Pulse signal monitoring system based on wearable pulse diagnosis bracelet

Abstract

The invention relates to a pulse signal monitoring system based on a wearable pulse diagnosis bracelet, which relates to the field of pulse signal monitoring, and comprises a sensing layer used for collecting motion information, temperature and humidity information and pulse monitoring signals; the index layer is used for collecting a pulse signal record value set and counting intervals in the pulse signal set, and the larger the disease number deviation is, the lower the disease similarity degree is; the anomaly analysis layer is used for adding one to an anomaly factor when the pulse monitoring signals do not belong to the pulse signal concentration interval, and subtracting one from the anomaly factor when the pulse monitoring signals belong to the pulse signal concentration interval in continuous preset duration, and the initial value of the anomaly factor is equal to 0; and the abnormity early warning layer is used for generating a pulse signal abnormity early warning and performing vibration prompt when the abnormal factor is greater than or equal to an abnormal factor threshold value. The technical problem that in the prior art, due to the fact that the threshold value of the pulse signal is fixedly set and cannot adapt to changes of interference factors, the false alarm rate is high is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of pulse signal monitoring, and in particular to a pulse signal monitoring system based on a wearable pulse diagnosis bracelet. Background Art

[0002] As a wearable device, the pulse diagnosis bracelet can monitor the user's pulse signal in real time and provide the user with health reference.

[0003] Current pulse signal monitoring technology usually uses an empirically determined fixed indicator threshold range to determine whether the pulse signal is abnormal. However, the pulse signal of the wearable user will change with the user's environment and activity type, causing the healthy threshold range to shift accordingly. The fixed threshold monitoring method cannot adapt to changes in interference factors, resulting in a high false alarm rate. Summary of the invention

[0004] The present invention aims to solve the technical problem that in the prior art, the threshold of the pulse signal is fixed and cannot adapt to the changes of interference factors, resulting in a high false alarm rate. A pulse signal monitoring system based on a wearable pulse diagnosis bracelet is provided to solve the problem.

[0005] The technical solution of the present invention to solve the above technical problems is as follows: The invention provides a pulse signal monitoring system based on a wearable pulse diagnosis bracelet, which is embedded in the wearable pulse diagnosis bracelet. The wearable pulse diagnosis bracelet includes an acceleration sensor, a temperature and humidity sensor and a pulse signal sensor, and includes: a perception layer, which is used to collect motion information through the acceleration sensor, collect temperature and humidity information through the temperature and humidity sensor, and collect pulse monitoring signals through the pulse signal sensor; an index layer, which is used to collect a set of pulse signal recording values ​​of healthy users with the same modality through networking with motion information, temperature and humidity information and pre-stored disease numbers as modal constraints, and to count pulse signal concentration intervals, wherein the larger the deviation of the disease number, the lower the degree of disease similarity; an abnormal analysis layer, which is used to increase an abnormal factor by one when a pulse monitoring signal does not belong to a pulse signal concentration interval, and to reduce the abnormal factor by one when a pulse monitoring signal belongs to a pulse signal concentration interval for a continuous preset time, and the initial value of the abnormal factor is equal to 0; an abnormal warning layer, which is used to generate a pulse signal abnormal warning and vibrate when the abnormal factor is greater than or equal to an abnormal factor threshold.

[0006] The beneficial effect of the present invention is: by deploying multiple types of sensors on the pulse diagnosis bracelet to monitor the user's movement information and temperature and humidity information in real time, and then combining the user's disease portrait, the pulse signal concentration interval of healthy users in the same state is counted, and then it is determined whether the pulse monitoring signal is abnormal and then an alarm is executed. The alarm stability of the pulse signal monitoring is guaranteed by the dynamic adjustment of the pulse signal concentration interval, and the technical effect of reducing the false alarm rate is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 A schematic diagram of the structure of a pulse signal monitoring system based on a wearable pulse diagnosis bracelet provided by the present invention; Figure 2 A schematic diagram of the motion information collection process of a pulse signal monitoring system based on a wearable pulse diagnosis bracelet provided by the present invention.

[0008] In the accompanying drawings, the components represented by the reference numerals are described as follows: perception layer 100, indexing layer 200, anomaly analysis layer 300, anomaly warning layer 400. DETAILED DESCRIPTION

[0009] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0010] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0011] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present invention.

