Severe heatstroke gait pre-training method based on rising bed angle dynamic optimization

Abstract

The invention discloses a severe heatstroke gait pre-training method based on rising bed angle dynamic optimization, and relates to the technical field of rehabilitation medical instruments.The severe heatstroke gait pre-training method comprises the following steps that a patient is safely fixed to an electric rising bed and connected with multi-modal physiological signal monitoring equipment, and a patient examination result is obtained; the central processing unit receives the collected initial data and initializes setting of angle parameters of the standing bed; the electric standing bed is controlled to lift the tentative angle and maintain the tentative angle, and all physiological signal data are collected and recorded; in the standing bed angle maintaining process, physiological signal data flow is obtained, physiological data are input into the tolerance prediction model, and a tolerance index is obtained; according to the tolerance index value, the tolerance level of the patient is judged, and different bed rising instruction operations are executed; the real-time data acquisition process, the tolerance index calculation process and the angle adjustment decision making process are repeatedly executed until the end; after training is finished, a training report is generated, and all data are encrypted and stored in a database.

Description

Technical Field

[0001] This invention belongs to the field of rehabilitation medical device technology, and in particular relates to a method for gait pre-training for severe heatstroke patients with dynamic optimization of the angle of the standing bed. Background Technology

[0002] In the rehabilitation treatment of patients with severe heatstroke, gait pre-training is a crucial step. Standing beds, as commonly used rehabilitation equipment, are used to help patients gradually adapt to an upright posture. However, traditional standing bed training methods often employ a fixed angle adjustment mode, lacking individualized responses to the patient's real-time physiological state. This makes it difficult to effectively avoid risks such as blood pressure fluctuations and abnormal cerebral blood flow caused by changes in body position during training. Although some physiological monitoring devices have been introduced in existing technologies, the integration and analysis of multimodal physiological signals and laboratory data has not yet been achieved, making it impossible to establish a dynamic optimization model to achieve intelligent adjustment of the standing bed angle. This results in insufficient training safety and limited effectiveness. Therefore, there is a need in this field for a pre-training method that can dynamically adjust the standing bed angle based on the patient's real-time tolerance. We propose a gait pre-training method for severe heatstroke with dynamic optimization of the standing bed angle. Summary of the Invention

[0003] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution: This invention relates to a gait pre-training method for severe heatstroke patients with dynamic optimization of the standing bed angle, comprising the following steps: Step S1: Secure the patient to the electric standing bed and connect the multimodal physiological signal monitoring device, while acquiring the patient's recent coagulation function laboratory test results to complete the initial data acquisition and device connection; Step S2: The central processing unit receives the collected initial data, calls the built-in individualized tolerance prediction model and sets the tolerance index threshold range of its output, and initializes the standing bed angle to complete the initialization and parameter setting of the individualized tolerance prediction model. Step S3: Control the electric lifting bed to slowly raise it to a low safe probing angle and maintain this angle for a period of time. During this period, collect and record all physiological signal data to establish baseline reference values, thus completing the low-angle probing and the establishment of baseline physiological data. Step S4: During the process of maintaining or changing the angle of the standing bed, acquire multimodal physiological signal data streams in real time at a fixed frequency, and input the latest physiological data and recent laboratory data into the individualized tolerance prediction model to calculate the current tolerance index; Step S5: Based on the tolerance index value calculated in real time, determine the tolerance level of the patient's current state, and select to execute different bed-standing instructions according to the decision logic corresponding to different preset levels. Step S6: Repeat the real-time data acquisition, tolerance index calculation and angle adjustment decision process until the standing bed angle reaches the preset maximum target training angle, or the angle descent operation is triggered more than the preset number of times, or the training is actively interrupted. Step S7: After training, control the standing bed to smoothly return to the horizontal position and automatically generate a training report containing the angle changes, physiological data and decision events throughout the process. Encrypt all data and store it in the patient's electronic medical record database to complete the training and data archiving.

