the art that, unless specifically stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the terms "including" and "comprising" used in the embodiments of the present application refer to that the corresponding features can be implemented as the presented features, information, data, steps, operations, elements and / or components, but do not exclude the implementation as other features, information, data, steps, operations, elements, components and / or combinations thereof supported by the technical field. It should be understood that when we say that an element is "connected" or "coupled" to another element, the one element may be directly connected or coupled to the other element, or it may refer to that the one element and the other element establish a connection relationship through an intermediate element. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein indicates at least one of the items defined by the term, for example, "A and / or B" may be implemented as "A", or as "B", or as "A and B".

[0014] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below in conjunction with the accompanying drawings.

[0015] The following describes several exemplary embodiments to illustrate the technical solutions of the embodiments of the present application and the technical effects produced by the technical solutions of the present application. It should be noted that the following embodiments can refer to, draw on or combine with each other, and the same terms, similar features and similar implementation steps in different embodiments will not be described repeatedly.

[0016] The intelligent oral and nasal mask multi-mode physiological parameter monitoring system for closed-loop control of a ventilator provided in the present application can be composed of a physiological signal monitoring system device and a ventilator adjustment method applied to the monitoring system device. Figure 1 A schematic diagram of a physiological signal monitoring system is provided for an embodiment of the present application. The schematic diagram of the structure of the physiological signal monitoring system is as follows: Figure 1 As shown, it includes: a breathing mask, an elastic pressure sensor, at least two EEG electrodes, an optical sensor, a body temperature sensor, an air pressure sensor, an airflow sensor, an airflow temperature sensor, a vibroacoustic sensor and at least two EOG electrodes.

[0017] In the present application embodiment, Figure 1 As shown, the respiratory mask can be an oronasal mask as shown in the figure (i.e., a mask that is worn on the mouth and nose at the same time).

[0018] The elastic pressure sensor can be an elastic headband, (i.e. Figure 1The middle gray diagonal part) uses a flexible pressure sensing material. The original breathing mask is in the form of a mouth and nose mask plus an ordinary headband. In the embodiment of the present application, an elastic pressure sensor in the form of an elastic headband is used to replace the ordinary headband. The elastic pressure sensor is used to monitor the pressure distribution of the target object's head in real time (i.e., the brain pressure distribution signal), ensure the fit between the breathing mask and the face, and record the pressure changes caused by the displacement or movement of the target object's head.

[0019] The EEG electrodes are mounted on elastic pressure sensors (i.e. Figure 1 The pentagonal part in the middle), the function of this electrode is to record the brain electrical activity signal (i.e., EEG signal) of the target object, which is suitable for monitoring the target object's sleep stage, wakefulness state and its relationship with the breathing pattern. In order to monitor the target object more comprehensively and accurately, at least two EEG electrodes are generally set on the elastic pressure sensor.

[0020] The optical sensor and the body temperature sensor are mounted at the same position on the elastic pressure sensor (i.e. Figure 1 The optical sensor is used to measure the pulse wave signal of the target object, using single-wavelength, multi-wavelength or two-dimensional photoplethysmography technology (for example, photoplethysmography (PPG) signal). The pulse wave signal can be used to estimate physiological parameters such as blood pressure changes and blood oxygen. The body temperature sensor is used to monitor the epidermal temperature (i.e., temperature signal) of the target object. Since it is installed near the main arteries of the face, it can provide accurate blood flow and temperature change information.

[0021] The air pressure sensor, air flow sensor and air flow temperature sensor are installed at the same position on the breathing mask (i.e. Figure 1 The air pressure sensor is used to record the target object's breathing pressure changes (i.e., the breathing air pressure signal). Since the airflow sensor is generally installed close to the nostrils or mouth, it can more accurately capture the target object's inhalation and exhalation pattern and rate (i.e., the breathing airflow signal). The airflow temperature sensor can detect changes in the target object's exhaled gas temperature (i.e., the breathing airflow temperature signal).

[0022] The vibroacoustic sensor is mounted on the breathing mask ( Figure 1 The black square in the middle is used to detect vibrations and sounds related to the target subject's breathing, such as snoring or irregular breathing. This position design can effectively capture the jaw vibration or acoustic signal (i.e., breathing rate signal) caused by the target subject's breathing.

