An intelligent monitoring and analysis system for multi-source critical illness parameters

Abstract

The present invention relates to the field of medical information technology processing technology, and in particular to an intelligent monitoring and analysis system for multi-source critical illness parameters. The system includes the following modules: a multi-source critical illness parameter acquisition module, a multi-source critical illness parameter noise reduction module, a multi-source critical illness parameter feature analysis module, a multi-source critical illness correlation trend feature analysis module, a multi-source critical illness abnormal condition analysis module, and a critical illness pathology auxiliary medical decision feedback module, which is used to monitor multi-source critical illness parameters through a medical sensor device driven by dynamic perception, perform noise reduction processing based on the time-frequency monitoring characteristics of the multi-source critical illness parameters, and generate noise-reduced multi-source critical illness parameters; analyze multi-source critical illness abnormal conditions through the nonlinear multi-source critical illness correlation trend characteristics of the noise-reduced multi-source critical illness parameters, and perform intelligent analysis of multi-source critical illness pathology through the multi-source critical illness abnormal conditions, and feed back the analysis results to the terminal. The present invention realizes auxiliary treatment decision support for multi-source critical illness pathology through intelligent multi-source critical illness parameter monitoring and analysis.

Description

Technical Field

[0001] The present invention relates to the field of medical information technology processing technology, and in particular to an intelligent monitoring and analysis system for multi-source critical illness parameters. Background Art

[0002] Multi-source critical care parameters refer to multi-dimensional data obtained from different vital signs and physiological indicators of critically ill patients. These data come from a variety of medical devices and sensors, covering key indicators such as heart rate, blood pressure, respiratory rate, blood oxygen saturation, body temperature, blood sugar, etc., reflecting the health status of patients. Due to the complexity and rapid changes of critically ill patients, a single parameter is often not enough to fully evaluate their condition. Intelligent monitoring and analysis of multi-source critical care parameters is an intelligent technology that integrates multiple data sources and monitors the vital signs and physiological parameters of critically ill patients in real time. It uses data fusion and artificial intelligence algorithms to analyze these data, provide intelligent diagnosis and treatment suggestions for medical staff, help adjust treatment plans in time, optimize the intensive care process, and improve patients' survival rate and medical quality. However, the traditional monitoring and analysis technology of multi-source critical care parameters collects multi-source critical care parameters through a single frequency method, and the dynamic changes of the monitored multi-source critical care parameters are large, which are interfered by various noises, resulting in low analysis efficiency and accuracy of multi-source critical care parameters, and it is impossible to use the potential information of multi-source critical care parameters for in-depth intelligent analysis, which is easy to miss some potential trends in disease changes and lacks support for auxiliary medical decision-making. Summary of the invention

[0003] Based on this, the present invention provides an intelligent monitoring and analysis system for multi-source critical illness parameters to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, an intelligent monitoring and analysis system for multi-source critical illness parameters includes the following modules:

[0005] The multi-source critical illness parameter acquisition module is used to perform intelligent adjustment processing on the dynamic perception monitoring parameters of the medical sensor equipment to obtain the intelligently adjusted medical sensor equipment; the acquisition parameters of the preliminary multi-source critical illness parameters are updated and optimized according to the intelligently adjusted medical sensor equipment to generate the multi-source critical illness parameters;

[0006] The multi-source critical illness parameter denoising module is used to perform time-series frequency monitoring feature analysis processing based on the multi-source critical illness parameters to generate time-frequency monitoring and analysis multi-source critical illness parameters; based on the time-frequency monitoring and analysis multi-source critical illness parameters, the time-frequency difference adaptive filtering denoising processing of the multi-source critical illness parameters is performed to generate denoised multi-source critical illness parameters;

[0007] Multi-source critical illness parameter feature analysis module, used for extracting multi-source critical illness parameter features based on denoised multi-source critical illness parameters for multi-scale entropy evaluation, and generating multi-source critical illness parameter feature data;

[0008] A multi-source critical illness correlation trend characteristic analysis module is used to perform multi-source critical illness parameter correlation analysis based on multi-source critical illness parameter characteristic data to generate multi-source critical illness parameter correlation data; perform non-linear multi-source critical illness correlation trend characteristic analysis based on multi-source critical illness parameter correlation data and multi-source critical illness parameter characteristic data to generate non-linear multi-source critical illness correlation trend characteristic data;

[0009] The multi-source critical illness abnormal condition analysis module is used to obtain historical multi-source critical illness routine diagnosis data; perform critical illness diagnosis health pattern modeling processing based on historical multi-source critical illness routine diagnosis data to generate a critical illness diagnosis health pattern model; perform multi-source critical illness abnormal condition analysis on the nonlinear multi-source critical illness correlation trend feature data based on the critical illness diagnosis health pattern model to generate multi-source critical illness abnormal condition data;

[0010] The critical illness pathology assisted medical decision-making feedback module is used to perform critical illness pathology diagnosis-probability intelligent reasoning processing based on multi-source critical illness abnormal condition data to generate critical illness pathology diagnosis-probability data; perform assisted critical illness pathology clinical decision-making intelligent prediction processing based on critical illness pathology diagnosis-probability data to generate assisted critical illness pathology clinical decision-making data; and feed back the critical illness pathology diagnosis-probability data and assisted critical illness pathology clinical decision-making data to the terminal to execute the critical illness pathology assisted medical decision-making feedback operation.

[0011] The present invention uses a multi-source critical care parameter acquisition module to dynamically sense and monitor and intelligently adjust the medical sensor equipment, and can optimize the monitoring parameters of the sensor equipment according to actual conditions, making it more suitable for the changes in the condition of a specific patient, improving the accuracy and flexibility of data acquisition, ensuring that the generated multi-source critical care parameters are more reliable and effective, reducing data redundancy and unnecessary interference, and significantly improving the accuracy of subsequent data analysis. The timing frequency monitoring feature analysis and adaptive filtering noise reduction of the multi-source critical care parameter noise reduction module can effectively remove noise interference based on different critical care parameter monitoring timing frequencies, such as targeted processing of noise interference of multi-source critical care parameters monitored at high or low frequencies, making the collected multi-source critical care parameters clearer and more accurate, not only improving data quality, but also better capturing subtle changes in patients' vital signs, helping to discover potential abnormal conditions earlier, and the noise-reduced data provides a high-quality foundation for subsequent feature analysis and trend prediction. The multi-source critical illness parameter feature analysis module performs multi-scale entropy evaluation and feature extraction on the multi-source critical illness parameters after noise reduction. It can extract key features from complex critical illness parameter data, accurately identify the main factors affecting the patient's condition, simplify the data dimension, and more accurately reflect the patient's physiological state, helping medical staff and systems to better understand the patient's current health status and provide in-depth insights into changes in the condition. The multi-source critical illness correlation trend feature analysis module performs correlation analysis and nonlinear trend analysis on critical illness parameter feature data, which can reveal the correlation and change trend between various physiological parameters, predict the future development direction of the disease, and provide more systematic and intelligent support for monitoring and early warning, allowing medical staff to take measures in advance to avoid worsening of the disease. In particular, through nonlinear analysis, it discovers the complex trends of critical illness that are difficult to capture, improves the early warning of multi-source critical illness abnormal condition analysis module to build a multi-source critical illness health model model, and combines the current nonlinear multi-source critical illness correlation trend feature data for abnormal condition analysis, which can provide personalized health status assessment for critically ill patients, identify the deviation between the patient's current status and historical health model, detect abnormal conditions in time, and generate abnormal condition data, providing medical staff with intuitive abnormal warning information, and improving the sensitivity and accuracy of abnormal detection. The critical illness pathology auxiliary medical decision feedback module performs intelligent prediction of pathological diagnosis and clinical decision-making based on critical illness pathology abnormal condition data, which can not only provide specific diagnostic suggestions for medical staff, but also continuously optimize the decision model through reinforcement learning, making the diagnostic results more accurate. At the same time, the system can feed back these results to the terminal device to realize automated medical decision support, greatly reducing the burden on medical staff and improving the efficiency and accuracy of diagnosis and treatment.

[0012] The beneficial effects of the present application are that the intelligent monitoring and analysis system of multi-source critical illness parameters of the present invention analyzes the patient's physical signs through a multi-source critical illness parameter acquisition module, intelligently adjusts the monitoring time series frequency of multi-source critical illness parameters, collects multi-source critical illness parameters in real time according to the patient's personalized critical illness condition, and can accurately obtain the potential information of multi-source critical illness parameters, and analyzes the time-frequency monitoring characteristics of multi-source critical illness parameters through a multi-source critical illness parameter noise reduction module to set adaptive filtering parameters, adaptively adjusts the filtering parameters for the time-frequency monitoring differences of multi-source data, removes noise and retains the effective characteristics of the data, so that the analysis efficiency and accuracy of multi-source critical illness parameters are excellent. Through the multi-source critical illness parameter feature analysis module, complex physiological parameters are decomposed in the time-frequency domain, and the characteristic changes of data are revealed from different time scales to extract the key information of multi-source critical illness data. The nonlinear multi-source critical illness correlation trend feature data of multi-source critical illness parameters is analyzed through the multi-source critical illness correlation trend feature analysis module to reveal the abnormal conditions of multi-source critical illness, and the probability of each critical illness pathology and the corresponding clinical decision are intelligently analyzed through the critical illness pathology auxiliary medical decision feedback module to provide intelligent support for auxiliary medical decision. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 This is a module flow diagram of an intelligent monitoring and analysis system for multi-source critical illness parameters of the present invention;

[0014] Figure 2 for Figure 1 Detailed functional flow chart of the multi-source critical parameter noise reduction module;

[0015] Figure 3 for Figure 1 Detailed functional flow chart of the multi-source severe disease correlation trend feature analysis module;

[0016] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0017] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0018] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0019] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0020] To achieve this, please refer to Figures 1 to 3 The present invention provides an intelligent monitoring and analysis system for multi-source critical illness parameters, comprising the following modules:

[0021] The multi-source critical illness parameter acquisition module is used to perform intelligent adjustment processing on the dynamic perception monitoring parameters of the medical sensor equipment to obtain the intelligently adjusted medical sensor equipment; the acquisition parameters of the preliminary multi-source critical illness parameters are updated and optimized according to the intelligently adjusted medical sensor equipment to generate the multi-source critical illness parameters;

