Intelligent control method and system for intensive care unit equipment

Abstract

The application provides an intensive care unit equipment intelligent control method and system, which is applied to the technical field of medical equipment intelligent control. The treatment intervention intensity information, the first physiological parameter and the second physiological parameter are acquired, the damping characteristic value is extracted, the consistency of the first physiological parameter is verified in combination with the second physiological parameter, the real-time change trend of the first physiological parameter is compared with the expected change trend to obtain a response deviation evaluation result, and whether the preset arbitration triggering condition is met is judged according to the damping characteristic value, the data reliability evaluation result and the response deviation evaluation result. If it is met, the adjustment instruction of the treatment intervention intensity is subjected to constraint processing, so that the treatment intervention intensity is kept within the preset safety threshold range or maintained in the current state, which has the advantages that the monitoring data distortion caused by physiological complexity or equipment physical characteristics can be effectively identified and processed, excessive intervention is avoided, and the treatment safety and effectiveness are improved.

Description

Technical Field

[0001] This application relates to the field of intelligent control technology for medical equipment, and in particular to an intelligent control method and system for intensive care unit equipment. Background Technology

[0002] In the intensive care unit, although intelligent control systems can automatically manage equipment such as ventilators and infusion pumps and adjust treatment plans based on real-time data to improve treatment efficiency, they may still face challenges in practical applications due to the interplay between the physiological complexity of patients and the physical characteristics of the equipment.

[0003] Take blood pressure management in patients with septic shock as an example: the system automatically adjusts the dosage of vasopressors based on the mean arterial pressure (MAP) target. However, when the patient's blood vessels are sluggish in response to the drug due to severe inflammation, the blood pressure may not rise as expected after the system increases the dosage, leading to a dilemma of continuously increasing the dosage but failing to achieve the target.

[0004] Meanwhile, prolonged invasive blood pressure monitoring may lead to the formation of microthrombi or fibrin adhesions at the tip of the ductus arteriosus, causing signal damping. This physical change distorts the monitoring data in two ways: first, the measured MAP value is systematically lower than the actual blood pressure; second, the pressure waveform transmission is delayed, causing a lag in the data time axis.

[0005] When the intelligent control system makes decisions based on this distorted data, it may misjudge that the patient's actual blood pressure is far below the target and that the patient's response to medication is abnormally slow. This double-misdriven system continuously issues commands to increase the dosage of vasopressors, but fails to obtain accurate feedback on the therapeutic effect.

[0006] Ultimately, in an attempt to correct a false low blood pressure caused by a sensor malfunction, the system continuously increases the drug infusion rate, leading to an excessively high actual blood pressure in the patient, potentially causing serious complications such as myocardial ischemia and cerebral hemorrhage. This process reveals how a single physical fault (signal damping) can couple with complex physiological responses and system decision-making logic, creating an iatrogenic treatment dilemma that endangers patient safety.

[0007] Therefore, existing technologies urgently need to be improved to address the aforementioned problems. Summary of the Invention

[0008] In view of the shortcomings of the prior art, this application provides an intelligent control method and system for critical care equipment, which has the advantages of effectively identifying and processing the distortion of monitoring data caused by physiological complexity or physical characteristics of equipment, avoiding excessive intervention, and improving the safety and effectiveness of treatment.

[0009] In a first aspect, a method for intelligent control of critical care equipment, the method comprising the following steps:

[0010] S1: Obtain information on the intensity of the treatment intervention for the controlled subject, and simultaneously obtain the first physiological parameter and at least one second physiological parameter fed back by the pressure monitoring channel;

[0011] S2: Analyze the waveform characteristics of the first physiological parameter and extract the damping characteristic value that characterizes the physical transmission characteristics of the pressure monitoring channel;

[0012] S3: Combine the second physiological parameter with the first physiological parameter to perform a consistency check and determine the data reliability evaluation result of the first physiological parameter;

[0013] S4: Compare the real-time change trend of the first physiological parameter with the expected change trend obtained based on the treatment intervention intensity information and the preset response model to obtain the response deviation evaluation result;

[0014] S5: Determine whether the damping characteristic value, the data reliability evaluation result, and the response deviation evaluation result meet the preset arbitration triggering conditions;

[0015] S6: If the arbitration triggering condition is met, the adjustment instruction for the intensity of the treatment intervention is constrained to keep the intensity of the treatment intervention within a preset safety threshold range or maintain the current state.

[0016] Furthermore, step S2 includes:

[0017] S21: Identify the slope of the initial ascending branch and the dicrotic notch features in the arterial pressure waveform of the first physiological parameter;

[0018] S22: Calculate the resonant frequency and damping coefficient of the pressure monitoring channel based on the initial rising slope and the dicrotic notch characteristics, and use the resonant frequency and the damping coefficient as the damping characteristic value.

[0019] Furthermore, step S3 includes:

[0020] S31: Obtain the first rate of change of the first physiological parameter and the second rate of change of the second physiological parameter, wherein the second physiological parameter includes at least one of heart rate, urine volume and blood lactate;

[0021] S32: Calculate the physiological correlation between the first rate of change and the second rate of change;

[0022] S33: When the physiological correlation is lower than the preset correlation threshold, the data reliability evaluation result of the first physiological parameter is determined to be low reliability.

[0023] Furthermore, step S4 includes:

[0024] S41: Input the treatment intervention intensity information into the individualized pharmacodynamic model for the controlled object to obtain the expected increase in physiological parameters under the current intervention intensity;

[0025] S42: Calculate the difference between the actual increase in the first physiological parameter and the expected increase in the physiological parameter;

[0026] S43: Determine the response deviation evaluation result based on the difference, wherein the larger the difference, the higher the degree of deviation represented by the response deviation evaluation result.

[0027] Furthermore, in step S5, the arbitration triggering conditions include: the damping characteristic value exceeds a preset damping threshold; or, the data reliability evaluation result is low reliability; or, the response deviation evaluation result exceeds a preset deviation threshold.

[0028] Furthermore, step S6 includes:

[0029] S61: Intercept the incremental portion of the adjustment instruction regarding increasing the intensity of the treatment intervention, so that the intensity of the treatment intervention remains at the current level;

[0030] Alternatively, based on the weighted value of the damping characteristic value, the data reliability evaluation result, and the response deviation evaluation result, the output weight corresponding to the adjustment command can be reduced to limit the rate of increase of the treatment intervention intensity, so that the treatment intervention intensity is kept within a preset safe threshold range.

[0031] Furthermore, step S6 includes the following:

[0032] S7: Continuously monitor the state changes of the damping characteristic value, the data reliability evaluation result, and the response deviation evaluation result;

[0033] S8: When it is determined that the damping characteristic value has recovered to the preset normal range and the data reliability evaluation result has turned to high reliability, the recovery boot program is started.