[0004] Furthermore, step S1 includes the following steps: Step S11: Secure the patient in a supine position on the electric standing bed. The multimodal physiological signal monitoring device includes at least a non-invasive continuous blood pressure monitor, an electrocardiogram monitor, an electroencephalogram monitor, and a transcranial Doppler cerebral blood flow monitor. Step S12: The non-invasive continuous blood pressure monitor is used to monitor the patient's real-time arterial blood pressure waveform and value; the electrocardiogram monitor is used to collect the patient's electrocardiogram signal and heart rate variability index; the electroencephalogram monitor is used to acquire the patient's brain electrical activity signal; and the transcranial Doppler cerebral blood flow monitor is used to monitor the blood flow velocity of the patient's middle cerebral artery. Step S13: Simultaneously acquire the patient's recent coagulation function laboratory test results, including prothrombin time, activated partial thromboplastin time, D-dimer concentration, and platelet count.

[0005] Furthermore, step S2 includes the following steps: Step S21: The central processing unit receives and preprocesses the collected physiological signals and laboratory data, and calls the built-in individualized tolerance prediction model based on a feedforward neural network. Step S22: The input feature vector of the model includes real-time measurements of systolic blood pressure, diastolic blood pressure, mean arterial pressure, heart rate, and cerebral blood flow velocity, as well as laboratory values ​​of prothrombin time, activated partial thromboplastin time, D-dimer, and platelet count. Step S23: Set the tolerance index output by the model. The continuous numerical range and the judgment threshold, among which The closer the value is to 1, the better the postural tolerance; the closer it is to 0, the worse the tolerance. Initialize the standing bed angle. That is, the initial horizontal position.

[0006] Furthermore, step S3 includes the following steps: Step S31: Control the electric erecting bed to The rate slowly increases to a preset low-safety probe angle. ; Step S32: Maintain the angle of the standing bed stable at... continued ,in, In the exploratory angle The duration of the stability; Step S33: While maintaining the elevation of the standing bed angle, continuously collect all access physiological signals at a sampling frequency of 1Hz, calculate the arithmetic mean and standard deviation of each signal during this time period, and establish the multimodal physiological baseline reference value of the patient in the low-angle state.

[0007] Furthermore, step S4 includes the following steps: Step S41: During the process of adjusting and maintaining the angle of the erecting bed, with... The frequency of continuous synchronous acquisition of real-time data streams of non-invasive continuous blood pressure, electrocardiogram, electroencephalogram, and transcranial Doppler cerebral blood flow velocity, among which, The sampling frequency of physiological signal data; Step S42: For each sampling time The latest instantaneous values ​​of physiological signals are combined with the most recently acquired laboratory data on coagulation function to form the input feature vector. The data is input into the individualized tolerance prediction model. Step S43: The model uses the network weights that have been trained internally. With bias After the hidden layer activation function and output layer activation function The nonlinear transformation is used to calculate and output the tolerance index at the current moment in real time. The formula is as follows: In the formula, To be at a specific sampling time The calculated real-time tolerance index, In time The input feature vector represents a column matrix containing the values ​​of all input signals; The weight matrix connecting the input layer and hidden layer or the hidden layer and output layer of a neural network; The bias vector is obtained through model training; This is the activation function for the output layer of the neural network, used to map the computation results to the final output range.

[0008] Furthermore, in step S5, the central processing unit calculates the results in real time. The numerical values ​​are used to execute the following non-linear decision-making logic: when If the patient's current condition is deemed to be well-tolerated, then the use of the standing bed should be controlled. The rate of upward elevation angle; when If the patient's current condition is deemed to be tolerable, then the use of the standing bed should be controlled. The standard safe rate of upward tilt angle; when If the patient is deemed to be at the critical point of tolerance, the angle change should be immediately stopped, and the current angle should be maintained. Unchanged, and continuously monitored At least Second; During this observation period If the price continues to rise and stabilizes at >0.5 for more than 60 seconds, then switch to a steady upward decision; otherwise, switch to a downward decision. like If the patient exhibits poor tolerance and a high risk, the patient should immediately stop using the standing bed. The rate decreases at the angle until Or return to the angle Safe position.

[0009] Furthermore, step S6 includes the following steps: Step S61: Repeat the data acquisition, model calculation, and decision control process described in steps S4 and S5 until one of the following termination conditions is met: The angle of the standing bed should reach the doctor's preset maximum target training angle for the day. ; The cumulative number of times the electronically controlled descent decision was triggered during a single training session exceeded the preset upper limit. The operating physician manually issues an interrupt command to terminate the training. Step S62: When the bed angle reaches... Afterwards, maintain this angle for gait pre-training, and continuously monitor during this phase. The numerical value indicates that when the standing bed descends and triggers the critical or poor tolerance decision logic, the angle is adjusted or the training is terminated according to the corresponding logic.