[0022] The multi-source critical illness parameter denoising module is used to perform time-series frequency monitoring feature analysis processing based on the multi-source critical illness parameters to generate time-frequency monitoring and analysis multi-source critical illness parameters; based on the time-frequency monitoring and analysis multi-source critical illness parameters, the time-frequency difference adaptive filtering denoising processing of the multi-source critical illness parameters is performed to generate denoised multi-source critical illness parameters;

[0023] Multi-source critical illness parameter feature analysis module, used for extracting multi-source critical illness parameter features based on denoised multi-source critical illness parameters for multi-scale entropy evaluation, and generating multi-source critical illness parameter feature data;

[0024] A multi-source critical illness correlation trend characteristic analysis module is used to perform multi-source critical illness parameter correlation analysis based on multi-source critical illness parameter characteristic data to generate multi-source critical illness parameter correlation data; perform non-linear multi-source critical illness correlation trend characteristic analysis based on multi-source critical illness parameter correlation data and multi-source critical illness parameter characteristic data to generate non-linear multi-source critical illness correlation trend characteristic data;

[0025] The multi-source critical illness abnormal condition analysis module is used to obtain historical multi-source critical illness routine diagnosis data; perform critical illness diagnosis health pattern modeling processing based on historical multi-source critical illness routine diagnosis data to generate a critical illness diagnosis health pattern model; perform multi-source critical illness abnormal condition analysis on the nonlinear multi-source critical illness correlation trend feature data based on the critical illness diagnosis health pattern model to generate multi-source critical illness abnormal condition data;

[0026] The critical illness pathology assisted medical decision-making feedback module is used to perform critical illness pathology diagnosis-probability intelligent reasoning processing based on multi-source critical illness abnormal condition data to generate critical illness pathology diagnosis-probability data; perform assisted critical illness pathology clinical decision-making intelligent prediction processing based on critical illness pathology diagnosis-probability data to generate assisted critical illness pathology clinical decision-making data; and feed back the critical illness pathology diagnosis-probability data and assisted critical illness pathology clinical decision-making data to the terminal to execute the critical illness pathology assisted medical decision-making feedback operation.

[0027] The present invention uses a multi-source critical care parameter acquisition module to dynamically sense and monitor and intelligently adjust the medical sensor equipment, and can optimize the monitoring parameters of the sensor equipment according to actual conditions, making it more suitable for the changes in the condition of a specific patient, improving the accuracy and flexibility of data acquisition, ensuring that the generated multi-source critical care parameters are more reliable and effective, reducing data redundancy and unnecessary interference, and significantly improving the accuracy of subsequent data analysis. The timing frequency monitoring feature analysis and adaptive filtering noise reduction of the multi-source critical care parameter noise reduction module can effectively remove noise interference based on different critical care parameter monitoring timing frequencies, such as targeted processing of noise interference of multi-source critical care parameters monitored at high or low frequencies, making the collected multi-source critical care parameters clearer and more accurate, not only improving data quality, but also better capturing subtle changes in patients' vital signs, helping to discover potential abnormal conditions earlier, and the noise-reduced data provides a high-quality foundation for subsequent feature analysis and trend prediction. The multi-source critical illness parameter feature analysis module performs multi-scale entropy evaluation and feature extraction on the multi-source critical illness parameters after noise reduction. It can extract key features from complex critical illness parameter data, accurately identify the main factors affecting the patient's condition, simplify the data dimension, and more accurately reflect the patient's physiological state, helping medical staff and systems to better understand the patient's current health status and provide in-depth insights into changes in the condition. The multi-source critical illness correlation trend feature analysis module performs correlation analysis and nonlinear trend analysis on critical illness parameter feature data, which can reveal the correlation and change trend between various physiological parameters, predict the future development direction of the disease, and provide more systematic and intelligent support for monitoring and early warning, allowing medical staff to take measures in advance to avoid worsening of the disease. In particular, through nonlinear analysis, it discovers the complex trends of critical illness that are difficult to capture, improves the early warning of multi-source critical illness abnormal condition analysis module to build a multi-source critical illness health model model, and combines the current nonlinear multi-source critical illness correlation trend feature data for abnormal condition analysis, which can provide personalized health status assessment for critically ill patients, identify the deviation between the patient's current status and historical health model, detect abnormal conditions in time, and generate abnormal condition data, providing medical staff with intuitive abnormal warning information, and improving the sensitivity and accuracy of abnormal detection. The critical illness pathology auxiliary medical decision feedback module performs intelligent prediction of pathological diagnosis and clinical decision-making based on critical illness pathology abnormal condition data, which can not only provide specific diagnostic suggestions for medical staff, but also continuously optimize the decision model through reinforcement learning, making the diagnostic results more accurate. At the same time, the system can feed back these results to the terminal device to realize automated medical decision support, greatly reducing the burden on medical staff and improving the efficiency and accuracy of diagnosis and treatment.

[0028] In the embodiment of the present invention, reference Figure 1The above is a module flow diagram of an intelligent monitoring and analysis system for multi-source critical illness parameters of the present invention. In this embodiment, the intelligent monitoring and analysis system for multi-source critical illness parameters includes:

[0029] S1: A multi-source critical illness parameter acquisition module, which is used to perform intelligent adjustment processing on the dynamic perception monitoring parameters of the medical sensor device to obtain the intelligently adjusted medical sensor device; perform acquisition parameter update and optimization processing on the preliminary multi-source critical illness parameters according to the intelligently adjusted medical sensor device to generate the multi-source critical illness parameters;

[0030] In an embodiment of the present invention, data collection is performed through multiple medical sensor devices. These sensors include, but are not limited to, electrocardiogram sensors, blood pressure sensors, blood oxygen sensors, respiratory rate sensors, etc., which monitor the physiological parameters of critically ill patients in real time, analyze the preliminarily collected multi-source critical parameters through a dynamic sensing monitoring mechanism, and adjust the monitoring frequency and sensitivity of the sensor. For example, when the heart rate fluctuates greatly, the sampling frequency of the electrocardiogram sensor will automatically increase to capture subtle changes in order to set dynamic sensing monitoring parameters. Based on the dynamic sensing monitoring parameters, the system intelligently adjusts the sensor device, such as adjusting the acquisition threshold and optimizing the data filtering rules, so as to ensure the accuracy and timeliness of data collection. The adjusted sensor device monitors the patient's multi-source critical parameters again to generate optimized multi-source critical parameters, and transmits them to the system in real time for subsequent processing.

[0031] S2: Multi-source critical illness parameter denoising module, used to perform time-series frequency monitoring feature analysis processing according to multi-source critical illness parameters, and generate time-frequency monitoring and analysis multi-source critical illness parameters; based on time-frequency monitoring and analysis of multi-source critical illness parameters, perform time-frequency difference adaptive filtering denoising processing of multi-source critical illness parameters, and generate denoised multi-source critical illness parameters;

[0032] In an embodiment of the present invention, multi-source critical care parameters are processed in a time-series synchronization manner to ensure that the data collected by different sensors are consistent on the time axis. For example, the timestamps of the electrocardiogram and blood pressure data are accurately aligned to ensure synchronous monitoring. The data that has undergone time-series synchronization will be analyzed for time-series frequency characteristics. The system will perform time-series frequency domain analysis on parameters such as heart rate, blood pressure, and respiratory signals, and extract time-series frequency characteristic data, such as frequency distribution and amplitude changes per unit time. Based on the analysis results, the system further uses a time-frequency difference adaptive filtering algorithm to perform noise filtering on the frequency characteristics of different physiological parameters. For example, the power frequency interference (50Hz) in the electrocardiogram signal will be removed by adaptive filtering, while the heart rate signal will be retained, and the noise-reduced multi-source critical care parameters will be output to provide high-quality data for subsequent analysis.

[0033] S3: Multi-source critical illness parameter feature analysis module, used for extracting multi-source critical illness parameter features based on denoised multi-source critical illness parameters for multi-scale entropy evaluation, and generating multi-source critical illness parameter feature data;

[0034] In an embodiment of the present invention, the multi-source critical illness parameters after denoising will be subjected to multi-scale time-frequency discretization processing. The system discretizes the critical illness parameters according to the frequency characteristics of different time scales. For example, the changing characteristics of the electrocardiogram data in the high-frequency domain and the key information of the blood pressure data in the low-frequency domain can be captured. Based on the discretized multi-source critical illness parameters, multi-scale entropy evaluation processing is performed. By evaluating the entropy value of each parameter, the complexity and change of the data can be quantified. For complex ECG signals, the system will evaluate its multi-scale entropy value to extract important features under specific pathological conditions. According to the entropy evaluation results, the system extracts key feature data. For example, features such as heart rate variability, blood pressure fluctuation amplitude, and abnormal respiratory rate will be extracted from the data of multi-source critical illness parameters to generate the final multi-source critical illness parameter feature data.

[0035] S4: Multi-source critical illness correlation trend characteristic analysis module, used to perform multi-source critical illness parameter correlation analysis based on multi-source critical illness parameter characteristic data to generate multi-source critical illness parameter correlation data; perform nonlinear multi-source critical illness correlation trend characteristic analysis based on multi-source critical illness parameter correlation data and multi-source critical illness parameter characteristic data to generate nonlinear multi-source critical illness correlation trend characteristic data;

[0036] In an embodiment of the present invention, the extracted multi-source critical illness parameter feature data is subjected to correlation analysis. For example, by analyzing the changing trends of heart rate and blood oxygen level, the correlation between them can be revealed, and the corresponding correlation data can be generated. The multi-source critical illness parameter feature data is used as a node, and the correlation data is used as an edge to construct a multi-source critical illness feature correlation map. For example, the electrocardiogram signal is used as a node, blood pressure and blood oxygen level are used as other nodes, and the edges between the nodes represent the correlation strength between the parameters, and the map is further subjected to time-series dependency embedding processing. For example, through time series analysis, the system identifies the correlation change trend of certain physiological characteristics at different time points, and based on the time series correlation map, the system analyzes the nonlinear correlation trend. For example, the electrocardiogram signal and the respiratory rate may show nonlinear correlation trend change characteristics under specific pathological conditions. The system identifies this complex target information to generate nonlinear multi-source critical illness correlation trend feature data.