[0012] like Figure 1 As shown, an embodiment of the present invention provides a pulse signal monitoring system based on a wearable pulse diagnosis bracelet, which is embedded in the wearable pulse diagnosis bracelet. The wearable pulse diagnosis bracelet includes an acceleration sensor, a temperature and humidity sensor, and a pulse signal sensor, including: The sensing layer 100 is used to collect motion information through an acceleration sensor, collect temperature and humidity information through a temperature and humidity sensor, and collect a pulse monitoring signal through a pulse signal sensor; Specifically, the perception layer 100 is the basic module of the entire pulse signal monitoring system. Its core function is to collect the user's physiological and environmental information through a variety of sensors. Specifically, the acceleration sensor is used to collect motion information, mainly reflecting the user's hand motion status, such as acceleration changes, motion trajectory, etc.; the temperature and humidity sensor is used to collect temperature and humidity information, reflecting the temperature and humidity conditions of the user's environment; the pulse signal sensor is used to collect pulse monitoring signals to reflect the user's heartbeat and pulse conditions. These sensors collect data in real time to provide basic information for subsequent analysis and judgment.

[0013] The index layer 200 is used to collect the pulse signal recording value set of healthy users of the same modality online with motion information, temperature and humidity information and pre-stored disease numbers as modal constraints, and to count the pulse signal concentration intervals, wherein the greater the deviation of the disease number, the lower the degree of disease similarity; Specifically, the index layer 200 is one of the core modules of the pulse signal monitoring system. Its main function is to perform modal constraints based on the user's motion information, temperature and humidity information and pre-stored disease numbers, thereby collecting a set of pulse signal record values ​​of healthy users with the same modality and counting the pulse signal concentration intervals.

[0014] In detail, modal constraint refers to screening out the pulse signal data of healthy users with similar motion state, environmental conditions and health status to the user's current motion state (motion speed, motion acceleration), environmental conditions (temperature and humidity) and health status (symptom type and symptom severity level) according to the user's motion state (motion speed, motion acceleration), environmental conditions (temperature and humidity) and health status; the pulse signal concentration interval refers to the concentrated distribution interval of the pulse signal values ​​of the aforementioned screened healthy users. Preferably, the concentrated distribution interval refers to the distribution interval of the eigenvalue set after deleting the outlier eigenvalues ​​of each attribute. Furthermore, the screening method of the outlier eigenvalues ​​can adopt quartile Method to determine, for example: suppose we have a set of pulse frequency data (unit: times / minute): 60, 62, 65, 68, 70, 72, 75, 78, 80, 82, 85, 88, 90, 92, 95, 98, 100, 102, 105, 110; data sorting: the data has been sorted; determine the quartiles: Q1 (first quartile): the 5th value (70); Q2 (second quartile): the 10th value (82); Q3 (third quartile): the 15th value (95); calculate the interquartile range (IQR): IQR = Q3 - Q1 = 95 - 70 = 25; Determine the outlier range: Lower bound = Q1 - 1.5 * IQR = 70 - 1.5 * 25 = 70 - 37.5 = 32.5; Upper bound = Q3 + 1.5* IQR = 95 + 1.5 * 25 = 95 + 37.5 = 132.5; According to the above calculation, any pulse signal value below 32.5 beats / minute or above 132.5 beats / minute is considered an outlier. In this example, all values ​​in the data set are between 32.5 and 132.5, so there are no outliers, and the pulse frequency concentration interval is [60,110]; in particular, the embodiment of the present application uses the doctor to pre-set the symptom number to represent the symptom type. The larger the deviation of the symptom number, the lower the similarity of the user's symptoms. The type difference data is quantified into the numerical difference of the number to realize the quantitative analysis of the modal constraint, which is beneficial to improving the calculation efficiency.

[0015] The abnormality analysis layer 300 is used to increase the abnormality factor by one when the pulse monitoring signal does not belong to the pulse signal concentration interval, and decrease the abnormality factor by one when the pulse monitoring signal belongs to the pulse signal concentration interval for a continuous preset time, and the initial value of the abnormality factor is equal to 0; Specifically, the abnormal analysis layer 300 is one of the key modules of the pulse signal monitoring system. Its main function is to dynamically adjust the value of the abnormal factor based on the comparison result between the pulse signal and the pulse signal concentration interval; the abnormal factor is an indicator used to evaluate the degree of abnormality of the pulse signal, and its initial value is 0; when the pulse monitoring signal is not in the pulse signal concentration interval, the abnormal factor will increase; when the pulse monitoring signal continues in the pulse signal concentration interval for a certain period of time, the abnormal factor will decrease.