[0034] Furthermore, in step S8, starting the recovery bootloader includes the following steps:

[0035] S81: Obtain the real-time values ​​of the physiological parameters of the controlled object at the current moment, and use them as the safety starting benchmark;

[0036] S82: Calculate the allowable recovery increment per unit time based on the difference between the safety starting benchmark and the preset treatment target value, combined with the individualized pharmacodynamic model;

[0037] S83: Increase the upper limit of the treatment intervention intensity in stages according to the allowed recovery increment, until the treatment intervention intensity is restored to the real-time target level determined by the automatic dose adjustment logic.

[0038] Furthermore, step S1 includes:

[0039] S11: The first physiological parameter, the second physiological parameter, and the treatment intervention intensity information are collected in real time via a data bus at a preset sampling frequency, wherein the preset sampling frequency is not less than 10Hz.

[0040] Secondly, an intelligent control system for critical care equipment, the system being used to implement any of the methods described above, the system comprising:

[0041] The information acquisition module is used to acquire information on the intensity of the treatment intervention for the controlled object, and simultaneously acquire the first physiological parameter and at least one second physiological parameter fed back by the pressure monitoring channel;

[0042] The feature analysis module is used to analyze the waveform features of the first physiological parameter and extract the damping feature value that characterizes the physical transmission characteristics of the pressure monitoring channel.

[0043] The consistency verification module is used to perform consistency verification on the first physiological parameter in conjunction with the second physiological parameter, and to determine the data reliability evaluation result of the first physiological parameter.

[0044] The trend comparison module is used to compare the real-time change trend of the first physiological parameter with the expected change trend obtained based on the treatment intervention intensity information and the preset response model to obtain the response deviation evaluation result.

[0045] The arbitration judgment module is used to determine whether the damping characteristic value, the data reliability evaluation result, and the response deviation evaluation result meet the preset arbitration triggering conditions;

[0046] The constraint execution module is used to perform constraint processing on the adjustment command of the treatment intervention intensity if the arbitration triggering condition is met, so as to keep the treatment intervention intensity within a preset safety threshold range or maintain the current state.

[0047] Beneficial Effects: The intelligent control method and system for critical care equipment proposed in this application acquires treatment intervention intensity information, a first physiological parameter, and a second physiological parameter. It analyzes the waveform characteristics of the first physiological parameter to extract damping characteristic values, combines the second physiological parameter to perform consistency verification on the first physiological parameter to determine the data reliability evaluation result, compares the real-time change trend of the first physiological parameter with the expected change trend to obtain the response deviation evaluation result, and determines whether a preset arbitration trigger condition is met based on the damping characteristic value, the data reliability evaluation result, and the response deviation evaluation result. If met, it performs constraint processing on the adjustment command of the treatment intervention intensity to keep the treatment intervention intensity within a preset safety threshold range or maintain the current state. This method effectively identifies and handles monitoring data distortion caused by physiological complexity or equipment physical characteristics, avoids over-intervention, and improves treatment safety and effectiveness. Attached Figure Description

[0048] Figure 1 This is a flowchart of an intelligent control method for intensive care unit equipment proposed in this application.

[0049] Figure 2 This is a structural diagram of an intelligent control system for critical care equipment proposed in this application.

[0050] Figure 3 This is a schematic diagram of an intelligent control system for critical care equipment proposed in this application.

[0051] Labeling Explanation: 201 Information Acquisition Module; 202 Feature Parsing Module; 203 Consistency Verification Module; 204 Trend Comparison Module; 205 Arbitration Judgment Module; 206 Constraint Execution Module. Detailed Implementation

[0052] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and marked in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0053] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0054] Please refer to Figure 1 A method for intelligent control of critical care equipment, the method comprising the following steps:

[0055] S1: Obtain information on the intensity of the treatment intervention for the controlled subject, and simultaneously obtain the first physiological parameter and at least one second physiological parameter fed back by the pressure monitoring channel;

[0056] S2: Analyze the waveform characteristics of the first physiological parameter and extract the damping characteristic value that characterizes the physical transmission characteristics of the pressure monitoring channel;

[0057] S3: Combine the second physiological parameter with the first physiological parameter to verify the consistency and determine the data reliability evaluation result of the first physiological parameter;

[0058] S4: Compare the real-time change trend of the first physiological parameter with the expected change trend obtained based on the treatment intervention intensity information and the preset response model to obtain the response deviation evaluation result;

[0059] S5: Determine whether the damping characteristic value, data reliability evaluation result, and response deviation evaluation result meet the preset arbitration triggering conditions;

[0060] S6: If the arbitration triggering condition is met, the adjustment instruction for the intensity of treatment intervention will be constrained to keep the intensity of treatment intervention within the preset safety threshold range or maintain the current state.

[0061] The core of this method lies in introducing a high-level safety arbitration mechanism into the intelligent control logic, enabling it to examine the reliability of the decision-making basis from multiple dimensions before executing treatment instructions. When it detects a risk of distortion in core physiological data, or when the patient's physiological response is significantly inconsistent with expectations, the mechanism will proactively intervene to constrain the treatment instructions, thereby preventing dangerous treatment cycles that may be caused by data distortion.

[0062] In step S1, the controlled subject refers to the patient to be treated. In a typical intensive care setting, the controlled subject could be a patient with septic shock, whose treatment goal is to maintain a mean arterial pressure above 65 mmHg. At this time, a central control unit, such as an embedded computer equipped with a high-performance microprocessor like the ARM Cortex-M series, communicates in real-time with various medical devices at the bedside via a high-speed, reliable data bus, such as a medical-grade controller area network bus or Ethernet. The central control unit continuously collects multi-source information. Specifically, the treatment intervention intensity information refers to the real-time infusion rate of vasopressors, such as norepinephrine, as fed back by the intelligent infusion pump, typically measured in micrograms per kilogram of body weight per minute.

[0063] The first physiological parameter refers to the core hemodynamic index fed back by the invasive arterial pressure monitoring module. This is not just a single mean arterial pressure value, but a continuous pressure signal data stream containing complete waveform information, which originates from the indwelling catheter in the patient's radial or femoral artery.

[0064] At least one second physiological parameter acquired simultaneously is other vital sign information used for cross-validation and comprehensive judgment, such as heart rate obtained through electrocardiogram monitoring electrodes, hourly urine volume obtained through a urinary catheter with sensors, and blood lactate concentration measured regularly by bedside blood gas analysis equipment or a central laboratory.

[0065] To ensure the accuracy of subsequent analysis, the data acquisition process requires a high degree of real-time performance. Specifically, step S1 includes:

[0066] S11: Collect the first physiological parameter, the second physiological parameter, and the intensity of treatment intervention in real time via the data bus at a preset sampling frequency, wherein the preset sampling frequency is not less than 10Hz.