[0037] S5: Multi-source critical illness abnormal condition analysis module, used to obtain historical multi-source critical illness routine diagnosis data; perform critical illness diagnosis health pattern modeling processing based on historical multi-source critical illness routine diagnosis data to generate a critical illness diagnosis health pattern model; perform multi-source critical illness abnormal condition analysis on nonlinear multi-source critical illness correlation trend feature data based on the critical illness diagnosis health pattern model to generate multi-source critical illness abnormal condition data;

[0038] In an embodiment of the present invention, historical multi-source critical illness routine diagnosis data is obtained, such as routine data of electrocardiograms, blood pressure, and blood oxygen levels of past patients, and health pattern modeling is performed based on these historical data to establish a critical illness diagnosis health pattern model. The model can reflect the characteristic patterns of healthy patients and changes in their condition, compare the nonlinear multi-source critical illness correlation trend characteristic data with the health pattern model, and analyze the deviation of the current patient's data from the health pattern. For example, the fluctuation of the current electrocardiogram signal deviates from the normal waveform in the health pattern, and further abnormal distribution detection processing is performed on this deviation to generate abnormal distribution data for the current patient, identify abnormal heart rate or abnormal breathing, and thus generate multi-source critical illness abnormal condition data, which describes the patient's current abnormal condition in detail.

[0039] S6: Critical pathology assisted medical decision feedback module is used to perform critical pathology diagnosis-probability intelligent reasoning processing based on multi-source critical abnormal condition data to generate critical pathology diagnosis-probability data; perform assisted critical pathology clinical decision intelligent prediction processing based on critical pathology diagnosis-probability data to generate assisted critical pathology clinical decision data; feed back critical pathology diagnosis-probability data and assisted critical pathology clinical decision data to the terminal to execute critical pathology assisted medical decision feedback operation.

[0040] In an embodiment of the present invention, the critical pathology diagnosis data of historical patients is obtained, and intelligent auxiliary medical decision analysis is performed in combination with the critical medicine knowledge graph data. For example, the system uses the diagnostic information about cardiovascular diseases in the knowledge graph to assist in the diagnosis of the current patient. Based on the Bayesian reasoning algorithm and the critical pathology diagnosis data of historical patients, the possible pathological diagnosis results of the patient are inferred based on the multi-source critical abnormal condition data, and the probability data of critical pathology diagnosis is generated. For example, the system may infer that the probability of myocardial ischemia in the patient is 80%, and the probability of health status is 20%. Further based on the reinforcement learning algorithm, it learns how to optimize clinical decisions through a large amount of historical data, and uses it to predict auxiliary treatment plans. According to the generated auxiliary critical pathology clinical decision data, it is provided to medical staff for assisting clinical decision-making and treatment. Finally, all diagnosis and decision data will be fed back to the terminal device for real-time reference and execution by medical staff.

[0041] Preferably, the multi-source critical care parameter acquisition module includes the following functions:

[0042] Perform preliminary collection of multi-source critical illness parameters according to medical sensor equipment to obtain preliminary multi-source critical illness parameters;

[0043] Based on the preliminary multi-source critical illness parameters, the dynamic perception monitoring parameters of medical sensors are analyzed to generate dynamic perception monitoring parameters;

[0044] Performing intelligent adjustment processing of dynamic sensing monitoring parameters on the medical sensor device according to the dynamic sensing monitoring parameters to obtain an intelligently adjusted medical sensor device;

[0045] The intelligently adjusted medical sensor equipment collects and updates the preliminary multi-source critical care parameters to generate multi-source critical care parameters.

[0046] The present invention collects preliminary data from a variety of medical sensor devices, and the system can timely obtain the multi-source critical parameters of the patient, comprehensively reflect the patient's real-time vital signs, ensure that all key information can be quickly captured, and lay the foundation for subsequent data analysis and processing. The preliminary collection covers a variety of physiological indicators to ensure the comprehensiveness of the monitoring data. By analyzing the preliminary collected multi-source critical parameters, the patient's current health status and the working status of the medical sensor can be dynamically understood. According to these analysis results, the system generates dynamic perception monitoring parameters, so that the sensor can better adapt to the individual differences of the patient, improve the flexibility and sensitivity of the monitoring system, and enable the monitoring of the sensor to be adjusted at any time. The sensor device is intelligently adjusted, automatically adjusted according to the specific needs and status of the patient, and the acquisition frequency, accuracy and other key parameters of the sensor are optimized, so that the device is always in an excellent state, thereby improving the accuracy and timeliness of data collection, reducing the need for intervention by medical staff, and significantly improving the personalization and adaptability of monitoring. After the sensor device is intelligently adjusted, the preliminary collected multi-source critical parameters are updated and optimized, and higher-quality multi-source critical parameters can be generated according to the latest device status to ensure the accuracy and integrity of the data. The optimized data better reflects the multi-source critical care monitoring status of patients and provides a more reliable basis for subsequent data analysis.

[0047] In an embodiment of the present invention, multiple medical sensor devices are used to collect physiological parameters of critically ill patients in real time. For example, an electrocardiogram sensor is used to monitor the patient's heart rate, a blood oxygen sensor is used to detect blood oxygen saturation, a blood pressure sensor is used to collect blood pressure data of the patient, and a respiratory sensor is used to measure respiratory rate, etc. These sensors interact with the system through a wireless network or a wired connection, and the preliminary collected data is transmitted to the system to form preliminary multi-source critical parameters. The preliminary multi-source parameters may include a timestamp, a sensor type, and the original physiological data collected. The system analyzes the data monitored by the sensor based on the preliminary collected multi-source critical parameters. For example, the system detects whether the heart rate fluctuates rapidly, whether the respiratory rate changes significantly, or whether the blood oxygen saturation is lower than the safety threshold. Through these data analyses, the system generates dynamic perception monitoring parameters to reflect the patient's current condition. For example, if the system detects an irregular heart rate or abnormal breathing, the system adjusts the monitoring frequency of the electrocardiogram and respiratory sensors to capture the patient's state changes more carefully. According to the dynamic perception monitoring parameters, the system intelligently adjusts the acquisition frequency and sensitivity of the sensor. For example, when the heart rate fluctuates greatly, the sampling rate of the ECG sensor is automatically increased to obtain more data points. When the blood oxygen saturation remains stable, the system can reduce the sampling frequency of the blood oxygen sensor to save device energy consumption, adjust the sensor threshold or filter noise parameters. For example, for the ECG sensor, the system may adjust the baseline noise filtering algorithm to eliminate environmental interference and ensure the accuracy of data collection. The intelligently adjusted sensor device continues to monitor the patient and re-collects more accurate data based on the updated acquisition frequency and adjusted acquisition parameters. For example, the ECG sensor will capture more detailed heart rate change patterns, and the blood oxygen sensor will continue to monitor the patient's blood oxygen level changes in a low-energy state. By updating and optimizing the acquisition parameters, multi-source critical care parameters are generated. These data provide high-quality input for subsequent analysis, diagnosis, and decision support.

[0048] Preferably, the dynamic sensing monitoring parameter analysis of medical sensors based on preliminary multi-source critical illness parameters includes:

[0049] Perform heterogeneous characteristic analysis of multi-source critical illness parameters based on preliminary multi-source critical illness parameters to generate heterogeneous characteristic data of multi-source critical illness parameters;

[0050] Conduct self-organizing network collaboration analysis of monitoring nodes based on heterogeneous characteristic data of multi-source critical illness parameters to generate self-organizing network collaboration data of monitoring nodes;

[0051] Perform condition-driven adaptive monitoring drive frequency analysis based on preliminary multi-source critical illness parameters to generate adaptive monitoring drive frequency data;

[0052] According to the monitoring node self-organizing network collaboration data and the adaptive monitoring driving frequency data, the dynamic perception monitoring parameters of the medical sensor are analyzed to generate dynamic perception monitoring parameters.

[0053] The present invention performs heterogeneous feature analysis on the initially collected multi-source critical illness parameters, identifies the differences in structure and data features of different types of physiological parameters, ensures that the system can process various heterogeneous data from different sensor devices, and uniformly processes them to generate standardized heterogeneous feature data, improves the compatibility of multi-source data, enables the system to efficiently integrate different data sources, and overcomes the analysis obstacles brought by data heterogeneity. Based on the heterogeneous feature data of multi-source critical illness parameters, self-organizing network collaborative analysis is performed on each monitoring node, and collaboration and communication between different sensor nodes are realized through self-organizing network technology, thereby improving the efficiency of data transmission and processing. Through automatic collaboration between nodes, the monitoring tasks of each sensor are dynamically adjusted to achieve more effective resource allocation and information sharing, enhance the intelligence level of the monitoring system, and improve the accuracy and flexibility of data collection. The frequency of monitoring drive is dynamically adjusted according to the patient's condition, so that the system can flexibly adjust the monitoring frequency of the sensor according to different condition requirements. Through this adaptive drive frequency analysis, it can respond to changes in the condition in a timely manner, avoid resource waste or monitoring lag caused by fixed frequency monitoring, and the adjustment of the adaptive drive frequency improves the accuracy and real-time performance of monitoring, ensuring that the system can respond quickly to changes in the condition at critical moments. The self-organizing network collaboration data and adaptive monitoring drive frequency data of the comprehensive monitoring nodes can perform comprehensive dynamic perception monitoring parameter analysis, ensuring that the medical sensor equipment can automatically adjust the parameter configuration according to the current patient's condition and monitoring needs to achieve the best monitoring effect. The generated dynamic perception monitoring parameters not only enhance the flexibility of the sensor, but also optimize the utilization of monitoring resources to the greatest extent, ensuring efficient and accurate data collection.