[0023] The electrooculogram electrodes are mounted on elastic pressure sensors ( Figure 1The middle diamond-shaped part) is used to monitor the target object's eye movements (i.e., electrooculogram eye movement signals), which can provide an indication of whether the target object has entered the sleep stage and help analyze the correlation between the target object's eye movements and apnea or hypopnea events.

[0024] Optionally, an ear calibration module (not shown in the figure) can be provided on the ear side of the elastic pressure sensor, which can be used to calibrate the blood pressure in the supine position before sleeping, and realize accurate blood pressure monitoring by combining the blood pressure change signal. In addition, a cuff-type sphygmomanometer can also be used to monitor blood pressure, which is not limited in the embodiments of the present application.

[0025] The physiological signal monitoring system provided in the present application, by setting various sensors for collecting the physiological signals of the target object on a breathing mask or an elastic pressure sensor bound to the breathing mask, allows the target object to only wear the breathing mask connected to the ventilator and the elastic pressure sensor bound to the breathing mask, without the need to wear other additional monitoring equipment, thereby effectively reducing the burden on the target object caused by wearing equipment.

[0026] Based on the above embodiments, as an optional embodiment, the number of EEG electrodes is multiple, and they are respectively located in the frontal lobe area, parietal lobe area, temporal lobe area and occipital lobe area of ​​the target object's brain; The optical sensor and the body temperature sensor are located in the temporal artery area on both sides of the target subject's face; The air pressure sensor, air flow sensor, and air flow temperature sensor are located inside the breathing mask and in the mouth and nose area of ​​the target subject; The vibroacoustic sensor is located inside the breathing mask and in the nose area of ​​the target subject; The EOG electrodes are placed in the periocular area of ​​the target subject.

[0027] It should be noted that the installation positions of the physiological signal acquisition devices such as sensors or electrodes provided in the embodiments of the present application are not completely fixed and can be adaptively adjusted within a certain range according to the brain size of the target object.

[0028] Based on the above embodiments, as an optional embodiment, the respiratory mask in the physiological signal monitoring system provided by the present application can also be a nasal mask or a nasal pillow or other mask types; the schematic diagram of the device structure when the respiratory mask is a nasal mask can be as follows: Figure 2 As shown, the schematic diagram of the device structure when the breathing mask is a nasal pillow can be as follows Figure 3 As shown, the color and Figure 1 Same as described in detail above Figure 1 Similarly, the specific installation location can be as follows Figure 2 or Figure 3 As shown, no further elaboration is given here.

[0029] Figure 4 A flow chart of a ventilator adjustment method is provided for an embodiment of the present application. The execution subject of the method may be a terminal (such as a computer, a mobile phone, a watch, etc.) or a server, etc. Figure 4 As shown, the method may include: Step S401 : various physiological signals of the target object are collected in real time through the physiological signal monitoring system provided by the above embodiments.

[0030] The various physiological signals of the embodiments of the present application include at least one of the following: brain pressure distribution signal, EEG signal, pulse wave signal, body temperature signal, respiratory pressure signal, respiratory airflow signal, respiratory airflow temperature signal, respiratory frequency signal and electrooculogram signal.

[0031] The physiological signal monitoring system of the embodiment of the present application can be generally composed of a breathing mask and an elastic headband. The breathing mask is directly connected to the ventilator and can be an oral and nasal mask, a nasal mask, a nasal pillow or other mask types, which are not limited in this embodiment. The physiological signal monitoring system also carries various micro sensors and electrodes, which are respectively arranged on the breathing mask or the elastic headband. The target object can be a target patient who needs to use a ventilator.

[0032] Specifically, when the target object needs to use a ventilator, the breathing mask in the physiological signal monitoring system can be worn on the mouth and nose of the target object, and the elastic headband can be bound to the head of the target object in a surrounding form, so that the ventilator can transport oxygen to the target object through the breathing mask directly connected to it. At the same time, the sensors or electrodes carried in the physiological signal monitoring system can collect various physiological signals of the target object in real time.

[0033] Step S402: extracting features from each physiological signal to obtain initial features of each physiological signal.