[0010] Furthermore, step S7 includes the following steps: Step S71: After training, control the electric lifting bed to return to the horizontal position at a slow and steady speed; Step S72: The system automatically summarizes the angle change curves, all collected physiological signal data, and tolerance index during this training process. The time-series change curves and records of all triggered decision events are used to generate a structured training report; Step S73: Encrypt all process data and generated reports and store them in the patient's electronic medical record database.

[0011] The present invention has the following beneficial effects: 1. This invention collects multimodal physiological signals from patients in real time, including non-invasive continuous blood pressure, electrocardiogram, electroencephalogram, and transcranial Doppler cerebral blood flow velocity. Combined with coagulation function laboratory data, it dynamically calculates the tolerance index using an individualized tolerance prediction model based on a feedforward neural network. This allows for intelligent adjustment of the standing bed angle according to the patient's real-time condition. This method completely overcomes the drawbacks of traditional fixed-angle adjustment, effectively preventing potential safety hazards such as sudden drops in blood pressure and insufficient cerebral blood flow during training. It improves the safety and reliability of training and enhances the individualization and overall effectiveness of rehabilitation training.

[0012] 2. This invention achieves a multi-dimensional and precise assessment of patient positional tolerance by real-time acquisition of multiple physiological parameters and simultaneous acquisition of coagulation function laboratory data, inputting these multi-source data into an individualized tolerance prediction model. This integrated analysis fully utilizes multiple information sources, avoiding the shortcomings of single-data assessment in traditional methods, and can more timely and accurately detect subtle changes in the patient's physiological state, thus providing more scientific and reliable decision support for the dynamic optimization of the standing bed angle.

[0013] 3. This invention uses a central processing unit to collect multimodal physiological signal data in real time at a fixed frequency, and calculates the tolerance index using an individualized tolerance prediction model. Based on a preset nonlinear decision-making logic, the system automatically adjusts the angle of the standing bed. The system automatically executes different operational instructions according to the numerical range of the tolerance index; for example, it raises the bed at a relatively fast rate when the index is good, pauses observation when the index is critical, and rapidly lowers it when the index is poor, until the termination condition is met. This automated control ensures the efficiency and adaptability of the training process, reducing the workload of medical staff and subjective errors in human judgment. Simultaneously, after training, the system automatically generates a structured report and stores it encrypted, providing data support for medical decision-making and comprehensively improving the intelligent management level and operational convenience of rehabilitation training.

[0014] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating a method for gait pre-training in severe heatstroke patients based on dynamic optimization of the standing bed angle, as described in this invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figure 1 As shown, this invention is a gait pre-training method for severe heatstroke patients with dynamic optimization of the standing bed angle, comprising the following steps: Step S1: Secure the patient to the electric standing bed and connect the multimodal physiological signal monitoring device, while acquiring the patient's recent coagulation function laboratory test results to complete the initial data acquisition and device connection; Step S2: The central processing unit receives the collected initial data, calls the built-in individualized tolerance prediction model and sets the tolerance index threshold range of its output, and initializes the standing bed angle to complete the initialization and parameter setting of the individualized tolerance prediction model. Step S3: Control the electric lifting bed to slowly raise it to a low safe probing angle and maintain this angle for a period of time. During this period, collect and record all physiological signal data to establish baseline reference values, thus completing the low-angle probing and the establishment of baseline physiological data. Step S4: During the process of maintaining or changing the angle of the standing bed, acquire multimodal physiological signal data streams in real time at a fixed frequency, and input the latest physiological data and recent laboratory data into the individualized tolerance prediction model to calculate the current tolerance index; Step S5: Based on the tolerance index value calculated in real time, determine the tolerance level of the patient's current state, and select to execute different bed-standing instructions according to the decision logic corresponding to different preset levels. Step S6: Repeat the real-time data acquisition, tolerance index calculation and angle adjustment decision process until the standing bed angle reaches the preset maximum target training angle, or the angle descent operation is triggered more than the preset number of times, or the training is actively interrupted. Step S7: After training, control the standing bed to smoothly return to the horizontal position and automatically generate a training report containing the angle changes, physiological data and decision events throughout the process. Encrypt all data and store it in the patient's electronic medical record database to complete the training and data archiving.

