Integrated dynamic monitoring platform based on overall space-time synchronization respiratory cycle metabolism function

Abstract

The invention discloses an integrated dynamic monitoring platform based on an integral space-time synchronization respiratory cycle metabolism function, which relates to the technical field of intelligent monitoring and comprises a multi-modal data acquisition system, a data synchronization system, a data analysis system and a report generation system, the multi-modal data acquisition system is used for acquiring respiration, circulation, metabolism and motion data and preprocessing the acquired data to generate preprocessed data; the data synchronization system is used for correcting the preprocessed data and carrying out space-time alignment on the corrected preprocessed data to obtain space-time synchronization data; the data analysis system is used for inputting the time-space synchronization data into a trained dynamic analysis model and outputting an exercise risk level; the report generation system is used for automatically generating image-text reports and exercise suggestions according to the time-space synchronization data and the exercise risk levels, real-time safety protection is provided for patients, and occurrence of exercise-related adverse events is effectively prevented.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent monitoring, and more particularly to a whole spatiotemporal synchronization respiratory circulatory metabolic function integrated dynamic monitoring platform. BACKGROUND

[0002] At present, the number of chronic disease patients continues to grow worldwide, posing a major challenge to public health. With the progress of medical technology and the development of health management concepts, exercise function assessment, as an important part of disease management, has become increasingly important. However, current exercise function assessment methods mostly rely on static detection or single parameter monitoring (such as heart rate, blood oxygen, etc.), which can provide some physiological state information, but cannot dynamically reflect the physiological state changes of patients during the entire exercise cycle. This limitation leads to an incomplete and in-depth understanding of the actual physical condition of patients, limiting the effective development of personalized treatment plans.

[0003] Traditional cardiopulmonary exercise test (CPET) can comprehensively evaluate cardiopulmonary metabolic function, but also has obvious shortcomings. First, its application range is mostly limited to laboratory environments, making it difficult to achieve widespread application in daily life. Second, in terms of data collection and analysis, traditional CPET lacks spatiotemporal synchronization and cannot achieve continuous dynamic monitoring of patients. In addition, existing technical means have limited integration and processing capabilities for multi-source heterogeneous data (such as respiration, electrocardiogram, metabolism, exercise posture, etc.), often only allowing single-dimensional data analysis, and cannot achieve collaborative analysis between multi-dimensional data. This directly leads to fragmented evaluation results, making it difficult to accurately reflect the true health status of patients, and making diagnosis and treatment recommendations based on these data may not be precise enough. Therefore, there is an urgent need for an integrated platform that can integrate spatiotemporal synchronization data collection, CPET intelligent analysis, and dynamic feedback. SUMMARY

[0004] Therefore, the present application provides a whole spatiotemporal synchronization respiratory circulatory metabolic function integrated dynamic monitoring platform, which overcomes the above-mentioned defects.

[0005] To achieve the above-mentioned purposes, the present application adopts the following technical solutions:

[0006] A whole spatiotemporal synchronization respiratory circulatory metabolic function integrated dynamic monitoring platform, comprising: a multi-modal data acquisition system, a data synchronization system, a data analysis system, and a report generation system;

[0007] The multi-modal data acquisition system is configured to acquire respiratory, circulatory, metabolic, and exercise data, and to pre-process the acquired data to generate pre-processed data.

[0008] The data synchronization system is configured to correct the preprocessed data and perform spatiotemporal alignment on the corrected preprocessed data to obtain spatiotemporal synchronized data.

[0009] The data analysis system is configured to input the spatiotemporal synchronized data into a trained dynamic analysis model and output a movement risk level.

[0010] The report generation system is configured to automatically generate a graphic-text report and a movement suggestion based on the spatiotemporal synchronized data and the movement risk level.

[0011] Optionally, the multi-modal data acquisition system comprises:

[0012] A respiration monitoring module is configured to acquire ventilation volume, oxygen uptake, carbon dioxide output, and respiratory rate.

[0013] A circulatory function module is configured to synchronously acquire heart rate, blood pressure, and cardiac output.

[0014] A metabolism monitoring module is configured to monitor muscle oxygen saturation, blood glucose fluctuation data, and blood lactic acid concentration.

