Trend prediction method and system based on multi-physiological signal fusion, terminal and storage medium

Abstract

The invention relates to the field of biomedical engineering technology and information processing, and discloses a trend prediction method and system based on multi-physiological signal fusion, a terminal and a storage medium, and the method comprises the steps: obtaining a venous pressure analog signal collected by a sensor or a radar, and carrying out the automatic calibration and digital processing of the venous pressure analog signal, obtaining central venous pressure data; processing the central venous pressure data based on a waveform feature automatic deconstruction algorithm to obtain hemodynamic feature parameters; and predicting the hemodynamic characteristic parameters according to a preset clinical knowledge rule to obtain clinical decision prompt information. According to the central venous pressure monitoring system and method, intelligent analysis and fusion prediction are conducted on the continuously-collected venous pressure signals, hierarchical clinical decision prompt information is obtained, and the automation, continuity and clinical decision support capacity of central venous pressure monitoring are greatly improved.

Description

Technical Field

[0001] This invention relates to the fields of biomedical engineering technology and information processing, and in particular to a trend prediction method, system, terminal, and computer-readable storage medium based on the fusion of multiple physiological signals. Background Technology

[0002] Currently, the widely used manual or semi-automatic water column method and single-point pressure sensor measurement method in clinical practice have the following inherent drawbacks: Intermittent: They only provide values ​​at discrete time points and cannot capture rapidly changing hemodynamic states. Cumbersome and error-prone: They heavily rely on manual zeroing (ensuring the transducer is level with the right atrium) and readings, introducing subjective errors and infection risks. Isolated information: CVP values ​​are disconnected from synchronous contexts such as heart rate (HR), invasive arterial pressure (IBP), and respiration, making interpretation highly dependent on clinical experience. Lack of predictive ability: They cannot predict in advance a patient's potential response to treatments such as fluid resuscitation or diuresis.

[0003] Existing methods for monitoring central venous pressure (CVP) include manual measurement based on the principle of water column pressure, single-point measurement based on external pressure sensors, and methods based on waveform display on traditional monitors. These methods are limited by discontinuous measurement processes, heavy reliance on manual operation and interpretation, lack of multi-physiological signal collaborative analysis capabilities, and insufficient real-time predictive functions. They struggle to achieve continuous, automated monitoring and intelligent early warning of CVP, resulting in a reliance on empirical and lagging numerical judgments in the hemodynamic management of critically ill patients, hindering prospective assessment and intervention guidance of volume status and cardiac function changes. Therefore, existing CVP monitoring and assessment technologies require further improvement and optimization. Summary of the Invention

[0004] The main objective of this invention is to provide a trend prediction method, system, terminal, and computer-readable storage medium based on the fusion of multiple physiological signals, aiming to solve the problem that it is difficult to achieve continuous, automated monitoring and intelligent early warning of CVP using existing central venous pressure (CVP) monitoring technology.

[0005] To achieve the above objectives, the present invention provides a trend prediction method based on the fusion of multiple physiological signals, the trend prediction method based on the fusion of multiple physiological signals comprising the following steps: Acquire simulated venous pressure signals from sensors or radar, and automatically calibrate and digitize the simulated venous pressure signals to obtain central venous pressure data; The central venous pressure data is processed using an automatic waveform feature deconstruction algorithm to obtain hemodynamic characteristic parameters; Based on preset clinical knowledge rules, the hemodynamic characteristic parameters are predicted to obtain clinical decision-making prompts.

[0006] Optionally, the trend prediction method based on multi-physiological signal fusion, wherein acquiring the simulated venous pressure signal collected by a sensor or radar, and automatically calibrating and digitizing the simulated venous pressure signal to obtain central venous pressure data, specifically includes: Collect the simulated venous pressure signal acquired by the millimeter-wave radar array of the sensor or radar, and perform noise reduction processing on the simulated venous pressure signal to obtain the processed simulated venous pressure signal; Based on a MEMS pressure chip, the processed venous pressure analog signal is digitally processed to obtain central venous pressure data. The noise reduction process includes jitter reduction and debubbling reduction.

