Body position hypotension intervention system based on dynamic risk assessment

Abstract

The invention provides a dynamic risk assessment-based postural hypotension intervention system, which comprises a postural monitoring module for acquiring postural information and postural change parameters of a user in real time, and judging whether the user has rapid postural change possibly inducing hypotension or not according to a preset threshold value; the blood pressure monitoring module is used for monitoring blood pressure parameters of the user in real time to judge whether the blood pressure drops abnormally and send an alarm signal; the environment sensing module is used for providing environment factor parameters; the central control unit communicates with the body position monitoring module, the blood pressure monitoring module and the environment sensing module and is used for fusing and processing characteristic parameters collected by all the modules and carrying out risk score R calculation; and the stimulation execution module is used for performing electrical stimulation intervention according to the instruction sent by the central control unit. Through multi-dimensional data fusion, real-time dynamic monitoring and personalized intervention, a complete solution is provided for management of postural hypotension.

Description

Technical Field

[0001] The present invention belongs to the field of orthostatic hypotension treatment, and particularly relates to an orthostatic hypotension intervention system based on dynamic risk assessment. Background Art

[0002] Orthostatic Hypotension (OH) is a syndrome characterized by a significant drop in blood pressure caused by a change in body position, commonly seen in the elderly, patients with autonomic neuropathy diseases, and during certain drug treatments. According to the classical definition, orthostatic hypotension is manifested as a systolic blood pressure drop of ≥20 mmHg and / or a diastolic blood pressure drop of ≥10 mmHg within 3 minutes of standing. This blood pressure regulation disorder can lead to dizziness, fatigue, blurred vision, and in severe cases, syncope, falls, and even endanger life. The harms of orthostatic hypotension are mainly reflected in the following aspects: In daily life, when patients change their body positions (such as from lying to standing), a sudden drop in blood pressure may occur, leading to sudden syncope and falls. Especially for elderly patients, such situations often result in serious consequences such as fractures. In addition, orthostatic hypotension usually coexists with chronic diseases such as cardiovascular diseases and diabetic autonomic neuropathy, which may accelerate the deterioration of the patient's condition. Complications such as falls and syncope caused by orthostatic hypotension will also significantly increase the consumption of medical resources and at the same time reduce the patient's quality of life.

[0003] Currently, the monitoring and intervention methods for orthostatic hypotension mainly focus on two aspects: drug treatment and wearable devices. Drug treatment is an important means of orthostatic hypotension intervention, and commonly used drugs include vasopressors and drugs that increase blood volume. These drugs can relieve symptoms in the short term, but it is difficult to effectively control the dynamically changing orthostatic hypotension. There are two main problems with drug treatment: one is the large individual differences, the drug efficacy varies from person to person, and at the same time, adverse reactions may occur, such as causing too high blood pressure; the other is the lack of dynamic regulation. Drug treatment cannot effectively respond to the instantaneous blood pressure drop caused by rapid body position changes, resulting in a lag in intervention and making it difficult to prevent acute events.

[0004] In recent years, wearable devices have gradually been applied to the monitoring of blood pressure, heart rate, and body position changes. These devices usually integrate sensors such as accelerometers and gyroscopes, which can collect the patient's movement data. Some devices also incorporate blood pressure measurement functions to achieve early monitoring of orthostatic hypotension. Although wearable devices have made certain progress in the monitoring and intervention of orthostatic hypotension, there are still significant deficiencies and cannot meet the needs of orthostatic hypotension patients for daily dynamic monitoring and immediate intervention.

[0005] The occurrence of orthostatic hypotension is the result of the combined action of various factors such as body position changes, blood pressure fluctuations, and neural regulation. Existing technologies usually only focus on data in a single dimension (such as blood pressure, heart rate, or body position), ignoring the correlation between parameters, resulting in the inability to accurately predict the risk of orthostatic hypotension. Specifically, existing technologies have the following deficiencies: First, the combination of body position monitoring and blood pressure monitoring is insufficient. Separate body position change monitoring cannot identify the magnitude of blood pressure drop, and separate blood pressure monitoring cannot correlate with the speed and pattern of body position changes. Second, the neglect of environmental factors is another major problem. Environmental factors such as air pressure, temperature, and humidity, which have a significant impact on blood pressure and nerve function, are not included in the monitoring scope, resulting in poor prediction effects of existing devices in special environments.

[0006] The risk of orthostatic hypotension mainly lies in acute events caused by a rapid drop in blood pressure. However, existing technologies are mostly used for post-event recording and lack the ability of real-time dynamic risk assessment and prediction. First, existing devices generally lack a real-time warning mechanism and cannot issue a risk warning when blood pressure begins to drop, resulting in patients possibly fainting without any warning. Second, existing algorithms rely on static threshold judgments and cannot dynamically adjust the risk assessment model according to the patient's real-time status and historical data, resulting in the lag of the algorithm.

[0007] In addition to the deficiencies in monitoring, current wearable devices mostly focus on passively recording events and lack active intervention measures. For example, the level of hierarchical intervention is insufficient. Existing devices cannot provide personalized intervention strategies according to the patient's risk level, such as adjusting heart rate, prompting posture changes, or triggering drug release. In addition, the instantaneous response ability of existing devices is weak. Orthostatic hypotension usually occurs during rapid body position changes. Existing devices fail to combine real-time monitoring and intervention, resulting in lagged or inaccurate intervention.

[0008] Insufficient environmental adaptability is also one of the main limitations of existing technologies. Patients with orthostatic hypotension are more likely to develop the disease in extreme temperature and humidity environments, but existing technologies generally do not consider the impact of the environment on blood pressure regulation and patient health. For example, temperature and humidity have a significant impact on human blood circulation and nerve function, but existing technologies rarely include environmental parameters in the monitoring and intervention scope.

[0009] Insufficient portability and battery life limit the popularization and application of long-term management devices for orthostatic hypotension. Some existing devices are large in size and not convenient for daily wearing, affecting patient compliance. In addition, the battery life is insufficient, making it difficult for the device to meet the needs of long-term dynamic monitoring and restricting its use in daily life scenarios.

[0010] Based on the deficiencies of the existing technology, it is of great significance to propose a postural hypotension intervention device based on dynamic risk assessment. This device aims to provide a complete solution for the management of postural hypotension through multi-dimensional data fusion, real-time dynamic monitoring, and personalized intervention. Such a device can not only comprehensively monitor body position, blood pressure, and environmental factors, but also dynamically adjust the risk assessment model according to real-time data, and combine with a hierarchical intervention strategy to provide timely and effective protection for patients, especially in high-risk situations or extreme environments, providing new technical support for the health management of postural hypotension patients. Summary of the Invention

[0011] The object of the present invention is to provide a postural hypotension intervention system based on dynamic risk assessment. The present invention aims to provide a complete solution for the management of postural hypotension through multi-dimensional data fusion, real-time dynamic monitoring, and personalized intervention.

[0012] To achieve the above object, the present invention provides the following technical solution: A postural hypotension intervention system based on dynamic risk assessment, comprising: A body position monitoring module, which is used to collect the posture information and body position change parameters of the user in real time, detect the posture angle θ and angular velocity ω by using a MEMS inertial sensor, and judge whether the user has a rapid body position change that may induce hypotension according to a preset threshold; A blood pressure monitoring module, which is used to monitor the blood pressure parameters of the user in real time, collect the pulse wave signal through a PPG sensor and extract the blood pressure value, calculate the blood pressure change amount ΔBP and the mean arterial pressure MAP, so as to determine whether the blood pressure drops abnormally and send an alarm signal; An environment perception module, which is used to collect the environmental temperature T and humidity H, calculate the environmental stress index E through a normalized weighted formula, and provide an environmental factor correction parameter for the risk score; A central control unit, which communicates with the body position monitoring module, the blood pressure monitoring module, and the environment perception module, and is used to fuse and process the characteristic parameters collected by each module and calculate the risk score R; A stimulation execution module, which is used to perform electrical stimulation intervention according to the instruction issued by the central control unit, and its stimulation parameters include stimulation frequency f and intensity I, and the stimulation parameters are determined according to the risk score R.

