Blood pressure measurement calibration method, wearable device and storage medium

Abstract

The invention discloses a blood pressure measurement calibration method, wearable equipment and a storage medium, and relates to the technical field of wearable equipment.The blood pressure measurement calibration method comprises the steps that detection environment parameters detected by a target user based on the wearable equipment are obtained in real time; under the condition that the human body state is recognized as the standard human body state, collecting a baseline physiological signal, and determining a blood pressure baseline feature and a blood pressure baseline state based on the baseline physiological signal; under the condition that the blood pressure base line is changed, outputting calibration prompt information to indicate a target user to synchronously collect a gold-labeled blood pressure value and detect a to-be-calibrated blood pressure value; and optimizing and adjusting model parameters of the target blood pressure detection model based on the gold-labeled blood pressure value. The objective change of the blood pressure baseline state is used as a calibration triggering condition, unnecessary user calibration operation is reduced, calibration convenience is improved, meanwhile, the model parameters are corrected in combination with the gold label blood pressure value, calibration can more accurately aim at the long-term physiological state change of the user, and calibration accuracy is improved.

Description

Technical Field

[0001] This application relates to the field of wearable device technology, and in particular to a blood pressure measurement calibration method, a wearable device, and a storage medium. Background Technology

[0002] Blood pressure is a key indicator of human physiological health. Regular blood pressure measurement helps people detect potential cardiovascular disease risks early, take corresponding preventive measures, and reduce the incidence of diseases. To meet users' needs for non-intrusive, high-frequency, and even continuous blood pressure monitoring, cuffless blood pressure measurement technology has been widely researched and applied.

[0003] However, current cuffless blood pressure measurement technology typically requires periodic calibration of wearable devices, and the current calibration process suffers from two main problems. First, the calibration process requires users to manually adjust the settings at fixed intervals, which hinders normal use of the wearable device. Furthermore, the manual calibration process is cumbersome, severely impacting the convenience of using wearable devices for cuffless blood pressure measurement. Second, blood pressure measurement is affected by both long-term and short-term factors. Long-term factors include vascular stiffness, which increases slowly with age. Short-term factors include changes in cardiac output and peripheral vascular resistance related to sympathetic nervous system regulation caused by diet, exercise, and sleep. Short-term factors often significantly interfere with the calibration process. Summary of the Invention

[0004] The main objective of this application is to provide a blood pressure measurement calibration method, wearable device, and storage medium, aiming to solve the technical problem of how to simultaneously improve the convenience and accuracy of cuffless blood pressure measurement calibration during the calibration process.

[0005] To achieve the above objectives, this application proposes a blood pressure measurement calibration method, which includes: The system acquires in real time the detection environment parameters of the target user based on the wearable device at the current time period, including IMU data that reflects the human body state; When the target user's human state is identified as a standard human state based on the detection environment parameters, baseline physiological signals are collected, blood pressure baseline characteristics are determined based on the baseline physiological signals, and blood pressure baseline state is determined based on the blood pressure baseline characteristics. The baseline physiological signals include at least two of the following: baseline electrocardiogram signals, baseline heart sound signals, and baseline pulse signals. The blood pressure baseline characteristics are reference values ​​corresponding to each physiological characteristic under the standard human state, which is either a sleep state or a resting state. The physiological characteristics are used to characterize blood pressure levels. The blood pressure baseline state is either a change in blood pressure baseline or no change in blood pressure baseline. Specifically, if the difference between the blood pressure baseline characteristics and the target baseline characteristics exceeds a preset threshold, the blood pressure baseline state is considered a change in blood pressure baseline. When the blood pressure baseline status is changed, a first calibration prompt message is output. The first calibration prompt message is used to instruct the target user to simultaneously collect the gold standard blood pressure value through the gold standard blood pressure measurement device and detect the blood pressure value to be calibrated through the wearable device. The gold standard blood pressure value collected by the gold standard blood pressure measurement device is obtained, and the model parameters of the target blood pressure detection model of the wearable device are optimized and adjusted based on the gold standard blood pressure value, so as to adjust the blood pressure value to be calibrated output by the target blood pressure detection model to the gold standard blood pressure value.

[0006] In one embodiment, prior to the step of acquiring in real-time detection environment parameters detected by the wearable device based on the target user in the current time period, the method further includes: When it is detected that the target user is wearing the wearable device for the first time, the detection environment parameters of the target user based on the wearable device are acquired for the first time; If the target user's human body state is identified as a standard human body state based on the initially acquired detection environment parameters, the baseline physiological signal is collected for the first time, and the collection time of the initial baseline physiological signal is accumulated. If the collection time exceeds a preset time threshold, the collection time is stopped from being accumulated and a second calibration prompt message is output. The second calibration prompt message is used to instruct the target user to simultaneously collect the gold standard blood pressure value for the first time through the gold standard blood pressure measurement device and to detect the blood pressure value to be calibrated for the first time through the wearable device. The gold standard blood pressure value collected for the first time by the gold standard blood pressure measurement device is obtained, and the model parameters of the target blood pressure detection model are adjusted for the first time based on the gold standard blood pressure value collected for the first time, so as to adjust the blood pressure value to be calibrated detected by the wearable device for the first time to the gold standard blood pressure value collected for the first time.

[0007] In one embodiment, before the step of outputting the second calibration prompt information, the method further includes: A second calibration prompt message is generated using a preset model calibration unit; The target user's human body status is identified based on the detection environment parameters using a preset human body status recognition unit. If the human body is identified as being in a preset calibration requirement state, the step of outputting the second calibration prompt information is executed, wherein the calibration requirement state includes a resting state.

[0008] In one embodiment, the detection environment parameters further include pressure data applied by the wearable device to the target user's body surface measurement location, and temperature data of the target user at a preset body part; the blood pressure measurement calibration method further includes: The system uses a preset human body state recognition unit to determine whether the detection environment parameters meet the preset blood pressure measurement conditions. The human body state recognition unit includes an inertial sensor, a pressure sensor, and a temperature sensor. When the detection environment parameters meet the blood pressure measurement conditions, the step of optimizing and adjusting the model parameters of the target blood pressure detection model of the wearable device based on the gold standard blood pressure value is performed.

[0009] In one embodiment, after the step of outputting the first calibration prompt information, the method further includes: The system acquires in real time the target physiological signals detected by the wearable device during the current time period, wherein the target physiological signals include at least two of the target electrocardiogram signals, target heart sound signals, and target pulse signals. The target physiological signal, the detection environment parameters, the blood pressure baseline characteristics, and the blood pressure baseline status are input into the target blood pressure detection model to obtain the blood pressure value to be calibrated. The step of optimizing and adjusting the model parameters of the target blood pressure detection model of the wearable device based on the gold standard blood pressure value includes: The gold standard blood pressure value is used as the model output target and input into the target blood pressure detection model so that the target blood pressure detection model can adjust the blood pressure value to be calibrated to the gold standard blood pressure value before outputting it by adjusting the model parameters.

[0010] In one embodiment, the step of acquiring the target physiological signal detected by the wearable device in real time based on the target user during the current time period includes: The wearable device collects the target user's raw physiological signals in real time during the current period, wherein the raw physiological signals include at least two of the following: raw electrocardiogram signals, raw heart sound signals, and raw pulse signals. The original physiological signal is preprocessed by a preset signal preprocessing unit to obtain a preprocessed physiological signal. The preprocessing includes at least one of filtering, baseline correction, normalization and effective signal extraction. The target physiological signal is determined based on the preprocessed physiological signal.

[0011] In one embodiment, before the step of inputting the target physiological signal, the detection environment parameters, the blood pressure baseline characteristics, and the blood pressure baseline state into the target blood pressure detection model to obtain the blood pressure value to be calibrated, the method further includes: The target physiological signal is input into a preset blood pressure variability identification model, which outputs multiple blood pressure variability types and the corresponding weight of each blood pressure variability. The blood pressure variability identification model is a deep learning model based on prior physiological knowledge. The target blood pressure detection model is determined from multiple preset blood pressure detection models by mapping the types of multiple blood pressure variability factors and the corresponding weight of each blood pressure variability factor.

[0012] In one embodiment, the blood pressure measurement calibration method further includes: Acquire training data, which includes sample physiological signals, sample environmental parameters, sample baseline features, sample baseline status, and reference blood pressure values. The training data is collected from users corresponding to the same type of blood pressure variability factor. Establish a blood pressure detection model corresponding to the aforementioned blood pressure mutating factor type, wherein the blood pressure detection model includes a deep learning model and a physiological model, and the physiological model is used to provide a theoretical basis for blood pressure detection for the deep learning model; The training data is used as input to the blood pressure detection model to train the blood pressure detection model and obtain a blood pressure detection model that maps the types of blood pressure mutagenic factors.

[0013] In addition, to achieve the above objectives, this application also proposes a wearable device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the blood pressure measurement calibration method as described above.

[0014] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the blood pressure measurement calibration method described above.