[0034] Specifically, the physiological signals of the embodiments of the present application are usually not directly interpreted by the model, so these physiological signals need to be converted into initial features that can be used by the model before subsequent adjustment calculations can be facilitated. Specifically, for example, when the pulse wave signal is collected, since the pulse wave signal is an analog type signal, it can first be preprocessed, including filtering, signal segmentation, and peak detection. On this basis, the pulse wave signal is feature extracted to obtain key time and amplitude parameters, such as the rise time, amplitude ratio, and pulse wave conduction time of the pulse wave. After these characteristic parameters are mapped by a specific regression model or calibration relationship, they can be further used to estimate SBP (Systolic Blood Pressure) and DBP (Diastolic Blood Pressure) values ​​as the initial features corresponding to the signal.

[0035] In addition, for other physiological signals, such as respiratory airflow signals, features such as airflow change rate or airflow cycle peak can be extracted; for blood oxygen signals, feature quantities such as time-dependent changes or pulse amplitude ratio can be extracted. These physiological signals are converted into initial features through the feature extraction process, providing a standardized data basis for subsequent machine learning model input.

[0036] Step S403, inputting each initial feature into the multimodal learning network model in real time, and outputting the real-time physiological features of the target object.

[0037] The real-time physiological features of the embodiments of the present application can characterize the real-time physical state of the target object. The multimodal learning network model can be divided into two levels, wherein the first level contains multiple trained sub-models, which are used to process each initial feature and output the corresponding prediction value. The second level can be used to integrate data from different sensors, and the second level contains a fusion model, which can be used to fuse the outputs of each sub-model in the first level to obtain cross-modal deep features, which can also be called real-time physiological features of the target object, and can characterize the real-time physiological state of the target object.

[0038] Specifically, after the initial features of each physiological signal are extracted, these initial features are input into the multimodal learning network model. First, each initial feature can be processed by multiple sub-models contained in the first layer to output the corresponding prediction value. Then, the prediction value output by each model is input into the second layer. The fusion model in the second layer fuses the outputs of each sub-model in the first layer. Finally, the fused output is the real-time physiological feature that characterizes the real-time physical state of the target object.

[0039] It should be noted that the sub-models provided in the embodiments of the present application can also use cross-validation and other methods to evaluate the performance of each sub-model during training (cross-validation means taking each data set as a validation set, and taking other data sets other than the validation set of this training as training sets to train the model each time), and adjust the model's hyperparameters, such as the learning rate and the number of network layers, when necessary. At the same time, when judging the termination conditions of model training, select a suitable loss function, such as using the mean square error method for regression tasks, and using the cross entropy loss method for classification tasks, etc. At the same time, use optimization algorithms such as Adam (an optimizer) or stochastic gradient descent (SGD, Stochastic Gradient Descent) to minimize the loss function to improve the prediction accuracy of the sub-model.

[0040] Furthermore, since the sub-models in the physiological signal monitoring system provided in the embodiment of the present application need to have a certain processing capability for each initial feature data, the trained sub-models should not be overfitted during the sub-model training process (i.e., they only have a strong processing capability for one or several types of data, but a poor processing capability for other data). In this regard, the present embodiment also introduces regularization technology when training each sub-model, such as: the dropout algorithm, that is, randomly discarding (or "discarding") a part of the neurons in the neural network during the training process, so that the network structure seen by each training sample is different, thereby increasing the generalization ability of the model; the L1 / L2 regularization algorithm, which realizes feature selection and sparsification by making the model parameters as small as possible, avoiding excessive complexity of the model, thereby reducing the risk of overfitting; or other regularization algorithms, which are not limited in the embodiment of the present application.

[0041] Optionally, if the model is found to have insufficient or limited data when training in a certain environment, the model can be migrated to a new similar environment for learning and training, or data enhancement (i.e., modification of the existing training data set) can be used to increase the diversity of training data. The above methods can also increase the generalization ability of the model, so that it has a certain degree of accuracy when facing different groups of objects.

[0042] Step S404, based on the difference value between the pre-acquired first standard physiological characteristic and the real-time physiological characteristic, the preset parameters of the ventilator are adjusted in real time until the difference value approaches the preset threshold value.

[0043] The first standard physiological feature of the embodiment of the present application can be obtained by extracting various physiological signals of the target object group in a healthy state and calculating them. The difference value can be used to characterize the health of the target object. Generally speaking, the lower the difference value, the healthier the physical condition of the target object, and the higher the difference value, the worse the physical condition of the target object. The preset threshold value can be used to determine whether the target object is within the standard range of health.