[0013] Furthermore, the risk score R is obtained by combining a fuzzy logic algorithm and a weighted algorithm: The fuzzy logic algorithm includes fuzzy processing of the posture angle θ, angular velocity ω, blood pressure change ΔBP, blood pressure drop rate, and environmental factors, and uses the minimum method (min) and maximum method (max) for reasoning based on a preset rule base to obtain a risk fuzzy set, and uses the centroid method for defuzzification calculation to obtain the risk score R F ; The weighted algorithm is based on the risk score R F to respectively determine the postural factor sub-score R P , the blood pressure factor sub-score R B , and the environmental factor sub-score R E , and through the weight coefficients ω P , ω B , ω E perform weighted calculation to obtain the comprehensive risk score R: R = R P ·ω P + R B ·ω B + R E ·ω E .

[0014] Furthermore, the stimulation parameters in the stimulation execution module include the stimulation frequency f and intensity I, and the stimulation parameters are determined according to the risk score R: Stimulation frequency: f = f min +(f max -f min )·R; Stimulation intensity: I = I min +(I max -I min )·R; Meanwhile, for heart rate regulation, the central control unit determines the target heart rate Ht according to the real-time feedback error e BP of blood pressure and the risk score R: Ht = H base +K BP ·e BP +(H max - H base )·R.

[0015] Furthermore, the central control unit adopts a closed-loop feedback mechanism to adjust the stimulation parameters in real time through a PI controller, where the control output increment ΔI is calculated as: ΔI = K p ·e BP +Ki∫e BP d t ; Meanwhile, a fuzzy control strategy is adopted to dynamically adjust the PI controller parameters K p and K i or directly adjust the control output to form an adaptive closed-loop feedback regulation integrating PI control and fuzzy control.

[0016] Furthermore, the postural monitoring module sets the attitude angle threshold, defines θ < 30° as standing, 30° ≤ θ ≤ 70° as sitting, and θ > 70° as lying flat.

[0017] Further, after the PPG signal in the blood pressure monitoring module is processed by a 0.5 - 8 Hz band - pass filter, the blood pressure value is estimated through a pre - calibrated linear model BP = k·PTT + b or an empirical formula.

[0018] Further, the environmental stress index E is calculated by a temperature - humidity normalization and weighting formula: E = ω T ·(T / T max ) + ω H ·(H / H max ), where ω T +ω H = 1.

[0019] Further, the stimulation site of the stimulation execution module is selected as the lower spinal cord segments T6 - T12 or the lower limb skeletal muscles, and the heart rate is adjusted by electrically stimulating the peripheral nerves, muscles or directly adjusting the pacemaker.

[0020] Further, the fuzzy - logic control strategy of the central control unit includes the following typical rules: When the blood pressure is significantly lower than the target and the risk score is high, maintain a high - intensity stimulation; When the blood pressure is close to the target and the risk score decreases, gradually reduce the stimulation intensity; When the blood pressure exceeds the target, quickly weaken or stop the stimulation.

[0021] Further, the closed - loop feedback regulation mechanism has an anti - integral saturation strategy to avoid overshoot of blood pressure caused by excessive stimulation.

[0022] The orthostatic hypotension intervention device and method based on dynamic risk assessment proposed by the present invention have the following remarkable advantages: By the posture monitoring module, the patient's posture angle and the rate of posture change are accurately monitored in real - time, and combined with the blood pressure monitoring module to continuously detect the patient's blood pressure change, realizing precise early warning of hypotensive events caused by rapid posture changes, effectively overcoming the problems of inaccurate posture recognition and lagging intervention start in the prior art.

[0023] By innovatively adopting a risk - scoring strategy that combines fuzzy logic and weighted algorithms, various influencing factors such as the patient's real - time posture, the amplitude and rate of blood pressure drop, and environmental factors can be comprehensively considered, realizing multi - dimensional precise risk assessment, and effectively solving the problem of insufficient risk assessment of a single physiological index in the prior art. The specific formulas and models used in the risk - scoring algorithm, such as the fuzzy inference and weighted formula of the risk score R, ensure the smoothness, accuracy and clinical applicability of the risk score.

[0024] The central control unit integrates fuzzy adaptive control and PI control algorithms, and accurately adjusts the electrical stimulation intensity (I), frequency (f), and target heart rate (Ht) according to the real-time risk score R and blood pressure feedback, achieving adaptive dynamic control of patient individual differences and non-linear physiological responses. The clear control parameter mapping formulas (such as stimulation frequency, stimulation intensity, and heart rate adjustment formula) given in the present invention make the intervention accurate, reliable, and easy to implement clinically.

[0025] (4) The introduction of the closed-loop feedback mechanism enables the system to automatically adjust the intervention intensity and frequency in real time according to the intervention effect, ensuring the stability and safety of the patient's blood pressure during rapid body position changes, and making it less likely to produce overshoot or insufficiency. Compared with traditional single control algorithms, this system greatly improves the guarantee of patient safety and significantly reduces the risk of falls caused by orthostatic hypotension.

[0026] In summary, the orthostatic hypotension prevention system provided by the present invention has the advantages of accurate early warning, comprehensive risk assessment, real-time adaptability of intervention control, etc., significantly improves the prevention effect and safety of patients against orthostatic hypotension, and has outstanding clinical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is a schematic diagram of the system of the present invention; Figure 2 is a schematic diagram of the data flow of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0028] An embodiment of the present invention provides an orthostatic hypotension intervention system based on dynamic risk assessment, as Figure 1 shown, including a body position monitoring module, a blood pressure monitoring module, an environmental perception module, a central control unit, and a stimulation execution module. The functions of each module are as follows: Body position monitoring module: used to collect the posture information and body position change parameters of the user. This module can use MEMS inertial sensors (such as a three-axis accelerometer and a three-axis gyroscope) fixed to the user's body (such as the waist or chest) to detect the body posture angle and movement rate in real time. The accelerometer measures the components of the gravitational acceleration on each axis, and calculates the inclination angle of the body relative to the vertical direction, that is, the posture angle θ; the gyroscope measures the angular velocity ω, reflecting the speed of posture change. For example, when the accelerometer outputs (a x , a y , a z ) in the stationary state, the angle between the longitudinal axis of the body and the vertical direction can be calculated by the formula . At the same time, the angular velocity signals (ω x , ω y , ω z)After filtering out high-frequency noise, it can be used to characterize the rate of attitude change. The magnitude of the angular velocity can be expressed as: , which is a scalar reflecting the overall speed of body rotation. The rate of body position change v can also be obtained by taking the difference of the attitude angles at consecutive moments θ , for example: v θ = , which is used to quantify the degree of change in the attitude angle per unit time.

[0029] Based on the processed attitude angle θ and angular velocity ω, the body position monitoring module uses a threshold model to identify body position states and change events. Angle thresholds for different body positions are preset. For example, the standing posture is defined as θ < 30°, the sitting posture is defined as θ between 30° and 70°, and the lying flat posture is defined as θ > 70°. When it is detected that the patient changes from the lying flat or sitting posture to the upright posture (i.e., θ rapidly drops from greater than 70° to less than 30°), and at the same time the angular velocity ω exceeds the set threshold ω ° / s (e.g., 100 ° / s, indicating a rapid getting-up movement), the module generates a body position change event signal. This indicates that the patient has had a rapid body position change, which may induce orthostatic hypotension. Through the above data collection and processing, this module can know in real time whether the patient has made actions such as a rapid stand-up that trigger orthostatic hypotension, providing reliable attitude change parameters for the central control unit.

[0030] The blood pressure monitoring module performs raw signal processing and blood pressure value extraction on the collected pulse waveform signal. First, a band-pass filter (e.g., 0.5–8 Hz, covering the typical heart rate band) is applied to the PPG raw waveform to remove the DC offset and high-frequency noise, obtaining a smooth pulse wave curve. Within each cardiac cycle, the module uses an algorithm to identify the characteristic points of the pulse wave (such as the systolic wave peak and diastolic wave trough), and estimates the current systolic blood pressure (SBP) and diastolic blood pressure (DBP) accordingly. For example, the blood pressure can be calculated through a pre-calibrated linear model: BP = k ⋅ PTT + b, where PTT is the conduction time of the pulse wave from the heart to the periphery, and k and b are calibration coefficients (determined according to the patient's individual baseline). In the absence of PTT measurement, the system can also approximately convert to obtain SBP and DBP values through empirical formulas based on the steepness and amplitude change of the rising edge of the pulse wave. Through such a mathematical model, the continuous pulse wave signal is converted into an instantaneous blood pressure reading to achieve continuous monitoring of the patient's blood pressure.