[0015] One or more technical solutions proposed in this application have at least the following technical effects: This application acquires IMU data and other detection environment parameters reflecting the human body's state in real time. Then, based on these detection environment parameters, it collects baseline physiological signals under standard human conditions and determines the blood pressure baseline characteristics and blood pressure baseline state. The determination of the blood pressure baseline state depends on whether the difference between the blood pressure baseline characteristics and the target baseline characteristics exceeds a preset threshold. Subsequently, by continuously monitoring the blood pressure baseline state, when the blood pressure baseline state indicates a change in the blood pressure baseline, a calibration prompt message is output to introduce the gold standard blood pressure value collected by the gold standard blood pressure measurement device. The model parameters of the target blood pressure detection model are then adjusted. Thus, the objective change in the blood pressure baseline state serves as the condition for triggering calibration, replacing the original fixed-cycle calibration method. This significantly reduces unnecessary user operations and improves the convenience of the calibration process.

[0016] Meanwhile, by limiting baseline acquisition to standard human conditions to filter out interference from short-term changes, and by setting a threshold for the degree of difference to ensure that calibration is only initiated when significant long-term physiological changes that may affect the model occur, and by combining this with gold standard blood pressure values ​​to accurately correct model parameters, the interference of short-term fluctuations on the calibration process is effectively isolated. This allows the calibration to more accurately target the user's real long-term physiological changes, thereby reducing the calibration frequency while improving the accuracy of calibration and the overall precision of blood pressure measurement. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating an embodiment of the blood pressure measurement calibration method of this application. Figure 2 This is a flowchart illustrating Embodiment 2 of the blood pressure measurement calibration method of this application; Figure 3 This is a flowchart illustrating Embodiment 3 of the blood pressure measurement calibration method of this application; Figure 4 This is a schematic diagram of the framework modules of the blood pressure measurement calibration method provided in Embodiment 3 of this application; Figure 5 This is a schematic diagram of the hardware operating environment involved in the blood pressure measurement calibration method in this application embodiment.

[0020] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0022] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0023] The main solution of this application embodiment is: real-time acquisition of detection environment parameters of the target user based on the wearable device in the current time period, the detection environment parameters including IMU data reflecting the human body state; when the human body state of the target user is identified as a standard human body state according to the detection environment parameters, baseline physiological signals are collected, blood pressure baseline characteristics are determined based on the baseline physiological signals, and blood pressure baseline state is determined based on the blood pressure baseline characteristics, wherein the baseline physiological signals include at least two of baseline electrocardiogram signals, baseline heart sound signals, and baseline pulse signals, the blood pressure baseline characteristics are reference values ​​corresponding to each physiological characteristic in the standard human body state, the standard human body state is a sleep state or a resting state, and the physiological characteristics are used to characterize the blood pressure level. The blood pressure baseline state is defined as either a changed blood pressure baseline or no changed blood pressure baseline. Specifically, a changed blood pressure baseline state occurs when the difference between the blood pressure baseline characteristics and the target baseline characteristics exceeds a preset threshold. In the case of a changed blood pressure baseline state, a first calibration prompt is output. This first calibration prompt instructs the target user to simultaneously collect a gold standard blood pressure value using a gold standard blood pressure measurement device and to detect a blood pressure value to be calibrated using the wearable device. The gold standard blood pressure value collected by the gold standard blood pressure measurement device is obtained, and based on this gold standard blood pressure value, the model parameters of the target blood pressure detection model of the wearable device are optimized and adjusted to adjust the blood pressure value to be calibrated output by the target blood pressure detection model to the gold standard blood pressure value.

[0024] Current cuffless blood pressure measurement technology typically requires periodic calibration of wearable devices, and the existing calibration process suffers from two main problems. First, the calibration process necessitates manual adjustment by the user at fixed intervals, hindering normal use of the wearable device and proving cumbersome, significantly impacting the convenience of cuffless blood pressure measurement. Second, blood pressure measurement is affected by both long-term and short-term factors. Long-term factors include vascular stiffness, which increases slowly with age, while short-term factors include changes in cardiac output and peripheral vascular resistance related to sympathetic nervous system regulation due to diet, exercise, and sleep. Short-term factors often significantly interfere with the calibration process.

[0025] This application provides a solution that acquires real-time detection environment parameters, such as IMU data reflecting the human body's state, and then collects baseline physiological signals under standard human conditions based on these detection environment parameters, determining the blood pressure baseline characteristics and blood pressure baseline state. The determination of the blood pressure baseline state depends on whether the difference between the blood pressure baseline characteristics and the target baseline characteristics exceeds a preset threshold. Subsequently, by continuously monitoring the blood pressure baseline state, when the blood pressure baseline state indicates a change, a calibration prompt is output to introduce the gold standard blood pressure value collected by the gold standard blood pressure measurement device, adjusting the model parameters of the target blood pressure detection model. Thus, objective changes in the blood pressure baseline state serve as the condition for triggering calibration, replacing the original fixed-cycle calibration method, significantly reducing unnecessary user operations and improving the convenience of the calibration process. Meanwhile, by limiting baseline acquisition to standard human conditions to filter out interference from short-term changes, and by setting a threshold for the degree of difference to ensure that calibration is only initiated when significant long-term physiological changes that may affect the model occur, and by combining this with gold standard blood pressure values ​​to accurately correct model parameters, the interference of short-term fluctuations on the calibration process is effectively isolated. This allows the calibration to more accurately target the user's real long-term physiological changes, thereby reducing the calibration frequency while improving the accuracy of calibration and the overall precision of blood pressure measurement.

[0026] It should be noted that the executing entity in this embodiment can be a wearable device with data processing, network communication, and program execution functions, such as a smart bracelet, smartwatch, and head-mounted display device. The head-mounted display device can include, but is not limited to, wearable devices such as smart headphones, Mixed Reality (MR) devices (e.g., MR glasses or MR helmets), Augmented Reality (AR) devices (e.g., AR glasses or AR helmets), Virtual Reality (VR) devices (e.g., VR glasses or VR helmets), Extended Reality (XR) devices, or some combination thereof. The following description uses a wearable device as the executing entity to illustrate this embodiment and the subsequent embodiments.

[0027] Based on this, the embodiments of this application provide a blood pressure measurement calibration method, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the blood pressure measurement calibration method of this application.

[0028] In this embodiment, the blood pressure measurement calibration method includes steps S10 to S40: Step S10: Real-time acquisition of detection environment parameters of the target user based on the wearable device in the current time period, the detection environment parameters including IMU data reflecting the human body state; It should be noted that the target user refers to the wearer of the wearable device; the current time period refers to the time window with start and end timestamps that begins after the wearable device triggers blood pressure measurement; the detection environment parameters refer to a set of auxiliary data collected synchronously with the original physiological signals, which describes the external physical conditions and the user's own physical behavior state during the collection of physiological signals, and serve as contextual feature inputs for subsequent signal processing, model detection, and model calibration. These parameters typically include, but are not limited to, IMU (Inertial Measurement Unit) data, and may also include skin temperature, pressure applied by the wearable device to the measurement location on the body surface, etc.

[0029] IMU data refers to information about the motion state of an object in space collected by inertial sensors, including data such as acceleration and angular velocity, which is used to reflect the state of the human body. For example, the raw IMU data can be processed by algorithms (such as activity recognition and posture estimation) to transform it into a semantic classification of the user's current body movements, such as classification as "stationary", "walking", "running", "upper arm raised" and other states.

[0030] Step S20: When the target user's human body state is identified as a standard human body state based on the detection environment parameters, baseline physiological signals are collected, blood pressure baseline characteristics are determined based on the baseline physiological signals, and blood pressure baseline state is determined based on the blood pressure baseline characteristics. The baseline physiological signals include at least two of the following: baseline electrocardiogram signals, baseline heart sound signals, and baseline pulse signals. The blood pressure baseline characteristics are reference values ​​corresponding to each physiological characteristic under the standard human body state, which is either a sleep state or a resting state. The physiological characteristics are used to characterize blood pressure levels. The blood pressure baseline state is either a change in blood pressure baseline or no change in blood pressure baseline. Specifically, if the difference between the blood pressure baseline characteristics and the target baseline characteristics exceeds a preset threshold, the blood pressure baseline state is considered a change in blood pressure baseline. Among them, the blood pressure baseline features are the reference values ​​corresponding to each physiological feature under standard human conditions. The standard human conditions are sleep or rest. The physiological features are used to characterize the blood pressure level. The blood pressure baseline status is either a change in the blood pressure baseline or no change in the blood pressure baseline. In the case where the difference between the blood pressure baseline features and the target baseline features stored in the target blood pressure calculation model is greater than a preset threshold, the blood pressure baseline status is a change in the blood pressure baseline.

[0031] It should be noted that baseline blood pressure characteristics refer to the values ​​of various physiological characteristics detected in a target user under standard human conditions (such as sleep or rest), used to characterize the normal level of blood pressure in a target user under standard human conditions. Physiological characteristics refer to various measurable signals or indicators that can reflect the physiological state and blood pressure level of the human body, such as heart rate, the position and amplitude of S1 and S2 in the PCG signal, pulse wave, pre-ejection time, blood oxygen saturation, etc.

[0032] Standard human condition refers to predefined specific human condition conditions used to obtain baseline blood pressure characteristics. In baseline characteristic acquisition, sleep or resting state is used as the standard because the physiological state of the human body is relatively stable in these two states, allowing for more accurate acquisition of physiological characteristic reference values ​​reflecting blood pressure levels. Sleep refers to a natural physiological cycle state in which the body's conscious activity is relatively reduced, and the ability to respond to external stimuli is decreased. Due to the excitation of the parasympathetic nervous system and the inhibition of the sympathetic nervous system during sleep, blood pressure during sleep is 10%-20% lower than blood pressure when awake. The model needs to incorporate this difference when calculating blood pressure. Resting state refers to a state of quiet, relaxed state when awake, without strenuous exercise or significant mental stress. At this time, the body's activity level is low, energy consumption is low, and the functions of various organ systems are in a relatively stable state. Blood pressure values ​​measured at this time are usually close to normal baseline blood pressure levels.