[0019] Step S1 includes the following steps: Step S11: Secure the patient in a supine position on the electric standing bed. The multimodal physiological signal monitoring equipment includes at least a non-invasive continuous blood pressure monitor, an electrocardiogram monitor, an electroencephalogram monitor, and a transcranial Doppler cerebral blood flow monitor. Step S12: The non-invasive continuous blood pressure monitor is used to monitor the patient's real-time arterial blood pressure waveform and value; the electrocardiogram monitor is used to collect the patient's electrocardiogram signal and heart rate variability index; the electroencephalogram monitor is used to acquire the patient's brain electrical activity signal; and the transcranial Doppler cerebral blood flow monitor is used to monitor the blood flow velocity of the patient's middle cerebral artery. Step S13: Simultaneously obtain the patient's recent coagulation function laboratory test results, including prothrombin time, activated partial thromboplastin time, D-dimer concentration, and platelet count.

[0020] Step S2 includes the following steps: Step S21: The central processing unit receives and preprocesses the collected physiological signals and laboratory data, and calls the built-in individualized tolerance prediction model based on a feedforward neural network. Step S22: The input feature vector of the model includes real-time measurements of systolic blood pressure, diastolic blood pressure, mean arterial pressure, heart rate, and cerebral blood flow velocity, as well as laboratory values ​​of prothrombin time, activated partial thromboplastin time, D-dimer, and platelet count. Step S23: Set the tolerance index of the model output. The continuous numerical range and the judgment threshold, among which The closer the value is to 1, the better the postural tolerance; the closer it is to 0, the worse the tolerance. Initialize the standing bed angle. That is, the initial horizontal position.

[0021] Step S3 includes the following steps: Step S31: Control the electric erecting bed to The rate slowly increases to a preset low-safety probe angle. ; Step S32: Maintain the angle of the standing bed stable at... continued ,in, In the exploratory angle The duration of the stability; Step S33: While maintaining the elevation of the standing bed angle, continuously collect all access physiological signals at a sampling frequency of 1Hz, calculate the arithmetic mean and standard deviation of each signal during this time period, and establish the multimodal physiological baseline reference value for the patient in the low-angle state.

[0022] Step S4 includes the following steps: Step S41: During the process of adjusting and maintaining the angle of the erecting bed, with... The frequency of continuous synchronous acquisition of real-time data streams of non-invasive continuous blood pressure, electrocardiogram, electroencephalogram, and transcranial Doppler cerebral blood flow velocity, among which, The sampling frequency of physiological signal data; Step S42: For each sampling time The latest instantaneous values ​​of physiological signals are combined with the most recently acquired laboratory data on coagulation function to form the input feature vector. The data is then input into the individualized tolerance prediction model. Step S43: The model uses the network weights that have been trained internally. With bias After the hidden layer activation function and output layer activation function The nonlinear transformation is used to calculate and output the tolerance index at the current moment in real time. The formula is as follows: In the formula, To be at a specific sampling time The calculated real-time tolerance index, In time The input feature vector represents a column matrix containing the values ​​of all input signals; The weight matrix connecting the input layer and hidden layer or the hidden layer and output layer of a neural network; The bias vector is obtained through model training; This is the activation function for the output layer of the neural network, used to map the computation results to the final output range.

[0023] In step S5, the central processing unit calculates the results in real time. The numerical values ​​are used to execute the following non-linear decision-making logic: when If the patient's current condition is deemed to be well-tolerated, then the use of the standing bed should be controlled. The rate of upward elevation angle; when If the patient's current condition is deemed to be tolerable, then the use of the standing bed should be controlled. The standard safe rate of upward tilt angle; when If the patient is deemed to be at the critical point of tolerance, the angle change should be immediately stopped, and the current angle should be maintained. Unchanged, and continuously monitored At least Second; During this observation period If the price continues to rise and stabilizes at >0.5 for more than 60 seconds, then switch to a steady upward decision; otherwise, switch to a downward decision. like If the patient exhibits poor tolerance and a high risk, the patient should immediately stop using the standing bed. The rate decreases at the angle until Or return to the angle Safe position.