[0054] In an embodiment of the present invention, heterogeneous feature analysis is performed on the preliminarily collected multi-source critical parameters. These parameters come from different types of sensor devices (such as electrocardiogram, blood pressure, blood oxygen, respiratory rate, etc.), and the data characteristics collected by each sensor are different. For example, electrocardiogram data is a high-frequency waveform, while blood pressure data is a low-frequency change value. The heterogeneous characteristics of each parameter, including the sampling frequency, fluctuation amplitude, time dependence, etc. of the data, are analyzed to generate heterogeneous feature data of multi-source critical parameters, reflecting the characteristics and change patterns of various physiological signals. Based on the heterogeneous feature data, a self-organizing network collaboration analysis is performed on each sensor node. For example, different sensors may form a collaborative network through a wireless network or a wired connection to coordinate data collection and transmission, analyze the collaboration mode of different nodes, and ensure that the sensor nodes can communicate efficiently through the self-organizing network protocol. For electrocardiogram and respiratory rate sensors, let them share a synchronous data stream to ensure the collaborative collection between heart rate and respiratory rate. The generated self-organizing network collaboration data helps to improve the communication efficiency of the sensor network and reduce data delay and packet loss rate. According to the changes in the patient's condition, the monitoring frequency of each sensor is automatically adjusted. For example, when the system detects rapid fluctuations in heart rate, the system will increase the sampling frequency of the ECG sensor and appropriately reduce the sampling frequency of other sensors (such as blood oxygen sensors) to reduce the collection of irrelevant data. Adaptive monitoring drive frequency analysis can dynamically adjust the working frequency of each sensor according to the patient's condition to ensure the accuracy of data collection and efficient use of resources. For example, when the system detects that the patient's breathing rate is stable, the sampling frequency of the respiratory sensor can be reduced to reduce energy consumption. The generated adaptive monitoring drive frequency data is used to optimize the working mode of each sensor. The self-organizing network collaboration data of the monitoring node and the adaptive monitoring drive frequency data are combined to perform dynamic perception monitoring analysis. For example, according to the collaborative working conditions of the ECG sensor and the respiratory sensor, the monitoring frequency and data transmission mode between them are adjusted to improve the timeliness and accuracy of the data. The dynamic perception monitoring parameters are used to provide real-time feedback on the patient's current condition, such as the fluctuation amplitude of the ECG signal, the changing trend of the respiratory rate, etc., and the dynamic perception monitoring parameters are generated.

[0055] Preferably, the multi-source critical illness parameter noise reduction module includes the following functions:

[0056] Perform time-series synchronization and alignment processing on multi-source critical illness parameters to generate synchronized multi-source critical illness parameters;

[0057] Perform parameter monitoring time series frequency characteristic analysis and processing based on synchronous multi-source critical illness parameters to generate parameter monitoring time series frequency characteristic data;

[0058] According to the parameter monitoring time series frequency characteristic data, the synchronous multi-source critical illness parameters are analyzed and processed with time-frequency monitoring characteristics to generate time-frequency monitoring and analysis multi-source critical illness parameters;

[0059] Based on time-frequency monitoring and analysis of multi-source critical illness parameters, time-frequency difference adaptive filtering and denoising of multi-source critical illness parameters are performed to generate denoised multi-source critical illness parameters.

[0060] The present invention performs time-series synchronization and alignment processing on multi-source critical parameters to ensure that data from different sensors can be accurately matched and unified in the time dimension, effectively solves the time misalignment problem caused by different sensor sampling frequencies or data transmission delays, and ensures the synchronization and consistency of multi-source data. The generated synchronized multi-source critical parameters provide an accurate data basis for subsequent time-series analysis, which helps to improve the overall accuracy and real-time performance of monitoring. By performing time-series frequency feature analysis on synchronized multi-source critical parameters, the frequency characteristics and change laws of each monitoring parameter can be captured, and the potential periodicity or mutation characteristics in the parameters can be found, helping medical staff to better understand the dynamic changes of patients' vital signs. The generated time-series frequency feature data provides an effective reference for subsequent time-frequency monitoring analysis and filtering noise reduction. Using time-frequency analysis technology to deeply analyze the time-series frequency characteristics of parameter monitoring can reveal the changing trends of multi-source critical parameters in different time-frequency domains. Through time-frequency monitoring feature analysis, the abnormal fluctuations or specific patterns of different parameters in the time-frequency domain can be identified. The generated time-frequency monitoring analysis multi-source critical parameters provide reliable data support for subsequent adaptive filtering noise reduction, further enhancing the accuracy of the data and dynamic monitoring capabilities. Through time-frequency difference adaptive filtering and noise reduction processing, according to the characteristic changes of multi-source critical illness parameters in different frequency ranges, the filtering parameters are automatically adjusted to remove noise and retain key physiological signals. Adaptive filtering can dynamically adapt to the time-frequency characteristics of the parameters, making the noise reduction process more accurate and effective. The generated noise-reduced multi-source critical illness parameters not only greatly improve the clarity of the data, but also enhance the reliability of data analysis, providing more accurate basic data for subsequent diagnosis and prediction.

[0061] As an example of the present invention, refer to Figure 2 As shown, Figure 1 The functional flow chart of the multi-source critical illness parameter noise reduction module in this example includes:

[0062] S201: Performing time-series synchronization and alignment processing on multi-source critical illness parameters to generate synchronized multi-source critical illness parameters;

[0063] In an embodiment of the present invention, in multi-source critical care monitoring, the sampling frequencies and timestamps of different sensors may differ. In order to ensure the consistency and synchronization of the data, the system first performs time-series synchronization alignment processing on the multi-source critical care parameters. For example, the electrocardiogram sensor collects data 1000 times per second, and the blood oxygen sensor collects data 10 times per second. The data of different sensors are aligned on the time axis, and the electrocardiogram and blood oxygen data are made to correspond at the same time point through interpolation or data resampling methods. After the synchronization alignment process, synchronized multi-source critical care parameters are generated. These parameters include electrocardiogram, blood oxygen, respiratory rate and other data, and ensure that they correspond accurately in the time dimension.

[0064] S202: Analyze and process the time series frequency characteristics of parameter monitoring according to the synchronous multi-source critical illness parameters to generate time series frequency characteristic data of parameter monitoring;

[0065] In an embodiment of the present invention, after completing the timing synchronization alignment, the timing frequency characteristic analysis is performed on the synchronized multi-source critical care parameters. For example, the system performs timing frequency domain analysis on the electrocardiogram data to identify the frequency components under the same time unit in the electrocardiogram signal, such as baseline drift, QRS complex, and characteristic frequencies of arrhythmia. For blood oxygen saturation data, the system will analyze the frequency changes of its timing fluctuations to capture possible low-frequency or high-frequency abnormalities. Parameter monitoring timing frequency characteristic data is generated through timing frequency characteristic analysis to reflect the time-frequency characteristics of each physiological signal, providing a basis for subsequent noise filtering and noise reduction processing.

[0066] S203: performing time-frequency monitoring feature analysis processing on the synchronous multi-source critical illness parameters according to the parameter monitoring time series frequency feature data to generate time-frequency monitoring analysis multi-source critical illness parameters;

[0067] In an embodiment of the present invention, based on the time-series frequency characteristic data of the previous step, time-frequency monitoring characteristic analysis processing is performed on multi-source critical illness parameters. Through time-frequency analysis techniques, such as wavelet transform, the system can analyze the signal frequency components of different time periods. For example, the system performs a detailed analysis of the frequency distribution of specific time periods in the electrocardiogram signal (such as when the heartbeat accelerates), identifies the high-frequency and low-frequency components in the signal, and further detects abnormalities. At the same time, the respiratory frequency signal is analyzed for its frequency changes at different time points, especially when the condition fluctuates. The analyzed time-frequency data reveals the dynamic change patterns of each physiological signal in more detail, and generates multi-source critical illness parameters for time-frequency monitoring analysis.

[0068] S204: Based on the time-frequency monitoring and analysis of the multi-source critical illness parameters, the time-frequency difference adaptive filtering and denoising of the multi-source critical illness parameters are performed to generate denoised multi-source critical illness parameters.

[0069] In an embodiment of the present invention, the results of time-frequency monitoring and analysis are used to apply an adaptive filtering algorithm to perform noise reduction processing for the specific time-frequency characteristics of each physiological parameter. Through adaptive filtering, the system can dynamically adjust the filtering parameters to eliminate noise in different frequency bands. For example, in the electrocardiogram signal, power frequency interference (50Hz or 60Hz) is a common noise. The system removes the noise in this frequency band through adaptive filtering while retaining the valid electrocardiogram signal. For the respiratory rate signal, the low-frequency drift is removed by low-pass filtering, while maintaining the normal respiratory rhythm characteristics. After adaptive filtering of time-frequency differences, noise-reduced multi-source critical parameters are generated, and these data have a higher signal-to-noise ratio.

[0070] Preferably, the time-frequency difference adaptive filtering and denoising processing of multi-source critical illness parameters based on time-frequency monitoring and analysis includes the following functions:

[0071] Based on time-frequency monitoring and analysis of multi-source critical illness parameters, time-frequency difference noise analysis of multi-source critical illness parameters is performed to generate time-frequency difference noise data of multi-source critical illness parameters;

[0072] According to the multi-source critical illness parameter time-frequency difference noise data, the adaptive filtering parameter analysis of time-frequency monitoring difference is performed to generate the time-frequency difference adaptive filtering parameters;

[0073] According to the time-frequency difference adaptive filtering parameters, multi-source critical illness parameters are denoised and reconstructed for time-frequency monitoring and analysis to generate denoised multi-source critical illness parameters.

[0074] The present invention identifies the noise characteristics existing in different time and frequency domains in multi-source critical illness parameters through time-frequency difference noise analysis, accurately locates the source of unstable or interfering signals in the data, ensures that the classification of noise is more detailed and effective, and the generated time-frequency difference noise data provides an accurate noise characteristic reference for subsequent adaptive filtering, which helps the system to perform targeted noise reduction processing according to the characteristics of the noise, thereby improving the accuracy and effectiveness of noise removal. A detailed analysis of the time-frequency difference noise data can automatically generate adaptive filtering parameters for different time and frequency noise characteristics, adaptively adjust the filter parameters according to the distribution and intensity of different noises, and perform noise elimination in the most appropriate way, thereby improving the accuracy of the filtering process, avoiding the damage of effective information in the monitoring signal of time-frequency difference by traditional fixed filtering methods, and ensuring that important physiological signals are retained while noise reduction. Adaptive filtering parameters are used to reduce the noise of multi-source critical illness parameters analyzed by time-frequency monitoring, and important physiological signals are reconstructed in the process. The combination of noise reduction and signal reconstruction not only effectively removes noise, but also retains and even enhances key physiological information. The resulting reduced-noise multi-source critical illness parameters have higher signal quality and accuracy, providing more reliable and clear data support for subsequent medical analysis and diagnosis.