[0033] Blood pressure baseline status refers to a status indicator determined based on the degree of difference between blood pressure baseline characteristics and target baseline characteristics. It is used to indicate whether the blood pressure baseline has changed and provides additional reference information for the model to detect blood pressure values.

[0034] The target baseline features refer to the blood pressure baseline features that were calculated and saved in the previous blood pressure measurement process under standard human conditions and stored in the target blood pressure detection model.

[0035] The degree of difference refers to the magnitude of the difference between the currently measured blood pressure baseline characteristics and the target baseline characteristics. It can be represented by calculating the difference between the two, the relative rate of change, the Euclidean distance, or the weighted comprehensive score for the difference in each feature classification. The larger the value, the greater the difference between the newly measured blood pressure baseline characteristics and the previously stored target baseline characteristics.

[0036] The preset threshold is a pre-set critical value used to determine whether the difference is significant.

[0037] For example, when a target user is watching TV in the evening or sleeping, their body may be in a standard human state. The blood pressure baseline monitoring unit is activated, and the ECG sensor, heart sound sensor, and pulse sensor begin to collect baseline physiological signals. Based on the collected baseline ECG signal, baseline PCG signal, and baseline PPG signal, blood pressure baseline characteristics, such as heart rate reference value and pulse wave velocity reference value, are determined. The currently measured blood pressure baseline characteristics are compared with the target baseline characteristics stored in the target blood pressure calculation model. The degree of difference between the two is calculated using methods such as Euclidean distance or weighted summation, and it is determined whether the degree of difference is greater than a preset threshold. If the degree of difference is greater than the preset threshold, the blood pressure baseline state is determined to be changed; if the degree of difference is less than or equal to the preset threshold, the blood pressure baseline state is determined to be unchanged.

[0038] Step S30: When the blood pressure baseline status is changed, output the first calibration prompt information. The first calibration prompt information is used to instruct the target user to simultaneously collect the gold standard blood pressure value through the gold standard blood pressure measurement device and detect the blood pressure value to be calibrated through the wearable device. The first calibration prompt is a user guidance signal generated and output after a change in the blood pressure baseline is detected. Its purpose is to proactively notify the user to initiate the calibration process when a significant change in the user's long-term physiological state is detected that may affect the accuracy of the blood pressure measurement model. Specifically, this message instructs the user to simultaneously perform two actions: first, to collect a standard blood pressure value (i.e., the gold standard blood pressure value) using a gold standard blood pressure measuring device; and second, to simultaneously perform a blood pressure measurement using a wearable device to obtain the blood pressure value to be calibrated. This transforms the calibration action from being triggered by a fixed period to being triggered by changes in objective physiological state, a key human-computer interaction element for achieving calibration when necessary rather than periodically.

[0039] Gold standard blood pressure measurement devices refer to a type of independent hardware device that is pre-designated or recognized by the medical community to provide reference values ​​for blood pressure measurement. The accuracy of blood pressure measurement is higher than the preset accuracy threshold, and it is used to provide high-precision blood pressure reference values. Examples include ambulatory blood pressure monitors (ABPM) and upper arm oscillometric electronic blood pressure monitors. The blood pressure values ​​measured by these devices can be input into wearable devices as gold standard blood pressure values ​​through wireless connection or manual input.

[0040] In this context, the "gold standard blood pressure value" refers to the blood pressure measurement obtained by the user in response to the first calibration prompt using an external, medically compliant, or certified blood pressure measurement device—that is, a gold standard blood pressure measurement device. In this embodiment, this value is considered the true value or reference standard of blood pressure at the current moment. Its function is to serve as a benchmark for optimizing and adjusting the target blood pressure detection model: adjusting the model parameters aims to make the blood pressure value output by the wearable device infinitely close to or equal to this gold standard blood pressure value. Introducing the gold standard blood pressure value is an essential external reference for achieving calibration accuracy and aligning the model with the user's current actual state.

[0041] Step S40: Obtain the gold standard blood pressure value collected by the gold standard blood pressure measurement device, and based on the gold standard blood pressure value, optimize and adjust the model parameters of the target blood pressure detection model of the wearable device to adjust the blood pressure value to be calibrated output by the target blood pressure detection model to the gold standard blood pressure value.

[0042] The target blood pressure detection model is an algorithmic model embedded in or independent of a wearable device, used to calculate a user's blood pressure value. In this embodiment, the model's input includes at least the real-time acquired target physiological signal, detection environment parameters, blood pressure baseline features, and blood pressure baseline status. This model is a pre-trained mathematical model capable of characterizing the complex mapping relationship between physiological signals and blood pressure. It can be a specific model instance currently selected or in use for the user based on the current input data, and serves as the direct object for blood pressure detection and subsequent parameter optimization.

[0043] Understandably, existing cuffless blood pressure measurement technologies, due to their fixed-period calibration, can be inconvenient for users, and the calibration process is easily affected by short-term physiological fluctuations, impacting accuracy. Therefore, this embodiment identifies users in standard human states such as sleep or rest by real-time monitoring of environmental parameters. It collects baseline physiological signals and determines blood pressure baseline characteristics. Then, it objectively determines whether the blood pressure baseline state has changed based on whether the difference between these characteristics and the target baseline characteristics exceeds a threshold. Thus, the first calibration prompt is triggered only when the state is confirmed to have changed, guiding the user to simultaneously collect both the gold standard blood pressure value and the blood pressure value to be calibrated. The gold standard blood pressure value is then used to optimize and adjust the parameters of the target blood pressure detection model. This embodiment avoids unnecessary frequent calibrations and improves convenience by using objective physiological state changes instead of fixed periods as the sole triggering condition for calibration. Simultaneously, by limiting the evaluation of baseline characteristics to standard human states, it effectively filters out the interference of short-term fluctuations such as diet and exercise on the calibration criteria, avoiding misjudgments of calibration timing and ensuring that calibration only targets real long-term physiological changes. This significantly reduces the user's operational burden while simultaneously improving calibration accuracy and blood pressure measurement precision.

[0044] Model parameters are the set of internal adjustable variables or weights that constitute the target blood pressure monitoring model. These parameters determine how the model processes input signals and ultimately outputs blood pressure values. Optimizing and adjusting model parameters refers to modifying these internal variables using specific algorithms (such as backpropagation, least squares, etc.) with the gold standard blood pressure as the optimization target, so that when the model receives the same or similar input data, its output value (i.e., the blood pressure value to be calibrated) can be brought closer to or aligned with the gold standard blood pressure value. Adjusting model parameters is a direct means of adapting a generalized or outdated model to the user's current specific long-term physiological state; it is the concrete implementation of calibration at the algorithmic level.

[0045] For example, firstly, the wearable device continuously monitors the target user's motion data in real time using its built-in IMU as the core detection environment parameter. When the pre-set human state recognition unit in the wearable device determines that the user has entered a standard human state such as sleep or rest based on the detection environment parameter, it automatically triggers the blood pressure baseline monitoring unit. Simultaneously, it collects the user's electrocardiogram signal and photoplethysmography (PPG) signal as baseline physiological signals, and extracts key physiological features such as pulse wave transit time and heart rate variability from them, calculating their average value as the current blood pressure baseline feature. Subsequently, this blood pressure baseline feature is compared with the target baseline feature pre-stored in the device. If the difference obtained by the preset algorithm (such as calculating Euclidean distance) exceeds a preset threshold, the comparison is performed. If the value is not found, the blood pressure baseline status is determined to be "blood pressure baseline changed". Once the status is confirmed to be changed, the wearable device immediately outputs the first calibration prompt message through its display screen or associated mobile application, clearly instructing the user to use the other arm to measure blood pressure with a certified upper arm electronic blood pressure monitor while maintaining the current posture to obtain the gold standard blood pressure value. At the same time, the device also completes a blood pressure measurement to obtain the blood pressure value to be calibrated. Finally, the wearable device obtains the gold standard blood pressure value input by the user and uses this value as the optimization target. Through parameter optimization algorithms such as gradient descent, it adjusts the weight parameters of its internal target blood pressure detection model so that the blood pressure value to be calibrated calculated by the model for this synchronous measurement input is consistent with the gold standard blood pressure value, thereby completing this calibration.