[0031] To detect blood pressure changes related to body position, the blood pressure monitoring module calculates the change amount ΔBP of the relative blood pressure value in real time. The stable blood pressure value of the patient when lying flat (or sitting) is recorded as BP （0） before the getting-up action occurs, and the blood pressure measured at any time point t after the body position change is recorded as BP (t) , then the blood pressure change amount is defined as: ΔBP = BP (t)-BP （0） . Orthostatic hypotension is manifested as a decrease in blood pressure after a change in posture, i.e., ΔBP (t) is negative and its absolute value is large. The system sets a threshold blood pressure target value (BP 目标 ) to determine significant hypotensive events. (BP 目标 ) is defined as: systolic blood pressure target value = baseline systolic blood pressure - 20 mmHg, diastolic blood pressure target value = baseline diastolic blood pressure - 10 mmHg. The module determines that the blood pressure has dropped abnormally and sends a hypotensive alarm signal to the central control unit. In addition, for the auxiliary control strategy, the module can further calculate the mean arterial pressure MAP as an indicator of the overall blood pressure level. The calculation formula is . Continuous SBP, DBP, and MAP data will be sent to the central control unit for subsequent data fusion and feedback control.

[0032] Environmental perception module: Used to obtain external parameters of the user's environment. This module can include a temperature sensor and a humidity sensor for measuring environmental temperature and humidity. Information such as the current time period or the type of activity performed by the user can also be recorded when necessary. The external environment can affect the body's blood pressure regulation ability. For example, a high-temperature environment can cause peripheral vasodilation and sweating dehydration, thereby increasing the probability of orthostatic hypotension; when getting up in the morning, the autonomic regulation has not been fully activated, and hypotension is more likely to occur. Therefore, the environmental information provided by the environmental perception module can be used as a correction factor for risk scoring. In hot, high-humidity, or other adverse environments, the central control unit can increase the weight of environmental factors in risk assessment to more carefully prevent hypotensive events. The sampling frequency of the environmental sensor is relatively low, such as 1 Hz or once per second, which can meet the requirements because environmental changes are usually slow. The sensors can be arranged on the outer shell of the device worn by the patient or in the surrounding space of the patient to continuously sense the physical parameters of the current environment. All collected environmental data are transmitted to the environmental perception module for processing via wired / wireless means.

[0033] In the environmental perception module, the raw environmental data will be subjected to smoothing filtering and normalization to improve reliability and facilitate fusion calculations. For slowly changing quantities such as temperature and humidity, a moving average filtering method can be used to reduce the impact of instantaneous fluctuations, thereby obtaining stable readings of environmental parameters. Subsequently, the module compares the current environmental parameters with preset comfort thresholds to evaluate the potential load on blood pressure regulation. For example, if the preset comfortable temperature T0 is 25 °C and it is detected that the current environmental temperature T is much higher than this value (such as reaching 34 °C), it is determined that the environment is in a high-temperature state, which may lead to peripheral vasodilation and an increased risk of orthostatic hypotension. Similarly, a comfortable value H0 for relative humidity can be set (for example, 40%), and when the actual humidity H is significantly high (such as above 80%), the environment is considered stuffy. With these evaluations, the module can further calculate an environmental impact factor E through a mathematical model. For example, a normalized weighted combination of temperature and humidity is used: . Here, T max and H max represent the tolerable maximum temperature and humidity thresholds respectively (such as 40 °C and 100%), and ω T and ω H are weight coefficients (satisfying ω T + ω H = 1). The above formula converts the current environmental conditions into a dimensionless environmental stress index E in the range of 0 to 1: the closer E is to 1, the worse the environment (high temperature and high humidity), which may have an adverse impact on the blood pressure stability of the patient; a lower E indicates a comfortable environment with little additional impact on blood pressure. In addition to this continuous quantification method, the environmental perception module can also adopt a simple threshold discrimination model: for example, when the temperature exceeds 30 °C or the humidity exceeds 80%, the environment is marked as "hot" or "stuffy and humid" state, and the warning level is increased in subsequent risk assessments. Through the settings of the environmental perception module, the system can timely obtain the external environmental conditions, providing an auxiliary basis for comprehensively evaluating the risk of orthostatic hypotension.

[0034] Such as Figure 2As shown in the figure, the central control unit: The central control unit is usually composed of a microcontroller or an embedded processor, which is respectively connected to the body position monitoring module, the blood pressure monitoring module, and the environmental perception module (through wired or wireless communication), and is connected to the stimulation execution module to send control instructions. After obtaining the data uploaded by each sensing module, the central control unit preprocesses and fuses the data, runs a risk scoring algorithm to evaluate the risk level of the current state, and generates corresponding stimulation control instructions according to a predetermined control strategy. Specifically, the central control unit obtains the attitude angle θ and angular velocity ω from the body position monitoring module, the current blood pressure value P(t), blood pressure change ΔBP, and physiological parameters such as heart rate HR (the heart rate can be deduced from the pulse interval of the blood pressure waveform) from the blood pressure monitoring module, and obtains the set of environmental parameters from the environmental perception module. Then, the central control unit filters and extracts features from the above data, and inputs them into the risk scoring algorithm to calculate the risk score R. According to the level of the risk score R and the built-in control strategy rules, the central control unit decides in real time whether to initiate stimulation intervention and the intensity of the intervention: for example, calculate the required electrical stimulation frequency f, stimulation intensity I, or the target heart rate H to be achieved, etc., and generate corresponding control signals to output to the stimulation execution module. The central control unit also continuously monitors the feedback to adjust the stimulation output and form a closed-loop control. Design basis: The central control unit acts as the brain of the system, fusing multi-source data for intelligent decision-making. Using embedded fuzzy logic reasoning and closed-loop control algorithms, the system can simulate the decision-making process of clinicians, adjust the intervention in real time and dynamically according to the patient's state, and achieve personalized and precise control. t The central control unit also continuously monitors the feedback to adjust the stimulation output and form a closed-loop control. Design basis: The central control unit acts as the brain of the system, fusing multi-source data for intelligent decision-making. Using embedded fuzzy logic reasoning and closed-loop control algorithms, the system can simulate the decision-making process of clinicians, adjust the intervention in real time and dynamically according to the patient's state, and achieve personalized and precise control.

[0035] Stimulation Execution Module: It is used to apply physical stimulation to the user according to the instructions issued by the Central Control Unit to adjust cardiovascular parameters and prevent or correct hypotensive states. The Stimulation Execution Module may include a programmable electrical stimulation generator, stimulation electrodes, a drive circuit, etc. In terms of specific implementation forms, it can be an implantable or surface-attached electrical stimulation device. For example, in this embodiment, the Stimulation Execution Module applies electrical stimulation to the peripheral nerves or muscles of the patient through surface electrodes, thereby triggering an increase in heart rate and peripheral vasoconstriction to elevate blood pressure. Preferably, the stimulation site can be selected on the neural pathway that affects blood pressure regulation, such as stimulating the sympathetic ganglion segment to increase cardiac sympathetic activity, or stimulating the lower limb skeletal muscles to promote venous blood return. If the patient has a cardiac pacemaker implanted, this module can also directly serve as a pacemaker interface to increase the heart rate by adjusting the pacing frequency. The stimulation parameters include the stimulation pulse frequency f (e.g., the number of electrical stimulation pulses per second) and the stimulation intensity I (e.g., the current amplitude or voltage magnitude), and the Central Control Unit will set the corresponding f and I according to the risk level. After receiving the control signal, the Stimulation Execution Module generates an electrical stimulation pulse sequence at the specified frequency and intensity, and acts on the target tissue via the electrodes. Design Basis: Enhancing the cardiovascular response through external stimulation is an effective measure to prevent orthostatic hypotension. For example, increasing the heart rate and contractility can directly increase cardiac output, and constricting the peripheral blood vessels and the skeletal muscle pump effect helps to increase the return blood volume and peripheral resistance, thereby raising blood pressure. This module executes closed-loop control instructions, simulates the body's baroreflex mechanism, intervenes in a timely manner when detecting a hypotensive risk, and quickly corrects it to maintain the blood pressure at a safe level.