[0007] Optionally, in the trend prediction method based on multi-physiological signal fusion, the process of processing the central venous pressure data using an automatic waveform feature deconstruction algorithm to obtain hemodynamic characteristic parameters specifically includes: The central venous pressure data is automatically identified based on the waveform feature automatic deconstruction algorithm to obtain the atrial systolic amplitude, closure amplitude, and late systolic amplitude; Acquire arterial pressure, electrocardiogram, and respiratory waveform. Evaluate the arterial pressure, electrocardiogram, respiratory waveform, atrial contraction amplitude, closure amplitude, and late systolic amplitude using a sensing algorithm to obtain hemodynamic characteristic parameters. The hemodynamic characteristic parameters include waveform characteristics, respiratory variability characteristics, and time series variation characteristics.

[0008] Optionally, the trend prediction method based on multi-physiological signal fusion, wherein the automatic identification of the central venous pressure data based on the waveform feature automatic deconstruction algorithm to obtain the atrial systolic amplitude, closure amplitude, and post-systolic amplitude specifically includes: The central venous pressure data is automatically identified based on a waveform feature automatic deconstruction algorithm to obtain multiple target waveforms; The automatic waveform feature deconstruction algorithm is used to analyze multiple target waveforms to obtain the atrial contraction amplitude, closure amplitude, and late contraction amplitude.

[0009] Optionally, the trend prediction method based on multi-physiological signal fusion, wherein the step of evaluating the atrial contraction amplitude, the closure amplitude, and the late-systolic amplitude using a sensing algorithm to obtain hemodynamic characteristic parameters, further includes: The target values ​​are obtained by analyzing the hemodynamic characteristic parameters based on the time series using an LSTM neural network. If the target value reaches the preset value, the hemodynamic characteristic parameters are inferred through the physiological model to obtain prediction prompt information.

[0010] Optionally, in the trend prediction method based on multi-physiological signal fusion, the preset clinical knowledge rules include hemodynamic knowledge graph rules; The step of identifying the hemodynamic characteristic parameters according to preset clinical knowledge rules to obtain clinical decision-making prompts specifically includes: The hemodynamic feature parameters are identified according to the hemodynamic knowledge graph rules to obtain the waveform area; Obtain heart rate and CVP values, and analyze the waveform area based on the heart rate and CVP values ​​to obtain patient characteristics; The patient characteristics are predicted using a physiological model to obtain clinical decision-making suggestions.

[0011] Optionally, the trend prediction method based on multi-physiological signal fusion, wherein predicting patient characteristics using a physiological model to obtain clinical decision-making prompts specifically includes: Acquire CVP change pattern data, input the CVP change pattern data into the physiological model for training, and obtain the target physiological model; The patient's characteristics are predicted using the target physiological model to obtain clinical decision-making suggestions.

[0012] Furthermore, to achieve the above objectives, the present invention also provides a trend prediction system based on the fusion of multiple physiological signals, wherein the trend prediction system based on the fusion of multiple physiological signals includes: The venous pressure analog signal acquisition module is used to acquire venous pressure analog signals collected by sensors or radar, and to automatically calibrate and digitize the venous pressure analog signals to obtain central venous pressure data; The feature parameter extraction module is used to process the central venous pressure data based on the waveform feature automatic deconstruction algorithm to obtain hemodynamic feature parameters; The data prediction module is used to predict the hemodynamic characteristic parameters based on preset clinical knowledge rules, and obtain clinical decision-making prompts.

[0013] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a trend prediction program based on multi-physiological signal fusion, and when the trend prediction program based on multi-physiological signal fusion is executed by a processor, it implements the steps of the trend prediction method based on multi-physiological signal fusion as described above.

[0014] In this invention, simulated venous pressure signals acquired by sensors or radar are automatically calibrated and digitized to obtain central venous pressure (CVP) data. The CVP data is then processed using an automatic waveform feature deconstruction algorithm to obtain hemodynamic characteristic parameters. These hemodynamic characteristic parameters are then predicted according to preset clinical knowledge rules to obtain clinical decision-making suggestions. This invention, through intelligent analysis and fusion prediction of continuously acquired venous pressure signals, obtains hierarchical clinical decision-making suggestions, significantly improving the automation, continuity, and clinical decision support capabilities of central venous pressure monitoring. Attached Figure Description