[0046] This embodiment provides a blood pressure measurement calibration method. It acquires real-time detection environment parameters, such as IMU data reflecting the human body's state, and then collects baseline physiological signals under standard human conditions based on these parameters. The method determines the blood pressure baseline characteristics and status. The determination of the blood pressure baseline status depends on whether the difference between the blood pressure baseline characteristics and the target baseline characteristics exceeds a preset threshold. Subsequently, by continuously monitoring the blood pressure baseline status, when a change in the blood pressure baseline is indicated, a calibration prompt is output to introduce the gold standard blood pressure value collected by the gold standard blood pressure measurement device. This adjusts the model parameters of the target blood pressure detection model. Thus, objective changes in the blood pressure baseline status serve as the trigger for calibration, replacing the original fixed-cycle calibration method. This significantly reduces unnecessary user operations and improves the convenience of the calibration process. Meanwhile, by limiting baseline acquisition to standard human conditions to filter out interference from short-term changes, and by setting a threshold for the degree of difference to ensure that calibration is only initiated when significant long-term physiological changes that may affect the model occur, and by combining this with gold standard blood pressure values ​​to accurately correct model parameters, the interference of short-term fluctuations on the calibration process is effectively isolated. This allows the calibration to more accurately target the user's real long-term physiological changes, thereby reducing the calibration frequency while improving the accuracy of calibration and the overall precision of blood pressure measurement.

[0047] In one feasible implementation, the environmental parameters to be detected further include pressure data applied by the wearable device to the target user's body surface measurement location, and temperature data of the target user at a preset body part. The blood pressure measurement calibration method may further include steps S100~S200: Step S100: Using a preset human body state recognition unit, determine whether the detection environment parameters meet the preset blood pressure measurement conditions. The human body state recognition unit includes an inertial sensor, a pressure sensor, and a temperature sensor. It should be noted that the human body state recognition unit refers to the module in a wearable device used to collect user state information, typically including inertial sensors, pressure sensors, and temperature sensors. An inertial sensor is a sensor capable of measuring an object's three-axis attitude angles (or angular rates) and acceleration. It can sense the object's motion state, including direction, speed, and acceleration, and is commonly used in motion monitoring and navigation positioning. A pressure sensor is a device or apparatus that can sense pressure signals and convert them into usable electrical signals according to a certain rule. It is mainly used to measure the pressure exerted on an object's surface. A temperature sensor is a sensor that can sense temperature and convert it into a usable output signal. It can measure the temperature of an object or environment and convert the temperature signal into an electrical signal or other forms of signal for transmission and processing.

[0048] In this embodiment, pressure data refers to the pressure value applied by the wearable device to the target user's body surface measurement location, which is collected by a pressure sensor to reflect the degree of contact between the wearable device and the skin and the pressure distribution; temperature data refers to the temperature value measured by a temperature sensor at a preset body part of the target user (such as behind the ear, the skin of the wrist where the wearable device contacts the target user, etc.).

[0049] Blood pressure measurement conditions refer to a set of parameters or standards pre-set in wearable devices to determine whether the user's current state is suitable for blood pressure measurement. These conditions are usually set based on medical standards and the measurement requirements of the device. For example, blood pressure measurement conditions may include the user being at rest (determined by IMU data), the device being worn correctly (determined by pressure data), and body temperature being within the normal range (determined by temperature data).

[0050] Step S200: When the detection environment parameters meet the blood pressure measurement conditions, perform the step of optimizing and adjusting the model parameters of the target blood pressure detection model of the wearable device based on the gold standard blood pressure value.

[0051] For example, after collecting environmental parameters through the human body state recognition unit, the IMU data can be analyzed. Based on changes in velocity and acceleration data, it can be identified whether the human body is in motion (such as walking or running) or at rest, and the duration of each state. If it is detected that the user has not rested for a sufficient period of time, failing to meet the blood pressure measurement calibration standards, the user is prompted to continue resting and recalibrate. Alternatively, if the user's state meets the blood pressure measurement calibration standards, but during the calibration period there are multiple instances of body movement or abnormal pressure from the wearable device and body contact, affecting signal quality, and if there are too many invalid signals, the user is prompted to remeasure and calibrate. If the valid signals are sufficient for blood pressure calculation, invalid data can be removed during signal preprocessing, combined with the state markers from the state recognition unit, and valid data can be used for blood pressure measurement calibration.

[0052] In this embodiment, by collecting detection environment parameters such as IMU data, pressure data, and temperature data, the physiological and environmental states of the user during blood pressure measurement are comprehensively captured to determine whether the current detection environment is suitable for blood pressure measurement calibration. By determining whether the detection environment parameters meet the preset blood pressure measurement conditions, environmental factors and human condition interference that are not conducive to accurate measurement can be eliminated. For example, blood pressure values ​​will fluctuate in high temperature and high humidity environments or when the human body is in a state of exercise or emotional excitement, and the blood pressure measurement results will be inaccurate at this time. The calibration process is only performed when the environmental parameters and human condition meet the conditions to obtain more realistic and accurate blood pressure data.

[0053] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 2 Before step S10, the blood pressure measurement calibration method further includes steps S01 to S04: Step S01: When it is detected that the target user is wearing the wearable device for the first time, the detection environment parameters of the target user based on the wearable device are obtained for the first time. It is understood that detecting the first time a target user wears the wearable device can be achieved through any or a combination of the following methods: including but not limited to detecting that the wearable device is powered on for the first time and has completed its hardware self-test; detecting that there is no historical user data or physiological signal record in the internal storage unit of the wearable device; recognizing that the wearable device and the mobile application have completed the pairing process for the first time; detecting a contact signal consistent with human characteristics for the first time through a pressure sensor or bioimpedance sensor; receiving a physiological signal measurement command triggered for the first time in the absence of any historical calibration records or personalized model parameters; or explicitly receiving a setting command from the user to select "new user" or "first use" through the user interface. This embodiment does not make specific limitations on the method of detecting the first time a target user wears the wearable device.

[0054] Step S02: If the target user's human body state is identified as a standard human body state based on the first acquired detection environment parameters, the baseline physiological signal is collected for the first time, and the collection time of the first collection of the baseline physiological signal is accumulated. The initial baseline physiological signal acquisition duration refers to the total time spent continuously and effectively acquiring baseline physiological signals when the target user first wears and uses the wearable device, in order to establish their personalized blood pressure monitoring baseline, while recognizing the user in a standard human state. This duration is a key indicator for determining whether sufficient data has been collected to reliably determine the user's initial blood pressure baseline characteristics. Only when this cumulative duration exceeds a preset duration threshold (e.g., several nights of sleep cycles) does the system consider stable and representative personal baseline data to have been obtained, thus allowing the subsequent initial calibration process to be triggered, thereby ensuring the accuracy of the initial personalization of the model.

[0055] Step S03: If the collection time exceeds the preset time threshold, stop accumulating the collection time and output a second calibration prompt message. The second calibration prompt message is used to instruct the target user to simultaneously collect the gold standard blood pressure value for the first time through the gold standard blood pressure measurement device and detect the blood pressure value to be calibrated for the first time through the wearable device. The second calibration prompt is a guiding message automatically generated and output by the system after the device completes its initial baseline acquisition. Its purpose is to guide the user in completing the initial calibration of the target blood pressure detection model during the initial use of the device. This message explicitly instructs the user to perform two operations simultaneously: first, perform an initial standard blood pressure measurement using a gold standard blood pressure measuring device to obtain a baseline gold standard blood pressure value; and second, perform an initial blood pressure measurement using this wearable device to obtain an initial blood pressure value to be calibrated. Through this process, the system can use the acquired first gold standard blood pressure value to make initial adjustments to the parameters of the general model, which has not yet undergone any personal adaptation, thereby initially adapting the model to the current user's physiological state and completing the initial setup from a general model to a personalized model. Furthermore, this second calibration prompt may or may not be consistent with the first calibration prompt; this embodiment does not specifically limit the content of the second calibration prompt.

[0056] Step S04: Obtain the gold standard blood pressure value collected for the first time by the gold standard blood pressure measuring device, and based on the gold standard blood pressure value collected for the first time, adjust the model parameters of the target blood pressure detection model for the first time, so as to adjust the blood pressure value to be calibrated detected by the wearable device for the first time to the gold standard blood pressure value collected for the first time.

[0057] Understandably, since first-time wearable device users haven't established any personalized blood pressure baseline characteristics and model parameters, directly applying a general model for measurement would result in severely insufficient initial accuracy. Therefore, this embodiment employs a technical solution of "first-time wear detection—cumulative baseline acquisition time—triggered first calibration after reaching the set time threshold." This forces the establishment of a reliable personal physiological baseline upon the user's first use. Specifically, the system first detects the first-time wear event, then continuously acquires baseline physiological signals under standard human conditions until the cumulative time meets a preset threshold, ensuring a stable initial blood pressure baseline characteristic. Only then does it output a second calibration prompt, guiding the user to initialize model parameters based on the gold standard blood pressure value. This solution avoids the problems of large initial measurement errors and low user trust caused by hastily operating the device without personal data. By ensuring personalized adaptation of the model from the outset, it lays an accurate benchmark for subsequent long-term dynamic calibration, thereby achieving measurement accuracy throughout the device's entire lifespan.

[0058] For example, when a user first activates and wears the wearable device, the device detects the absence of any historical user data in its internal memory and the first continuous contact with the body surface via the pressure sensor, confirming this as the first wearing event. Subsequently, the wearable device begins real-time monitoring of IMU data. When it recognizes the user entering sleep at night—a standard human state—it automatically initiates the initial baseline acquisition process, simultaneously recording electrocardiograms and photoplethysmography (PPG) waves as baseline physiological signals, and continuously accumulating effective acquisition time, such as accumulating at least 6 hours of sleep signals for three consecutive nights. When the accumulated time exceeds a preset 72 hours... After reaching the threshold, the device stops timing and displays a second calibration prompt on the screen, instructing the user to measure their blood pressure using an upper arm medical blood pressure monitor at rest the following morning to obtain the initial gold standard blood pressure value. The user then operates the wearable device to complete a blood pressure measurement to obtain the initial calibrated blood pressure value. After the user inputs the initial gold standard blood pressure value into the wearable device, the device uses this value as a target and employs a gradient descent algorithm to adjust the weight parameters of its neural network target blood pressure detection model, ensuring that the model outputs a calibrated blood pressure value consistent with the gold standard blood pressure value, thus completing the personalized initialization of the model.