[0036] The specific process is as follows: S1. Data Acquisition and Processing Process: Each module works together to achieve data acquisition, analysis, and feedback control. Refer to Figure 1 (Schematic Diagram of System Workflow), and its steps are as follows: Signal Acquisition: The Posture Monitoring Module acquires attitude sensor data at a high frequency (e.g., 10 Hz or higher), continuously outputs the current attitude angle θ and angular velocity ω, and sends the data to the Central Control Unit; the Blood Pressure Monitoring Module continuously measures blood pressure in units of the cardiac cycle, outputs the systolic blood pressure SBP(t), diastolic blood pressure DBP(t), and / or mean arterial pressure MAP(t) of each heartbeat, calculates the current blood pressure value P (t) , and sends it to the Central Control Unit (meanwhile, the pulse wave interval can be provided for calculating the current heart rate HR); the Environmental Sensing Module periodically reads environmental sensor data, such as temperature T, humidity H, etc., and sends it to the Central Control Unit.

[0037] Data preprocessing: The central control unit filters, denoises, and corrects the data of each channel. For attitude data, Kalman filtering can be used to fuse the information of the accelerometer and gyroscope to obtain smooth and accurate attitude angles; for blood pressure waveform data, digital filtering can be applied to eliminate motion artifacts and measurement noise, and smooth the blood pressure sequence between heartbeats. Environmental data such as temperature is subjected to unit conversion or simple filtering as needed. After preprocessing, the central control unit extracts key feature parameters, including the current attitude angle θ, angular velocity ω, baseline blood pressure P0 (which can be the reference value measured at supine rest), current blood pressure value P(t), blood pressure drop ΔBP = BP (t) −BP （0） , blood pressure drop rate, current heart rate HR, and the set of environmental parameters {T, H}. These feature parameters constitute the feature vector describing the user's current state.

[0038] Risk assessment: The central control unit inputs the extracted feature parameters into the built-in risk scoring algorithm module. Based on the pre-established fuzzy logic rules and weighted calculation model, this algorithm comprehensively evaluates the risk level of orthostatic hypotension induced by the current body position change and outputs a risk score value R. The specific principle and implementation of the risk scoring algorithm will be described in detail later. Briefly, the algorithm comprehensively considers factors such as the amplitude and speed of body position change, the amplitude and speed of blood pressure drop, and the degree of environmental stress to simulate clinical judgment rules and obtain the immediate risk level. The central control unit then determines whether intervention is needed currently and the intensity of intervention: if the risk score R is lower than the safety threshold, it is determined that there is no risk of orthostatic hypotension for the time being, and only continuous monitoring is continued; if R exceeds the threshold, it is considered that there is a potential risk of orthostatic hypotension or syncope, and the next step of intervention control is required.

[0039] Control decision-making and execution: When the risk score R exceeds the threshold, the central control unit calculates the required intervention parameters according to the control strategy and generates a stimulation control command to send to the stimulation execution module. The control command includes the set stimulation pulse frequency f, stimulation intensity I, and (if applicable) target heart rate H t etc. After receiving the command, the stimulation execution module immediately starts the electrical stimulation output and generates electrical stimulation signals at the specified frequency and intensity to act on the user's body. For peripheral nerve / muscle stimulation devices, this will cause physiological reactions such as increased heart rate and vasoconstriction; for pacemaker devices, it directly drives the heart to reach the target heart rate. Through this active intervention, it attempts to counteract the blood pressure drop caused by body position changes.

[0040] Closed-loop feedback regulation: During the operation of the stimulation execution module, the central control unit continuously monitors the changes in parameters such as blood pressure and posture to achieve closed-loop feedback control. Specifically, the central control unit re-collects blood pressure values and other sensing data at short intervals (such as every heartbeat or per second), and continuously updates the risk score R. According to the deviation between the latest blood pressure situation and the target value, the central control unit adjusts the stimulation parameters: if the blood pressure has risen after intervention and the risk score has dropped to the safe range, the central control unit gradually reduces the stimulation frequency and intensity to avoid blood pressure overshoot; if the blood pressure is still low (the risk score is still high), the central control unit can further increase the stimulation intensity or frequency within the safety limit. Such repeated regulation forms a negative feedback control loop until the blood pressure returns to near the baseline level and the risk is eliminated. Throughout the process, the central control unit ensures that the stimulation output matches the real-time demand, thereby achieving closed-loop automatic regulation and maintaining blood pressure stability. After the intervention is completed (for example, the user stands stably for a period of time and the blood pressure remains normal), the central control unit instructs the stimulation execution module to stop output, and the system returns to the monitoring standby state.

[0041] The above process realizes a complete closed-loop control from body position change detection, blood pressure drop risk assessment to stimulation intervention and feedback regulation, and can automatically provide timely physiological compensation in cases where the user changes from lying / sitting position to standing position, etc., reducing the incidence of postural change-related hypotension and syncope events.

[0042] S2. Risk scoring algorithm: The risk scoring algorithm is the core algorithm used by the central control unit to evaluate the risk of postural change-related hypotension. The present invention combines fuzzy logic and weighted calculation methods to comprehensively analyze multiple input factors and outputs a quantitative risk score R. The design of this algorithm aims to integrate clinical expert experience rules and data-driven weight adjustment to achieve accurate quantification of risks.

[0043] S21. Fuzzy logic algorithm: First, the key physiological parameters affecting the risk are fuzzified to establish a fuzzy rule inference system. The input parameters include: body position parameters (posture angle θ and angular velocity ω), blood pressure parameters (ΔBP and blood pressure drop rate), and environmental parameters (such as temperature, etc.). For each input, several fuzzy sets and their membership functions are defined to describe the semantic risk levels corresponding to different range values. Examples are as follows: Posture angle θ: It can be defined as three fuzzy sets: "small" (close to horizontal, such as 0°), "medium" (semi-inclined, such as 45°), and "large" (close to upright, such as close to 90°). Its membership function μ(θ) can adopt a trapezoidal or triangular function to achieve a gradual transition near the critical angle. For example, when θ < 30°, the membership degree μ 直立 (θ) = 0; when θ > 70°, μ 直立(θ) = 1; μ is between 30° and 70° 直立 increases linearly with θ. This means that the closer the body is to the upright position, the higher the tendency to trigger the risk.

[0044] Angular velocity ω: It can be defined as two fuzzy sets of "slow" and "fast" to describe the speed of body position change. The threshold can be set according to experience. For example, if the change rate exceeds a certain angle per second, it is regarded as "fast". The membership function can be designed similarly. When ω is low, μ 快 (ω) = 0, and when ω exceeds the threshold, μ 快 is close to 1. A rapid change in body position (such as standing up suddenly) will give a higher risk tendency.

[0045] Blood pressure drop ΔBP: It is defined as three fuzzy states of "slight", "moderate", and "severe". For example, ΔBP = 0 (no drop) corresponds to a "slight" membership degree of 1; when ΔBP increases to more than 20 mmHg, the "severe" membership degree gradually approaches 1. Near the threshold of 20 mmHg, the "severe" membership function can smoothly increase from 0 to 1, reflecting the fact that a 20 mmHg drop in systolic blood pressure is a significant event clinically. Similarly, the "moderate" membership function is 0 at a smaller drop, takes the maximum in the moderate range, and then decreases at a larger drop, forming a trapezoidal function (representing the moderate risk interval).

[0046] Blood pressure drop rate: It can be defined as two fuzzy sets of "gentle" and "rapid". If the slope of blood pressure drop over time exceeds a certain threshold, it is determined to be a "rapid" drop. The design of its membership function is similar to the fuzzification of angular velocity - the faster the drop, the higher the membership degree belonging to the "rapid" category. This indicator complements ΔBP because even if the total amount of ΔBP is not large, but if it drops suddenly in a very short time, it may also trigger syncope.

[0047] Environmental temperature T (for example): It can be defined as three fuzzy sets of "normal", "high", and "very high". At normal room temperature (such as around 20°C), the membership degree of "high" is 0; when the temperature exceeds 30°C, the membership degree of "high" or "very high" increases (the specific threshold is set according to clinical statistics. For example, a high-temperature environment is considered to have a significant impact on risk when it exceeds 35°C). Other factors of the environment (humidity, whether it is airtight and stuffy, etc.) can also be fuzzified similarly or considered qualitatively in the rules.