[0075] In an embodiment of the present invention, the time-frequency characteristics of multi-source critical parameters are obtained through time-frequency monitoring and analysis. For different physiological parameters (such as electrocardiogram, respiratory rate, blood oxygen level, etc.), their frequency distribution in different time periods is analyzed respectively, and the noise components therein are detected. For example, in an electrocardiogram signal, common noise comes from power frequency interference, baseline drift, and electromyographic interference, which appear in different frequency bands: power frequency interference is concentrated at 50Hz or 60Hz, baseline drift is usually in the low frequency range, and electromyographic interference appears in the high frequency band. Through time-frequency analysis techniques, such as wavelet transform or short-time Fourier transform (STFT), these noises are finely analyzed for time-frequency distribution to generate corresponding noise data. The generated time-frequency difference noise data identifies the frequency range and intensity of the noise. According to the time-frequency difference noise data, the system automatically performs filter parameter analysis and generates adaptive filter parameters for different noise components. The adaptive filter parameters can dynamically adjust the characteristics of the filter to adapt to the time-frequency changes of different physiological parameters and noise. For example, for low-frequency baseline drift in the electrocardiogram signal, an adaptive high-pass filter is used to remove the low-frequency noise. The cutoff frequency of the filter will be dynamically adjusted according to the specific noise situation. For high-frequency interference in the respiratory signal, an adaptive low-pass filter is used. The filter will flexibly adjust the bandwidth according to the frequency range and noise characteristics of the respiratory signal. The generation of adaptive filter parameters is accurately matched by the time-frequency distribution characteristics of the noise, so that the filter can effectively remove noise and retain important physiological signals. The generated time-frequency difference adaptive filter parameters provide the basis for the specific filtering process. After obtaining the adaptive filter parameters, the corresponding filtering process is applied to the data of each physiological parameter. For example, in the electrocardiogram signal, a high-pass filter is first used to remove the baseline drift, and then a bandpass filter is used to eliminate the electromyographic interference. For the blood oxygen signal, a specific low-pass filter needs to be selected according to the noise distribution to remove the high-frequency interference. After applying the filter to remove noise, the signal is parameter reconstructed to ensure that important physiological features are fully retained. For example, in an electrocardiogram, the reconstructed signal should completely retain key ECG features such as the P wave, QRS complex, and T wave, without causing signal distortion due to noise reduction. Through adaptive filtering and signal reconstruction of time-frequency differences, noise-reduced multi-source critical illness parameters are generated.

[0076] Preferably, the time-frequency difference noise analysis of multi-source critical illness parameters based on time-frequency monitoring and parsing of multi-source critical illness parameters includes:

[0077] Based on the empirical mode decomposition technology, the intrinsic mode function of the time-frequency monitoring difference of the multi-source critical parameters is analyzed to generate the time-frequency difference intrinsic mode function;

[0078] Calculating the energy density of the time-frequency difference intrinsic mode function according to the time-frequency difference intrinsic mode function to generate energy density data of the time-frequency difference intrinsic mode function;

[0079] Based on the energy density data of time-frequency difference intrinsic mode function, time-frequency difference noise analysis of multi-source critical illness parameters is performed to generate time-frequency difference noise data of multi-source critical illness parameters.

[0080] The present invention uses the empirical mode decomposition (EMD) technology to decompose the data of time-frequency monitoring analysis, and can decompose the complex multi-source critical parameter signals into several intrinsic mode functions (IMFs) with different frequency characteristics, effectively separating the noise in the data from the actual useful physiological signals, and the generated time-frequency difference intrinsic mode function accurately reflects the signal characteristics in different time and frequency domains, providing a more accurate data basis for subsequent energy density analysis and noise processing. The energy density of the intrinsic mode function is calculated to identify the distribution of energy in different modes, so as to accurately determine which parts of the signal are important physiological information and which parts may be noise or interference signals, retain important high-energy signals in the noise filtering process, and remove low-energy noise at the same time, thereby enhancing the system's ability to distinguish useful signals from noise and making noise reduction more accurate. The energy density data of the time-frequency difference intrinsic mode function is used to accurately analyze the time-frequency difference noise in the multi-source critical parameters, further refine the frequency and time distribution characteristics of the noise, and make the distinction between noise and useful signals clearer. The generated noise data provides accurate noise characteristics for subsequent adaptive filtering processing, ensuring that the noise removal process is more efficient and targeted while retaining key physiological signals to the greatest extent.

[0081] In an embodiment of the present invention, time-frequency monitoring and analysis of multi-source critical parameters (such as electrocardiogram, respiratory rate, etc.) are performed, and the empirical mode decomposition (EMD) technology is applied to decompose the complex signal into several intrinsic mode functions (IMF). For example, for an electrocardiogram signal, EMD decomposes the original waveform into multiple IMF components with different frequencies and amplitudes, and these components correspond to different frequency components in the electrocardiogram signal (such as low-frequency baseline drift, high-frequency myoelectric interference, etc.). Each IMF represents the dynamic changes in different frequency domains in the signal. After decomposition, these IMF components can more clearly show the time-frequency change characteristics in the signal, and generate time-frequency difference intrinsic mode functions for subsequent energy calculation and noise analysis. After generating the IMF, the energy density of each intrinsic mode function is further calculated. The energy density reflects the energy concentration of each IMF component within a specific frequency range, and reveals the strength of the effective signal and noise contained in different mode functions. For example, for low-frequency modal functions (IMF1), their energy density mainly reflects the intensity of baseline drift, while high-frequency modal functions (such as IMF3 or IMF4) are related to myoelectric interference. Through energy density calculation, high-energy areas (valid signals) and low-energy areas (noise) are distinguished. The system calculates the energy density of each IMF component in time and frequency, generates time-frequency difference intrinsic modal function energy density data, and provides a basis for subsequent noise detection and filtering. According to the calculated energy density data, the noise components in different IMF components are identified. Generally, modal functions with low energy density are more likely to contain noise, while modal functions with high energy density correspond to valid physiological signals. By analyzing these energy density data, noise detection of time-frequency difference is performed for different modal functions. For example, in electrocardiogram data, the low-energy component in the low-frequency IMF1 is regarded as baseline drift noise, and the low-energy part in the high-frequency IMF3 is myoelectric interference noise. The noise components are identified from the time-frequency signals to generate multi-source critical illness parameter time-frequency difference noise data, which can be used for subsequent adaptive filtering and noise reduction processing to ensure that the effective signal is retained and the noise is removed.

[0082] Preferably, the multi-source critical illness parameter feature analysis module includes the following functions:

[0083] Perform parameter multi-scale time-frequency discretization processing on the denoised multi-source critical illness parameters to generate multi-scale time-frequency discretized multi-source critical illness parameters;

[0084] Based on the multi-scale time-frequency discretization of multi-source critical illness parameters, the entropy value of multi-scale multi-source critical illness parameters is evaluated and processed to generate multi-scale multi-source critical illness parameter entropy data;

[0085] According to the multi-scale multi-source critical illness parameter entropy data, multi-source critical illness parameter features are extracted from the denoised multi-source critical illness parameters to generate multi-source critical illness parameter feature data.

[0086] The present invention performs multi-scale time-frequency discretization processing on the multi-source critical illness parameters after noise reduction, decomposes and discretizes complex physiological signals on different time and frequency scales, and captures the details of changes in the patient's physiological state on different scales. The entropy value of the multi-source critical illness parameters discretized in multi-scale time and frequency is evaluated to quantify the uncertainty and complexity in the data. The size of the entropy value can reflect the changing pattern of the patient's physiological state, help identify key physiological signals, and ensure that the direction of feature extraction is more accurate. Multi-scale entropy evaluation helps to mine deep-level information in the data and improves the sensitivity to changes in patient status. Feature extraction based on entropy data can identify the most diagnostically valuable features from the multi-source critical illness parameters after noise reduction. The feature extraction process can eliminate useless or redundant information and retain the key data that best reflects changes in the patient's condition, significantly improving the efficiency and accuracy of data processing.

[0087] In an embodiment of the present invention, multi-source critical parameters (such as electrocardiogram, blood oxygen, respiratory rate, etc.) after noise reduction are subjected to multi-scale time-frequency discretization processing, and the physiological signals are decomposed at different time and frequency scales to capture the changing characteristics of the signals at multiple scales. For example, an electrocardiogram signal may show a higher heart rate on a short time scale, while reflecting a more stable heart rhythm change on a longer time scale. Through time-frequency discretization processing, the system can simultaneously focus on the rapid changes and long-term trends of the electrocardiogram, thereby capturing the complex characteristics of different physiological signals at multiple scales. In the respiratory frequency signal, the short time scale can be used to analyze the instantaneous respiratory rhythm, while the longer time scale can be used to capture the slow changes in the respiratory pattern. The generated multi-scale time-frequency discretization multi-source critical parameters can provide a multi-dimensional data basis for subsequent entropy value evaluation and feature extraction. After completing the time-frequency discretization processing, the multi-source critical parameters on each time series frequency scale are subjected to entropy value evaluation, and the entropy value evaluation helps identify abnormal changes in the patient's physiological parameters by quantifying the uncertainty and complexity of the signal. For example, the sample entropy of the electrocardiogram signal is calculated to evaluate the complexity of the heart rate variability (HRV). A higher sample entropy may indicate that the heart rate has a strong variability, while a lower sample entropy may indicate that the heart rate tends to be regular or there is an abnormal arrhythmia. Similarly, the entropy evaluation of the respiratory signal can reveal the changes in the respiratory rhythm. If the respiratory rate changes significantly in a short period of time, the entropy value will be high, indicating irregular breathing; if the respiratory rate is stable, the entropy value is low. In this way, multi-scale multi-source critical parameter entropy data is generated, reflecting the complexity and abnormal characteristics of each physiological parameter at different scales. Based on the multi-scale multi-source critical parameter entropy data, the key features of the multi-source critical parameters are extracted, and the key and complex physiological signals are simplified into a set of important characteristic indicators reflecting the patient's health status. For example, for the electrocardiogram signal, characteristic indicators related to heart rate variability are extracted according to the entropy evaluation results, such as the heart rate mean, coefficient of variation, and energy distribution in a specific frequency band. These features can help identify pathological conditions such as arrhythmia and myocardial ischemia. For respiratory rate signals, the system may extract features related to respiratory fluctuations, such as the standard deviation of respiratory rate, the range of variation of respiratory intervals, etc., to help judge the stability of respiratory function or the risk of respiratory failure, and generate multi-source critical care parameter feature data. These data condense the key information in the original physiological signals and provide important reference for subsequent diagnosis, monitoring and decision support.