[0067] The preset sampling frequency is no less than 10 Hz. For rapidly changing signals such as arterial pressure waveforms, higher sampling frequencies, such as 100 Hz or even higher, are necessary because only sufficiently high time resolution can fully capture the key morphological features in the waveform, providing high-quality raw data for subsequent damping feature analysis.

[0068] After acquiring comprehensive data, the safety arbitration mechanism begins to assess the physical condition of the pressure monitoring channel itself. Specifically, the waveform characteristics of the first physiological parameter are analyzed to extract damping characteristic values ​​that characterize the physical transmission properties of the pressure monitoring channel. The fundamental purpose of this step is to determine whether the sensor is functioning correctly.

[0069] Arterial pressure waveforms are not smooth curves, but complex waveforms containing rich hemodynamic information. A normal arterial pressure waveform has a steep ascending limb, a clear dicrotic notch, and a relatively gentle descending limb. When the pressure monitoring channel is damped, these morphological features typically change; for example, the ascending limb becomes gentler, and the dicrotic notch becomes blurred or even disappears.

[0070] The signal processing algorithm inside the control unit analyzes the acquired continuous arterial pressure waveform data in real time. Through specific algorithms, it identifies and quantifies these morphological changes, thereby calculating one or more characteristic values ​​that characterize the channel's damping level. This damping characteristic value, like a health index of a pressure monitoring channel, directly reflects whether its physical transmission characteristics are intact.

[0071] Specifically, identifying and quantifying these morphological changes to calculate one or more eigenvalues ​​that characterize the degree of channel damping includes:

[0072] S21: Identify the slope of the initial ascending limb and the dicrotic notch features in the arterial pressure waveform of the first physiological parameter;

[0073] S22: Calculate the resonant frequency and damping coefficient of the pressure monitoring channel based on the slope of the initial rising branch and the characteristics of the diphtheria notch, and use the resonant frequency and damping coefficient as damping characteristic values.

[0074] In practical applications, the signal processing program within the control unit receives continuous arterial pressure waveform data acquired at a high sampling frequency. The program first uses a waveform recognition algorithm to automatically locate the starting point and key feature points of each cardiac cycle waveform.

[0075] The initial rising slope, which is the slope of the pressure waveform from the beginning of the rapid rise in pressure at the end of diastole, reflects the velocity of left ventricular ejection and aortic compliance. This slope will decrease significantly when damping occurs.

[0076] The dicrotic notch, a small notch formed on the middle descending branch of the aorta when the aortic valve closes, represents the high-frequency component of the pressure wave. Its clarity is closely related to the frequency response characteristics of the pressure monitoring system. When damping increases, the system's ability to respond to high-frequency signals decreases, and the dicrotic notch becomes blurred or even disappears completely.

[0077] The analysis process is achieved through joint time-domain and frequency-domain analysis of the acquired continuous arterial pressure signals. The identification of the slope of the initial ascending limb is accomplished by calculating the maximum value of the first derivative of the early systolic pressure change over time during the cardiac cycle. This value reflects the response speed of the pressure transmission channel to rapid pressure changes.

[0078] The identification of dicrotic notch features relies on locating the secondary pressure peaks appearing in the descending limb, and determining the depth and temporal position of the notch by detecting curvature change points on the descending limb. These morphological parameters are then substituted into a pre-defined physical mapping model, which simulates a second-order underdamped oscillating system in fluid dynamics.

[0079] By calculating the natural frequency of the waveform and the amplitude attenuation ratio, the resonant frequency and damping coefficient of the pressure monitoring channel are obtained in reverse. When there is thrombus or fibrin adhesion at the catheter tip, the energy of the signal in the high-frequency band will be significantly attenuated, which is manifested as a significant increase in the damping coefficient and a decrease in the resonant frequency, thus realizing a quantitative characterization of the physical transmission characteristics of the channel.

[0080] After identifying and quantifying these waveform features, such as calculating specific slope values ​​and scoring the morphology of the dicrotic notch, the program uses a pre-set mathematical model that describes the relationship between waveform morphology features and the second-order oscillation characteristic parameters of the pressure monitoring system.

[0081] Specifically, the pre-set mathematical model can be a transfer function model based on the theory of second-order underdamped oscillating systems. This model treats the pressure monitoring channel as a mechanical oscillating system composed of mass, springs, and dampers, whose dynamic response can be represented by the following differential equation: Where m represents the effective mass, c represents the damping coefficient, k represents the elastic coefficient, x represents the displacement, and F(t) represents the input pressure.

[0082] The differential equation can be transformed into a frequency domain transfer function using the Laplace transform: Among them, the resonant frequency Damping ratio In practical applications, Fourier analysis of the initial ascending slope and dicrotic notch characteristics of the arterial pressure waveform can extract the high-frequency components and attenuation characteristics of the waveform.

[0083] By fitting these frequency domain characteristics to the transfer function model, for example using the least squares method or the Kalman filter algorithm, the resonant frequency of the system can be solved in reverse. and the damping ratio ζ. These are the solutions. ζ and ζ are damping characteristic values ​​used to characterize the physical transmission properties of the pressure monitoring channel.

[0084] The resonant frequency reflects the system's inherent vibration tendency, while the damping coefficient directly quantifies the degree of energy dissipation within the system. An ideal pressure monitoring system should have a high resonant frequency and a moderate damping coefficient. When the conduit becomes blocked or kinked, the damping increases significantly, leading to a higher calculated damping coefficient value. These two calculated values, the resonant frequency and the damping coefficient, together constitute a precise quantitative description of the physical transmission characteristics of the pressure monitoring channel. They are used as damping characteristic values, providing an objective and quantitative basis for subsequent arbitration judgments.

[0085] Next, the security arbitration mechanism will assess the inherent logical consistency between the first physiological parameter and other physiological parameters. Specifically, it will combine the second physiological parameter to perform a consistency check on the first physiological parameter, and determine the data reliability evaluation result of the first physiological parameter.

[0086] The human body is an organic whole, with profound and complex internal connections between various physiological parameters. For example, in a state of hypotension caused by insufficient blood volume or heart failure, the body usually compensates by increasing the heart rate to maintain cardiac output. Simultaneously, the decrease in blood pressure leads to a reduction in renal perfusion pressure, resulting in decreased urine output. Therefore, a true state of hypotension is often accompanied by a series of coordinated physiological changes, such as increased heart rate and decreased urine output. If monitoring data shows that the first physiological parameter, mean arterial pressure, remains consistently low, but at the same time, the second physiological parameter, such as heart rate, remains stable or even slow, and urine output is ample, this constitutes a physiological contradiction.