[0048] For the above fuzzified inputs, a fuzzy rule base is established to infer the risk level output. The fuzzy rules are expressed in the form of _if...then..._, mapping multiple input conditions to the risk output. The formulation of the rules is based on clinical medical knowledge and expert experience. For example: Rule 1: IF the posture angle is "large" (upright), the angular velocity is "fast", and the blood pressure drop is "severe", THEN the risk is "high". (Explanation: When the patient stands up quickly and the blood pressure drops significantly, syncope is very likely to occur, and the risk is judged at the highest level.) Rule 2: IF the posture angle is "large", the blood pressure drop is "moderate", and the environmental temperature is "very high", THEN the risk is "moderately high". (Explanation: There is a moderate blood pressure drop when standing upright, and the risk is further increased due to vasodilation in a high-temperature environment.) Rule 3: IF the posture angle is "small" (lying flat or only slightly inclined), THEN the risk is "low". (Explanation: Without significant postural changes, there is almost no risk of orthostatic hypotension.) Rule 4: IF the blood pressure drop is "severe" and the angular velocity is "slow" (getting up slowly but with a large blood pressure drop), THEN the risk is "medium to high". (Explanation: Even if the movement is slow, but the blood pressure drop is severe, it may indicate potential cardiovascular regulation abnormalities and also requires vigilance.) The above rule base can cover various typical situations. In the reasoning process, the membership degrees of multiple input conditions are combined using the fuzzy logic operation "AND" for the premise part of the rule. Generally, the minimum value (min) of the membership degrees is used as the rule compliance degree; for conditions with an "OR" relationship, the maximum value (max) operation can be used. Each rule will obtain a corresponding fuzzy membership value of the risk output level. Then, the outputs of all rules are aggregated (using max to take the union) to obtain a comprehensive risk fuzzy set. Finally, through the defuzzification process, the fuzzy set is converted into a specific risk score value R F . Preferably, the centroid method (center of gravity method) is used for defuzzification to calculate the risk score to ensure that the output value changes smoothly with the input. The obtained risk score R f is usually normalized to 0 - 1 or represented as a percentage system of 0 - 100, where the larger the value, the higher the risk. Through fuzzy logic reasoning, the system can convert the complex influence relationship of physiological signals into a rule-based judgment process that can be understood by humans, thereby obtaining a preliminary risk quantification result.

[0049] S22. Weighting algorithm: While obtaining the basic risk assessment value of fuzzy reasoning, the risk scoring algorithm of the present invention further introduces a weighting algorithm to perform a linear combination of the contributions of different factors to correct and optimize the final risk score R. The weighting algorithm assigns weight coefficients to various risk factors to reflect their relative importance, thereby realizing a more flexible risk calculation.

[0050] Specifically, let R P represent the risk sub-score related to postural changes (determined by the posture angle θ and the angular velocity ω), RB represents the risk sub - score related to blood pressure changes (determined by ΔBP and the rate of blood pressure change), R E represents the risk sub - score related to environmental factors (determined by environmental parameters such as temperature). The central control unit can set the corresponding weight coefficients ω P , ω B , ω E , and then calculate the comprehensive risk score: R = R P ˙ω P + R B ˙ω B + R E ˙ω E ; where each sub - score R P , R B , R E can be obtained from their respective related fuzzy evaluation results or determined through a simple normalization function. For example, R P can be comprehensively determined according to the fuzzy evaluation of the attitude angle and angular velocity: if the attitude is quickly upright, the assigned value is close to 1, and if lying flat and stationary, the assigned value is close to 0; R B can be determined based on the blood pressure drop and the rate of decline: when ΔBP is large and the blood pressure decline rate is fast, the assigned value is close to 1, and conversely, when the blood pressure is stable, it is close to 0; R E is assigned according to the degree of environmental stress, with a higher value in a harsh high - temperature environment and a lower value in a comfortable environment. Through the above linear weighted formula, the obtained R superimposes each factor proportionally.

[0051] Assigning weights to different factors allows the system to adjust the emphasis of risk assessment according to individual conditions and situations. For example, for a patient with autonomic insufficiency, whose blood pressure regulation ability is poor, the weight ω B of the blood pressure drop itself R B should be set larger to highlight the importance of the blood pressure factor; for a healthy young person, the blood pressure may drop quickly and be corrected by physiological reflexes, but if standing up too quickly, there may also be a brief dizziness. At this time, the weight ω P of the body position factor can be appropriately increased. The weight ω E of the environmental factor can be enabled in special situations, such as when the patient is in a high - temperature environment or has just taken a bath, and temporarily increase ω E to more strictly evaluate the risk. The weight values can be determined offline through clinical trial data or adjusted through an adaptive algorithm during the operation of the system. In short, the combination of the weighted algorithm and fuzzy logic enables the risk score to consider both the non - linear effects of multiple factors and provides an adjustable quantitative framework, thus more reliably reflecting the actual risk.

[0052] For the convenience of subsequent decision-making on control strategies, the central control unit makes discrete risk level judgments based on the obtained continuous risk score R. Specifically, the risk score can be divided into several level intervals: for example, when 0 ≤ R < 0.4, it is judged as a low risk level (L1), indicating that although the body position changes, the blood pressure drops slightly and the environmental impact is small; when 0.4 ≤ R < 0.7, it is judged as a medium risk level (L2), indicating that moderate intervention is required; when R ≥ 0.7R, it is judged as a high risk level (L3), indicating that the blood pressure drops significantly and the external conditions may exacerbate, and strong intervention measures are needed. The thresholds of the risk levels can be adjusted and optimized according to clinical data and the specific conditions of the patient. In another implementation, if fuzzy inference is used to directly output discrete risk level results (such as directly giving "high", "medium", "low" through rules), the level conclusion can be obtained directly without numerical threshold comparison. In any way, the central control unit can convert complex continuous multi-sensor information into clear risk levels, providing a clear basis for treatment control.

[0053] Based on the evaluated risk level and score R, the central control unit calculates the corresponding treatment control parameters to drive the stimulation execution module to intervene and adjust the patient. This system includes two main intervention approaches: one is the control of the spinal cord electrical stimulation device (to change the peripheral vascular resistance and venous return blood volume), and the other is the control of the heart rate regulation device (directly affecting the cardiac output). The central control unit uses a pre-set parameter model to map the risk score R to specific stimulation frequencies, stimulation intensities, and target heart rate values.

[0054] Here, to achieve the intervention of spinal cord electrical stimulation and heart rate regulation, the stimulation execution module is an important part of this system. Spinal cord electrical stimulation module: The electrodes of this module are usually fixed in the epidural space of the patient's spine through minimally invasive surgery. The optimal installation position is the lower part of the spine (such as the T6 to T12 segments) to stimulate the spinal cord sympathetic nerves and then adjust the peripheral vasoconstriction. The positioning of the electrodes should ensure stable and effective transmission of electrical pulses while taking into account the comfort and safety of the patient. Heart rate regulation module: The electrodes or pacemaker modules used to adjust the heart rate are usually implanted in the patient's chest cavity, close to the heart position (such as the right atrium or right ventricle) to ensure precise regulation of the cardiac electrical activity. The pacemaker and its control unit should be fixed subcutaneously for subsequent parameter adjustment and maintenance, while ensuring stable data communication with the central control unit.

[0055] Based on the risk score R calculated in real time, the central control unit decides on the specific parameters and methods of stimulation intervention according to the pre-designed control strategy. The control strategy of the present invention includes stimulation parameter calculation (frequency and intensity), target heart rate calculation, and a closed-loop feedback regulation mechanism to achieve dynamic stable control of the cardiovascular state.

[0056] When intervention needs to be initiated (i.e., when the risk score reaches the threshold), the central control unit first determines the intensity and frequency of the electrical stimulation. These two parameters determine the intensity and rhythm of the intervention and should be adjusted according to the degree of risk. Generally speaking, the higher the risk, the stronger and more frequent the required stimulation.

[0057] In this embodiment, for the spinal cord electrical stimulation part, a method of mapping the risk score R to the stimulation frequency f and intensity I is adopted. Assume that the minimum value f of the stimulation frequency is preset min (for example, 0 Hz means no stimulation) and the maximum value f max (determined by the device limit or safety limit, for example, 100 Hz), as well as the minimum value I of the stimulation current intensity min and the maximum value I max . The central control unit converts the normalized risk score (between 0 and 1) into stimulation parameters: Stimulation frequency: . When R = 0 (no risk), f = f min (it can be 0, no stimulation), when R = 1 (highest risk), f = f max (apply stimulation at the highest allowable frequency). For R values between 0 and 1, linear interpolation is used to obtain the corresponding frequency. Such a design makes the risk value proportional to the stimulation frequency. The higher the risk, the more frequent the stimulation pulses, and the intervention can take effect more quickly.