[0044] Specifically, while the real-time physiological characteristics of the target object are output in real time through the multimodal learning network model, the output of the multimodal learning network model can also be compared with the first standard physiological characteristics of the target object in a standard healthy state in real time, and the physical state of the target object can be judged in real time by calculating the difference value in real time. It can be understood that the amount of oxygen supply required by the target object is different under different physical conditions. Therefore, the parameters of the ventilator itself can also be dynamically adjusted in real time according to the real-time changing difference value (the parameters of the ventilator itself can include the oxygen flow rate, oxygen supply frequency, etc.) to dynamically adjust the oxygen supply to the target object so that the oxygen supply level for the target object remains in the most appropriate state.

[0045] The solution provided in the embodiment of the present application collects and extracts the physiological signals of the target object in real time through the physiological signal monitoring system, and then interprets the multimodal physiological parameters through the integrated multimodal machine learning network model to judge the real-time physiological characteristics of the target object, so that the ventilator can intelligently adjust the relevant parameters such as the air pressure / airflow output by itself in real time according to the real-time physiological characteristics output by the model, thereby realizing the closed-loop feedback control of the ventilator. While assisting in maintaining smooth breathing, it can carry out all-round multimodal physiological parameter monitoring, and carry out personalized breathing regulation according to the patient's real-time feedback characteristics, realizing intelligent synchronization between man and machine.

[0046] Based on the above embodiments, as an optional embodiment, before extracting features from each physiological signal, the following steps are further specifically included: Pre-process various physiological signals; The preprocessing operation includes at least one of the following: noise reduction processing, Z value standardization processing, and normalization processing.

[0047] It is understandable that when extracting physiological signals, there may be obvious abnormal signals in the extracted physiological signals due to some external interference (such as the target object's improper behavior causing the device to fall off, etc.). Therefore, in order to reduce subsequent calculation errors, the extracted physiological signals need to be preprocessed before feature extraction. The preprocessing method can be noise reduction processing, Z value standardization processing, normalization processing, etc., which is not limited in the embodiments of this application.

[0048] Based on the above embodiments, as an optional embodiment, the multimodal learning network model includes multiple prediction value calculation sub-models; Input each initial feature into the multimodal learning network model in real time, and output the real-time physiological features of the target object, including: For each initial feature, obtain the data type corresponding to the initial feature, input the initial feature into the corresponding prediction value calculation sub-model, and output the prediction value of the initial feature; Each predicted value is input into the fusion sub-model, and the real-time physiological characteristics are output.

[0049] The data types of the initial features of the embodiments of the present application can be divided into time series data and non-time series data. Time series data can be a series of values ​​collected in chronological order, each value is associated with a specific time point or time period, such as continuously monitored blood pressure, heart rate and respiration. Non-time series data can be collected data that is not continuous or sequential in time, such as mask fit and pressure distribution images. Since the type characteristics of the above two types of data are different, in the embodiments of the present application, the data will be input into the corresponding sub-models for processing according to the distribution of the type characteristics of the data.

[0050] Optionally, when data inconsistency is found after feature extraction of various physiological signals, data reconstruction / segmentation processing can be performed on the extracted initial features to split the incoherent data into multiple parts / merge some incoherent data into one part, etc., to ensure the consistency of the overall data.

[0051] Specifically, after extracting each initial feature, each initial feature can be first input into the first level of the multimodal learning network model. In the first level, the corresponding prediction value calculation sub-model can be matched for each initial feature according to the data type of each initial feature, and each initial feature can be processed separately by the corresponding prediction value calculation sub-model to output the prediction value corresponding to each initial feature. These prediction values ​​are then input into the second level of the multimodal learning network model, and these prediction values ​​are fused and integrated through the fusion sub-model in the second level. Finally, the output of the fusion sub-model is used as the model output of the multimodal learning network model, that is, the real-time physiological feature.

[0052] Based on the above embodiments, as an optional embodiment, the type of the prediction value calculation sub-model includes a time series data calculation sub-model and a non-time series data calculation sub-model; Input the initial features into the corresponding prediction value calculation sub-model, including: If the data type is time series data, the initial features are input into the time series data calculation sub-model; If the data type is non-time series data, the initial features are input into the non-time series data calculation sub-model.