[0024] Step S6 includes the following steps: Step S61: Repeat steps S4 and S5 for data acquisition, model calculation, and decision control until one of the following termination conditions is met: The angle of the standing bed should reach the doctor's preset maximum target training angle for the day. ; The cumulative number of times the electronically controlled descent decision was triggered during a single training session exceeded the preset upper limit. The operating physician manually issues an interrupt command to terminate the training. Step S62: When the bed angle reaches... Afterwards, maintain this angle for gait pre-training, and continuously monitor during this phase. The numerical value indicates that when the standing bed descends and triggers the critical or poor tolerance decision logic, the angle will be adjusted or the training will be terminated according to the corresponding logic.

[0025] Step S7 includes the following steps: Step S71: After training, control the electric lifting bed to return to the horizontal position at a slow and steady speed; Step S72: The system automatically summarizes the angle change curves, all collected physiological signal data, and tolerance index during this training process. The time-series change curves and records of all triggered decision events are used to generate a structured training report; Step S73: Encrypt all process data and generated reports and store them in the patient's electronic medical record database.

[0026] One specific application of this embodiment is: Implementation Background: 1. We selected one patient recovering from severe heatstroke (heat exhaustion). Specific information is as follows: Gender: Male; Age: 45; Medical history: Heatstroke caused by working in high-temperature environments. After emergency treatment, the body temperature returned to normal (36.8℃). The patient was conscious but had a tendency for orthostatic hypotension (systolic blood pressure of 125 mmHg when lying down and dropping to 105 mmHg when sitting up). There was no history of coagulation dysfunction and the patient had not taken any medications that affect blood pressure or coagulation recently. Inclusion criteria: Stable vital signs (heart rate 65-80 beats / min, respiratory rate 18-22 breaths / min, blood oxygen saturation ≥95%), meeting the criteria for gait pre-training during the recovery period of severe heatstroke; 2. Electric standing bed: Utilizing a brand-name intelligent rehabilitation standing bed (model: RK-Q100), it supports angle adjustment from 0° to 80°, with an angle control accuracy of ±0.5°, and features a smooth lifting drive system. Multimodal physiological signal monitoring equipment: Non-invasive continuous blood pressure monitor (model: Philips IntelliVue MP70): sampling frequency 1Hz, can output arterial blood pressure waveform, systolic blood pressure (SBP), diastolic blood pressure (DBP), and mean arterial pressure (MAP) in real time. Electrocardiogram monitor (model: GE MAC 2000): Acquires standard 12-lead electrocardiogram signals and outputs heart rate (HR) and heart rate variability (HRV); Electroencephalogram (EEG) monitor (model: Nihon Kohden EEG-1200): uses 8-lead electrodes to monitor brain electrical activity signals (alpha wave and beta wave amplitude). Transcranial Doppler cerebral blood flow monitor (model: EME TC2020): Monitors blood flow velocity (Vm) in the middle cerebral artery (MCA), with a sampling frequency of 1Hz; Individualized tolerance prediction model: Based on a feedforward neural network, the input layer contains 8 features (SBP, DBP, MAP, HR, MCA-Vm, PT, APTT, D-dimer, platelet count), and the hidden layer has 2 layers (16 neurons per layer). The activation function is ReLU, and the output layer activation function is Sigmoid (mapping output Tᵢ∈[0,1]). The model was trained with historical data from 200 patients recovering from severe heatstroke, and the cross-validation accuracy was 92.3%. It has been integrated into the central processing unit (CPU model: Intel Core i7-12700K). Implementation steps: Step 1: 1. Patient fixation and equipment connection: The patient is placed in a supine position (the initial angle of the bed when standing up is θ=0°). The trunk and lower limbs are fixed by the bed restraint straps (the tightness should be such that one finger can be inserted). The non-invasive blood pressure cuff is tied to the left arm (2cm above the elbow). The electrocardiogram electrodes are attached to the chest (V1-V6, limb leads). The electroencephalogram electrodes are attached to the frontal and parietal lobes according to the international 10-20 system. The transcranial Doppler probe is fixed to the right temporal window (aligned with the middle cerebral artery). 2. Simultaneously retrieve the patient's coagulation function laboratory report within 24 hours. The data is as follows: Prothrombin time (PT): 12.5s (reference range 11-14s); Activated partial thromboplastin time (APTT): 34.2s (reference range 25-38s); D-dimer concentration: 0.45 mg / L (reference value <0.5 mg / L); Platelet count (PLT): 215 × 10⁻⁶ 9 / L (Reference value 100-300×10) 9 / L); Step Two: 1. Data preprocessing and model invocation: The central processing unit receives the collected physiological signals (SBP=122mmHg, DBP=81mmHg, MAP=95mmHg, HR=72 beats / min, MCA-Vm=52cm / s at the initial supine position) and coagulation data. After filtering (removing power frequency interference) and normalization (mapping the values ​​to the [0,1] interval), the built-in feedforward neural network model is invoked. 