[0087] The control unit runs a consistency check algorithm that incorporates various physiological models to assess the correlation between trends in different parameters. These include cardiovascular compensation models, renal perfusion models, and metabolic state models. The cardiovascular compensation model describes a negative feedback relationship between mean arterial pressure (MAP) and heart rate; that is, when MAP decreases, heart rate typically increases compensatorily. The renal perfusion model describes a positive correlation between MAP and urine output; that is, when MAP is maintained above a certain level, renal perfusion is adequate, and urine output remains normal. The metabolic state model describes the relationship between tissue perfusion and blood lactate levels; that is, when tissue perfusion is insufficient, anaerobic metabolism increases, leading to elevated blood lactate levels. These models can characterize the dynamic correlation between physiological parameters using mathematical functions or rule sets. For example, the cardiovascular compensation model can be a linear or nonlinear function that maps the rate of change in MAP to the expected rate of change in heart rate. The renal perfusion model can define a threshold below which the rate of change in urine output should be negative. The metabolic state model can establish an inverse correlation between the rate of change in blood lactate and the rate of change in MAP. When the consistency verification algorithm is executed, these models are run simultaneously. The actual rate of change of the first and second physiological parameters is input into the corresponding models, and the consistency between the actual changes and the model predictions is compared.

[0088] When the algorithm detects that the trend of the first physiological parameter deviates significantly from the normal physiological logic along with the trends of one or more other second physiological parameters, it determines that the data for the first physiological parameter has low reliability and generates a low-reliability evaluation result. This step cross-validates the authenticity of the core data from a physiological logic perspective.

[0089] Furthermore, step S3 includes:

[0090] S31: Obtain a first rate of change of a first physiological parameter and a second rate of change of a second physiological parameter, wherein the second physiological parameter includes at least one of heart rate, urine volume and blood lactate;

[0091] S32: Calculate the physiological correlation between the first rate of change and the second rate of change;

[0092] S33: When the physiological correlation is lower than the preset correlation threshold, the data reliability evaluation result of the first physiological parameter is determined to be low reliability.

[0093] This step is a concrete implementation of how to assess data reliability through multi-parameter correlation analysis.

[0094] Specifically, the control unit maintains a short-term data buffer, storing various physiological parameters over a past period, such as the past hour. The program calculates the rate of change (i.e., slope) of the first physiological parameter, such as mean arterial pressure, within a specific time window, such as 5 minutes. Simultaneously, the program also calculates the rate of change of second physiological parameters, such as heart rate, hourly urine output, and blood lactate concentration.

[0095] The program then evaluates whether the rates of change of these different parameters are logically consistent. The specific evaluation method is as follows:

[0096] The calculation of physiological correlation is based on vector analysis of changes in multidimensional physiological parameters. The first rate of change represents the trend of mean arterial pressure within a specific sliding time window. The second rate of change covers the trends of parameters such as heart rate, urine output, or blood lactate within the same or physiologically lagging time windows.

[0097] The calculation employs a weighted correlation analysis method. First, a positive and negative correlation matrix is ​​established based on clinical circulatory dynamics logic. For example, a decrease in blood pressure accompanied by a compensatory increase in heart rate is defined as a normal physiological correlation. After vectorizing the rate of change of each parameter, the cosine of the angle between the first and second rate of change vectors in the physiological logic space is calculated. If the calculated correlation score is close to one, it indicates that the changes in various vital signs conform to the inherent coupling law of the human circulatory system. If the correlation score is close to zero or even negative, it indicates that the change in the first physiological parameter is isolated from the overall physiological state, and the risk of physical interference or sensor failure in the data source is extremely high. The corresponding reliability evaluation result is marked as low reliability.

[0098] In a specific embodiment, a multivariate correlation assessment model can be constructed. This model can be implemented as a rule-based and statistical analysis-based software module. This module contains a physiological rule base that defines the expected correlation patterns between key physiological parameters under different treatment interventions. For example, when the dose of vasopressors increases, the rule base expects the rate of change in mean arterial pressure to be positive, the rate of change in heart rate to be negative or zero, and the rate of change in urine volume to be positive. The program acquires the rates of change of mean arterial pressure, heart rate, and urine volume in real time over the past five minutes, forming an actual rate of change vector. Simultaneously, based on the current treatment intervention type, the corresponding expected rate of change vector is extracted from the rule base. Subsequently, the program calculates the Pearson correlation coefficient between these two vectors as the physiological correlation. For example, if the actual rate of change vector is [+2 mmHg / min, -5 bpm / min, +1 mL / min], and the expected rate of change vector is [+3 mmHg / min, -8 bpm / min, +2 mL / min], the correlation coefficient is calculated. If the correlation coefficient is lower than the preset threshold of 0.3, the program will determine that the mean arterial pressure data point is likely inaccurate because it contradicts the trends of other physiological indicators in a way that cannot be explained physiologically. Therefore, the program will classify the reliability evaluation result of the first physiological parameter as low and output this conclusion to the arbitration judgment module.

[0099] After evaluating the physical state of the sensors and the physiological logic of the data, the safety arbitration mechanism assesses whether the patient's actual response to the treatment intervention meets expectations. Specifically, it compares the real-time trend of the first physiological parameter with the expected trend based on the treatment intervention intensity information and a preset response model to obtain a response deviation evaluation result. For the same patient, the response to vasopressors is not constant, but it usually follows individual pharmacological patterns. The control unit establishes and continuously updates an individualized pharmacodynamic response model for each patient. This model learns and records the magnitude and rate of blood pressure changes caused by different doses of vasopressors during recent treatments. When the intelligent control logic executes a treatment intervention, such as increasing the infusion rate of vasopressors by 0.02 micrograms per kilogram of body weight per minute, this individualized pharmacodynamic model predicts, based on historical data, the expected trajectory of mean arterial pressure within the next time window, for example, an increase of 3 mmHg within 5 minutes. The control unit closely monitors the actual magnitude and rate of increase of the first physiological parameter and compares it with the expected value predicted by the model in real time. If the actual blood pressure response is significantly slower or weaker than expected, the difference between the two will increase. This difference, after calculation and normalization, forms the response bias assessment result. A high response bias assessment result indicates that the patient's response to treatment is abnormal. Possible reasons include the patient developing drug resistance, or, more alarmingly, the actual blood pressure is not low, but the measurement value fails to reflect the true increase due to signal damping.

[0100] Furthermore, step S4 includes:

[0101] S41: Input the treatment intervention intensity information into the individualized pharmacodynamic model for the controlled subject to obtain the expected increase in physiological parameters under the current intervention intensity;

[0102] S42: Calculate the difference between the actual increase in the first physiological parameter and the expected increase in the physiological parameter;

[0103] S43: Determine the response deviation evaluation result based on the difference, where the larger the difference, the higher the degree of deviation represented by the response deviation evaluation result.