[0058] Stimulation intensity: . Similarly, the lowest intensity or no stimulation is used when R = 0, and the maximum safe intensity is used when R = 1. The intensity can refer to the amplitude of the electrical stimulation current or the voltage amplitude, depending on the device type. The intensity is positively correlated with the risk, so that greater stimulation energy is provided in high-risk situations to achieve more significant physiological effects.

[0059] The above calculation methods for frequency and intensity are simple and effective. Of course, in other embodiments, a look-up table or hierarchical control method can also be adopted: for example, the risk score is divided into several levels (low, medium, high, extremely high), and each level corresponds to a fixed combination of stimulation frequency and intensity. Whether using continuous function mapping or discrete level control, the principle is that the higher the risk, the stronger the stimulation, so as to ensure timely and sufficient intervention. To balance the stimulation effect and comfort, the system can also limit the change rate of the stimulation parameters to avoid discomfort caused by sudden jumps in frequency or intensity. For example, when R rises rapidly, f and I can be gradually increased instead of jumping to the highest value at once, so as to give the physiological system a certain reaction time.

[0060] When dealing with orthostatic hypotension, increasing the heart rate is one of the effective means to restore blood pressure. For the heart rate regulation part, the central control unit communicates with the pacemaker (or other regulation devices) implanted in the patient's body through an interface to adjust the target heart rate Ht , so as to maintain sufficient cardiac output. The determination of the target heart rate takes into account both the real-time blood pressure feedback and the system risk score R to achieve a combination of rapid response and precise control.

[0061] Specifically, the central control unit first performs real-time feedback calculation based on the error between the current mean arterial pressure MAP (t) and the preset target mean arterial pressure MA Pref : In the formula, eBP>0 indicates that the patient's current blood pressure is low. Based on this blood pressure error, the system initially determines a basic heart rate increment ΔH BP : In the formula, K BP is the blood pressure-heart rate conversion coefficient set according to clinical experience (for example, the heart rate increment to be increased per 1 mmHg drop in blood pressure, unit: beats per minute·mmHg).

[0062] On this basis, the system further corrects the target heart rate according to the comprehensive risk score R to ensure a rapid response of the control strategy to the risk level:

[0063] In the formula: H base is the patient's basic resting heart rate (such as 70 beats per minute).

[0064] H max is the upper limit of the highest safe heart rate allowed for the patient (such as 120 beats per minute).

[0065] R is the risk score calculated in real time by the central control unit, with a value range of 0 to 1. The higher the value, the greater the risk, indicating that the patient's current state urgently requires an increase in heart rate to stabilize blood pressure.

[0066] Through this comprehensive formula, when the patient's blood pressure is significantly lower than the preset level (e BP is large), the basic increment ΔH BP will first give precise adjustment; at the same time, the addition of the risk score R enables rapid reaching of the target heart rate in high-risk (R close to 1) situations to prevent the expansion of risks, while in low-risk (R close to 0) situations, the basic blood pressure feedback is mainly used and only minor adjustments are made.

[0067] For example: Suppose the patient's current mean arterial pressure MAP (t) is 80 mmHg and the target is 90 mmHg, then: Take KBP = 1 beat / min·mmHg, the basic heart rate increment is: If the current patient's resting heart rate Hbase = 70 beats / min and the risk score is R = 0.5 (moderate risk), and the maximum allowable heart rate is Hmax = 120 beats / min, then:

[0068] After the central control unit calculates the target heart rate, it will be transmitted to the pacemaker controller in real time, enabling the patient's heart rate to quickly increase to the target value, thus effectively coping with the risk of orthostatic hypotension.

[0069] This comprehensive method takes into account both the real-time blood pressure feedback and the risk of rapid response ability, and can more precisely adapt to the clinical needs of different patients, improving the reliability and safety of the intervention strategy.

[0070] In one embodiment, the control strategy of the system of the present invention adopts a closed-loop feedback regulation mechanism to ensure the safety, efficiency and adaptability to individual differences of the intervention process. In the orthostatic hypotension (OH) intervention system, the main goal of closed-loop control is to maintain the patient's blood pressure and heart rate within a safe range, especially to quickly stabilize the blood pressure when the blood pressure drops suddenly due to the change from lying to standing position. The specific control objectives include: Maintain the stability of mean arterial pressure (MAP): Control the patient's mean arterial pressure near the target value to prevent persistent hypotension (for example, ensure that the systolic blood pressure does not drop by more than 20 mmHg). This helps to ensure the perfusion pressure of important organs and avoid symptoms such as dizziness and syncope.

[0071] Maintain the target heart rate (H t ) moderate: While increasing the blood pressure, control the heart rate within the target range to avoid excessive heart rate due to overcompensation. This can not only assist in maintaining blood pressure but also prevent risks such as tachycardia.

[0072] The system realizes closed-loop control regulation through multiple physiological feedback signals. The key feedback variables include: Real-time arterial pressure P (t) : Continuously monitor the immediate blood pressure value as the main basis for judging the degree of deviation of blood pressure from the target.

[0073] Heart rate (HR): Monitor the current heart rate to evaluate the body's autonomic nerve response to hypotension and as a feedback on the achievement of the heart rate control target.

[0074] Risk score (R): A risk indicator calculated based on the degree and duration of blood pressure drop and the patient's symptoms, etc., which reflects the risk level of syncope or hypoperfusion caused by hypotension. A high R value indicates that the patient is in a high-risk state and requires more aggressive intervention; a low R value indicates a relatively stable situation.

[0075] These feedback amounts are obtained in real time by sensors and then sent to the controller to determine the output intensity required for blood pressure regulation. The closed-loop system aims to maintain MAP and H t and performs feedback control on signals such as P (t) , HR, and R, thereby automatically adjusting the output of the intervention device (such as the tilt angle of the body position, the intensity of electrical stimulation, or the infusion rate of vasoconstrictor drugs, etc.) to achieve dynamic intervention for orthostatic hypotension.

[0076] PI control strategy design and clinical significance: The closed-loop control first uses a classical proportional-integral (PI) controller to continuously regulate blood pressure. The PI control calculates the increment of the intervention output according to the magnitude of the blood pressure error, and its control law can be expressed as: where represents the blood pressure error, that is, the gap between the current mean arterial pressure and the target value; ΔI represents the increment of the control output (such as the adjustment amount of the stimulation current or the infusion rate). The proportional gain K p and the integral gain K i are the controller parameters.

[0077] Proportional control (P) part: The item provides an immediate correction effect according to the magnitude of the instantaneous error. When the blood pressure is lower than the target value, the proportional part generates an output increment proportional to the error. The larger the error, the larger the output, just like the more obvious the low blood pressure clinically, the stronger the intervention measures are given. This part ensures a rapid response to the current deviation and promptly increases the blood pressure. For example, if a patient suddenly stands up and causes a significant drop in MAP, a large error will trigger a strong pressurization intervention (such as increasing the pump infusion speed or the amplitude of electrical stimulation) through the proportional part to quickly curb the blood pressure drop. This can shorten the duration of hypotension and prevent the patient from fainting. Conversely, when the error is small, the proportional output is also relatively small to avoid overreaction.

[0078] Integral control (I) part: The integral term accumulates past errors, providing the effect of eliminating steady-state errors. If blood pressure deviates from the target for a long time (even if the deviation is small but persistent), the integral link will gradually accumulate the output and gradually increase the intervention intensity until the blood pressure error is completely corrected. Clinically, this is equivalent to compensating for persistent mild hypotension: when the patient's blood pressure remains below the target value for a period of time, the integral term will continuously increase the output, slowly raising the blood pressure to the target level and eliminating the residual deviation. For example, the initial proportional control may raise the blood pressure to slightly below the target value, and then the integral control intervenes to further fine-tune and increase the output, ultimately maintaining the MAP precisely near the target value. Integral control ensures blood pressure stability on a long time scale and prevents the occurrence of persistent small-scale hypotension. However, it should be noted that if the integral action is too strong, it may cause overshoot, that is, the blood pressure rebounds above the target value. Therefore, the integral gain K i needs to be balanced during design - clinically known as the anti-integral saturation strategy - to avoid blood pressure overshoot caused by excessive accumulation.