2. Threshold and angle initialization: Set the threshold for the tolerance index Tᵢ: Good grade: Tᵢ≥0.8 Acceptable level: 0.5 ≤ Tᵢ < 0.8 Critical level: 0.3 ≤ Tᵢ < 0.5 Defect level: Tᵢ<0.3 Initialize the bed angle θ = 0° (horizontal position). Step 3: 1. The central processing unit controls the erecting bed to slowly rise at a rate of 1° / 10s, reaching a low safety testing angle after 50s. and maintain that angle for 120 seconds. =120s) 2. During the 120-second period of maintaining the standing bed angle at 5°, all accessed physiological signals were continuously acquired at a sampling frequency of 1Hz. The arithmetic mean and standard deviation of each signal during this time period were calculated to establish the multimodal physiological baseline reference values ​​for the patient in the low-angle state. The specific data are as follows: the mean value of systolic blood pressure (SBP) was 118 mmHg, and the standard deviation was 2.3 mmHg; the mean value of diastolic blood pressure (DBP) was 78 mmHg, and the standard deviation was 1.8 mmHg; the mean value of heart rate (HR) was 75 beats / min, and the standard deviation was 1.5 beats / min; the mean value of middle cerebral artery blood flow velocity (Vm) was 49 cm / s, and the standard deviation was 1.2 cm / s; the mean value of EEG alpha wave amplitude was 50 μV, and the standard deviation was 3.1 μV. Step Four: 1. The standing bed begins its dynamic adjustment phase from 5°, with continuous synchronous acquisition of SBP, DBP, HR, MCA-Vm, and EEG signals at a frequency of 1Hz; at each sampling time... (e.g., t1=50s, t2=51s...), the instantaneous values ​​of the physiological signals at that moment are combined with the coagulation data obtained in S1 to form the input feature vector. (like: =[119,79,76,50,12.5,34.2,0.45,215]); 2. The model uses a pre-trained weight matrix W (input layer → hidden layer: 8×16 matrix, hidden layer → output layer: 16×1 matrix) and a bias vector b (hidden layer b1=[0.12,0.08,...], output layer b2=0.05), which is then activated by the ReLU activation function and the Sigmoid output function (F). )calculate The formula is as follows: when =100s =[121,80,74,51,12.5,34.2,0.45,215]; Calculation yields =0.86; Step 5: According to real time The numerical execution of nonlinear decision-making logic involves the following adjustment process: When Tᵢ=0.86 (≥0.8, good grade), the lifting bed is controlled to rise rapidly at a rate of 1° / 5s, taking 25s to rise from 5° to 10°; during this period, Tᵢ is continuously monitored, and the average value is maintained between 0.82 and 0.87. When the angle rises to 10°, Tᵢ drops to 0.75 (0.5≤0.75<0.8, acceptable level), and switches to a standard rate of 1° / 10s for the rise, taking 50s to rise from 10° to 15°; during this stage, Tᵢ fluctuates between 0.72 and 0.78. When the angle increases to 15°, Tᵢ drops sharply to 0.42 (0.3≤0.42<0.5, critical level). Immediately stop the angle change, maintain θ=15° and continue monitoring for 180s. =180s); for the first 60s, Tᵢ remains at 0.43-0.45, and for the next 120s it gradually rises to 0.61 (>0.5) and remains stable for more than 60s, then jumps to "uniform speed increase decision"; The recovery process begins with an elevation increase at a rate of 1° / 10s, lasting 50 seconds from 15° to 20° (the preset maximum target training angle for the day). =20°); during this stage, Tᵢ stabilizes at 0.65-0.72; Step Six: 1. Angle reached =20°, and maintain this angle for gait pre-training (patients perform passive movements such as ankle dorsiflexion and knee flexion and extension under the guidance of medical staff) for 30 minutes; data were collected every 1 second during this period, and Tᵢ remained between 0.68 and 0.75, without triggering the critical or adverse grade; 2. The 30-minute pre-training ends without triggering "cumulative descent count exceeded limit" or "manual interruption", and meets the termination condition of "reaching the maximum target angle and completing the preset training time". Step Seven: 1. Reset of the standing bed: The central processing unit controls the standing bed to descend from 20° to 0° (horizontal position) at a slow rate of 1° / 15s, lasting 300s. During the process, the patient's SBP is maintained at 115-120mmHg, without dizziness, nausea or other discomfort. 2. Report Generation: The system automatically summarizes the training data and generates a structured training report, including: Angle change curve (0°→5°→10°→15° (pause)→20°→0°); Physiological signal time series diagram (trends of SBP, HR, MCA-Vm over time); Record decision events (e.g., "t=100s: Tᵢ=0.86, rapid rise" "t=300s: Tᵢ=0.42, pause observation"). 3. After encrypting the training report, raw physiological data, and coagulation data using the AES-256 encryption algorithm, upload them to the hospital's electronic medical record database (Patient ID: P2024051001). Authorized medical staff can access and view them after verification of permissions.