[0104] This step details the implementation of how to quantify the deviation between a patient's treatment response and expectations. The control unit maintains a dynamically updated, individualized pharmacodynamic model for each patient. Initially, this model may be a standard pharmacodynamic model based on population data, but as treatment progresses, it continuously learns and corrects parameters online using the patient's own treatment intervention data, such as the history of vasopressor dosage adjustments, and physiological parameter response data, such as the corresponding history of mean arterial pressure changes. For example, recursive least squares or Kalman filtering algorithms can be used to adjust key parameters characterizing the patient's drug sensitivity in the model in real time.

[0105] This step details the implementation of how to quantify the deviation between a patient's treatment response and expectations. The control unit maintains a dynamically updated, individualized pharmacodynamic model for each patient. Initially, this model may be a standard pharmacodynamic model based on population data, but as treatment progresses, it continuously learns and corrects parameters online using the patient's own treatment intervention data, such as the history of vasopressor dosage adjustments, and physiological parameter response data, such as the corresponding history of mean arterial pressure changes. For example, recursive least squares or Kalman filtering algorithms can be used to adjust key parameters characterizing the patient's drug sensitivity in the model in real time.

[0106] The personalized pharmacodynamic model is constructed using a nonlinear least-squares fitting algorithm based on the patient's historical response data. The model takes the drug administration intensity as input and the mean arterial pressure response as output, maintaining the patient's drug sensitivity coefficient in real time. When executing a dose adjustment command, the model combines the current dosing baseline with the planned dose increment to calculate the expected increase in blood pressure for the patient in their current physiological state. This prediction process considers the drug's distribution and kinetic characteristics in the body, establishing a nonlinear mapping relationship between drug concentration and physiological effects to predict the pressure level that should be reached at a specific time point after administration. Using this predicted value as a standard reference, false unresponsiveness caused by monitoring signal delays or damping can be accurately identified.

[0107] When a new treatment intervention occurs, for example, if the infusion pump reports an increase in the vasopressor dose from 0.1 micrograms per kilogram of body weight per minute to 0.12 micrograms per kilogram of body weight per minute, this increment is fed as input into the patient's current individualized pharmacodynamic model. The model then outputs a prediction of how much the patient's mean arterial pressure is expected to rise over the next 5 minutes; for example, the model predicts an expected increase of 4 mmHg in this physiological parameter.

[0108] Meanwhile, the control unit continuously monitors the actual changes in the primary physiological parameter. After a 5-minute observation window, the program calculates the actual increase in mean arterial pressure, for example, only 1 mmHg. The program then calculates the difference between these two values: 4 minus 1 equals 3 mmHg. This difference directly reflects the degree of deviation between the actual response and the individualized expectation. For ease of subsequent processing, this difference is normalized, for example, by dividing by the expected increase to obtain a relative deviation value, or mapped to a scoring range of 0 to 1, forming the final response deviation assessment result. In this example, a larger difference corresponds to a higher response deviation assessment value, clearly indicating that the patient's response to this treatment fell far short of expectations based on their historical patterns.

[0109] After independent evaluation across the three dimensions mentioned above, the control unit obtains three key evaluation indicators: a damping characteristic value reflecting the physical state of the sensor, a data reliability evaluation result reflecting the authenticity of the data, and a response deviation evaluation result reflecting the effectiveness of the treatment. Subsequently, the arbitration step proceeds to determine whether the damping characteristic value, data reliability evaluation result, and response deviation evaluation result meet the preset arbitration triggering conditions.

[0110] As a specific implementation method, the arbitration triggering condition determination logic adopts a multi-factor concurrent triggering mode. In highly dynamic intensive care scenarios, if the damping characteristic value shows that the resonant frequency of the monitoring channel has dropped below five Hz, it means that the monitoring system can no longer accurately capture the rapid pressure fluctuations of the cardiac cycle. At the same time, the consistency verification module finds that the patient's real-time urine output remains above fifty milliliters per hour and the blood lactate level is steadily decreasing, which creates a serious physiological logic conflict with the low blood pressure value shown by the first physiological parameter. Even if the response deviation evaluation has not exceeded the standard at this time, the arbitration determination module will still immediately output a trigger signal because the first two indicators have triggered the preset dangerous logic combination. This multi-dimensional complementary determination logic not only considers the absolute value of a single indicator, but also emphasizes the mutual verification between physical characteristics and physiological logic. Through this rigorous arbitration mechanism, it is possible to provide early warning and cut off potential dangerous drug administration cycles at the initial stage of signal damping, or even before the waveform morphology is completely distorted, based on cross-parameter logical contradictions, thereby minimizing iatrogenic harm.

[0111] Furthermore, in step S5, the arbitration triggering conditions include: the damping characteristic value exceeds a preset damping threshold; or, the data reliability evaluation result is low reliability; or, the response deviation evaluation result exceeds a preset deviation threshold.

[0112] This constraint is a clearly defined logical OR gate structure, embodying the principle of safety priority. The safety arbitration mechanism should be triggered if any of the three evaluation dimensions triggers a red light. Specifically, the control unit presets a series of thresholds. For example, the damping coefficient in the damping characteristic value might have a preset damping threshold of 0.7, which is generally considered in engineering to be the boundary between critically damped and overdamped states of the system. The data reliability evaluation result is a classification result; when it is judged as low reliability, the condition is met. The response deviation evaluation result, if quantified as a score from 0 to 1, might have a preset deviation threshold of 0.8, indicating that the actual response deviates significantly from the expected response. In actual operation, the control unit continuously compares the calculated real-time evaluation values ​​with these preset thresholds. Once any condition is found to be met—for example, if the real-time calculated damping coefficient rises to 0.75, exceeding the 0.7 threshold—the arbitration judgment module will immediately output a trigger signal, initiating the subsequent constraint processing flow.

[0113] If any one or more of the above three evaluation indicators reach a preset risk threshold, thus fulfilling the arbitration trigger condition, the method will execute the most critical protective action: constraining the adjustment command for the treatment intervention intensity to keep it within a preset safety threshold or maintain its current state. This means that the main control logic of the intelligent control system, the automatic dose adjustment program designed to raise blood pressure to the target value, will have its output commands intercepted or restricted by the safety arbitration mechanism. For example, even if the main control logic calculates that the vasopressor dose needs to be further increased based on the received false hypotension data, this instruction to increase the dose will be blocked by the constraint execution module. In this way, the method effectively cuts off the dangerous positive feedback loop that may be formed due to data distortion, avoids the unlimited accumulation of vasopressor drugs, thereby controlling the patient's actual blood pressure within a safe range and preventing iatrogenic hypertension and its complications.

[0114] Furthermore, step S6 includes:

[0115] S61: Intercept the incremental part of the adjustment instruction regarding increasing the intensity of treatment intervention, so that the intensity of treatment intervention is maintained at the current level;

[0116] Alternatively, based on the weighted values ​​of the damping characteristic value, data reliability evaluation results, and response deviation evaluation results, the output weight corresponding to the adjustment command can be reduced to limit the rate of increase in the intensity of the treatment intervention and keep the intensity of the treatment intervention within a preset safe threshold range.