[0079] By reasonably adjusting K p and K i , the PI controller achieves a trade-off between response speed and stability clinically: a higher K p allows for rapid correction when blood pressure drops, but being too high may cause blood pressure oscillations or even repeated rises and falls; an appropriate K i eliminates steady-state errors, but being too large may lead to blood pressure overshoot. Therefore, in specific implementations, these two parameters are usually adjusted based on clinical data so that for most patients, PI control can maintain blood pressure stability while avoiding hypotension. However, due to different patients' sensitivities to interventions, fixed-gain PI control may exhibit problems of insufficient or excessive response in some extreme cases (such as special constitutions or sudden drastic changes). Therefore, this system introduces fuzzy control on the basis of PI control for adaptive adjustment to further improve the control performance and safety margin in various situations.

[0080] To enhance the system's adaptability to individual differences and non - linear complex situations, we integrated a fuzzy logic control strategy. The fuzzy controller, as the advanced layer of PI control, online adjusts control parameters and outputs based on expert experience rules. This fuzzy logic layer incorporates clinical knowledge into control decisions, such as modifying the intervention intensity according to blood pressure trends and risk levels, enabling the system to take appropriate measures in different situations. Fuzzy control is defined in the form of linguistic rules (IF - THEN rules), using fuzzy sets to represent the states of input and output variables, thus dealing with fuzzy situations where precise thresholds are difficult to determine. Its inputs can include: the magnitude and change trend of the current blood pressure error, the level of the risk score R, and even the degree of deviation of the heart rate, etc.; the output is the correction of the control action, such as adjusting the PI gain or directly intervening in the output intensity. Through fuzzy inference, artificial experience judgment is introduced to achieve the adaptive optimization of the control strategy.

[0081] Rule design example: Fuzzy control rules cover various possible physiological states to determine corresponding control strategies. Typical rules are as follows: If the blood pressure is still significantly lower than the target and the risk score R is very high, maintain a strong stimulation output (keep the intervention device working at high power) until the blood pressure significantly rebounds. This corresponds to the situation where the patient has severe hypotension and a high - risk syncope. The controller will continuously provide the maximum - intensity intervention to rapidly increase the blood pressure.

[0082] If the blood pressure is low and the risk score R is medium, moderately increase the output intensity. That is, apply a medium - intensity intervention in this case: the intensity is higher than normal but lower than the maximum, which can correct the low blood pressure without over - stimulating and observe the gradual increase in the blood pressure trend.

[0083] If the blood pressure is close to the target and the risk score R is low, gradually reduce the output to smoothly transition to the steady state. That is, when the hypotensive situation is relieved, the controller orderly reduces the intervention intensity, avoids overshooting of the blood pressure, and ensures a smooth transition of control to maintain the blood pressure fluctuating near the target.

[0084] If the blood pressure overshoots above the target, rapidly reduce or pause the intervention to prevent the blood pressure from being too high. Similarly, if the heart rate is too high beyond the target heart rate range, the load can be reduced by reducing the stimulation intensity, etc., so as not to induce other cardiovascular events.

[0085] The above - mentioned fuzzy rules exemplify the decision - making methods of the system in different situations. Through such a rule set, the fuzzy controller can adjust the parameters of the PI controller or directly modify the output according to the magnitude, change of the blood pressure error, and the risk level: Dynamically adjust the PI gain: Fuzzy logic can modify K p and K iThe value of can realize the function similar to gain scheduling. For example, when the blood pressure error is large and the R value is high, the proportional gain K is increased. p (or integral gain K i ) to speed up the response; conversely, when the error is small or the blood pressure tends to be stable and the R value decreases, the gain is appropriately reduced to avoid overshoot. In this way, the sensitivity of the controller changes adaptively with the physiological state, improving the control effect at each stage.

[0086] Directly correct the control output: In some cases, the fuzzy controller can bypass the PI calculation and intervene directly in the output. For example, when the blood pressure falls below the dangerous threshold, the fuzzy rule can make the output immediately rise to the preset high level (such as instantly increasing the stimulation current to the maximum safe value), which is equivalent to preempting PI control for a short time to quickly curb the deterioration of hypotension. For example, when the blood pressure is about to reach the target and there is a tendency to overshoot, the fuzzy control can weaken the output in advance or even temporarily set it to zero to prevent the blood pressure from overshooting. Such direct regulation is equivalent to limiting or adding bias to the control amount based on fuzzy judgment, making the system more robust.

[0087] Through the above mechanisms, fuzzy control acts as an "intelligent regulator" and "safety net" for PI control. It converts clinical experience into rules to modify and compensate for control, enabling the system to understand complex physiological situations and make corresponding adjustments. For example, fuzzy control can "understand" when a patient is in an abnormally dangerous state (high R) and decisively strengthen intervention, and also know when the situation is stable and can be cautiously converged. This knowledge-based regulation greatly enhances the robustness and adaptability of the system, making the control strategy both flexible and reliable. As pointed out by relevant research, fuzzy logic control is good at interpreting and processing parameter trends of multi-factor coupling such as blood pressure, and optimizes the control strategy by combining the patient's medical information. The fuzzy control part of this system uses this feature to integrate patient risks and status into control decisions, thereby providing customized control parameter adjustments for different individuals.

[0088] In summary, the basic idea of fusion control logic is: within a stable and controllable range, it mainly relies on PI control for fine adjustment, and when there are drastic fluctuations or high-risk conditions, the nonlinear strategy of fuzzy control is triggered first. The specific operation logic is as follows: Normal state (low risk, small fluctuation): When the patient's blood pressure fluctuation is within the controllable range (for example, the error is small and the R value is low), the system mainly runs the PI controller to maintain the blood pressure. At this time, the fuzzy control is in monitoring or fine adjustment mode, and no drastic intervention is performed. Due to the small error, PI control is capable of smooth regulation, and its linear correction ensures smooth output changes and stable blood pressure. The fuzzy layer monitors in the background, and if there is no special situation, the PI output is directly used. This ensures the smoothness of the control and avoids unnecessary complex adjustments.

[0089] Abnormal state (high risk, large fluctuations): When it is detected that the patient is at high risk or there are significant fluctuations in blood pressure (e.g., a sharp drop in P(t) within a short period, a sudden increase in error, and an increase in the R value), the system will elevate the priority of fuzzy control. The fuzzy controller immediately intervenes according to pre-established rules and adjusts the control parameters or outputs. For example, in the case of a high R value, fuzzy logic may temporarily increase Kp and Ki (increase the response gain), or directly output a large control action to rapidly increase blood pressure. At the same time, the conventional regulation of the PI controller serves as a fundamental role in this situation and jointly determines the final control quantity with the fuzzy output. However, fuzzy control dominates to ensure strong and timely intervention. This strategy is equivalent to turning on the "acceleration mode" in an emergency to prioritize ensuring that the blood pressure quickly rebounds to a safe level.

[0090] Fuzzy-PI switching and coordination: When the blood pressure recovers close to the target value and the risk decreases from the abnormal state, the control strategy will smoothly transition back to mainly PI control. The fuzzy control gradually weakens the intervention intensity to prevent rebound oscillations caused by over-action. By introducing hysteresis and threshold determination, the system avoids frequent jitter between the two modes. For example, set the R value threshold and error threshold, and switch back to the conventional PI regulation when the R value drops to a safe level and the error returns to a small range. Throughout the process, PI and fuzzy control do not operate separately, but calculate their respective outputs simultaneously and adopt the weighted or optimal selection according to the situation. This ensures the coherence and stability of the switching. Obtain the fast and strong response of fuzzy control at high risk, and maintain the precise and smooth characteristics of PI control after returning to stability.

[0091] The fused control algorithm fully utilizes the advantages of the two methods: PI control provides a stable basic regulation ability, while fuzzy control provides an "intelligent" regulation ability to adjust in real time according to the patient's state. Research shows that such a dual-input fuzzy controller can achieve an excellent balance between control strength and stability. Therefore, in the case of significant fluctuations, this system can act both "boldly" and "prudently" - boldly means outputting a strong enough intervention force when needed, and prudently means maintaining the stability of the system without losing control throughout the process. When the patient's condition changes, the control strategy will automatically adjust between the PI and fuzzy modes, avoiding the limitations of a single fixed-gain control in various situations and preventing instability caused by blind overcorrection. This closed-loop strategy is particularly suitable for the highly individualized and dynamic physiological responses in the OH scenario and can more effectively prevent the risk of orthostatic hypotension caused by body position changes.