[0027] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0028] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for gait pre-training in severe heatstroke patients with dynamic optimization of the angle of the standing bed, characterized in that: Includes the following steps: Step S1: Secure the patient to the electric standing bed and connect the multimodal physiological signal monitoring device, while acquiring the patient's recent coagulation function laboratory test results to complete the initial data acquisition and device connection; Step S2: The central processing unit receives the collected initial data, calls the built-in individualized tolerance prediction model and sets the tolerance index threshold range of its output, and initializes the standing bed angle to complete the initialization and parameter setting of the individualized tolerance prediction model. Step S3: Control the electric lifting bed to slowly raise it to a low safe probing angle and maintain this angle for a period of time. During this period, collect and record all physiological signal data to establish baseline reference values, thus completing the low-angle probing and the establishment of baseline physiological data. Step S4: During the process of maintaining or changing the angle of the standing bed, acquire multimodal physiological signal data streams in real time at a fixed frequency, and input the latest physiological data and recent laboratory data into the individualized tolerance prediction model to calculate the current tolerance index; Step S5: Based on the tolerance index value calculated in real time, determine the tolerance level of the patient's current state, and select to execute different bed-standing instructions according to the decision logic corresponding to different preset levels. Step S6: Repeat the real-time data acquisition, tolerance index calculation and angle adjustment decision process until the standing bed angle reaches the preset maximum target training angle, or the angle descent operation is triggered more than the preset number of times, or the training is actively interrupted. Step S7: After training, control the standing bed to smoothly return to the horizontal position and automatically generate a training report containing the angle changes, physiological data and decision events throughout the process. Encrypt all data and store it in the patient's electronic medical record database to complete the training and data archiving.

2. The method for gait pre-training in severe heatstroke with dynamic optimization of the standing bed angle according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Secure the patient in a supine position on the electric standing bed. The multimodal physiological signal monitoring device includes at least a non-invasive continuous blood pressure monitor, an electrocardiogram monitor, an electroencephalogram monitor, and a transcranial Doppler cerebral blood flow monitor. Step S12: The non-invasive continuous blood pressure monitor is used to monitor the patient's real-time arterial blood pressure waveform and value; the electrocardiogram monitor is used to collect the patient's electrocardiogram signal and heart rate variability index; the electroencephalogram monitor is used to acquire the patient's brain electrical activity signal; and the transcranial Doppler cerebral blood flow monitor is used to monitor the blood flow velocity of the patient's middle cerebral artery. Step S13: Simultaneously acquire the patient's recent coagulation function laboratory test results, including prothrombin time, activated partial thromboplastin time, D-dimer concentration, and platelet count.

3. The method for gait pre-training in severe heatstroke with dynamic optimization of the standing bed angle according to claim 1, characterized in that, Step S2 includes the following steps: Step S21: The central processing unit receives and preprocesses the collected physiological signals and laboratory data, and calls the built-in individualized tolerance prediction model based on a feedforward neural network. Step S22: The input feature vector of the model includes real-time measurements of systolic blood pressure, diastolic blood pressure, mean arterial pressure, heart rate, and cerebral blood flow velocity, as well as laboratory values ​​of prothrombin time, activated partial thromboplastin time, D-dimer, and platelet count. Step S23: Set the tolerance index output by the model. The continuous numerical range and the judgment threshold, among which The closer the value is to 1, the better the postural tolerance; the closer it is to 0, the worse the tolerance. Initialize the standing bed angle. That is, the initial horizontal position.