[0059] In this embodiment, the wearable device detects the user's state upon first wear and accumulates baseline physiological signal collection time under standard human conditions. After the time reaches the target, it triggers the first calibration process. This avoids the problem of large initial measurement errors and low user trust caused by directly applying a general model when the wearable device is first used due to the lack of a user's personal physiological benchmark. By ensuring that the model must be based on a sufficiently long period of stable personal baseline data before being put into use, and completing the initial adaptation based on the gold standard blood pressure, a high-precision personalized measurement starting point is established for the device. This lays a solid foundation for subsequent long-term reliable dynamic monitoring and calibration, thereby comprehensively improving the initial measurement accuracy of the device.

[0060] In one feasible implementation, before the step of outputting the second calibration prompt information in step S03, steps A01 to A03 may also be included: Step A01: Generate a second calibration prompt message through a preset model calibration unit; Step A02: The target user's human body status is identified by a preset human body status recognition unit based on the detection environment parameters. Step A03: When the human body is identified as being in a preset calibration requirement state, the step of outputting the second calibration prompt information is executed, wherein the calibration requirement state includes a resting state.

[0061] It should be noted that the calibration requirement state refers to a set of specific human states suitable for the user to actively perform calibration operations, determined by the human state recognition unit based on the detection environment parameters before the initial model calibration, i.e., before responding to the second calibration prompt. The core characteristic of this state is that the user is awake, able to cooperate, and physiologically relatively stable. For example, the explicitly included "resting state" refers to a state where the user is awake but keeps their body still without significant movement. This ensures that the user can simultaneously operate the gold standard device and the wearable device, while their physiological signals are not interfered with by vigorous activity, thereby guaranteeing the validity of the data collected during the initial calibration and the feasibility of the calibration process. This standard requirement state may not be exactly the same as the standard human state used for passively acquiring baseline signals; the emphasis is on the user's awareness and cooperation required for active calibration participation.

[0062] Understandably, since the initial calibration requires the user's active cooperation and synchronized operation, if the user is in a state where they cannot respond or are not suitable for accurate measurement (such as being asleep) when the calibration prompt is issued, the calibration process will not be able to be executed or the data will be invalid. Therefore, this embodiment further determines whether the user is in a preset calibration requirement state, such as a conscious resting state, before outputting the second calibration prompt information. This ensures that the calibration prompt is issued only when the user is able and suitable to perform synchronized measurement, avoiding problems such as calibration operation interruption, user cooperation difficulties, or inaccurate measurement data due to inappropriate prompt timing. Thus, while ensuring the quality of the initial calibration data, it improves the smoothness of the calibration process and the user experience, effectively increasing the success rate of the initial calibration.

[0063] For example, when the wearable device completes its initial baseline acquisition and is ready to trigger calibration, the preset human state recognition unit analyzes the detection environment parameters from sensors such as IMU and heart rate in real time to determine whether the user is currently in a calibration requirement state suitable for the first active calibration. This calibration requirement state is predefined as a series of specific situations in which the user is conscious, physically still, and has stable physiological indicators, such as: lying still in bed after waking up in the morning, sitting still while working or reading for a long time during the day, transitional resting state after waking up from a nap but not engaging in any activity, or pre-sleep state of relaxation and no physical movement before going to sleep at night. Only when the system recognizes that the user is in one of these states will it generate and output a second calibration prompt message through the model calibration unit, thereby ensuring that the user can cooperate in a timely and effective manner to complete the subsequent synchronous measurement and calibration operations.

[0064] In this embodiment, before outputting the second calibration prompt, the user's human body state is identified in real time by the human body state recognition unit. The prompt is only triggered when the state meets the preset calibration requirements. This avoids the problems of calibration process failure, user operation difficulty, or invalid measurement data caused by the user being asleep or in an unstable physiological state when the calibration prompt is issued. This ensures that the first calibration process is started only when the user is awake, cooperative, and has stable physiological indicators. This significantly improves the success rate of the first calibration operation and the validity of calibration data, thereby enhancing the user experience and the reliability of the calibration process.

[0065] Based on the first and / or second embodiments of this application, in the third embodiment of this application, the content that is the same as or similar to that in embodiments one and two above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 3 After step S30, which involves outputting the first calibration prompt information, steps S301-S302 may also be included: Step S301: Real-time acquisition of target physiological signals detected by the wearable device based on the target user during the current time period, wherein the target physiological signals include at least two of the target electrocardiogram signals, target heart sound signals, and target pulse signals; As those skilled in the art will know, a physiological signal is a measurable physical quantity that reflects the physiological state or functional activity of an organism. Common types include electrical signals (such as electrocardiogram and electroencephalogram), acoustic signals (such as heart sounds and breath sounds), and mechanical signals (such as pulse and blood pressure), which are used for non-invasive monitoring of human health status.

[0066] Electrocardiogram (ECG) refers to the cardiac electrical activity signal collected by sensors such as ECG surface electrodes. It reflects the depolarization and repolarization process of the myocardium. Its typical waveforms include the P wave, QRS complex (including Q wave and R wave), and T wave. The R wave is the highest amplitude positive wave in the QRS complex and is often used to mark the onset of cardiac electrical activity. The Q wave is the first negative wave in the QRS complex before the R wave and is often used to mark the onset of ventricular depolarization and the entry into the pre-ejection phase.

[0067] Phonocardiogram (PCG) refers to the acoustic signal collected by sensors such as microphones and accelerometers, which is generated by mechanical vibrations such as the closure of heart valves and blood flow turbulence. It mainly includes the first heart sound (S1, corresponding to the closure of the mitral and tricuspid valves) and the second heart sound (S2, corresponding to the closure of the aortic and pulmonary valves). Its internal multi-peak structure contains information on valvular dynamics and cardiac function status.

[0068] Pulse signal refers to the peripheral arterial pulsation signal collected by photoplethysmography (PPG), pressure sensor, etc., which reflects the changes in blood volume caused by the heart pumping blood. Its periodic trough point usually corresponds to the start of ventricular diastole and can be used to calibrate the end of pulse wave transit time (PTT). PPG signal refers to photoplethysmography pulse wave signal (pulse signal).

[0069] It should be noted that the target physiological signal refers to the multimodal physiological signal synchronously acquired within the current time period, including at least two of the target electrocardiogram signal, target heart sound signal, and target pulse signal. Each physiological signal is strictly aligned in time to ensure physiological consistency in subsequent joint analysis based on temporal relationships.

[0070] Step S302: Input the target physiological signal, the detection environment parameters, the blood pressure baseline characteristics, and the blood pressure baseline state into the target blood pressure detection model to obtain the blood pressure value to be calibrated.

[0071] Understandably, by continuously updating and introducing blood pressure baseline features, the blood pressure detection model can track the slow physiological changes caused by age, training, medication, etc., and use the latest personal benchmark during detection. This is equivalent to providing a dynamically updated "zero calibration point" for general or scenario-based models, thereby solving the problem of model accuracy decreasing due to user physiological changes (i.e., model drift) in long-term use, effectively reducing interference from individual differences, and improving the accuracy of blood pressure measurement.

[0072] The step S40, which involves optimizing and adjusting the model parameters of the target blood pressure detection model of the wearable device based on the gold standard blood pressure value, may further include step S41: Step S41: The gold standard blood pressure value is used as the model output target and input into the target blood pressure detection model, so that the target blood pressure detection model adjusts the blood pressure value to be calibrated to the gold standard blood pressure value for output by adjusting the model parameters.

[0073] It should be noted that the model output target refers to the specific numerical benchmark or ideal value that the model output value needs to approximate during the process of optimizing and adjusting the parameters of the target blood pressure detection model. That is, the gold standard blood pressure value synchronously collected by the user through the gold standard blood pressure measurement device. Its role is to provide a clear and objective optimization direction for the model parameter optimization process: when the model receives the current input data (including target physiological signals, detection environment parameters, etc.) and obtains a blood pressure value to be calibrated for the upcoming output, the system will use the gold standard blood pressure value as the target that the output should achieve. By adjusting the internal parameters of the model, the model's output value can be directly equal to or as close as possible to the gold standard blood pressure value for this set of input data.