[0092] Through the above fusion of PI control and fuzzy control, the adaptive closed-loop system can finally output a stable, smooth, and individualized control signal to ensure safe and rapid blood pressure regulation and adapt to the needs of different patients. The specific manifestations are as follows: Stable and smooth control output: Thanks to the regulation of PI control and the anti-overshoot strategy of fuzzy logic, the output changes are generally continuous and stable, without obvious sharp switching or oscillation. During blood pressure regulation, the output of the controller (such as stimulation intensity or drug administration rate) is gradually adjusted according to the error and smoothly converges when the blood pressure approaches the target, thus avoiding large overshoots or fluctuations in the patient's blood pressure. The simulation and experimental results also verify this advantage: compared with the PI controller with fixed parameters, the control with fuzzy adaptation can significantly shorten the time required for blood pressure stabilization and reduce the overshoot amplitude at the same time. For example, a study compared the performance of different control schemes and found that the fuzzy PI controller shortened the blood pressure convergence time from about 400 seconds to 71 seconds and basically eliminated the blood pressure overshoot (the peak overshoot dropped to 0 in the sensitive patient model). This means that the closed-loop system can pull the blood pressure back to the target in a shorter time and will not rebound too high, keeping the blood pressure stable near the target value. For patients, this stable control reduces the duration of hypotensive symptoms and also avoids the discomfort of fluctuating blood pressure.

[0093] Safe and rapid correction of hypotension: The fusion control strategy ensures that the output can be rapidly enhanced to correct hypotension when needed, but always remains within a safe range. The intervention of fuzzy control at high risk makes the system respond very quickly, significantly shortening the duration of hypotension and improving the safety under acute OH events; while the constraint of the PI link and the amplitude limit of fuzzy rules prevent the output from being too strong, resulting in hypertension or too fast heart rate. In this way, the patient's blood pressure can quickly recover when the body position changes, and the whole process is strictly controlled, without causing new risks due to excessive intervention. Clinically, this means reducing the risk of syncope and falls caused by orthostatic hypotension and ensuring perfusion of vital organs.

[0094] Individualized dynamic adjustment: This closed-loop control has the ability to adapt and can be adjusted according to the physiological characteristics and real-time status of different patients, thus realizing individualized treatment. The introduction of the risk score R enables the controller to judge the patient's tolerance to hypotension and the risk level "person by person", and then adjust the control strategy. For example, patients with a sensitive constitution may have a high R alarm when their blood pressure drops slightly, and the system will quickly strengthen the intervention for this situation; while for patients with better tolerance, the system allows the blood pressure to recover slowly to avoid overcorrection. The fuzzy rules and control parameters can also be adapted and optimized based on the patient's historical data. As pointed out in the literature, incorporating the patient's clinical status into the control algorithm and dynamically updating the parameters is the key to achieving the best control effect. This system achieves this through risk indicators and fuzzy reasoning: it can automatically adjust the control output as the patient's state changes, fully considering individual differences. In practice, each patient will receive the intervention intensity tailored to their own needs, avoiding both insufficient blood pressure elevation and excessive blood pressure elevation.

[0095] In summary, the closed-loop control strategy of the present invention organically integrates the robustness and precision of classical PI control with the intelligent adaptability of fuzzy logic control to form a closed-loop dynamic regulation mechanism. This mechanism can provide safe, rapid, and individualized blood pressure intervention when orthostatic hypotension occurs: it can not only quickly correct hypotension and stabilize the patient's condition, but also maintain smooth output without violent fluctuations, and automatically adjust the strategy according to the physiological responses of different patients. In this way, the system significantly improves the effectiveness and safety of orthostatic hypotension intervention.

Claims

1. A postural hypotension intervention system based on dynamic risk assessment, characterized in that, Including: A body position monitoring module, which is used to collect the posture information and body position change parameters of the user in real time, uses a MEMS inertial sensor to detect the posture angle θ and angular velocity ω, and judges whether the user has a rapid body position change that may induce hypotension according to a preset threshold; A blood pressure monitoring module, which is used to monitor the blood pressure parameters of the user in real time, collects the pulse wave signal through a PPG sensor and extracts the blood pressure value, calculates the blood pressure change amount ΔBP and the mean arterial pressure MAP to determine whether the blood pressure drops abnormally and sends an alarm signal; An environment perception module, which is used to collect the environmental temperature T and humidity H, calculates the environmental stress index E through a normalized weighted formula, and provides environmental factor parameters; A central control unit, which communicates with the body position monitoring module, the blood pressure monitoring module and the environment perception module, is used to fuse and process the characteristic parameters collected by each module, and calculate the risk score R; A stimulation execution module, which is used to perform electrical stimulation intervention according to the instruction issued by the central control unit, and its stimulation parameters include stimulation frequency f and intensity I, and the stimulation parameters are determined according to the risk score R.

2. The orthostatic hypotension intervention system based on dynamic risk assessment according to claim 1, wherein, The risk score R is obtained by combining a fuzzy logic algorithm and a weighted algorithm: The fuzzy logic algorithm includes performing fuzzy processing on the attitude angle θ, angular velocity ω, blood pressure change ΔBP, blood pressure drop rate, and environmental factors, and performing inference using the minimum method (min) and maximum method (max) based on a preset rule base to obtain a risk fuzzy set, and using the centroid method to defuzzify and calculate the risk score R F ; The weighted algorithm is based on the risk score R F to respectively determine the postural factor sub-score R P , the blood pressure factor sub-score R B , and the environmental factor sub-score R E , and calculate the comprehensive risk score R through weighted calculation using the weight coefficients ω P , ω B , and ω E : R = R P ·ω P + R B ·ω B + R E ·ω E 。 3. The orthostatic hypotension intervention system based on dynamic risk assessment according to claim 1, wherein The stimulation parameters in the stimulation execution module include stimulation frequency f and intensity I, and the stimulation parameters are determined according to the risk score R: Stimulation frequency: f = f min +(f max -f min )·R; Stimulation intensity: I = I min +(I max -I min )·R; Meanwhile, for heart rate regulation, the central control unit determines the target heart rate Ht based on the real-time feedback error e of blood pressure BP and the risk score R: H t = H base +K BP ·e BP +(H max - H base )·R。 4. The postural hypotension intervention system based on dynamic risk assessment according to claim 3, characterized in that, The central control unit adopts a closed-loop feedback mechanism to adjust the stimulation parameters in real time through a PI controller, and the control output increment ΔI is calculated as: ΔI = K p ·e BP + Ki∫e BP d t ; At the same time, a fuzzy control strategy is adopted to dynamically adjust the PI controller parameters K p and K i or directly adjust the control output to form an adaptive closed-loop feedback regulation integrating PI control and fuzzy control.

5. The orthostatic hypotension intervention system based on dynamic risk assessment according to claim 1, wherein The body position monitoring module sets a posture angle threshold, defines θ < 30° as standing, 30° ≤ θ ≤ 70° as sitting, and θ > 70° as lying flat.

6. The postural hypotension intervention system based on dynamic risk assessment according to claim 1, characterized in that After the PPG signal in the blood pressure monitoring module is processed by a 0.5 - 8Hz band-pass filter, the blood pressure value is estimated through a pre-calibrated linear model BP = k·PTT + b or an empirical formula.

7. The orthostatic hypotension intervention system based on dynamic risk assessment according to claim 1, wherein The environmental stress index E is calculated through a temperature and humidity normalized weighted formula: E = ω T ·(T / T max ) + ω H ·(H / H max ), where ω T + ω H = 1.

8. The postural hypotension intervention system based on dynamic risk assessment according to claim 1, characterized in that, The stimulation site of the stimulation execution module is selected from the lower spinal cord segments T6 to T12 or the lower limb skeletal muscles, and the heart rate is adjusted by electrically stimulating the peripheral nerves, muscles or directly adjusting the pacemaker.

9. The postural hypotension intervention system based on dynamic risk assessment according to claim 4, wherein, The fuzzy logic control strategy of the central control unit includes the following typical rules: When the blood pressure is significantly lower than the target and the risk score is high, maintain high-intensity stimulation; When the blood pressure is close to the target and the risk score decreases, gradually reduce the stimulation intensity; When the blood pressure exceeds the target, quickly weaken or stop the stimulation.

10. The postural hypotension intervention system based on dynamic risk assessment according to claim 4, characterized in that, The closed-loop feedback regulation mechanism has an anti-integral saturation strategy to avoid overshoot of blood pressure caused by excessive stimulation.

Images