[0088] Preferably, the multi-source severe disease correlation trend feature analysis module includes the following functions:

[0089] Perform multi-source critical illness parameter correlation analysis based on multi-source critical illness parameter characteristic data to generate multi-source critical illness parameter correlation data;

[0090] The multi-source critical illness parameter feature data is used as a multi-source critical illness feature association map node, and the multi-source critical illness parameter correlation data is used as a multi-source critical illness feature association map edge, and a multi-source critical illness feature association map is established according to the multi-source critical illness feature association map nodes and the corresponding multi-source critical illness feature association map edges;

[0091] Perform graph temporal dependency embedding processing on the multi-source critical illness feature association graph to generate a time-series multi-source critical illness feature association graph;

[0092] According to the time series multi-source severe disease characteristic association map, nonlinear multi-source severe disease association trend characteristic analysis is performed to generate nonlinear multi-source severe disease association trend characteristic data.

[0093] The present invention performs correlation analysis on multi-source critical illness parameter feature data, identifies and quantifies the mutual influence and correlation degree between different physiological parameters, reveals the intrinsic relationship between various vital signs of patients, especially in complex critical illness environments, identifies potential pathological correlations, and improves the depth of understanding of the disease and global perception ability. Taking multi-source critical illness parameter feature data as nodes and correlation data as edges, a comprehensive multi-source critical illness feature correlation map can be constructed, which can intuitively display the complex relationship between different parameters. The system can fully grasp the change pattern of the patient's physiological state through this map. The graph time-series dependency embedding processing dynamically tracks and predicts the changes of physiological parameters over time by introducing the time dimension into the feature correlation map, revealing the change trend and correlation between physiological parameters at different time points, which helps the system identify long-term trends and short-term fluctuations. Nonlinear trend analysis of the time-series multi-source critical illness feature correlation map can identify complex correlations and trend changes that are difficult to capture with traditional linear methods. Nonlinear analysis can better reflect the complex dynamic characteristics in the physiological state of critically ill patients and capture potential trends of worsening conditions.

[0094] As an example of the present invention, refer to Figure 3 As shown, Figure 1 Schematic diagram of the functional flow of the multi-source severe disease correlation trend feature analysis module. In this example, the functions of the multi-source severe disease correlation trend feature analysis module include:

[0095] S401: performing multi-source critical illness parameter correlation analysis according to multi-source critical illness parameter characteristic data to generate multi-source critical illness parameter correlation data;

[0096] In an embodiment of the present invention, characteristic data of multi-source critical care parameters are obtained, including characteristics of multiple physiological signals such as electrocardiogram, blood oxygen, blood pressure, respiratory rate, etc. For example, electrocardiogram characteristics may include the amplitude and duration of P waves and QRS complexes, and blood oxygen characteristics include the range of saturation changes, etc. The correlation between different physiological parameters is calculated through statistical analysis and correlation analysis. For example, the system can calculate the correlation coefficient between the electrocardiogram signal and the respiratory rate, evaluate whether the two show synchronous fluctuations in time, or the linear or nonlinear correlation between blood pressure and blood oxygen saturation, identify the potential relationship between multiple physiological parameters, and generate multi-source critical care parameter correlation data, which reflect the correlation characteristics of each parameter in pathological changes.

[0097] S402: Using the multi-source critical illness parameter feature data as a multi-source critical illness feature association map node, and using the multi-source critical illness parameter correlation data as a multi-source critical illness feature association map edge, and establishing a multi-source critical illness feature association map according to the multi-source critical illness feature association map nodes and the corresponding multi-source critical illness feature association map edges;

[0098] In an embodiment of the present invention, the feature of each multi-source critical illness parameter is used as a graph node. For example, features such as the QRS complex of the electrocardiogram, the fluctuation amplitude of the respiratory rate, and the blood oxygen level exist as independent nodes. Then, the multi-source critical illness parameter correlation data is used as an edge in the graph to represent the relationship between different physiological parameter features. For example, the high correlation between the electrocardiogram and the respiratory rate will be represented by a strongly connected edge, while the weaker correlation between blood pressure and blood oxygen saturation may be connected by a weaker edge to construct a complete multi-source critical illness feature correlation graph, in which the nodes represent physiological features and the edges represent the strength of the correlation between these features. This graph provides a structured data representation for subsequent time series dependency analysis and trend prediction.

[0099] S403: performing graph temporal dependency embedding processing on the multi-source severe disease feature association graph to generate a temporal multi-source severe disease feature association graph;

[0100] In an embodiment of the present invention, after establishing a feature association map, the system introduces a time dimension and embeds the changes in physiological characteristics at different time points into the map, so that the map not only shows static correlations, but also can reflect the dynamic changes of physiological parameters in the time series. For example, track changes in the electrocardiogram, and analyze whether the correlation strength between the electrocardiogram features and the respiratory rate changes at different time points. Under certain pathological conditions, the correlation between the ECG signal and the respiratory rate may change significantly over time, and this time series embedding can capture this trend. Through the time-series dependency embedding processing of the map, a time-series multi-source critical feature association map is generated, which not only shows the static correlation between the various physiological parameters, but also reflects the changes in these correlations at different time points, predicting the future development of the disease.

[0101] S404: Perform nonlinear multi-source severe disease correlation trend feature analysis based on the time series multi-source severe disease feature correlation map to generate nonlinear multi-source severe disease correlation trend feature data.

[0102] In an embodiment of the present invention, nonlinear analysis techniques (such as support vector machines, nonlinear regression models or recursive neural networks) are applied to perform nonlinear multi-source critical illness correlation trend feature analysis on the time series multi-source critical illness feature association map. These techniques can identify complex dynamic relationships that cannot be captured by traditional linear analysis methods. The nonlinear change trend of the complex correlation pattern of multi-source critical illness features in the time series multi-source critical illness feature association map is analyzed by nonlinear analysis technology. For example, it is found that the patient's electrocardiogram and blood oxygen saturation do not present a simple linear relationship, but the nonlinear model identifies that under a specific heart rate threshold, the relationship between the two is sharply enhanced, reflecting that the physiological state has a sudden change under specific conditions. This nonlinear association can help the system detect potential acute changes in the condition. Finally, feature data that can reflect the dynamic changes in the physiological state of critically ill patients are extracted, including complex time-dependent correlation changes, such as nonlinear multi-source critical illness correlation trend feature data generated by the nonlinear relationship between heart rate and respiratory rate, blood pressure and blood oxygen saturation. For example, when the patient's heart rate exceeds a certain threshold, the change in respiratory rate is no longer stable, but fluctuates sharply; or when the blood oxygen level drops, the change in blood pressure may reflect the signal of worsening of the condition in advance.

[0103] Preferably, the multi-source severe abnormal condition analysis module includes the following functions:

[0104] Obtain historical multi-source critical care routine diagnosis data;

[0105] Based on the historical multi-source critical illness routine diagnosis data, critical illness diagnosis health pattern modeling is performed to generate a critical illness diagnosis health pattern model;

[0106] According to the critical illness diagnosis health pattern model, the nonlinear multi-source critical illness correlation trend feature data is analyzed for multi-source critical illness health pattern deviation characteristics to generate multi-source critical illness health pattern deviation feature data;

[0107] Perform multi-source severe illness abnormal distribution detection and processing based on multi-source severe illness health pattern deviation feature data to generate multi-source severe illness abnormal distribution data;

[0108] Multi-source severe abnormal condition analysis is performed based on multi-source severe abnormal condition distribution data to generate multi-source severe abnormal condition data.

[0109] The present invention obtains historical multi-source critical illness routine diagnosis data to obtain a large amount of previous diagnosis records and medical data. The acquisition of historical data provides reliable empirical data for subsequent health model modeling and abnormality detection, and enhances the system's accurate assessment of the current critical illness condition. Modeling of historical multi-source critical illness routine diagnosis data generates a critical illness diagnosis health model model, which represents the standard physiological parameter pattern of patients in different health states. The health model can provide a baseline for analyzing the data of the current patient and help the system evaluate the abnormality of the current data. The nonlinear multi-source critical illness association trend feature data of the current patient is compared and analyzed with the health model model to identify the deviation between the patient data and the health state, reveal the degree of difference between the patient's current state and the health model, and timely discover changes in the condition or potential risks. Based on the health model deviation feature data, the area of ​​abnormal distribution is identified and detected, the abnormal trend or change in the current data is revealed, the location of the disease change is further refined, and the potential pathological characteristics or disease risks are more accurately discovered. According to the multi-source critical illness health model deviation feature data, multi-source critical illness abnormal distribution detection processing is performed to reveal the health risks or disease deterioration trends currently faced by the patient, and provide early warning information. The abnormal condition data generated provides a reliable basis for medical decision-making, improves response efficiency and treatment accuracy, helps to intervene in the disease earlier, and reduces the possibility of disease deterioration.