[0074] It is understandable that simply triggering calibration and obtaining the gold standard blood pressure value may not be enough to automatically and accurately optimize and adjust the model parameters. This results in an unclear mapping between the model output (i.e., the blood pressure value to be calibrated) and the calibration target (i.e., the gold standard blood pressure value), and a lack of direct guidance for parameter adjustment. To address this, this embodiment further employs a technical solution that acquires the target physiological signal in real time after the output prompt and inputs it into the model to obtain the blood pressure value to be calibrated. This explicitly sets the gold standard blood pressure value as the model output target to drive parameter adjustment, constructing a complete closed-loop calibration circuit. Specifically, after prompting the user to take a synchronous measurement, the system actively collects data such as the target physiological signal at that moment and calculates the blood pressure value to be calibrated through the model. Then, the gold standard blood pressure value is used as the sole optimization target to guide the algorithm in reverse adjustment of the model parameters until the model can output a blood pressure value consistent with the gold standard value for the same input. This effectively avoids the shortcomings of the calibration process remaining at the level of manual triggering and lacking automated and precise parameter tuning. It achieves a fully automated closed loop from prompting to precise correction in the calibration process, significantly improving the targeting and efficiency of parameter adjustment, thereby ensuring that each calibration ensures that the model output accurately and reliably converges to the standard reference value.

[0075] For example, when the wearable device outputs the first calibration prompt message due to a change in the blood pressure baseline, the device immediately and simultaneously acquires the target electrocardiogram signal and target pulse wave signal of the user at this moment through its PPG and ECG sensors as target physiological signals. These signals, along with real-time IMU data and other detection environment parameters, as well as the determined blood pressure baseline characteristics and status, are input into a preset target blood pressure detection model, such as a neural network model. The model then calculates a specific blood pressure value to be calibrated. After the user simultaneously measures the gold standard blood pressure value using an upper arm electronic blood pressure monitor and inputs it into the device, the system explicitly sets this gold standard blood pressure value as the sole model output target for this model optimization. Through the backpropagation algorithm, the internal weight parameters of the target blood pressure detection model are automatically adjusted so that the model's output value is adjusted to be consistent with the gold standard blood pressure value while keeping the current input data unchanged, thereby completing the calibration process.

[0076] In this embodiment, the target physiological signal is acquired in real time after calibration is triggered, and a blood pressure value to be calibrated is generated. The gold standard blood pressure value is then explicitly set as the model output target to drive the automatic adjustment of model parameters. This avoids the problems of subjective operation, unclear adjustment direction, low calibration efficiency, and difficulty in guaranteeing accuracy caused by the lack of a specific and automated parameter adjustment mechanism after calibration is triggered. The entire calibration process, from triggering and measurement to parameter optimization, is automated and closed-loop. By using the gold standard blood pressure value as a clear optimization target to directly guide the iteration of model parameters, it is ensured that the model output can quickly and accurately converge to the standard reference value in each calibration, thereby significantly improving the accuracy, reliability, and overall efficiency of blood pressure measurement calibration.

[0077] In one feasible implementation, step S301 may further include steps A301 to A303: Step A301: Based on the wearable device, the target user's raw physiological signals in the current time period are collected in real time, wherein the raw physiological signals include at least two of the raw electrocardiogram signals, raw heart sound signals and raw pulse signals; It should be noted that the raw physiological signal refers to the physiological signal in analog or digital form directly acquired by the sensor without any preprocessing. It retains all time and frequency domain information of the physiological signal, but also includes non-physiological components such as environmental noise, motion artifacts, and power supply interference. In this embodiment, the raw physiological signal includes at least two of the following: raw electrocardiogram signal, raw pulse signal, and raw heart sound signal, and is consistent with the target physiological signal.

[0078] For example, a multimodal sensor array (including ECG surface electrodes, PCG vibration sensors, PPG photoelectric sensors, etc.) on a wearable device can be used to synchronously acquire the raw physiological signals of the target user at the current time period, ensuring that each physiological signal is strictly aligned on the time axis. For instance, in a smart bracelet or smartwatch, ECG electrodes, PCG vibration sensors, and PPG photoelectric sensors can be integrated to achieve hardware-level synchronous acquisition of the three types of raw physiological signals: ECG, PCG, and PPG. The sampling frequency can be set according to the characteristics of each physiological signal, such as acquiring ECG and PCG signals at a sampling frequency of 500Hz and acquiring PPG signals at a sampling frequency of 100Hz. The sampling duration can be set to 3 minutes to ensure that a set of signals includes multiple cardiac cycles.

[0079] For example, after initiating blood pressure measurement, the user places the arm with the wearable device in front of their chest, transmitting the PCG signal through the arm to the device, where it is received by the bone conduction VPU (Video Processing Unit) sensor and IMU sensor inside the wearable device; the other finger presses the ECG electrode on the device to measure the single-lead ECG signal, while the PPG signal is collected by the photoelectric sensor on the side of the wearable device close to the wrist.

[0080] Step A302: The original physiological signal is preprocessed by a preset signal preprocessing unit to obtain a preprocessed physiological signal. The preprocessing includes at least one of filtering, baseline correction, normalization and effective signal extraction. It should be noted that the signal preprocessing unit refers to the module in a wearable device used to preprocess the acquired raw physiological signals. It typically contains a series of algorithms and hardware circuits to perform operations such as filtering, correction, and normalization.

[0081] Filtering refers to using specific filtering algorithms or filters to suppress or enhance specific frequency components in a signal in order to remove noise and interference and retain useful signal components. Common filtering methods include low-pass filtering, high-pass filtering, and band-pass filtering.

[0082] Baseline correction refers to estimating and subtracting the low-frequency slow-changing trend (baseline) in a signal using algorithms (such as moving average, polynomial fitting, morphological operations, etc.) to restore the zero line or center line of the signal to a normal level.

[0083] Normalization refers to scaling the values ​​of different signals to a specific range to eliminate the influence of differences in dimensions and numerical ranges between different signals.

[0084] Effective signal extraction refers to identifying and separating useful signal components that reflect blood pressure levels from the original signal, and removing signal parts that are irrelevant to changes in blood pressure or interfere with the analysis. For example, effective pulse waves can be identified from pulse waves based on amplitude thresholds.

[0085] For example, the signal preprocessing unit can use a high-pass filter or polynomial fitting to remove baseline drift of the original physiological signal, and use a low-pass filter or adaptive filter to remove muscle noise in the original physiological signal; then, it can perform minimum-maximum normalization on the amplitude of each physiological signal to scale the signal amplitude range to the [0,1] interval; and then, based on the signal-to-noise ratio of each physiological signal or physiological prior knowledge, it can remove invalid physiological signal data and signal data that are unrelated to changes in blood pressure.

[0086] Optionally, after the detection environment parameters are collected, they can also be input to the signal preprocessing unit for preprocessing, such as filtering and calibration, to improve the quality and reliability of the detection environment parameters.

[0087] Understandably, by filtering out noise, correcting the baseline to eliminate drift, normalizing to unify the scale, and extracting effective signals to remove interference, the detectability of key feature points (such as R-wave peaks, pulse signal troughs, S1 / S2 main peaks, etc.) is enhanced, which can more accurately reflect the user's physiological state and thus improve the accuracy of blood pressure measurement.

[0088] Step A303: Determine the target physiological signal based on the preprocessed physiological signal.

[0089] For example, the preprocessed physiological signal can be directly identified as the target physiological signal of the target user. Alternatively, it can be combined with the pressure data of the pressure sensor in the wearable device, which is applied to the body surface measurement location. Data that does not meet the requirements can be discarded to avoid interfering with the accuracy of blood pressure calculation. For example, if the pressure value during the physiological signal acquisition process is higher or lower than a certain range, the blood pressure measurement data may not meet the requirements of the model calculation, and such data will be discarded.

[0090] For example, the preprocessed physiological signals can also be divided into cardiac cycles. For instance, the time interval between two R waves in the ECG signal can be defined as a cardiac cycle, and the PCG and PPG signals can be divided simultaneously according to the division of the cardiac cycle of the ECG signal, so as to compare and analyze the changing trends of ECG, PCG and PPG signals in different cardiac cycles.

[0091] In this embodiment, by synchronously acquiring raw physiological signals, the timing consistency of multiple signals such as electrocardiogram signals, heart sound signals, and pulse signals is ensured, so as to accurately calculate key features such as pulse wave propagation time and improve the underlying accuracy of blood pressure measurement. Furthermore, through operations such as filtering, baseline correction, and normalization, the signal-to-noise ratio and quality reliability of multimodal signals are improved, effectively overcoming the interference problem in users' daily use scenarios. This is an indispensable preliminary step for achieving high-precision blood pressure measurement.

[0092] In one feasible implementation, steps A100 to A200 may be included before step S302: Step A100: Input the target physiological signal into a preset blood pressure variability identification model, and output multiple blood pressure variability types and the corresponding proportion of each blood pressure variability, wherein the blood pressure variability identification model is a deep learning model based on prior physiological knowledge. It should be noted that the blood pressure variability identification model refers to a pre-trained deep learning model (such as a convolutional neural network model, a recurrent neural network model, or a multi-head attention model), whose input is at least two physiological signals, and whose output is the type and proportion of blood pressure variability factors. This model incorporates prior physiological knowledge during training, enabling it to extract and analyze features from the input target physiological signals, and identify the types of blood pressure variability factors and the corresponding proportions of each factor.

[0093] It is understandable that the reasons for blood pressure changes vary from person to person under different conditions. This implementation method identifies blood pressure-causing factors based on the specific circumstances of each target user and selects a matching target blood pressure detection model. This allows for a more accurate consideration of various factors affecting blood pressure, avoiding the errors that may result from using a single general model, thereby improving the accuracy of blood pressure measurement.

[0094] Step A200: From multiple preset blood pressure detection models, determine the target blood pressure detection model that maps the multiple types of blood pressure mutagenic factors and the corresponding weight of each blood pressure mutagenic factor.