[0015] Figure 1 This is a flowchart of a preferred embodiment of the trend prediction method based on multi-physiological signal fusion of the present invention; Figure 2 This is a structural diagram of a preferred embodiment of the trend prediction system based on multi-physiological signal fusion of the present invention; Figure 3 This is a structural diagram of a preferred embodiment of the terminal of the device of the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0017] Existing clinically widely used manual or semi-automatic water column methods and single-point pressure sensor measurements have the following inherent drawbacks: Intermittency: They only provide values ​​at discrete time points, failing to capture rapidly changing hemodynamic states. Cumbersome and error-prone operation: They heavily rely on manual zeroing (ensuring the transducer is level with the right atrium) and readings, introducing subjective errors and infection risks. Information isolation: CVP values ​​are disconnected from synchronous contexts such as heart rate (HR), invasive arterial pressure (IBP), and respiration, making interpretation highly dependent on clinical experience. Lack of predictive ability: They cannot predict a patient's potential response to treatments such as fluid resuscitation or diuresis. Therefore, a trend prediction method based on multi-physiological signal fusion is needed. By intelligently analyzing and fusing continuously acquired venous pressure signals, hierarchical clinical decision-making prompts can be obtained, significantly improving the automation, continuity, and clinical decision support capabilities of central venous pressure monitoring.

[0018] The trend prediction method based on multi-physiological signal fusion described in the preferred embodiment of the present invention, such as... Figure 1 As shown, the trend prediction method based on multi-physiological signal fusion includes the following steps: Step S10: Acquire the venous pressure simulation signal collected by the sensor or radar, and automatically calibrate and digitize the venous pressure simulation signal to obtain central venous pressure data.

[0019] Step S10 includes: Step S11: Collect the venous pressure simulation signal acquired by the millimeter-wave radar array of the sensor or radar, and perform noise reduction processing on the venous pressure simulation signal to obtain the processed venous pressure simulation signal; Step S12: Based on the MEMS pressure chip, the processed venous pressure analog signal is digitally processed to obtain central venous pressure data.

[0020] Specifically, the process involves collecting simulated venous pressure signals from millimeter-wave radar arrays of sensors or radars, denoising these simulated venous pressure signals to obtain processed simulated venous pressure signals (generating high-fidelity digital signals directly at the pressure source within the blood vessel, discarding fluid-filled tubing, and eliminating interference from tubing vibrations, bubbles, etc.). Based on a MEMS pressure chip, the processed simulated venous pressure signals are then digitized to obtain central venous pressure data (integrating a miniature fiber optic pressure sensor or a MEMS (microelectromechanical system) pressure chip). The denoising process includes vibration reduction and defoaming.

[0021] As an example, continuous closed-loop monitoring, with sensors operating continuously, achieves true continuous waveform output. This continuous waveform specifically includes a real-time, uninterrupted central venous pressure-time curve. It not only includes the characteristic a, c, and v waves generated with the cardiac cycle, but also fully records the periodic fluctuations caused by respiration (changes in intrathoracic pressure), as well as any instantaneous waveform morphology and pressure baseline changes caused by arrhythmias (such as the disappearance of a wave in atrial fibrillation), cardiac tamponade (waveforms appearing "flat-topped" or "downward-sloping"), or rapid changes in volume status. Compared to single-point values, continuous waveforms provide information on the trend, rate, and rhythm of pressure changes.

[0022] Step S20: Process the central venous pressure data based on the waveform feature automatic deconstruction algorithm to obtain hemodynamic characteristic parameters.

[0023] Step S20 includes: Step S21: The central venous pressure data is automatically identified based on the waveform feature automatic deconstruction algorithm to obtain the atrial contraction amplitude, closure amplitude, and late systolic amplitude; Step S22: Obtain the arterial pressure map, electrocardiogram, and respiratory waveform. Evaluate the arterial pressure map, electrocardiogram, respiratory waveform, atrial contraction amplitude, closure amplitude, and late systolic amplitude according to the sensing algorithm to obtain hemodynamic characteristic parameters.