[0053] It should be noted that the time series data calculation submodel or non-time series data calculation submodel of the embodiment of the present application may include multiple submodels, and the types of the submodels may be the same or different. For example, for the time series data calculation submodel, the model types included may include recurrent neural network (RNN, Recurrent Neural Network) or long short-term memory network (LSTM, Long Short-Term Memory), etc. For the non-time series data calculation submodel, the model types included may include convolutional neural network (CNN, Convolutional Neural Networks), etc., and the present application does not limit this. Therefore, after determining that the data type is time series data or non-time series data, the corresponding model type can be further selected according to the specific data.

[0054] It is understandable that in order to further improve the prediction degree of the model for the initial features, the same initial features can also be predicted by multiple sub-models. On this basis, further, it is also possible to select corresponding multiple sub-models to predict the initial features according to the type of object group to which the target object belongs, and the multiple sub-models can adopt a serial prediction method (that is, the output of the previous sub-model is used as the output of the next sub-model until the output of the last sub-model is obtained as the prediction value), a parallel prediction method (that is, the initial features are input into each sub-model and the output is obtained, and then the output of each sub-model is integrated to obtain the prediction value corresponding to the initial feature) or a method of performing serial prediction and parallel prediction simultaneously, etc., which is not limited to the embodiments of the present application.

[0055] Based on the above embodiments, as an optional embodiment, obtaining the first standard physiological feature of the target object includes: Acquire the object group type corresponding to the target object and the second standard physiological feature corresponding to the object group type; The second standard physiological feature is used as the first standard physiological feature.

[0056] Specifically, since different object groups have different criteria for determining health status, all object groups can be pre-classified in the embodiments of the present application. The classification method can be, for example, "old people, young people and children"; "athletes and ordinary people", etc. Different types of object groups have their own standard physiological characteristics (i.e., second standard physiological characteristics, and the physiological characteristics can be reflected in the form of a matrix, and each matrix element in the matrix can represent the physiological state of the target object in one aspect). These group types and second standard physiological characteristics can be stored in a preset database. When the target object uses the physiological signal monitoring system, the group type to which the target object belongs can be matched from the preset database according to the basic information of the input target object, and the second standard physiological characteristics corresponding to the type can be obtained as the first standard physiological characteristics that need to be referenced during this ventilator adjustment process.

[0057] Combine the following Figure 5 , the ventilator adjustment method provided in the embodiment of the present application is generally introduced, such as Figure 5 As shown, first, the physiological signals (multimodal signals) of the target object are collected through the physiological signal monitoring system, and then the collected physiological signals are preprocessed. The preprocessing operations may include noise reduction, Z value standardization / normalization, etc., to eliminate abnormal data, and then the remaining data is feature encoded and feature extracted. After feature extraction, if incoherent data is found, the extracted initial features are reconstructed / segmented to obtain coherent data, and then the preprocessed initial features are input into the multimodal learning network model (i.e., the multimodal machine learning network in the figure). The multimodal learning network model predicts each initial feature and feeds the result back to the ventilator. The ventilator can make adaptive adjustments to the control parameters such as the oxygen pressure / airflow of the oxygen supply according to the predicted value output by the model.

[0058] When training a multimodal learning network model, first determine the type of each sub-model and design the architecture between the sub-models. At the same time, introduce a regularization method during model training to prevent the trained model from overfitting. After training, perform a performance test on the model and adjust the model parameters appropriately. Finally, migrate the trained model to a new similar environment for further intensive training to improve the generalization ability of the model.

[0059] Figure 6 A structural block diagram of a ventilator adjustment device provided in an embodiment of the present application, such as Figure 6 As shown, the ventilator adjustment device 600 may include: a physiological signal acquisition module 601, a feature extraction module 602, a real-time feature output module 603 and an adjustment module 604, wherein: The physiological signal acquisition module 601 is used to acquire various physiological signals of the target object in real time through the physiological signal monitoring system provided by the above-mentioned various embodiments; The feature extraction module 602 is used to extract features from various physiological signals to obtain initial features of various physiological signals; The real-time feature output module 603 is used to input each initial feature into the multimodal learning network model in real time and output the real-time physiological features of the target object; The adjustment module 604 is used to adjust the preset parameters of the ventilator in real time based on the difference value between the pre-acquired first standard physiological characteristics and the real-time physiological characteristics, until the difference value is less than a preset threshold.