[0095] For example, this blood pressure variability identification model can compare the target physiological features corresponding to the target physiological signal with the pre-stored blood pressure baseline features in the model to determine the physiological features with differences and their corresponding deviation values. Then, based on the deviation values ​​of each physiological feature, it can determine the type of blood pressure variability factor causing the deviation and the weight of each blood pressure variability factor. Furthermore, the blood pressure calculation model mapped to the blood pressure variability factor with the largest weight can be determined as the target blood pressure detection model; alternatively, the target blood pressure detection model can be determined from multiple preset blood pressure detection models based on the weights of each blood pressure variability factor and the preset mapping relationship between the weight intervals of each blood pressure variability factor and the blood pressure calculation model. For example, when the weight of emotional tension is 80% and the weight of temperature is 20%, model A can be determined as the target blood pressure detection model, while when the weight of emotional tension is 60% and the weight of temperature is 40%, model B can be determined as the target blood pressure detection model.

[0096] In this embodiment, a deep learning model is used to dynamically identify blood pressure variability factors and their proportions based on the target user's physiological signals, thereby achieving personalized health monitoring. By combining prior physiological knowledge, the model can more accurately explain the relationship between physiological signals and blood pressure changes, improve the accuracy of judging blood pressure variability factors, and thus improve the accuracy of blood pressure measurement.

[0097] In one feasible implementation, the blood pressure measurement calibration method may further include steps A201-A203: Step A201: Obtain training data, wherein the training data includes sample physiological signals, sample environmental parameters, sample baseline features, sample baseline status and reference blood pressure values, and the training data is collected from users corresponding to the same type of blood pressure mutagenic factor. It should be noted that training data refers to the dataset used to train the blood pressure detection model, which includes input features and corresponding output labels. The input features include sample physiological signals, sample environmental parameters, sample baseline features, and sample baseline states, which correspond to the target physiological signals, detection environmental parameters, blood pressure baseline features, and blood pressure baseline states, respectively. The output labels include reference blood pressure values, which are obtained through high-precision equipment (such as gold standard blood pressure measurement devices).

[0098] For example, users affected by a specific type of blood pressure variability factor are selected as data collection subjects. The blood pressure baseline characteristics and blood pressure baseline status of the users in standard human condition are collected as sample baseline characteristics and sample baseline status. The physiological signals, detection environment parameters, and blood pressure values ​​measured by gold standard blood pressure measurement devices in other daily states of the users are collected as sample physiological signals, sample environment parameters, and reference blood pressure values ​​to obtain training data.

[0099] Step A202: Establish a blood pressure detection model corresponding to the type of blood pressure mutating factor, wherein the blood pressure detection model includes a deep learning model and a physiological model, and the physiological model is used to provide a theoretical basis for blood pressure detection for the deep learning model; It should be noted that physiological models refer to mathematical models built upon the working mechanisms of human physiological systems (such as the cardiovascular system) and the relationships between physiological parameters, providing a theoretical framework and computational methods for blood pressure calculation based on human physiological mechanisms. Deep learning models can employ neural network structures (such as fully connected networks, convolutional neural networks, recurrent neural networks, etc.) or other model architectures to automatically learn the complex, nonlinear mapping relationship between a user's physiological characteristics and blood pressure from large amounts of training data.

[0100] Step A203: Use the training data as input to the blood pressure detection model to train the blood pressure detection model and obtain the blood pressure detection model mapped by the type of blood pressure mutagenic factors.

[0101] For example, for the type of target mutagenic factor (such as temperature or mental state), the physiological pathways of blood pressure changes dominated by it (such as increased cardiac output and dynamic regulation of peripheral resistance) are analyzed. A suitable core physiological equation (such as a blood pressure calculation formula based on stroke volume, heart rate, and total peripheral resistance) is selected as the physiological model to provide the rationale for blood pressure calculation. Based on this, a neural network is designed, whose input layer receives input features from the training data mentioned above. Some nodes in the network are designed to correspond to key parameters in the physiological model (such as peripheral resistance). The output layer combines these intermediate physiological parameters and the physiological model to output the predicted blood pressure value. During model training, its loss function, in addition to minimizing the blood pressure prediction error, adds a physiological rationality constraint to penalize situations where the intermediate physiological parameters or final blood pressure value predicted by the network deviate too much from the theoretical prediction value of the physiological model.

[0102] In this embodiment, training data based on the same type of variable factor is acquired, enabling the model to focus on learning the most direct and stable causal relationship between physiological characteristics and blood pressure under specific variable factors. At the same time, by establishing a blood pressure detection model that integrates deep learning and physiological models, the model combines the powerful nonlinear fitting capability of deep learning with the interpretable theoretical framework and safety boundary provided by physiological models. Based on this, training is performed to obtain a dedicated blood pressure detection model that accurately maps to each type of blood pressure variable factor, which is beneficial for high-precision blood pressure detection.

[0103] For example, to help understand the implementation flow of the blood pressure measurement calibration method obtained by combining Embodiment 1 and Embodiment 2 above, please refer to... Figure 4 , Figure 4 A schematic diagram of the framework modules for a blood pressure measurement calibration method is provided, specifically: The wearable device for implementing the above-mentioned blood pressure measurement method includes a signal preprocessing unit, a human body state recognition unit, a blood pressure variability factor recognition unit, a blood pressure baseline monitoring unit, a blood pressure model base unit, and a model calibration unit. The human body state recognition unit collects environmental parameters of the wearable device during the current time period, such as IMU data, pressure data applied to the target user's body surface measurement location, and temperature data, to determine whether the current detection environment meets the preset blood pressure measurement conditions. If it does, the signal preprocessing unit further preprocesses the raw physiological signals of the target user during the current time period, such as raw electrocardiogram signals, raw heart sound signals, and raw pulse signals, removing high-frequency noise and baseline drift. Based on the preprocessed physiological signals, the target physiological signal is determined. This target physiological signal is then input to the blood pressure variability factor recognition unit, which uses a blood pressure variability factor recognition model to identify the blood pressure variability factors (including multiple types of blood pressure variability factors and their respective weights) that cause changes in the target user's blood pressure level. Furthermore, the human body state recognition unit detects that the target user is in a standard human body state. In this scenario, the blood pressure baseline monitoring unit acquires blood pressure baseline characteristics. Specifically, it can collect baseline physiological signals through sensors such as ECG sensors, heart sound sensors, and pulse sensors, and input them into the preprocessing unit for preprocessing before feature extraction to determine the blood pressure baseline characteristics. Based on the degree of difference between these blood pressure baseline characteristics and the target baseline characteristics uniformly stored in each preset blood pressure calculation model, the blood pressure baseline state is confirmed. Then, the aforementioned target physiological signals, blood pressure variability information, blood pressure baseline characteristics, blood pressure baseline state, and detection environment parameters are input into the blood pressure model base unit. The blood pressure model base unit then determines the target blood pressure calculation model mapped by the blood pressure variability information from multiple preset blood pressure calculation models, and inputs the target physiological signals, blood pressure baseline characteristics, blood pressure baseline state, and detection environment parameters into the target blood pressure calculation model. The target blood pressure calculation model then performs forward calculation to calculate the target user's blood pressure value to be calibrated.

[0104] During the blood pressure detection process described above, if the target user is wearing the wearable device for the first time, and the baseline physiological signal acquisition time exceeds a preset time threshold, the model calibration unit outputs a second calibration prompt to instruct the user to simultaneously acquire the gold standard blood pressure value and the blood pressure value to be calibrated, thereby calibrating the model parameters based on the blood pressure value to be calibrated. After the initial calibration, the model calibration unit continues to monitor the blood pressure baseline state, so that when the blood pressure baseline state changes, the first calibration prompt is output to remind the user to recalibrate the model parameters, thus enabling the blood pressure detection model to gradually better fit the user's actual physiological condition.

[0105] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the blood pressure measurement calibration method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0106] This application provides a wearable device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform the blood pressure measurement calibration method in Embodiment 1 above.

[0107] The following is for reference. Figure 5 This document illustrates a structural schematic diagram suitable for implementing wearable devices according to embodiments of this application. Wearable devices in these embodiments may include, but are not limited to, smart bracelets, smartwatches, and head-mounted displays. Specifically, head-mounted displays may include, but are not limited to, wearable devices such as smart headphones, mixed reality (MR) devices (e.g., MR glasses or MR helmets), augmented reality (AR) devices (e.g., AR glasses or AR helmets), virtual reality (VR) devices (e.g., VR glasses or VR helmets), extended reality (XR) devices, or some combination thereof. Figure 5 The wearable device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of this application.

[0108] like Figure 5As shown, the wearable device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 or a program loaded from a storage device 1003 into a random access memory 1004. The random access memory 1004 also stores various programs and data required for the operation of the wearable device. The processing unit 1001, the read-only memory 1002, and the random access memory 1004 are interconnected via a bus 1005. An input / output interface 1006 is also connected to the bus. Typically, the following systems can be connected to the input / output interface 1006: input devices 1007 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. Communication device 1009 allows the wearable device to communicate wirelessly or wiredly with other devices to exchange data. While wearable devices with various systems are shown in the figures, it should be understood that implementation or possession of all the systems shown is not required. More or fewer systems may be implemented alternatively.