[0015] A movement posture capturing module is configured to acquire movement intensity, posture, and energy consumption data through an inertial measurement unit and a three-dimensional motion capture system.

[0016] Optionally, the data synchronization system comprises:

[0017] A heterogeneous clock synchronization module is configured to perform data time synchronization on the preprocessed data through a dynamic delay compensation algorithm to generate time synchronized data.

[0018] A movement artifact correction module is configured to perform distortion correction on the time synchronized data using a trained deformity correction model.

[0019] A spatiotemporal alignment module is configured to establish a dynamic biomechanics coordinate system with the human body center of gravity as the origin, and map the corrected data to a specific muscle group movement state through Bayesian network fusion to obtain the spatiotemporal synchronized data.

[0020] Optionally, the heterogeneous clock synchronization module comprises a master node and a slave node; the master node is configured with a rubidium atomic clock, and the slave node is configured with a temperature-compensated crystal oscillator; a hybrid feedback channel is established between the master node and the slave node for clock signal synchronization and dynamic delay compensation between the master node and the slave node.

[0021] Optionally, the movement artifact correction module comprises:

[0022] A filtering unit is configured to perform filtering and dynamic noise suppression processing on the time synchronized data to generate filtered data.

[0023] a correction unit, configured to input the filtered data into a trained deformity correction model, and output corrected data.

[0024] Optionally, the deformity correction model is constructed based on a multi-modal generative adversarial network, including a generator and a discriminator; the generator takes the filtered data as input; the discriminator evaluates physiological reasonableness in combination with metabolic data, and corrects signal distortion of a movement period through a spatio-temporal attention mechanism to generate the corrected data.

[0025] Optionally, the data analysis system comprises:

[0026] a feature extraction module, configured to perform multi-modal feature extraction on the received spatio-temporal synchronous data, the multi-modal features including respiratory features, circulatory features, metabolic features, and movement features;

[0027] a risk analysis module, configured to input the respiratory features, the circulatory features, the metabolic features, and the movement features into a trained dynamic analysis model, and output a probability distribution of a movement risk level;

[0028] a level correction module, configured to correct the probability distribution based on a preset risk threshold, and generate a final movement risk level.

[0029] Optionally, the dynamic analysis model is constructed using a dual-flow deep neural network, including:

[0030] a time-series analysis branch, configured to analyze the respiratory features, the circulatory features, and the metabolic features using an encoder-decoder structure, and obtain a first risk value;

[0031] a spatial analysis branch, configured to parse the movement features using a three-dimensional convolutional network, and obtain a second risk value;

[0032] a feature fusion layer, configured to generate an output movement risk level probability distribution based on the first risk value and the second risk value.

[0033] Optionally, the dynamic analysis model is trained using a multi-stage optimization strategy.

[0034] Optionally, the report generation system comprises:

[0035] a risk traceability module, configured to visualize the contribution of each physiological parameter to the risk level through a backpropagation algorithm;

[0036] an adaptive exercise recommendation generation module, configured to dynamically adjust an exercise recommendation threshold based on user historical data;

[0037] a risk alert module, configured to issue an alert signal when the risk level exceeds a preset threshold risk value.

[0038] Via the technical solutions described above, compared with the prior art, the application discloses a kind of dynamic monitoring platform based on whole space-time synchronous respiratory circulatory metabolic function integration, with the following beneficial effects:

[0039] Through the integration of multi-modal data acquisition system (respiration, circulation, metabolism, motion posture), the platform can collect data in real time during the patient's daily activities or rehabilitation training in natural state, completely breaking the dependence of traditional cardiopulmonary exercise test (CPET) on laboratory environment, and realizing continuous, dynamic and real scene evaluation of patient physiological function.

[0040] The data synchronization system ensures the high-precision synchronization of multi-source heterogeneous data in time and space dimensions, realizes the precise correlation of different physiological system data in the same motion moment and the same body part state, and fundamentally solves the problems of data island and analysis fragmentation in the prior art.

[0041] The motion artifact correction module constructed based on multi-modal generative adversarial network, combined with filtering preprocessing and space-time attention mechanism, can intelligently identify and effectively correct signal distortion caused by body motion, evaluate physiological rationality combined with metabolic data, and significantly improve the signal-to-noise ratio and physiological credibility of data collected in dynamic activity state.