[0024] Specifically, step S20 further includes analyzing hemodynamic characteristic parameters based on time series data using an LSTM neural network to obtain target values. If the target value reaches a preset value, a physiological model is used to infer the hemodynamic characteristic parameters and obtain predictive prompts (predictive warning: before the absolute value of CVP reaches the alarm threshold, a "trend warning" is issued in advance based on its rate of increase / decrease and acceleration (the target value includes rate and acceleration); volume therapy simulation: based on the current hemodynamic state, a predictive prompt is given based on the physiological model to provide "the possible range of CVP changes if 250ml of fluid is rapidly administered"). The central venous pressure data is automatically identified based on a waveform feature automatic deconstruction algorithm. The system obtains the atrial contraction amplitude, closure amplitude, and post-systolic amplitude (automatically identifying and quantifying the duration, area, and ratio of the a wave (atrial contraction amplitude), c wave (tricuspid valve closure, i.e., closure amplitude), and v wave (ventricular post-systolic amplitude) in the CVP waveform), acquires invasive arterial pressure, electrocardiogram, and respiratory waveform (real-time synchronous integration of invasive arterial pressure, electrocardiogram, and respiratory waveform (from ventilator or impedance)), and evaluates the invasive arterial pressure, electrocardiogram, respiratory waveform, atrial contraction amplitude, closure amplitude, and post-systolic amplitude according to the sensing algorithm to obtain hemodynamic characteristic parameters, including waveform characteristics, respiratory variability characteristics, and time series change characteristics.

[0025] For example, the system integrates multi-source signals from the patient monitoring network in real time, including: invasive arterial pressure mapping (providing systolic pressure, diastolic pressure, pulse pressure, and waveform), electrocardiogram (providing heart rate, rhythm, P wave, and QRS duration), respiratory waveform (from ventilator pressure / flow curves or thoracic impedance methods, providing respiratory cycle, tidal volume, and inspiratory / expiratory phase markers), blood oxygen saturation (SpO2), and body temperature. For instance, it automatically calculates the respiratory variability of CVP and arterial pulse pressure to directly assess volume responsiveness; it identifies characteristic changes in CVP with the respiratory cycle to help determine the interaction between spontaneous breathing and mechanical ventilation.

[0026] In this embodiment, the predictive prompts are generated by the system based on input treatment parameters (such as fluid volume, rate, and diuretic dosage) and the current hemodynamic status, using built-in physiological models (such as the Guyton venous return and cardiac function curve model, and a digital approximation of Starling's law) for simulation calculations. Output results may include: 1) Numerical prediction: e.g., "CVP is expected to rise by 3-5 mmHg within 10 minutes and stabilize in the 8-10 mmHg range within 30 minutes." 2) Risk / benefit prompts: e.g., "According to the model, the current position is in the flat segment of the Starling curve; fluid resuscitation is expected to have a limited increase in cardiac output (<10%), and there is a risk of volume overload (predicted CVP > 12 mmHg)." 3) Comparison of suggested treatment plans: e.g., "Compared to rapid fluid resuscitation, slow fluid resuscitation (100 ml / h) is expected to result in a more gradual rise in CVP and a lower risk of volume overload."

[0027] Step S30: Predict the hemodynamic characteristic parameters according to preset clinical knowledge rules to obtain clinical decision prompts.

[0028] Step S30 includes: Step S31: Identify the hemodynamic feature parameters according to the hemodynamic knowledge graph rules to obtain the waveform area; Step S32: Obtain heart rate and CVP values, and analyze the waveform area based on the heart rate and CVP values ​​to obtain patient characteristics; Step S33: Predict the patient characteristics using a physiological model to obtain clinical decision-making suggestions.

[0029] Specifically, the hemodynamic feature parameters are identified according to the hemodynamic knowledge graph rules (built-in structured hemodynamic knowledge graph rules) to obtain the waveform area, heart rate value and CVP value, and the waveform area is analyzed according to the heart rate value and CVP value to obtain patient characteristics. The patient characteristics are then predicted through a physiological model to obtain clinical decision-making prompts.

[0030] In this embodiment, for example: waveform characteristics: v / a wave area ratio = 2.5 (normal <1.5), CVP respiratory variability (ΔCVP) = 8%. Fusion analysis results: volume responsiveness prediction: positive (based on consistency between PPV and ΔCVP), right ventricular function assessment: moderate tricuspid regurgitation or right ventricular diastolic dysfunction may be present. Trend prediction: CVP upward trend slope over the past 1 hour: +1.5 mmHg / h, predicting that it may reach the 12 mmHg threshold after 2 hours.