[0109] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0110] The wearable device provided in this application, employing the blood pressure measurement calibration method described in the above embodiments, can solve the technical problem of how to simultaneously improve the convenience and accuracy of cuffless blood pressure measurement calibration during the calibration process. Compared with the prior art, the beneficial effects of the wearable device provided in this application are the same as those of the blood pressure measurement calibration method provided in the above embodiments, and other technical features of the wearable device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0111] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0112] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0113] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to perform the blood pressure measurement calibration method in the above embodiments.

[0114] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0115] The aforementioned computer-readable storage medium may be included in the wearable device; or it may exist independently and not assembled into the wearable device.

[0116] The aforementioned computer-readable storage medium carries one or more programs that, when executed by a wearable device, cause the wearable device to: acquire in real time detection environment parameters of the target user based on the wearable device's detection during the current time period, the detection environment parameters including IMU data reflecting the human body state; and, if the target user's human body state is identified as a standard human body state based on the detection environment parameters, acquire baseline physiological signals, determine blood pressure baseline characteristics based on the baseline physiological signals, and determine blood pressure baseline state based on the blood pressure baseline characteristics, wherein the baseline physiological signals include at least two of baseline electrocardiogram signals, baseline heart sound signals, and baseline pulse signals, the blood pressure baseline characteristics are reference values ​​corresponding to each physiological characteristic under the standard human body state, and the standard human body state is a sleep state or a resting state. The physiological characteristics are used to characterize blood pressure levels. The blood pressure baseline state is either a changed blood pressure baseline or no changed blood pressure baseline. Specifically, when the difference between the blood pressure baseline characteristics and the target baseline characteristics exceeds a preset threshold, the blood pressure baseline state is considered a changed blood pressure baseline. When the blood pressure baseline state is considered a changed blood pressure baseline, a first calibration prompt is output. This first calibration prompt instructs the target user to simultaneously collect a gold standard blood pressure value using a gold standard blood pressure measurement device and detect the blood pressure value to be calibrated using the wearable device. The gold standard blood pressure value collected by the gold standard blood pressure measurement device is obtained, and based on the gold standard blood pressure value, the model parameters of the target blood pressure detection model of the wearable device are optimized and adjusted to adjust the blood pressure value to be calibrated output by the target blood pressure detection model to the gold standard blood pressure value.

[0117] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0118] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0119] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0120] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described blood pressure measurement calibration method. This solves the technical problem of how to simultaneously improve the convenience and accuracy of cuffless blood pressure measurement calibration during the calibration process. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the blood pressure measurement calibration method provided in the above embodiments, and will not be repeated here.

[0121] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A blood pressure measurement calibration method, characterized in that, The blood pressure measurement calibration method includes: The system acquires in real time the detection environment parameters of the target user based on the wearable device at the current time period, including IMU data that reflects the human body state; When the target user's human state is identified as a standard human state based on the detection environment parameters, baseline physiological signals are collected, blood pressure baseline characteristics are determined based on the baseline physiological signals, and blood pressure baseline state is determined based on the blood pressure baseline characteristics. The baseline physiological signals include at least two of the following: baseline electrocardiogram signals, baseline heart sound signals, and baseline pulse signals. The blood pressure baseline characteristics are reference values ​​corresponding to each physiological characteristic under the standard human state, which is either a sleep state or a resting state. The physiological characteristics are used to characterize blood pressure levels. The blood pressure baseline state is either a change in blood pressure baseline or no change in blood pressure baseline. Specifically, if the difference between the blood pressure baseline characteristics and the target baseline characteristics exceeds a preset threshold, the blood pressure baseline state is considered a change in blood pressure baseline. When the blood pressure baseline status is changed, a first calibration prompt message is output. The first calibration prompt message is used to instruct the target user to simultaneously collect the gold standard blood pressure value through the gold standard blood pressure measurement device and detect the blood pressure value to be calibrated through the wearable device. The gold standard blood pressure value collected by the gold standard blood pressure measurement device is obtained, and the model parameters of the target blood pressure detection model of the wearable device are optimized and adjusted based on the gold standard blood pressure value, so as to adjust the blood pressure value to be calibrated output by the target blood pressure detection model to the gold standard blood pressure value.

2. The blood pressure measurement calibration method as described in claim 1, characterized in that, Before the step of acquiring the detection environment parameters of the target user based on the wearable device in the current time period in real time, the method further includes: When it is detected that the target user is wearing the wearable device for the first time, the detection environment parameters of the target user based on the wearable device are acquired for the first time; If the target user's human body state is identified as a standard human body state based on the initially acquired detection environment parameters, the baseline physiological signal is collected for the first time, and the collection time of the initial baseline physiological signal is accumulated. If the collection time exceeds a preset time threshold, the collection time is stopped from being accumulated and a second calibration prompt message is output. The second calibration prompt message is used to instruct the target user to simultaneously collect the gold standard blood pressure value for the first time through the gold standard blood pressure measurement device and to detect the blood pressure value to be calibrated for the first time through the wearable device. The gold standard blood pressure value collected for the first time by the gold standard blood pressure measurement device is obtained, and the model parameters of the target blood pressure detection model are adjusted for the first time based on the gold standard blood pressure value collected for the first time, so as to adjust the blood pressure value to be calibrated detected by the wearable device for the first time to the gold standard blood pressure value collected for the first time.

3. The blood pressure measurement calibration method as described in claim 2, characterized in that, Before the step of outputting the second calibration prompt information, the method further includes: A second calibration prompt message is generated using a preset model calibration unit; The target user's human body status is identified based on the detection environment parameters using a preset human body status recognition unit. If the human body is identified as being in a preset calibration requirement state, the step of outputting the second calibration prompt information is executed, wherein the calibration requirement state includes a resting state.

4. The blood pressure measurement calibration method as described in claim 1, characterized in that, The detection environment parameters also include pressure data applied by the wearable device to the target user's body surface measurement location, and temperature data of the target user at a preset body part. The blood pressure measurement calibration method further includes: The system uses a preset human body state recognition unit to determine whether the detection environment parameters meet the preset blood pressure measurement conditions. The human body state recognition unit includes an inertial sensor, a pressure sensor, and a temperature sensor. When the detection environment parameters meet the blood pressure measurement conditions, the step of optimizing and adjusting the model parameters of the target blood pressure detection model of the wearable device based on the gold standard blood pressure value is performed.

5. The blood pressure measurement calibration method as described in claim 1, characterized in that, After the step of outputting the first calibration prompt information, the method further includes: The system acquires in real time the target physiological signals detected by the wearable device during the current time period, wherein the target physiological signals include at least two of the target electrocardiogram signals, target heart sound signals, and target pulse signals. The target physiological signal, the detection environment parameters, the blood pressure baseline characteristics, and the blood pressure baseline status are input into the target blood pressure detection model to obtain the blood pressure value to be calibrated. The step of optimizing and adjusting the model parameters of the target blood pressure detection model of the wearable device based on the gold standard blood pressure value includes: The gold standard blood pressure value is used as the model output target and input into the target blood pressure detection model so that the target blood pressure detection model can adjust the blood pressure value to be calibrated to the gold standard blood pressure value before outputting it by adjusting the model parameters.

6. The blood pressure measurement calibration method as described in claim 5, characterized in that, The step of acquiring the target physiological signal detected by the wearable device in the current time period based on the target user in real time includes: The wearable device collects the target user's raw physiological signals in real time during the current period, wherein the raw physiological signals include at least two of the following: raw electrocardiogram signals, raw heart sound signals, and raw pulse signals. The original physiological signal is preprocessed by a preset signal preprocessing unit to obtain a preprocessed physiological signal. The preprocessing includes at least one of filtering, baseline correction, normalization and effective signal extraction. The target physiological signal is determined based on the preprocessed physiological signal.

7. The blood pressure measurement calibration method as described in claim 5, characterized in that, Before the step of inputting the target physiological signal, the detection environment parameters, the blood pressure baseline characteristics, and the blood pressure baseline state into the target blood pressure detection model to obtain the blood pressure value to be calibrated, the method further includes: The target physiological signal is input into a preset blood pressure variability identification model, which outputs multiple blood pressure variability types and the corresponding weight of each blood pressure variability. The blood pressure variability identification model is a deep learning model based on prior physiological knowledge. The target blood pressure detection model is determined from multiple preset blood pressure detection models by mapping the types of multiple blood pressure variability factors and the corresponding weight of each blood pressure variability factor.

8. The blood pressure measurement calibration method as described in claim 7, characterized in that, The blood pressure measurement calibration method also includes: Acquire training data, which includes sample physiological signals, sample environmental parameters, sample baseline features, sample baseline status, and reference blood pressure values. The training data is collected from users corresponding to the same type of blood pressure variability factor. Establish a blood pressure detection model corresponding to the aforementioned blood pressure mutating factor type, wherein the blood pressure detection model includes a deep learning model and a physiological model, and the physiological model is used to provide a theoretical basis for blood pressure detection for the deep learning model; The training data is used as input to the blood pressure detection model to train the blood pressure detection model and obtain a blood pressure detection model that maps the types of blood pressure mutagenic factors.

9. A wearable device, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the blood pressure measurement calibration method as described in any one of claims 1 to 8.

10. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the blood pressure measurement calibration method as described in any one of claims 1 to 8.

Images