[0042] It can analyze and integrate physiological state in real time, dynamically calculate and output the current exercise risk level, and when the risk level exceeds the preset threshold, the risk warning module can immediately issue a warning signal to provide real-time safety protection for the patient and effectively prevent the occurrence of exercise-related adverse events. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are only embodiments of the application, and those skilled in the art can obtain other drawings according to the provided drawings without creative labor.

[0044] Figure 1 The system structure schematic diagram provided by the application;

[0045] Figure 2 The motion artifact correction module structure schematic diagram provided by the application. DETAILED DESCRIPTION

[0046] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all the other embodiments obtained by those ordinarily skilled in the art without creative work fall within the scope of the present application.

[0047] The embodiment of the present application discloses a dynamic monitoring platform for integrated function of whole-body spatiotemporal synchronous respiratory, circulation and metabolism, as shown in the figure, comprising a multi-modal data acquisition system, a data synchronization system, a data analysis system and a report generation system. Figure 1

[0048] The multi-modal data acquisition system is used for acquiring respiratory, circulation, metabolism and exercise data, and pre-processing the acquired data to generate pre-processed data.

[0049] The data synchronization system is used for correcting the pre-processed data and temporally and spatially aligning the corrected pre-processed data to obtain spatiotemporal synchronous data.

[0050] The data analysis system is used for inputting the spatiotemporal synchronous data into a trained dynamic analysis model and outputting an exercise risk level.

[0051] The report generation system is used for automatically generating a graphic-text report and exercise suggestions according to the spatiotemporal synchronous data and the exercise risk level.

[0052] In an embodiment, the multi-modal data acquisition system comprises:

[0053] The respiratory monitoring module is used for acquiring ventilation, oxygen uptake, carbon dioxide output and respiratory rate.

[0054] The circulation function module is used for synchronously acquiring heart rate, blood pressure and cardiac output.

[0055] The metabolism monitoring module is used for monitoring muscle oxygen saturation, blood glucose fluctuation data and blood lactic acid concentration.

[0056] The exercise posture capture module is used for acquiring exercise intensity, posture and energy consumption data through an inertial measurement unit and a three-dimensional motion capture system.

[0057] ​Further, the chronic disease patient is under the supervision of a doctor, and the appropriate power increasing rate is selected according to the patient's age, gender, disease severity, etc.; the treadmill test is carried out under the monitoring of the monitoring platform, and the process includes 3 minutes of rest, 3 minutes of no load, 60r / min speed treadmill warm-up, then the speed is unchanged, and the load power is increased by 30W / min until the maximum exercise tolerance, then the recovery period is observed for 10 minutes. The monitoring platform collects respiratory, circulatory and metabolic data through wearable electrocardio devices and multiple sensors, and obtains motion data through an inertial measurement unit (IMU) and a three-dimensional motion capture system; wherein the respiratory data includes ventilation, oxygen uptake, carbon dioxide output and respiratory rate; the circulatory data includes heart rate, blood pressure and cardiac output; the metabolic data includes muscle oxygen saturation, blood glucose fluctuation data and blood lactic acid concentration; the motion data includes motion intensity, posture and energy consumption.

[0058] In an embodiment, the data synchronization system comprises:

[0059] a heterogeneous clock synchronization module for performing data time synchronization on preprocessed data through a dynamic delay compensation algorithm to generate time synchronization data;

[0060] a motion artifact correction module for correcting distortion of the time synchronization data using a trained distortion correction model;

[0061] a space-time alignment module for establishing a dynamic biomechanical coordinate system with the center of gravity of the human body as the origin, and mapping the corrected data to a specific muscle group motion state through a Bayesian network to obtain space-time synchronization data.

[0062] In an embodiment, the heterogeneous clock synchronization module comprises a master node and a slave node; the master node is configured with a rubidium atomic clock, and the slave node is configured with a temperature-compensated crystal oscillator; a hybrid feedback channel is established between the master node and the slave node to form a star topology for clock signal synchronization and dynamic delay compensation between the master node and the slave node.