[0031] As an example, the clinical decision-making process is as follows: A significant V wave was detected (the area of ​​the V wave is significantly larger than that of the A wave). Combined with the patient's atrial fibrillation rhythm (no effective atrial contraction, disappearance of the A wave) and the current CVP value of 10 mmHg, the high V wave height strongly suggests functional tricuspid regurgitation or decreased right ventricular compliance. Bedside echocardiography is performed to assess tricuspid and right ventricular function. The current volume responsiveness assessment is positive, but considering the significant V wave and the already elevated baseline CVP, rapid fluid resuscitation should be approached with caution. A small-volume rapid loading test (e.g., 100 ml of crystalloid solution infused over 10 minutes) is recommended, with close monitoring of the response. Optimization of positive end-expiratory pressure (PEEP) should be considered to reduce right ventricular afterload.

[0032] Step S33 includes: Step S331: Obtain CVP change pattern data, input the CVP change pattern data into the physiological model for training, and obtain the target physiological model; Step S332: Predict the patient characteristics using the target physiological model to obtain clinical decision-making prompts.

[0033] Specifically, CVP change pattern data is acquired, and the CVP change pattern data is input into a physiological model for training to obtain a target physiological model. The target physiological model is then used to predict the patient's characteristics to obtain clinical decision-making prompts.

[0034] In this embodiment, the device consists of: a disposable intelligent central venous catheter: the proximal end of the catheter integrates a miniature pressure sensor, a microchip, and a wireless transmitter, all in a sterile package. A bedside / wearable host unit: includes a wireless receiver, a high-performance processor, an AI analysis engine, and a touchscreen display. An adaptive calibration module: the host unit has a built-in algorithm that automatically completes calibration based on the initial pressure and orientation data provided by the sensors. The doctor inserts the intelligent CVC catheter according to standard procedures. After the catheter contacts the blood, the sensors are activated and wirelessly paired with the host unit. The host unit automatically completes the initial calibration and begins continuously displaying real-time CVP waveforms, values, and trend curves. When the AI ​​engine identifies abnormal waveforms or predictive risks, the interface highlights a warning and displays a graded clinical prompt.

[0035] Furthermore, such as Figure 2 As shown, based on the above-mentioned trend prediction method based on multi-physiological signal fusion, the present invention also provides a trend prediction system based on multi-physiological signal fusion, wherein the trend prediction system based on multi-physiological signal fusion includes: The venous pressure analog signal acquisition module 51 is used to acquire the venous pressure analog signal acquired by the sensor or radar, and to automatically calibrate and digitize the venous pressure analog signal to obtain central venous pressure data. The feature parameter extraction module 52 is used to process the central venous pressure data based on the waveform feature automatic deconstruction algorithm to obtain hemodynamic feature parameters; The data prediction module 53 is used to predict the hemodynamic characteristic parameters according to preset clinical knowledge rules, and obtain clinical decision prompt information.

[0036] Furthermore, such as Figure 3 As shown, based on the above-mentioned trend prediction method and system based on multi-physiological signal fusion, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 3 Only some of the terminal components are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0037] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory. In other embodiments, the memory 20 may be an external storage device of the terminal, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc. Further, the memory 20 may include both internal and external storage devices. The memory 20 is used to store application software and various types of data installed on the terminal, such as program code installed on the terminal. The memory 20 may also be used to temporarily store data that has been output or will be output. In one embodiment, the memory 20 stores a trend prediction program 40 based on multi-physiological signal fusion, which can be executed by the processor 10 to implement the trend prediction method based on multi-physiological signal fusion in this application.

[0038] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in the memory 20 or process data, such as executing the trend prediction method based on multi-physiological signal fusion.

[0039] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface. The terminals communicate with each other via a system bus.

[0040] In one embodiment, when the processor 10 executes the trend prediction program 40 based on multi-physiological signal fusion in the memory 20, the following steps are performed: Acquire simulated venous pressure signals from sensors or radar, and automatically calibrate and digitize the simulated venous pressure signals to obtain central venous pressure data; The central venous pressure data is processed using an automatic waveform feature deconstruction algorithm to obtain hemodynamic characteristic parameters; Based on preset clinical knowledge rules, the hemodynamic characteristic parameters are predicted to obtain clinical decision-making prompts.

[0041] Specifically, the acquisition of simulated venous pressure signals from sensors or radar, followed by automatic calibration and digitization of these signals to obtain central venous pressure data, includes: Collect the simulated venous pressure signal acquired by the millimeter-wave radar array of the sensor or radar, and perform noise reduction processing on the simulated venous pressure signal to obtain the processed simulated venous pressure signal; Based on a MEMS pressure chip, the processed venous pressure analog signal is digitally processed to obtain central venous pressure data. The noise reduction process includes jitter reduction and debubbling reduction.