[0060] The solution provided in the embodiment of the present application collects and extracts the physiological signals of the target object in real time through the physiological signal monitoring system, and then interprets the multimodal physiological parameters through the integrated multimodal machine learning network model to judge the real-time physiological characteristics of the target object, so that the ventilator can intelligently adjust the relevant parameters such as the air pressure / airflow output by itself in real time according to the real-time physiological characteristics output by the model, thereby realizing the closed-loop feedback control of the ventilator. While assisting in maintaining smooth breathing, it can carry out all-round multimodal physiological parameter monitoring, and carry out personalized breathing regulation according to the patient's real-time feedback characteristics, realizing intelligent synchronization between man and machine.

[0061] Based on the above embodiments, as an optional embodiment, the device further includes a preprocessing module, which is specifically used for: Pre-process various physiological signals; The preprocessing operation includes at least one of the following: noise reduction processing, Z value standardization processing, and normalization processing.

[0062] Based on the above embodiments, as an optional embodiment, the real-time feature output module is specifically used for: For each initial feature, obtain the data type corresponding to the initial feature, input the initial feature into the corresponding prediction value calculation sub-model, and output the prediction value of the initial feature; Each predicted value is input into the fusion sub-model, and the real-time physiological characteristics are output.

[0063] Based on the above embodiments, as an optional embodiment, the real-time feature output module is further used to: If the data type is time series data, the initial features are input into the time series data calculation sub-model; If the data type is non-time series data, the initial features are input into the non-time series data calculation sub-model.

[0064] Based on the above embodiments, as an optional embodiment, the adjustment module is specifically used for: Acquire the object group type corresponding to the target object and the second standard physiological feature corresponding to the object group type; The second standard physiological feature is used as the first standard physiological feature.

[0065] Reference below Figure 7 , which shows an electronic device suitable for implementing the embodiments of the present application (for example, Figure 4 The electronic device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle terminals (such as vehicle navigation terminals), wearable devices such as watches, etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 7 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0066] The electronic device includes: a memory and a processor, the memory is used to store a program for executing the method described in each of the above method embodiments; the processor is configured to execute the program stored in the memory. The processor here can be referred to as the processing device 701 described below, and the memory can include at least one of the read-only memory (ROM) 702, the random access memory (RAM) 703, and the storage device 708 described below, as shown below: like Figure 7 As shown, the electronic device 700 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage device 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the electronic device 700 are also stored. The processing device 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0067] Typically, the following devices may be connected to the I / O interface 705: an input device 706 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 707 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 708 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 709. The communication device 709 may allow the electronic device 700 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 7An electronic device having various devices is shown, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0068] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 709, or installed from the storage device 708, or installed from the ROM 702. When the computer program is executed by the processing device 701, the above-mentioned functions defined in the method of the embodiment of the present application are executed.

[0069] It should be noted that the computer-readable storage medium mentioned above in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer readable signal media may also be any computer readable medium other than computer readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0070] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0071] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0072] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device: Various physiological signals of the target object are collected in real time through the physiological signal monitoring system; wherein each physiological signal includes at least one of the following: brain pressure distribution signal, electroencephalogram signal, pulse wave signal, body temperature signal, respiratory air pressure signal, respiratory airflow signal, respiratory airflow temperature signal, respiratory frequency signal and electrooculogram signal; feature extraction is performed on each physiological signal to obtain initial features of each physiological signal; each initial feature is input into the multimodal learning network model in real time to output the real-time physiological features of the target object; based on the difference value between the pre-acquired first standard physiological feature and the real-time physiological feature, the preset parameters of the ventilator are adjusted in real time until the difference value is less than the preset threshold.

[0073] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Matlab, Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0074] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0075] The modules or units involved in the embodiments of the present application may be implemented by software or hardware. The name of a module or unit does not limit the unit itself in some cases. For example, the first constraint acquisition module may also be described as a "module for acquiring the first constraint".

[0076] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0077] In the context of the present application, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0078] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0079] The above descriptions are only partial implementations of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A physiological signal monitoring system, characterized in that: include: A breathing mask connected to a pre-set ventilator; an elastic pressure sensor bound to the breathing mask; as well as A plurality of physiological signal sensors are respectively arranged on the breathing mask and the elastic pressure sensor.