[0063] Further, the heterogeneous clock synchronization module introduces a dynamic delay compensation algorithm for parameter alignment, and its mathematical model is:

[0064]

[0065] In the formula, C is the compensation value, i is the number of delays; k is the weight coefficient of the delay; and d is the delay.

[0066] In an embodiment, the motion artifact correction module comprises:

[0067] a filtering unit for adaptively filtering the time synchronization data and suppressing non-stationary noise by adjusting a dynamic threshold to generate filtered data;

[0068] The correction unit is configured to input the filtered data into a trained deformity correction model, correct signal distortion of the motion period through a spatio-temporal attention mechanism, and output corrected data.

[0069] Further, the time synchronization data (including respiratory, circulatory, metabolic, and motion data) from the heterogeneous clock synchronization module is filtered through an adaptive filtering technique (other filtering techniques such as low-pass filtering can also be used) to remove noise, and a dynamic threshold adjustment method is used to suppress non-stationary noise caused by limb motion in real time.

[0070] In an embodiment, the deformity correction model is constructed based on a multi-modal generative adversarial network, including a generator and a discriminator; the generator takes the filtered data as input; the discriminator evaluates physiological reasonableness in combination with metabolic data, and corrects signal distortion of the motion period through a spatio-temporal attention mechanism to generate corrected data.

[0071] Further, the generator extracts spatio-temporal features through a convolutional neural network to generate a preliminary corrected signal; the discriminator receives the output signal of the generator and metabolic data, evaluates signal physiological reasonableness, and allocates correction weights through a spatio-temporal attention mechanism, wherein the spatio-temporal attention mechanism includes: a time attention module that focuses on periods with high incidence of motion artifacts (such as rapid limb swinging periods) to dynamically weight signal segments; and a spatial attention module that allocates correction weights for signal distortion of specific muscle groups or sensor channels (such as muscle tremor interference).

[0072] Still further, an adversarial loss function is used to guide the generator to optimize the correction result.

[0073] In an embodiment, the spatio-temporal alignment module includes:

[0074] The real-time coordinate system construction unit is configured to establish a dynamic biomechanical coordinate system according to the position of the body center of gravity, and update the origin of the coordinate system in real time through an inertial measurement unit.

[0075] The synchronized data generation unit is configured to use a Bayesian network to probabilistically fuse the corrected data, associate muscle group motion states with physiological data, and generate spatio-temporal synchronized data.

[0076] Further, the specific processing steps of the Bayesian network for probabilistically fusing the corrected data are as follows: first, define the prior probability distribution of the muscle group motion state; then update the posterior probability distribution based on the corrected data; and finally output the muscle group motion state mapping result corresponding to the maximum posterior probability to obtain the spatio-temporal synchronized data.

[0077] In an embodiment, the data analysis system includes:

[0078] a feature extraction module configured to perform multi-modal feature extraction on the received spatio-temporal synchronization data, the multi-modal features including respiratory features, circulatory features, metabolic features, and motion features;

[0079] a risk analysis module configured to input the respiratory features, the circulatory features, the metabolic features, and the motion features into a trained dynamic analysis model, and output a probability distribution of a motion risk level;

[0080] a level correction module configured to correct the probability distribution based on a preset risk threshold, and generate a final motion risk level.

[0081] In an embodiment, the step of obtaining the multi-modal features comprises:

[0082] extracting, from the respiratory data, oxygen uptake slope, carbon dioxide output to ventilation ratio, and other related features; calculating, from the circulatory data, heart rate slope, resting heart rate, cardiac output, and other related features; extracting, from the metabolic data, blood lactate concentration rise rate, spatial gradient of muscle oxygen saturation (SmO2), and other related data; and analyzing, from the motion module data, joint angle trajectory, center of gravity offset variance, and other related parameters.

[0083] In another embodiment, a spatio-temporal attention mechanism is embedded in the dynamic analysis model, which detects physiological parameter mutation points through a sliding window in the time dimension and assigns a high attention weight; in the spatial dimension, it identifies abnormal load muscle groups according to the motion posture data, correlates corresponding physiological feature data, and uses a multi-head self-attention layer to model the cross-modal spatio-temporal correlation.