[0117] Furthermore, after the arbitration triggering conditions are met, the steps for executing constraint processing on the adjustment command can specifically include two methods. The first method is to intercept the incremental portion of the adjustment command regarding increasing the intensity of the treatment intervention, maintaining the intensity of the treatment intervention at the current level. This is a hard-freeze strategy. When the arbitration mechanism is triggered, especially when a clear sensor physical fault is detected, such as an excessive damping characteristic value or a serious data logic contradiction, such as low data reliability, the control unit will immediately take the safest measures. At this time, the constraint execution module will completely block any instructions to increase the dose from the upper-level automatic dose adjustment logic. For example, if the current vasopressor infusion rate is 0.15 micrograms per kilogram of body weight per minute, and the main control logic calculates based on distorted data that it needs to be increased to 0.18, the constraint execution module will ignore this incremental instruction and continuously send instructions to the infusion pump to maintain the dose at 0.15. This method can stop dangerous positive feedback as quickly as possible, locking the treatment intensity at the level at which arbitration was triggered, buying valuable time for medical staff to investigate the problem.

[0118] The second approach involves adjusting the output weight of the adjustment command based on the weighted values ​​of the damping characteristic value, data reliability evaluation results, and response deviation evaluation results. This limits the rate of increase in treatment intervention intensity, keeping it within a preset safe threshold range. This is a softer constraint strategy. In some cases, such as when only the response deviation evaluation result slightly exceeds the limit, while the sensor's physical state and data consistency are acceptable, directly freezing the treatment might be too conservative. In this situation, the constraint execution module first quantifies the degree of exceeding the limit for the three evaluation indicators and then performs a weighted sum based on preset weighting factors to obtain a comprehensive risk score. For example, the damping characteristic value has the highest weight, followed by data reliability, and then the response deviation has the lowest weight.

[0119] This comprehensive risk score is mapped to an output weighting coefficient between 0 and 1. Subsequently, incremental instructions issued by the main control logic are multiplied by this output weighting coefficient before execution. For example, if the main logic requests an increase of 0.02 micrograms per kilogram of body weight per minute, and the calculated output weight is 0.2, the actual incremental instruction issued to the infusion pump will only be 0.004. In this way, without completely stopping treatment adjustments, the rate of increase in treatment intensity is significantly slowed, achieving dynamic and hierarchical risk management.

[0120] As a specific implementation method, in the context of bedside monitoring equipment with limited computing power, dynamic management of drug administration risks is achieved through weighted coefficient allocation logic. When the damping coefficient of the arterial pressure measuring catheter is detected to slowly rise from 0.35 to 0.68, although it has not completely reached the threshold for forced freezing, the calculated comprehensive risk weighted score will cross the preset yellow warning line due to fluctuations in the data reliability score and a gradually increasing trend in response deviation. At this point, the drug administration increment command originally calculated by the automatic control algorithm will no longer be executed in full, but will be corrected by an attenuation coefficient generated in real time based on the weighted score. If the attenuation coefficient is determined to be 0.3, the originally planned increase in the drug administration rate of 0.1 micrograms per kilogram of body weight per minute will be limited to 0.03 micrograms per kilogram of body weight per minute in the actual command output stage. This weighted soft constraint processing method significantly extends the drug dosage adjustment cycle without completely interrupting automated treatment. This gradual dosing restriction mechanism effectively buffers the impact of misjudgments caused by sensor data distortion, providing nurses with sufficient reaction time to check the status of tubing and eliminate physical interference on-site, ensuring that patients do not receive excessive amounts of vasoactive drugs in a short period of time.

[0121] After imposing constraints on the intensity of the treatment intervention, the process does not end but enters a phase of continuous monitoring and recovery guidance. Furthermore, step S6 onwards includes:

[0122] S7: Continuously monitor the state changes of damping characteristic values, data reliability evaluation results, and response deviation evaluation results;

[0123] S8: When the damping characteristic value is determined to return to the preset normal range and the data reliability evaluation result turns to high reliability, the recovery boot program is started.

[0124] When treatment intensity is frozen or limited, the control unit will issue a clear alert to healthcare professionals, indicating potential data distortion and suggesting checks. For example, a notification may appear on the monitor screen indicating that the arterial pressure monitoring signal may be damped, automatic adjustment of vasopressors is limited, and to check the arterial pressure measurement tubing. During this period, the control unit will not cease operation but will continuously monitor the status changes of damping characteristic values, data reliability evaluation results, and response deviation evaluation results at a higher frequency. Upon receiving the alert, healthcare professionals may flush the arterial pressure measurement tubing, adjust the patient's position, or replace the sensor. The control unit will capture the effects of these interventions in real time. For example, after successful tubing flushing, the arterial pressure waveform will quickly return to normal, and the calculated damping characteristic value will rapidly decrease to the normal range. Simultaneously, because accurate blood pressure data has been obtained, its consistency with other physiological parameters will be restored, and the data reliability evaluation result will turn high reliability.

[0125] The system will only consider releasing constraints when the two most critical issues—sensor physics and data logic—are clearly resolved. For example, the system will only initiate the recovery bootstrap procedure and prepare to safely return control to the main control logic when both conditions are met: the control unit determines that the damping characteristic value has returned to the preset normal range, such as a damping coefficient of less than 0.4, and the data reliability evaluation result has turned to high reliability.

[0126] To avoid shocking the patient with drastic changes in treatment intensity after restraint removal during the initiation of the recovery guidance procedure, a safe and gradual recovery strategy is employed. Specifically, step S8, initiating the recovery guidance procedure, includes the following steps:

[0127] S81: Obtain the real-time values ​​of the physiological parameters of the controlled object at the current moment and use them as the safety starting benchmark;

[0128] S82: Calculate the allowable recovery increment per unit time based on the difference between the safe starting baseline and the preset treatment target value, combined with an individualized pharmacodynamic model;

[0129] S83: Increase the upper limit of the treatment intervention intensity in stages according to the allowed recovery increment, until the treatment intervention intensity is restored to the real-time target level determined by the automatic dose adjustment logic.

[0130] In one specific implementation, after catheter flushing or sensor replacement, when the monitored pressure waveform shows a clear dicrotic notch again and the damping coefficient drops below 0.3, the recovery guidance logic is triggered. At this point, the real-time mean arterial pressure is 72 mmHg, which is recorded as the logical baseline for this recovery process. Considering the patient's sensitivity to circulatory fluctuations, the allowed recovery increment is set to ensure that the change in the dosing rate within each five-minute observation cycle does not exceed 10% of the initial frozen value. During the first observation cycle, the upper limit of the adjustment command output is slightly relaxed, allowing the control logic to adjust slowly while ensuring safety. If, during this process, the trend of the first physiological parameter remains highly consistent with the prediction of the individualized pharmacodynamic model, and the data reliability remains high, the output upper limit is linearly relaxed again in the next cycle. This step-by-step recovery path achieves a robust transition from the restrained protection state to the normal automatic control state, effectively avoiding sudden changes in dosing due to direct release of restraint, and ensuring a smooth transition of the patient's hemodynamic state.