[0110] In an embodiment of the present invention, a large amount of conventional diagnostic data of historical multi-source critically ill patients is extracted from a medical database, including various physiological parameters in intensive care (such as electrocardiogram, blood pressure, blood oxygen, respiratory rate, etc.). According to the historical multi-source critical routine diagnostic data, a critical diagnosis health pattern model is generated using machine learning or statistical modeling technology, which reflects the typical change pattern of each physiological parameter of critically ill patients under normal conditions. For example, by analyzing a large amount of electrocardiogram and blood pressure data, a standard physiological pattern when the heart function is normal can be constructed. Based on these typical patterns, the model further matches and analyzes the data of new patients. The model can not only capture the abnormality of a single physiological parameter, but also combine the joint features of multi-source physiological parameters to generate a more complex health pattern, so as to generate a critical diagnosis health pattern model, which provides a basis for subsequent data deviation analysis. Using the critical diagnosis health pattern model, the nonlinear multi-source critical association trend feature data of the current patient is compared and analyzed. By comparing the patient's current nonlinear multi-source critical association trend feature data with the standard health pattern, the deviation between the patient's physiological state and the health pattern can be identified. For example, the Mahalanobis distance is used to calculate the difference between the patient's current nonlinear multi-source critical illness correlation trend feature data and the standard health model model, and to evaluate whether the current physiological state deviates from the normal fluctuation range. If the patient's current electrocardiogram and respiratory rate deviate significantly from the standard health model, such as increased heart rate fluctuations and abnormally reduced respiratory rate, this deviation is identified and marked as a potential pathological feature. Deviation feature analysis helps the system discover subtle abnormal changes that may not be discovered in time during initial monitoring, and generates multi-source critical illness health model deviation feature data, which indicates the degree and characteristics of the patient's deviation from the normal health model. The deviation feature data is further processed for abnormal distribution detection. By analyzing the temporal changes of the patient's physiological parameters and their differences from the health model, the abnormal distribution of multi-source physiological parameters in space and time is detected. For example, it is detected that the patient's blood oxygen saturation and respiratory rate fluctuate abnormally at the same time in a specific time period, and these abnormal distributions are regular, indicating that the patient is at risk of impaired lung function. These abnormal points and abnormal areas are marked for further analysis. Through multi-source critical illness abnormal distribution detection, potential high-risk time periods and parameters are identified, and multi-source critical illness abnormal distribution data are generated. Based on the abnormal distribution data of multi-source critical illness, a comprehensive abnormal analysis of the patient's current health status is conducted, and the abnormal distribution characteristics of multi-source physiological parameters are integrated to give a complete abnormal diagnosis report. For example, the combined abnormal fluctuations of the patient's electrocardiogram, respiratory rate, and blood oxygen level in the past few hours are identified, and combined with the abnormal distribution, the abnormal conditions of multi-source critical illness are judged. These analysis results are summarized to generate multi-source critical illness abnormal condition data, which contains detailed abnormal conditions, abnormal nodes, time distribution, etc.

[0111] Preferably, the critical pathology auxiliary medical decision-making feedback module includes the following functions:

[0112] Obtain critical care pathology diagnosis data and critical care medicine knowledge graph data;

[0113] Based on the Bayesian algorithm and critical pathology diagnosis data, a mapping relationship between critical abnormal conditions and pathology diagnosis probability reasoning is established to generate an intelligent critical pathology diagnosis probability reasoning model;

[0114] Transmit multi-source critical illness abnormal condition data to the intelligent critical illness pathology diagnosis probability reasoning model for critical illness pathology diagnosis-probability intelligent reasoning processing to generate critical illness pathology diagnosis-probability data;

[0115] Based on reinforcement learning and critical care medicine knowledge graph data, the mapping relationship between critical care pathology diagnosis information and clinical medical decision-making is carried out to obtain an intelligent critical care clinical decision-making prediction model;

[0116] The intelligent assisted critical pathology diagnosis-probability data is transmitted to the intelligent critical pathology clinical decision prediction model for assisted critical pathology clinical decision intelligent prediction processing to generate assisted critical pathology clinical decision data;

[0117] Feedback the critical pathology diagnosis-probability data and the critical pathology auxiliary clinical decision-making data to the terminal to execute the critical pathology auxiliary medical decision-making feedback task.

[0118] The present invention obtains critical pathological diagnosis data and critical medical knowledge graph data, and uses rich clinical experience and medical knowledge to provide deep background support for critical pathological diagnosis, ensuring the accuracy and extensiveness of the diagnostic model, and helping the system to make more scientific inferences based on a large amount of medical knowledge when facing complex critical illnesses. The introduction of the knowledge graph also enhances the system's ability to manage and update knowledge, enabling it to cope with ever-changing medical information. Through the Bayesian algorithm, the system can build a probabilistic reasoning model based on critical pathological diagnosis data and abnormal conditions, which can handle uncertainty and perform intelligent reasoning based on the influence of multiple factors, thereby providing a more accurate probability assessment for pathological diagnosis. The use of the Bayesian model enables the system to dynamically update the diagnosis results. With the addition of new data, the diagnostic ability of the model will gradually improve, improving the accuracy and real-time nature of the diagnosis. Multi-source critical abnormal condition data is input into the probabilistic reasoning model, and accurate pathological diagnosis reasoning is performed based on the current patient's status, so that the system can quickly obtain probabilistic diagnosis results and help medical staff identify potential pathological changes. This intelligent reasoning not only reduces the workload of manual analysis, but also improves the speed and accuracy of diagnosis, especially in critical conditions, and can provide reliable support for rapid diagnosis and response to the disease. Reinforcement learning can optimize the mapping relationship between diagnosis and decision-making in clinical practice through continuous iteration and learning, and learn how to make the best clinical decisions under different pathological conditions based on the medical knowledge graph, thereby generating a more accurate decision prediction model. The probability data of intelligent assisted pathological diagnosis is transmitted to the clinical decision prediction model, and the most appropriate clinical treatment plan is inferred based on the diagnosis results, which helps medical staff make the best decision when facing complex and critical conditions, reduces decision-making pressure, and improves the scientificity and effectiveness of decision-making. Feeding back the diagnosis and decision data to the terminal can directly use the reasoning results in clinical practice, help medical staff take quick action, ensure the seamless connection between auxiliary diagnostic information and medical decision-making, improve the feasibility and efficiency of decision-making, and reduce the delay in information transmission. The final feedback operation provides closed-loop support for clinical decision-making, ensuring that the diagnosis and treatment process of critically ill patients is efficient and accurate.

[0119] In an embodiment of the present invention, the critical pathological diagnosis data of historical patients are obtained from the medical database, including historical specific pathological diagnosis records, such as arrhythmia, respiratory failure, acute lung injury, etc., and related physiological parameter records. In addition, a medical knowledge map is also obtained, which includes medical research, expert decision-making, treatment plans and clinical guidelines. For example, the knowledge map includes standard treatment methods for arrhythmia, the physiological basis of myocardial injury, etc. The Bayesian algorithm is applied to establish a mapping relationship between critical abnormal conditions and pathological diagnosis results. For example, the probability relationship between heart rate fluctuations and arrhythmias in electrocardiogram signals is calculated, or the probability of correlation between decreased blood oxygen levels and respiratory failure is calculated, and model parameter training processing should be performed on critical pathological diagnosis data. The Bayesian algorithm can handle uncertainty, and by combining the current abnormal condition data with the historical diagnosis data, the probability of a specific pathological state occurring in the current patient is inferred to construct an intelligent critical pathological diagnosis probability reasoning model, which can perform intelligent reasoning based on the input abnormal condition data and give a specific pathological diagnosis probability. The patient's real-time multi-source critical abnormal condition data (such as abnormal fluctuations in electrocardiogram, blood oxygen, respiratory rate, etc.) is input into the intelligent reasoning model constructed earlier. Through intelligent reasoning processing, the probability of the patient having various pathological conditions is calculated. According to these probability inference results, critical pathology diagnosis-probability data is generated for reference by medical staff. Through reinforcement learning technology, combined with critical medicine knowledge graph, a mapping relationship between pathological diagnosis information and clinical treatment plans is established. For example, according to natural language processing technology and critical medicine knowledge graph, the decision model is gradually optimized to learn how to choose the best treatment strategy under different pathological conditions to build an intelligent critical clinical decision prediction model, which can predict the best treatment plan based on the diagnosis results. The critical pathology diagnosis-probability data is input into the clinical decision prediction model. Based on the current patient's pathological condition, the clinical decision prediction model infers the optimal treatment plan through intelligent prediction processing and generates auxiliary critical pathology clinical decision data. These data will be provided to medical staff to help them make fast and accurate treatment decisions. Finally, the system feeds back critical pathology diagnosis-probability data and auxiliary critical pathology clinical decision-making data to the doctor's or nurse's terminal device (such as electronic medical record system, mobile medical equipment, etc.), so that medical staff can obtain diagnosis and decision-making suggestions in real time. Through real-time feedback, the entire medical decision-making closed-loop support is completed, ensuring that medical staff can make the best treatment decisions based on intelligent reasoning and prediction results.

[0120] The beneficial effects of the present application are that the intelligent monitoring and analysis system of multi-source critical illness parameters of the present invention analyzes the patient's physical signs through a multi-source critical illness parameter acquisition module, intelligently adjusts the monitoring time series frequency of multi-source critical illness parameters, collects multi-source critical illness parameters in real time according to the patient's personalized critical illness condition, and can accurately obtain the potential information of multi-source critical illness parameters, and analyzes the time-frequency monitoring characteristics of multi-source critical illness parameters through a multi-source critical illness parameter noise reduction module to set adaptive filtering parameters, adaptively adjusts the filtering parameters for the time-frequency monitoring differences of multi-source data, removes noise and retains the effective characteristics of the data, so that the analysis efficiency and accuracy of multi-source critical illness parameters are excellent. Through the multi-source critical illness parameter feature analysis module, complex physiological parameters are decomposed in the time-frequency domain, and the characteristic changes of data are revealed from different time scales to extract the key information of multi-source critical illness data. The nonlinear multi-source critical illness correlation trend feature data of multi-source critical illness parameters is analyzed through the multi-source critical illness correlation trend feature analysis module to reveal the abnormal conditions of multi-source critical illness, and the probability of each critical illness pathology and the corresponding clinical decision are intelligently analyzed through the critical illness pathology auxiliary medical decision feedback module to provide intelligent support for auxiliary medical decision.