[0084] In an embodiment, the dynamic analysis model is constructed using a dual-flow deep neural network, comprising:

[0085] a time series analysis branch configured to analyze the respiratory features, the circulatory features, and the metabolic features using an encoder-decoder structure, and obtain a first risk value;

[0086] a spatial analysis branch configured to analyze the motion features using a three-dimensional convolutional network, and obtain a second risk value;

[0087] a feature fusion layer configured to generate an output motion risk level probability distribution based on the first risk value and the second risk value.

[0088] Further, the encoder-decoder structure uses an encoder-decoder with a gated recurrent unit (GRU).

[0089] In an embodiment, the dynamic analysis model is trained using a multi-stage optimization strategy.

[0090] In a first stage, each branch network is pre-trained using a public motion physiological data set.

[0091] Second stage: Align the feature distribution of laboratory environment and real sports scene through Adversarial Domain Adaptation;

[0092] Third stage: Fine-tune the fusion layer parameters based on user personalized data.

[0093] In an embodiment, the report generation system comprises:

[0094] A risk traceability module for visualizing the contribution of each physiological parameter to the risk level through a backpropagation algorithm;

[0095] An adaptive exercise recommendation generation module for dynamically adjusting exercise recommendation thresholds based on user historical data;

[0096] A risk alert module for issuing an alert signal when the risk level exceeds a preset threshold risk value.

[0097] In an embodiment, the risk traceability module comprises:

[0098] A gradient calculation unit for quantifying the feature contribution weights of respiratory, circulatory, metabolic, and exercise data based on the gradient backpropagation of the dynamic analysis model output layer;

[0099] A space-time mapping unit for superimposing the contribution weights with a user exercise posture three-dimensional model to mark the body parts and physiological parameter abnormalities corresponding to the high-risk period.

[0100] In an embodiment, the data processing steps of the adaptive exercise recommendation generation module are:

[0101] Based on the user's exercise load data in the past 30 days, an exponential weighted moving average algorithm is used to predict the tolerance threshold;

[0102] Combined with real-time environmental temperature and altitude, the tolerance threshold is corrected by linear interpolation;

[0103] According to the corrected threshold, a stepwise exercise intensity recommendation is generated, such as exercise intensity, exercise type, exercise duration, etc.

[0104] In an embodiment, the risk alert module comprises

[0105] A multi-level triggering unit configured to:

[0106] When the risk level probability is 30%-60%, trigger a first-level alert (user terminal vibration prompt);

[0107] When the risk level probability is 60%-90%, trigger a second-level alert (voice broadcast emergency stop instruction);

[0108] When the risk level probability exceeds 90%, trigger level 3 alarm (send emergency stop control signal to smart sports equipment);

[0109] Device linkage interface, support sending deceleration instruction and locking instruction to treadmill and power bicycle through communication protocol.

[0110] Further, emergency medical interface, when triggering level 3 alarm, automatically send help information including user location, real-time physiological data and risk trace heat map to preset emergency center;

[0111] In an embodiment, it also includes a local data erasing unit, which performs AES-256 encryption covering on sensitive data of the user terminal after sending the help information.

[0112] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The same or similar parts between the various embodiments can be referred to each other. For the device disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant part can be referred to the method part.

[0113] The above description of disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to the embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An integrated dynamic monitoring platform for respiratory, circulatory and metabolic functions based on overall spatiotemporal synchronization, characterized by: include: Multimodal data acquisition system, data synchronization system, data analysis system and report generation system; The multimodal data acquisition system is used to collect respiratory, circulatory, metabolic and motion data, and preprocess the collected data to generate preprocessed data; The data synchronization system is used to correct the pre-processed data and perform spatiotemporal alignment on the corrected pre-processed data to obtain spatiotemporal synchronized data; The data analysis system is used to input the spatiotemporal synchronization data into a trained dynamic analysis model and output a motion risk level; The report generation system is used to automatically generate a graphic report and exercise recommendations based on the spatiotemporal synchronization data and the exercise risk level.

2. The integrated dynamic monitoring platform for respiratory, circulatory and metabolic functions based on overall spatiotemporal synchronization according to claim 1, characterized in that: The multimodal data acquisition system comprises: Respiratory monitoring module, used to collect ventilation, oxygen uptake, carbon dioxide output and respiratory rate; Circulatory function module, used to synchronously obtain heart rate, blood pressure and cardiac output; Metabolic monitoring module, used to monitor muscle oxygen saturation, blood sugar fluctuation data and blood lactate concentration; The motion posture capture module is used to obtain motion intensity, posture and energy consumption data through an inertial measurement unit and a three-dimensional motion capture system.