[0131] First, the moment the data source is confirmed to be reliably restored, the control unit immediately acquires a current, accurate physiological parameter value; for example, the current mean arterial pressure is 72 mmHg. This value is established as the safe starting point for this recovery process.

[0132] Next, the program calculates a safe, permissible recovery increment per unit time based on the difference between this safe starting baseline and a preset treatment target value, such as 65 mmHg, and again by calling the updated individualized pharmacodynamic model. In this example, because the current blood pressure is higher than the target, the permissible recovery increment is actually a negative value, i.e., the permissible rate of decrease. If the current blood pressure is lower than the target, a permissible rate of increase is calculated. This calculation process ensures that subsequent adjustments are small, slow, and controlled.

[0133] Next, the program calculates a safe, permissible recovery increment per unit time based on the difference between this safe starting baseline and a preset treatment target value, such as 65 mmHg, and again by calling the updated individualized pharmacodynamic model. In this example, because the current blood pressure is higher than the target, the permissible recovery increment is actually a negative value, i.e., the permissible rate of decrease. If the current blood pressure is lower than the target, a permissible rate of increase is calculated. This calculation process ensures that subsequent adjustments are small, slow, and controlled.

[0134] Finally, the program gradually and progressively loosens the output limit on the intensity of the treatment intervention. Instead of immediately relinquishing control to the main control logic, the program first sets a slightly relaxed output limit; for example, allowing the main logic to reduce the dosage by a maximum of 0.01 micrograms per kilogram of body weight per minute over the next 5 minutes. After observing for 5 minutes and confirming a stable patient response, this limit is further relaxed. This process continues, like a gradually opening valve, until the output limit is fully restored to the ideal target level determined by the automatic dose adjustment logic based on real-time, accurate physiological parameters. This smooth transition achieves a safe and seamless switch from a safety-constrained state to a normal intelligent control state.

[0135] Please refer to Figure 2 , Figure 3 This application also provides an intelligent control system for intensive care unit equipment, the system being used to implement any of the above methods, the system comprising:

[0136] The information acquisition module 201 is used to acquire information on the intensity of the treatment intervention for the controlled object, and simultaneously acquire the first physiological parameter and at least one second physiological parameter fed back by the pressure monitoring channel;

[0137] The feature analysis module 202 is used to analyze the waveform features of the first physiological parameter and extract the damping feature value that characterizes the physical transmission characteristics of the pressure monitoring channel.

[0138] Consistency verification module 203 is used to perform consistency verification on the first physiological parameter in conjunction with the second physiological parameter, and determine the data reliability evaluation result of the first physiological parameter;

[0139] The trend comparison module 204 is used to compare the real-time change trend of the first physiological parameter with the expected change trend obtained based on the treatment intervention intensity information and the preset response model to obtain the response deviation evaluation result.

[0140] Arbitration judgment module 205 is used to determine whether the damping characteristic value, data reliability evaluation result and response deviation evaluation result meet the preset arbitration triggering conditions;

[0141] The constraint execution module 206 is used to perform constraint processing on the adjustment command of the treatment intervention intensity if the arbitration triggering condition is met, so as to keep the treatment intervention intensity within a preset safety threshold range or maintain the current state.

[0142] Specifically, the information acquisition module 201 can be a data acquisition unit that connects to various medical devices (such as ventilators, infusion pumps, monitors, etc.) via data interfaces to collect patients' physiological data (such as arterial pressure, heart rate, urine output, blood lactate, etc.) and information on the intensity of currently applied treatment interventions in real time. This module aims to ensure that all necessary data can be collected in a timely and accurate manner, providing a foundation for subsequent analysis and processing.

[0143] The feature analysis module 202 can be a signal processing unit that receives first physiological parameters (such as arterial pressure waveform data) from the information acquisition module and analyzes them using digital signal processing algorithms to identify key features in the waveform, such as the slope of the initial ascending limb of the arterial pressure waveform and the dicrotic notch feature. Based on these features, the resonant frequency and damping coefficient of the pressure monitoring channel can be calculated. These parameters together constitute the damping characteristic value, used to assess whether the physical transmission characteristics of the pressure monitoring channel are abnormal.

[0144] In practical applications, the consistency verification module 203 can be a data verification unit that cross-compares the first physiological parameter with at least one second physiological parameter (such as heart rate, urine volume, blood lactate, etc.). For example, by analyzing the physiological correlation between them, it can determine whether there are abnormal fluctuations or measurement errors in the first physiological parameter. When the physiological correlation is lower than a preset correlation threshold, the data reliability evaluation result of the first physiological parameter is determined to be low reliability, thereby identifying potential data quality problems.

[0145] Specifically, the trend comparison module 204 can be a model prediction and comparison unit. It utilizes an individualized pharmacodynamic model and a pre-defined response model to predict the expected trend or magnitude of increase in the first physiological parameter based on the current treatment intervention intensity information. Subsequently, this module compares the actual monitored real-time trend of the first physiological parameter with the predicted expected trend, calculates the difference between the two, and determines the response deviation evaluation result based on the magnitude of the difference to assess whether there is a significant deviation between the patient's actual response to the treatment intervention and the expected response.

[0146] The arbitration decision module 205 can be a decision logic unit that continuously receives damping characteristic values ​​from the feature parsing module, data reliability evaluation results from the consistency verification module, and response deviation evaluation results from the trend comparison module. This module has built-in preset arbitration trigger conditions, such as whether the damping characteristic value exceeds a preset damping threshold, whether the data reliability evaluation result is low reliability, or whether the response deviation evaluation result exceeds a preset deviation threshold. Once any condition or combination of conditions is met, the arbitration decision module triggers the arbitration mechanism.

[0147] Specifically, the constraint execution module 206 can be an instruction intervention unit that intercepts or corrects treatment intervention intensity adjustment instructions issued by the doctor or automatic dosage adjustment logic when the arbitration judgment module triggers the arbitration condition. For example, it can intercept the incremental portion of increasing treatment intervention intensity to maintain the treatment intervention intensity at the current level; or, based on the weighted value of various evaluation results, it can lower the output weight corresponding to the adjustment instruction to limit the rate of increase of treatment intervention intensity, ensuring that the treatment intervention intensity does not exceed a preset safety threshold, thereby effectively avoiding potential medical risks.