4. The method for gait pre-training in severe heatstroke with dynamic optimization of the standing bed angle according to claim 1, characterized in that, Step S3 includes the following steps: Step S31: Control the electric erecting bed to The rate slowly increases to a preset low-safety probing angle. ; Step S32: Maintain the angle of the standing bed stable at... continued ,in, In the exploratory angle The duration of the stability; Step S33: While maintaining the elevation of the standing bed angle, continuously collect all access physiological signals at a sampling frequency of 1Hz, calculate the arithmetic mean and standard deviation of each signal during this time period, and establish the multimodal physiological baseline reference value of the patient in the low-angle state.

5. The method for gait pre-training in severe heatstroke with dynamic optimization of the standing bed angle according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: During the process of adjusting and maintaining the angle of the erecting bed, with... The frequency of continuous synchronous acquisition of real-time data streams of non-invasive continuous blood pressure, electrocardiogram, electroencephalogram, and transcranial Doppler cerebral blood flow velocity, among which, The sampling frequency of physiological signal data; Step S42: For each sampling time The latest instantaneous values ​​of physiological signals are combined with the most recently acquired laboratory data on coagulation function to form the input feature vector. The data is input into the individualized tolerance prediction model. Step S43: The model uses the network weights that have been trained internally. With bias After the hidden layer activation function and output layer activation function The nonlinear transformation is used to calculate and output the tolerance index at the current moment in real time. The formula is as follows: In the formula, To be at a specific sampling time The calculated real-time tolerance index, In time The input feature vector represents a column matrix containing the values ​​of all input signals; The weight matrix connecting the input layer and hidden layer or the hidden layer and output layer of a neural network; The bias vector is obtained through model training; This is the activation function for the output layer of the neural network, used to map the computation results to the final output range.

6. The method for gait pre-training in severe heatstroke with dynamic optimization of the standing bed angle according to claim 1, characterized in that, In step S5, the central processing unit calculates the results in real time. The numerical values ​​are used to execute the following non-linear decision-making logic: when If the patient's current condition is deemed to be well-tolerated, then the use of the standing bed should be controlled. The rate of upward elevation angle; when If the patient's current condition is deemed to be tolerable, then the use of the standing bed should be controlled. The standard safe rate of upward tilt angle; when If the patient is deemed to be at the critical point of tolerance, the angle change should be immediately stopped, and the current angle should be maintained. Unchanged, and continuously monitored At least Second; During this observation period If the price continues to rise and stabilizes at >0.5 for more than 60 seconds, then switch to a steady upward decision; otherwise, switch to a downward decision. like If the patient exhibits poor tolerance and a high risk, the patient should immediately stop using the standing bed. The rate decreases at the angle until Or the angle returns to Safe position.

7. The method for gait pre-training in severe heatstroke with dynamic optimization of the standing bed angle according to claim 1, characterized in that, Step S6 includes the following steps: Step S61: Repeat the data acquisition, model calculation, and decision control process described in steps S4 and S5 until one of the following termination conditions is met: The angle of the standing bed should reach the doctor's preset maximum target training angle for the day. ; The cumulative number of times the electronically controlled descent decision was triggered during a single training session exceeded the preset upper limit. The operating physician manually issues an interrupt command to terminate the training. Step S62: When the bed angle reaches... Afterwards, maintain this angle for gait pre-training, and continuously monitor during this phase. The numerical value indicates that when the standing bed descends and triggers the critical or poor tolerance decision logic, the angle is adjusted or the training is terminated according to the corresponding logic.

8. The method for gait pre-training in severe heatstroke with dynamic optimization of the standing bed angle according to claim 1, characterized in that, Step S7 includes the following steps: Step S71: After training, control the electric lifting bed to return to the horizontal position at a slow and steady speed; Step S72: The system automatically summarizes the angle change curves, all collected physiological signal data, and tolerance index during this training process. The time-series change curves and records of all triggered decision events are used to generate a structured training report; Step S73: Encrypt all process data and generated reports and store them in the patient's electronic medical record database.

Images