[0042] Specifically, the automatic waveform feature deconstruction algorithm is used to process the central venous pressure data to obtain hemodynamic characteristic parameters, including: The central venous pressure data is automatically identified based on the waveform feature automatic deconstruction algorithm to obtain the atrial systolic amplitude, closure amplitude, and late systolic amplitude; Acquire arterial pressure, electrocardiogram, and respiratory waveform. Evaluate the arterial pressure, electrocardiogram, respiratory waveform, atrial contraction amplitude, closure amplitude, and late systolic amplitude using a sensing algorithm to obtain hemodynamic characteristic parameters. The hemodynamic characteristic parameters include waveform characteristics, respiratory variability characteristics, and time series variation characteristics.

[0043] Specifically, the automatic waveform feature-based deconstruction algorithm automatically identifies the central venous pressure data to obtain the atrial systolic amplitude, closure amplitude, and post-systolic amplitude, including: The central venous pressure data is automatically identified based on a waveform feature automatic deconstruction algorithm to obtain multiple target waveforms; The automatic waveform feature deconstruction algorithm is used to analyze multiple target waveforms to obtain the atrial contraction amplitude, closure amplitude, and late contraction amplitude.

[0044] The step of evaluating the atrial contraction amplitude, the closure amplitude, and the late systolic amplitude using a sensing algorithm to obtain hemodynamic characteristic parameters further includes: The target values ​​are obtained by analyzing the hemodynamic characteristic parameters based on the time series using an LSTM neural network. If the target value reaches the preset value, the hemodynamic characteristic parameters are inferred through the physiological model to obtain prediction prompt information.

[0045] The preset clinical knowledge rules include hemodynamic knowledge graph rules; The step of identifying the hemodynamic characteristic parameters according to preset clinical knowledge rules to obtain clinical decision-making prompts specifically includes: The hemodynamic feature parameters are identified according to the hemodynamic knowledge graph rules to obtain the waveform area; Obtain heart rate and CVP values, and analyze the waveform area based on the heart rate and CVP values ​​to obtain patient characteristics; The patient characteristics are predicted using a physiological model to obtain clinical decision-making suggestions.

[0046] Specifically, the step of predicting patient characteristics using a physiological model to obtain clinical decision-making suggestions includes: Acquire CVP change pattern data, input the CVP change pattern data into the physiological model for training, and obtain the target physiological model; The patient's characteristics are predicted using the target physiological model to obtain clinical decision-making suggestions.

[0047] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a trend prediction program based on multi-physiological signal fusion, and the trend prediction program based on multi-physiological signal fusion, when executed by a processor, implements the steps of the trend prediction method based on multi-physiological signal fusion as described above.

[0048] In summary, this invention provides a trend prediction method, system, terminal, and storage medium based on multi-physiological signal fusion. The method includes: acquiring simulated venous pressure signals collected by sensors or radar; automatically calibrating and digitizing the simulated venous pressure signals to obtain central venous pressure data; processing the central venous pressure data based on an automatic waveform feature deconstruction algorithm to obtain hemodynamic characteristic parameters; and predicting the hemodynamic characteristic parameters according to preset clinical knowledge rules to obtain clinical decision-making prompts. This invention, through intelligent analysis and fusion prediction of continuously acquired venous pressure signals, obtains hierarchical clinical decision-making prompts, greatly improving the automation, continuity, and clinical decision support capabilities of central venous pressure monitoring.

[0049] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal system that includes that element.

[0050] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.). The program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The computer-readable storage medium can be a memory, magnetic disk, optical disk, etc.

[0051] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A trend prediction method based on the fusion of multiple physiological signals, characterized in that, The trend prediction method based on multi-physiological signal fusion includes: Acquire simulated venous pressure signals from sensors or radar, and automatically calibrate and digitize the simulated venous pressure signals to obtain central venous pressure data; The central venous pressure data is processed using an automatic waveform feature deconstruction algorithm to obtain hemodynamic characteristic parameters; Based on preset clinical knowledge rules, the hemodynamic characteristic parameters are predicted to obtain clinical decision-making prompts.