3. The integrated dynamic monitoring platform for respiratory, circulatory and metabolic functions based on overall spatiotemporal synchronization according to claim 1 is characterized in that: The data synchronization system includes: A heterogeneous clock synchronization module, configured to perform data time synchronization on the pre-processed data by using a dynamic delay compensation algorithm to generate time synchronization data; A motion artifact correction module, configured to perform distortion correction on the time-synchronized data using a trained distortion correction model; The spatiotemporal alignment module is used to establish a dynamic biomechanical coordinate system with the center of gravity of the human body as the origin, and to fuse the correction data through the Bayesian network, map it to the movement state of a specific muscle group, and obtain the spatiotemporal synchronization data.

4. The integrated dynamic monitoring platform for respiratory, circulatory and metabolic functions based on overall spatiotemporal synchronization according to claim 3 is characterized in that: The heterogeneous clock synchronization module includes a master control node and a sub-node; the master control node is configured with a rubidium atomic clock, and the sub-node is configured with a temperature-compensated crystal oscillator; a hybrid feedback channel is established between the master control node and the sub-node for clock signal synchronization and dynamic delay compensation between the master control node and the sub-node.

5. The integrated dynamic monitoring platform for respiratory, circulatory and metabolic functions based on overall spatiotemporal synchronization according to claim 3 is characterized in that: The motion artifact correction module includes: a filtering unit, configured to perform filtering and dynamic noise suppression processing on the time synchronization data to generate filtered data; The correction unit is used to input the filtered data into the trained deformity correction model and output the corrected data.

6. The integrated dynamic monitoring platform for respiratory, circulatory and metabolic functions based on overall spatiotemporal synchronization according to claim 5 is characterized in that: The deformity correction model is constructed based on a multimodal generative adversarial network, including a generator and a discriminator; the generator takes the filtered data as input; the discriminator combines metabolic data to evaluate physiological rationality, and corrects signal distortion during exercise periods through a spatiotemporal attention mechanism to generate corrected data.

7. The integrated dynamic monitoring platform for respiratory, circulatory and metabolic functions based on overall spatiotemporal synchronization according to claim 1, characterized in that: The data analysis system includes: a feature extraction module, configured to extract multimodal features from the received spatiotemporal synchronized data, wherein the multimodal features include respiratory features, circulatory features, metabolic features, and motion features; a risk analysis module, configured to input the respiratory characteristics, the circulatory characteristics, the metabolic characteristics, and the motion characteristics into a trained dynamic analysis model and output a probability distribution of motion risk levels; The level correction module is used to correct the probability distribution based on a preset risk threshold to generate a final sports risk level.

8. The integrated dynamic monitoring platform for respiratory, circulatory and metabolic functions based on overall spatiotemporal synchronization according to claim 7 is characterized in that: The dynamic analysis model is constructed using a two-stream deep neural network, including: a time series analysis branch, configured to analyze the respiratory feature, the circulatory feature, and the metabolic feature using an encoder-decoder structure to obtain a first risk value; a spatial analysis branch, configured to analyze the motion features using a three-dimensional convolutional network to obtain a second risk value; A feature fusion layer is used to generate an output motion risk level probability distribution according to the first risk value and the second risk value.

9. The integrated dynamic monitoring platform for respiratory, circulatory and metabolic functions based on overall spatiotemporal synchronization according to claim 8, characterized in that: The dynamic analysis model is trained using a multi-stage optimization strategy.

10. The integrated dynamic monitoring platform for respiratory, circulatory and metabolic functions based on overall spatiotemporal synchronization according to claim 1, characterized in that: The report generating system comprises: The risk tracing module is used to visualize the contribution of each physiological parameter to the risk level through the back-propagation algorithm; Adaptive exercise suggestion generation module, used to dynamically adjust the exercise suggestion threshold based on user historical data; The risk warning module sends an alarm signal when the risk level exceeds the preset threshold risk value.

Images