[0121] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0122] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. An intelligent monitoring and analysis system for multi-source critical illness parameters, characterized in that: Includes the following modules: The multi-source critical illness parameter acquisition module is used to perform intelligent adjustment processing on the dynamic perception monitoring parameters of the medical sensor equipment to obtain the intelligently adjusted medical sensor equipment; the acquisition parameters of the preliminary multi-source critical illness parameters are updated and optimized according to the intelligently adjusted medical sensor equipment to generate the multi-source critical illness parameters; The multi-source critical illness parameter denoising module is used to perform time-series frequency monitoring feature analysis processing based on the multi-source critical illness parameters to generate time-frequency monitoring and analysis multi-source critical illness parameters; based on the time-frequency monitoring and analysis multi-source critical illness parameters, the time-frequency difference adaptive filtering denoising processing of the multi-source critical illness parameters is performed to generate denoised multi-source critical illness parameters; Multi-source critical illness parameter feature analysis module, used for extracting multi-source critical illness parameter features based on denoised multi-source critical illness parameters for multi-scale entropy evaluation, and generating multi-source critical illness parameter feature data; A multi-source critical illness correlation trend characteristic analysis module is used to perform multi-source critical illness parameter correlation analysis based on multi-source critical illness parameter characteristic data to generate multi-source critical illness parameter correlation data; perform non-linear multi-source critical illness correlation trend characteristic analysis based on multi-source critical illness parameter correlation data and multi-source critical illness parameter characteristic data to generate non-linear multi-source critical illness correlation trend characteristic data; The multi-source critical illness abnormal condition analysis module is used to obtain historical multi-source critical illness routine diagnosis data; perform critical illness diagnosis health pattern modeling processing based on historical multi-source critical illness routine diagnosis data to generate a critical illness diagnosis health pattern model; perform multi-source critical illness abnormal condition analysis on the nonlinear multi-source critical illness correlation trend feature data based on the critical illness diagnosis health pattern model to generate multi-source critical illness abnormal condition data; The critical illness pathology assisted medical decision-making feedback module is used to perform critical illness pathology diagnosis-probability intelligent reasoning processing based on multi-source critical illness abnormal condition data to generate critical illness pathology diagnosis-probability data; perform assisted critical illness pathology clinical decision-making intelligent prediction processing based on critical illness pathology diagnosis-probability data to generate assisted critical illness pathology clinical decision-making data; and feed back the critical illness pathology diagnosis-probability data and assisted critical illness pathology clinical decision-making data to the terminal to execute the critical illness pathology assisted medical decision-making feedback operation.

2. The intelligent monitoring and analysis system for multi-source critical illness parameters according to claim 1 is characterized in that: The multi-source critical care parameter acquisition module includes the following functions: Perform preliminary collection of multi-source critical illness parameters according to medical sensor equipment to obtain preliminary multi-source critical illness parameters; Based on the preliminary multi-source critical illness parameters, the dynamic perception monitoring parameters of medical sensors are analyzed to generate dynamic perception monitoring parameters; Performing intelligent adjustment processing of dynamic sensing monitoring parameters on the medical sensor device according to the dynamic sensing monitoring parameters to obtain an intelligently adjusted medical sensor device; The intelligently adjusted medical sensor equipment collects and updates the preliminary multi-source critical care parameters to generate multi-source critical care parameters.

3. The intelligent monitoring and analysis system for multi-source critical illness parameters according to claim 2 is characterized in that: The dynamic perception monitoring parameter analysis of medical sensors based on preliminary multi-source critical illness parameters includes: Perform heterogeneous characteristic analysis of multi-source critical illness parameters based on preliminary multi-source critical illness parameters to generate heterogeneous characteristic data of multi-source critical illness parameters; Conduct self-organizing network collaboration analysis of monitoring nodes based on heterogeneous characteristic data of multi-source critical illness parameters to generate self-organizing network collaboration data of monitoring nodes; Perform condition-driven adaptive monitoring drive frequency analysis based on preliminary multi-source critical illness parameters to generate adaptive monitoring drive frequency data; According to the monitoring node self-organizing network collaboration data and the adaptive monitoring driving frequency data, the dynamic perception monitoring parameters of the medical sensor are analyzed to generate dynamic perception monitoring parameters.

4. The intelligent monitoring and analysis system for multi-source critical illness parameters according to claim 1 is characterized in that: The multi-source critical parameter noise reduction module includes the following functions: Perform time-series synchronization and alignment processing on multi-source critical illness parameters to generate synchronized multi-source critical illness parameters; Perform parameter monitoring time series frequency characteristic analysis and processing based on synchronous multi-source critical illness parameters to generate parameter monitoring time series frequency characteristic data; According to the parameter monitoring time series frequency characteristic data, the synchronous multi-source critical illness parameters are analyzed and processed with time-frequency monitoring characteristics to generate time-frequency monitoring and analysis multi-source critical illness parameters; Based on time-frequency monitoring and analysis of multi-source critical illness parameters, time-frequency difference adaptive filtering and denoising of multi-source critical illness parameters are performed to generate denoised multi-source critical illness parameters.

5. The intelligent monitoring and analysis system for multi-source critical illness parameters according to claim 4 is characterized in that: The time-frequency difference adaptive filtering and denoising processing of multi-source critical illness parameters based on time-frequency monitoring and analysis includes the following functions: Based on time-frequency monitoring and analysis of multi-source critical illness parameters, time-frequency difference noise analysis of multi-source critical illness parameters is performed to generate time-frequency difference noise data of multi-source critical illness parameters; According to the multi-source critical illness parameter time-frequency difference noise data, the adaptive filtering parameter analysis of time-frequency monitoring difference is performed to generate the time-frequency difference adaptive filtering parameters; According to the time-frequency difference adaptive filtering parameters, multi-source critical illness parameters are denoised and reconstructed for time-frequency monitoring and analysis to generate denoised multi-source critical illness parameters.

6. The intelligent monitoring and analysis system for multi-source critical illness parameters according to claim 5 is characterized in that: The time-frequency difference noise analysis of multi-source critical illness parameters based on time-frequency monitoring and parsing of multi-source critical illness parameters includes: Based on the empirical mode decomposition technology, the intrinsic mode function of the time-frequency monitoring difference of the multi-source critical parameters is analyzed to generate the time-frequency difference intrinsic mode function; Calculating the energy density of the time-frequency difference intrinsic mode function according to the time-frequency difference intrinsic mode function to generate energy density data of the time-frequency difference intrinsic mode function; Based on the energy density data of time-frequency difference intrinsic mode function, time-frequency difference noise analysis of multi-source critical illness parameters is performed to generate time-frequency difference noise data of multi-source critical illness parameters.

7. The intelligent monitoring and analysis system for multi-source critical illness parameters according to claim 1 is characterized in that: The multi-source critical illness parameter feature analysis module includes the following functions: Perform parameter multi-scale time-frequency discretization processing on the denoised multi-source critical illness parameters to generate multi-scale time-frequency discretized multi-source critical illness parameters; Based on the multi-scale time-frequency discretization of multi-source critical illness parameters, the entropy value of multi-scale multi-source critical illness parameters is evaluated and processed to generate multi-scale multi-source critical illness parameter entropy data; According to the multi-scale multi-source critical illness parameter entropy data, multi-source critical illness parameter features are extracted from the denoised multi-source critical illness parameters to generate multi-source critical illness parameter feature data.

8. The intelligent monitoring and analysis system for multi-source critical illness parameters according to claim 1, characterized in that: The multi-source severe disease correlation trend feature analysis module includes the following functions: Perform multi-source critical illness parameter correlation analysis based on multi-source critical illness parameter feature data to generate multi-source critical illness parameter correlation data; The multi-source critical illness parameter feature data is used as a multi-source critical illness feature association map node, and the multi-source critical illness parameter correlation data is used as a multi-source critical illness feature association map edge, and a multi-source critical illness feature association map is established according to the multi-source critical illness feature association map nodes and the corresponding multi-source critical illness feature association map edges; Perform graph temporal dependency embedding processing on the multi-source critical illness feature association graph to generate a time-series multi-source critical illness feature association graph; According to the time series multi-source severe disease characteristic association map, nonlinear multi-source severe disease association trend characteristic analysis is performed to generate nonlinear multi-source severe disease association trend characteristic data.

9. The intelligent monitoring and analysis system for multi-source critical illness parameters according to claim 1, characterized in that: The multi-source critical abnormal condition analysis module includes the following functions: Obtain historical multi-source critical care routine diagnosis data; Based on the historical multi-source critical illness routine diagnosis data, critical illness diagnosis health pattern modeling is performed to generate a critical illness diagnosis health pattern model; According to the critical illness diagnosis health pattern model, the nonlinear multi-source critical illness correlation trend feature data is analyzed for multi-source critical illness health pattern deviation characteristics to generate multi-source critical illness health pattern deviation feature data; Perform multi-source severe illness abnormal distribution detection and processing based on multi-source severe illness health pattern deviation feature data to generate multi-source severe illness abnormal distribution data; Multi-source severe abnormal condition analysis is performed based on multi-source severe abnormal condition distribution data to generate multi-source severe abnormal condition data.

10. The intelligent monitoring and analysis system for multi-source critical illness parameters according to claim 1, characterized in that: The critical care pathology auxiliary medical decision-making feedback module includes the following functions: Obtain critical care pathology diagnosis data and critical care medicine knowledge graph data; Based on the Bayesian algorithm and critical pathology diagnosis data, a mapping relationship between critical abnormal conditions and pathology diagnosis probability reasoning is established to generate an intelligent critical pathology diagnosis probability reasoning model; Transmit multi-source critical illness abnormal condition data to the intelligent critical illness pathology diagnosis probability reasoning model for critical illness pathology diagnosis-probability intelligent reasoning processing to generate critical illness pathology diagnosis-probability data; Based on reinforcement learning and critical care medicine knowledge graph data, the mapping relationship between critical care pathology diagnosis information and clinical medical decision-making is carried out to obtain an intelligent critical care clinical decision-making prediction model; The intelligent assisted critical pathology diagnosis-probability data is transmitted to the intelligent critical pathology clinical decision prediction model for assisted critical pathology clinical decision intelligent prediction processing to generate assisted critical pathology clinical decision data; Feedback the critical pathology diagnosis-probability data and the critical pathology auxiliary clinical decision-making data to the terminal to execute the critical pathology auxiliary medical decision-making feedback task.

Images