2. The physiological signal monitoring system according to claim 1, wherein the plurality of physiological signal sensors comprises: an EEG electrode located on the elastic pressure sensor; an optical sensor and a body temperature sensor located on the elastic pressure sensor; An air pressure sensor, an air flow sensor, an air flow temperature sensor, and a vibroacoustic sensor located on the breathing mask; an electrooculogram electrode located on the elastic pressure sensor; Among them, the elastic pressure sensor is used to collect the brain pressure distribution signal of the target object; the EEG electrode is used to collect the EEG signal of the target object; the optical sensor is used to collect the pulse wave signal of the target object; the body temperature sensor is used to collect the body temperature signal of the target object; the air pressure sensor is used to collect the respiratory air pressure signal of the target object; the airflow sensor is used to collect the respiratory airflow signal of the target object; the airflow temperature sensor is used to collect the respiratory airflow temperature signal of the target object; the vibroacoustic sensor is used to collect the respiratory frequency signal of the target object; and the electrooculogram electrode is used to collect the electrooculogram eye movement signal of the target object.

3. The physiological signal monitoring system according to claim 2, characterized in that: The number of the EEG electrodes is multiple, and they are respectively located in the frontal lobe area, parietal lobe area, temporal lobe area and occipital lobe area of ​​the brain of the target object; The optical sensor and the body temperature sensor are located in the temporal artery region on both sides of the target object's face; The air pressure sensor, the air flow sensor and the air flow temperature sensor are located inside the breathing mask and in the mouth and nose area of ​​the target object; The vibroacoustic sensor is located inside the breathing mask and in the nose wing area of ​​the target subject; The electrooculogram electrodes are located in the periocular area of ​​the target object.

4. A method for adjusting a ventilator, characterized in that: The method comprises: Collect various physiological signals of the target object in real time by using the physiological signal monitoring system as described in any one of claims 1 to 3; Extract features of various physiological signals to obtain initial features of various physiological signals; Inputting each initial feature into the multimodal learning network model in real time, and outputting the real-time physiological features of the target object; Based on the difference value between the pre-acquired first standard physiological characteristic and the real-time physiological characteristic, the preset parameters of the ventilator are adjusted in real time until the difference value is less than a preset threshold.

5. The method according to claim 4, characterized in that Before extracting features from various physiological signals, the method further includes: Pre-process various physiological signals; The preprocessing operation includes at least one of the following: noise reduction processing, Z value standardization processing, and normalization processing.

6. The method according to claim 4, characterized in that The multimodal learning network model includes a plurality of prediction value calculation sub-models; The real-time input of each initial feature into the multimodal learning network model to output the real-time physiological features of the target object includes: For each initial feature, obtain the data type corresponding to the initial feature, input the initial feature into the corresponding prediction value calculation sub-model, and output the prediction value of the initial feature; Each predicted value is input into the fusion sub-model, and the real-time physiological feature is output.

7. The method according to claim 6, characterized in that The types of the predicted value calculation submodels include time series data calculation submodels and non-time series data calculation submodels; The step of inputting the initial features into the corresponding prediction value calculation sub-model comprises: If the data type is time series data, inputting the initial features into the time series data calculation sub-model; If the data type is non-time series data, the initial features are input into the non-time series data calculation sub-model.

8. The method according to claim 4, characterized in that The step of acquiring the first standard physiological characteristic of the target object includes: Acquire an object group type corresponding to the target object and a second standard physiological feature corresponding to the object group type; The second standard physiological feature is used as the first standard physiological feature.

9. A ventilator adjustment device, characterized in that: include: A physiological signal acquisition module, used for acquiring various physiological signals of a target object in real time through the physiological signal monitoring system according to any one of claims 1 to 3; A feature extraction module is used to extract features of various physiological signals to obtain initial features of various physiological signals; A real-time feature output module, used to input each initial feature into the multimodal learning network model in real time, and output the real-time physiological features of the target object; The adjustment module is used to adjust the preset parameters of the ventilator in real time based on the difference value between the pre-acquired first standard physiological characteristics and the real-time physiological characteristics until the difference value is less than a preset threshold.

10. An electronic device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 4 to 8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 4 to 8 are implemented.

Images