[0148] The intelligent control system for critical care equipment in this application transforms the logical steps of the aforementioned methods into executable physical or software functions through an integrated modular design. The information acquisition module 201, serving as the data entry point, ensures the real-time and comprehensive collection of physiological parameters and treatment intervention intensity information, laying the foundation for subsequent intelligent analysis. The feature parsing module 202 and the consistency verification module 203 evaluate the quality and reliability of the first physiological parameter from two dimensions: physical transmission characteristics and physiological correlation, effectively identifying potential measurement errors or equipment malfunctions. The trend comparison module 204 reveals individual differences or abnormal responses in patients to treatment interventions by comparing actual responses with expected responses. The arbitration judgment module 205, as the core decision-making unit, comprehensively evaluates these multi-dimensional risk indicators and immediately triggers the arbitration mechanism upon detecting any abnormal situation that may endanger patient safety. Finally, the constraint execution module 206 can intervene and correct treatment intervention instructions in a timely and automatic manner, thereby transforming abstract control logic into concrete safety measures and effectively avoiding treatment risks caused by unreliable data, equipment malfunctions, or abnormal patient responses.

[0149] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for intelligent control of equipment in intensive care units, characterized in that, The method includes the following steps: S1: Obtain information on the intensity of the treatment intervention for the controlled subject, and simultaneously obtain the first physiological parameter and at least one second physiological parameter fed back by the pressure monitoring channel; S2: Analyze the waveform characteristics of the first physiological parameter and extract the damping characteristic value that characterizes the physical transmission characteristics of the pressure monitoring channel; S3: Combine the second physiological parameter with the first physiological parameter to perform a consistency check and determine the data reliability evaluation result of the first physiological parameter; S4: Compare the real-time change trend of the first physiological parameter with the expected change trend obtained based on the treatment intervention intensity information and the preset response model to obtain the response deviation evaluation result; S5: Determine whether the damping characteristic value, the data reliability evaluation result, and the response deviation evaluation result meet the preset arbitration triggering conditions; S6: If the arbitration triggering condition is met, the adjustment instruction for the intensity of the treatment intervention is constrained to keep the intensity of the treatment intervention within a preset safety threshold range or maintain the current state.

2. The intelligent control method for intensive care unit equipment according to claim 1, characterized in that, Step S2 includes: S21: Identify the slope of the initial ascending branch and the dicrotic notch features in the arterial pressure waveform of the first physiological parameter; S22: Calculate the resonant frequency and damping coefficient of the pressure monitoring channel based on the initial rising slope and the dicrotic notch characteristics, and use the resonant frequency and the damping coefficient as the damping characteristic value.

3. The intelligent control method for intensive care unit equipment according to claim 1, characterized in that, Step S3 includes: S31: Obtain the first rate of change of the first physiological parameter and the second rate of change of the second physiological parameter, wherein the second physiological parameter includes at least one of heart rate, urine volume and blood lactate; S32: Calculate the physiological correlation between the first rate of change and the second rate of change; S33: When the physiological correlation is lower than the preset correlation threshold, the data reliability evaluation result of the first physiological parameter is determined to be low reliability.

4. The intelligent control method for intensive care unit equipment according to claim 1, characterized in that, Step S4 includes: S41: Input the treatment intervention intensity information into the individualized pharmacodynamic model for the controlled object to obtain the expected increase in physiological parameters under the current intervention intensity; S42: Calculate the difference between the actual increase in the first physiological parameter and the expected increase in the physiological parameter; S43: Determine the response deviation evaluation result based on the difference, wherein the larger the difference, the higher the degree of deviation represented by the response deviation evaluation result.

5. The intelligent control method for intensive care unit equipment according to claim 1, characterized in that, In step S5, the arbitration triggering conditions include: the damping characteristic value exceeds a preset damping threshold; or, the data reliability evaluation result is low reliability; or, the response deviation evaluation result exceeds a preset deviation threshold.

6. The intelligent control method for intensive care unit equipment according to claim 1, characterized in that, Step S6 includes: S61: Intercept the incremental portion of the adjustment instruction regarding increasing the intensity of the treatment intervention, so that the intensity of the treatment intervention remains at the current level; Alternatively, based on the weighted value of the damping characteristic value, the data reliability evaluation result, and the response deviation evaluation result, the output weight corresponding to the adjustment command can be reduced to limit the rate of increase of the treatment intervention intensity, so that the treatment intervention intensity is kept within a preset safe threshold range.

7. The intelligent control method for intensive care unit equipment according to claim 1, characterized in that, Step S6 and following it include: S7: Continuously monitor the state changes of the damping characteristic value, the data reliability evaluation result, and the response deviation evaluation result; S8: When it is determined that the damping characteristic value has recovered to the preset normal range and the data reliability evaluation result has turned to high reliability, the recovery boot program is started.

8. The intelligent control method for intensive care unit equipment according to claim 1, characterized in that, In step S8, starting the recovery bootloader includes the following steps: S81: Obtain the real-time values ​​of the physiological parameters of the controlled object at the current moment, and use them as the safety starting benchmark; S82: Calculate the allowable recovery increment per unit time based on the difference between the safety starting benchmark and the preset treatment target value, combined with the individualized pharmacodynamic model; S83: Increase the upper limit of the treatment intervention intensity in stages according to the allowed recovery increment, until the treatment intervention intensity is restored to the real-time target level determined by the automatic dose adjustment logic.

9. The intelligent control method for intensive care unit equipment according to claim 1, characterized in that, Step S1 includes: S11: The first physiological parameter, the second physiological parameter, and the treatment intervention intensity information are collected in real time via a data bus at a preset sampling frequency, wherein the preset sampling frequency is not less than 10Hz.

10. An intelligent control system for critical care equipment, characterized in that, The system is used to implement the method as described in any one of claims 1-9, the system comprising: The information acquisition module is used to acquire information on the intensity of the treatment intervention for the controlled object, and simultaneously acquire the first physiological parameter and at least one second physiological parameter fed back by the pressure monitoring channel; The feature analysis module is used to analyze the waveform features of the first physiological parameter and extract the damping feature value that characterizes the physical transmission characteristics of the pressure monitoring channel. The consistency verification module is used to perform consistency verification on the first physiological parameter in conjunction with the second physiological parameter, and to determine the data reliability evaluation result of the first physiological parameter. The trend comparison module is used to compare the real-time change trend of the first physiological parameter with the expected change trend obtained based on the treatment intervention intensity information and the preset response model to obtain the response deviation evaluation result. The arbitration judgment module is used to determine whether the damping characteristic value, the data reliability evaluation result, and the response deviation evaluation result meet the preset arbitration triggering conditions; The constraint execution module is used to perform constraint processing on the adjustment command of the treatment intervention intensity if the arbitration triggering condition is met, so as to keep the treatment intervention intensity within a preset safety threshold range or maintain the current state.

Images