2. The trend prediction method based on multi-physiological signal fusion according to claim 1, characterized in that, The acquisition of simulated venous pressure signals from sensors or radar, followed by automatic calibration and digitization of these signals to obtain central venous pressure data, specifically includes: Collect the simulated venous pressure signal acquired by the millimeter-wave radar array of the sensor or radar, and perform noise reduction processing on the simulated venous pressure signal to obtain the processed simulated venous pressure signal; Based on a MEMS pressure chip, the processed venous pressure analog signal is digitally processed to obtain central venous pressure data. The noise reduction process includes jitter reduction and debubbling reduction.

3. The trend prediction method based on multi-physiological signal fusion according to claim 2, characterized in that, The waveform feature-based automatic deconstruction algorithm processes the central venous pressure data to obtain hemodynamic characteristic parameters, specifically including: The central venous pressure data is automatically identified based on the waveform feature automatic deconstruction algorithm to obtain the atrial systolic amplitude, closure amplitude, and late systolic amplitude; Acquire arterial pressure, electrocardiogram, and respiratory waveform. Evaluate the arterial pressure, electrocardiogram, respiratory waveform, atrial contraction amplitude, closure amplitude, and late systolic amplitude using a sensing algorithm to obtain hemodynamic characteristic parameters. The hemodynamic characteristic parameters include waveform characteristics, respiratory variability characteristics, and time series variation characteristics.

4. The trend prediction method based on multi-physiological signal fusion according to claim 3, characterized in that, The waveform feature-based automatic deconstruction algorithm automatically identifies the central venous pressure data to obtain the atrial systolic amplitude, closure amplitude, and post-systolic amplitude, specifically including: The central venous pressure data is automatically identified based on a waveform feature automatic deconstruction algorithm to obtain multiple target waveforms; The automatic waveform feature deconstruction algorithm is used to analyze multiple target waveforms to obtain the atrial contraction amplitude, closure amplitude, and late contraction amplitude.

5. The trend prediction method based on multi-physiological signal fusion according to claim 3, characterized in that, The step of evaluating the atrial contraction amplitude, the closure amplitude, and the late systolic amplitude using a sensing algorithm to obtain hemodynamic characteristic parameters further includes: The target values ​​are obtained by analyzing the hemodynamic characteristic parameters based on the time series using an LSTM neural network. If the target value reaches the preset value, the hemodynamic characteristic parameters are inferred through the physiological model to obtain prediction prompt information.

6. The trend prediction method based on multi-physiological signal fusion according to claim 5, characterized in that, The preset clinical knowledge rules include hemodynamic knowledge graph rules; The step of identifying the hemodynamic characteristic parameters according to preset clinical knowledge rules to obtain clinical decision-making prompts specifically includes: The hemodynamic feature parameters are identified according to the hemodynamic knowledge graph rules to obtain the waveform area; Obtain heart rate and CVP values, and analyze the waveform area based on the heart rate and CVP values ​​to obtain patient characteristics; The patient characteristics are predicted using a physiological model to obtain clinical decision-making suggestions.

7. The trend prediction method based on multi-physiological signal fusion according to claim 6, characterized in that, The process of predicting patient characteristics using a physiological model to obtain clinical decision-making suggestions specifically includes: Acquire CVP change pattern data, input the CVP change pattern data into the physiological model for training, and obtain the target physiological model; The patient's characteristics are predicted using the target physiological model to obtain clinical decision-making suggestions.

8. A trend prediction system based on multi-physiological signal fusion, characterized in that, The trend prediction system based on multi-physiological signal fusion includes: The venous pressure analog signal acquisition module is used to acquire venous pressure analog signals collected by sensors or radar, and to automatically calibrate and digitize the venous pressure analog signals to obtain central venous pressure data; The feature parameter extraction module is used to process the central venous pressure data based on the waveform feature automatic deconstruction algorithm to obtain hemodynamic feature parameters; The data prediction module is used to predict the hemodynamic characteristic parameters based on preset clinical knowledge rules, and obtain clinical decision-making prompts.

9. A terminal, characterized in that, The terminal includes: a memory, a processor, and a trend prediction program based on multi-physiological signal fusion stored in the memory and executable on the processor. When the trend prediction program based on multi-physiological signal fusion is executed by the processor, it implements the steps of the trend prediction method based on multi-physiological signal fusion as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a trend prediction program based on multi-physiological signal fusion, which, when executed by a processor, implements the steps of the trend prediction method based on multi-physiological signal fusion as described in any one of claims 1-7.

Images