[0016] The specific judgment conditions are: when the pulse monitoring signal does not belong to the pulse signal concentration interval, the abnormal factor is increased by one; when the pulse monitoring signal belongs to the pulse signal concentration interval for a preset period of time (for example, 10 minutes, 30 minutes, etc.), the abnormal factor is reduced by one; the initial value of the abnormal factor is 0, and it is dynamically updated through continuous addition and subtraction operations.

[0017] For example, assuming the preset duration is 10 minutes, when the user's pulse signal jumps out of the concentrated interval for the first time, the abnormal factor becomes 1. If the subsequent pulse signal continues to be abnormal, the abnormal factor will continue to increase. However, if the pulse signal remains in the concentrated interval within 10 minutes, the abnormal factor will decrease by 1. This dynamic adjustment mechanism can more accurately reflect the user's health status and avoid misjudgment due to short-term fluctuations. In this way, the abnormal analysis layer 300 can provide a more reliable basis for the abnormal warning layer and improve the accuracy and stability of the entire system.

[0018] The abnormal warning layer 400 is used to generate a pulse signal abnormal warning and vibrate to prompt when the abnormal factor is greater than or equal to the abnormal factor threshold.

[0019] Specifically, the abnormal warning layer 400 is an alarm module of the pulse signal monitoring system, and its core function is to determine whether a warning needs to be issued according to the value of the abnormal factor.

[0020] In detail, the abnormal factor threshold is a pre-set judgment standard. Preferably, the abnormal factor threshold is a judgment threshold that can be custom-set by the medical end, which is used to determine whether the pulse signal is in an abnormal state. When the value of the abnormal factor is greater than or equal to the threshold, the system considers that the pulse signal is abnormal and needs to remind the user. The warning method includes generating abnormal pulse signal warning information and prompting the user through vibration.

[0021] In a preferred embodiment, assuming that the abnormal factor threshold is set to 3, when the abnormal factor reaches 3 or higher through the dynamic adjustment of the abnormal analysis layer, the abnormal warning layer will immediately generate a warning message and remind the user through the vibration function of the bracelet. This design ensures that the user can perceive the abnormal situation in time and take corresponding measures.

[0022] Further, such as Figure 2 As shown, the wearable pulse diagnosis bracelet also includes a position sensor and a gyroscope, and collects motion information through an acceleration sensor, including: The position sensor is used to collect the motion position time series information, the acceleration sensor is used to collect the acceleration time series information, and the gyroscope is used to collect the angular velocity time series information; Determine the number of wrist position turns according to the angular velocity time series information, input the network integration quantity configuration function, and obtain the network integration quantity; Furthermore, the network integration quantity configuration function is: , in, Characterize the number of network integrations, Characterize the number of basic networks, ≥16, N represents the number of wrist position turns, Characterize the maximum number of channels, ≤256.

[0023] Select multiple motion intensity level assessment networks equal to the number of network integrations, perform mean calculation configuration on the outputs, analyze the motion position timing information, acceleration timing information, and angular velocity timing information, and obtain the motion intensity level, which is set as motion information, wherein the motion intensity level assessment network is obtained through multiple sets of data based on neural network training, and any set of the multiple sets of data includes: wrist motion position timing information, wrist acceleration timing information, wrist angular velocity timing information, and a label indicating the motion intensity level.

[0024] In a preferred embodiment, the present application uses the movement intensity level to characterize the movement state. The movement intensity level is the quantitative information of the intensity of the movement. Since the fluctuation of the pulse signal is affected by the movement mainly depends on the intensity of the movement, the embodiment of the present application directly uses the movement intensity level as one of the parameters of the modal constraint, which can ensure the effectiveness of the modal constraint of the movement dimension and avoid the interference of redundant data.

[0025] Furthermore, the exercise intensity level is affected by motion information of multiple dimensions, such as motion position timing information, acceleration timing information, and angular velocity timing information. In order to achieve accurate mapping of motion information of multiple dimensions to the exercise intensity level, the embodiment of the present application proposes to process the real-time monitored motion position timing information, acceleration timing information, and angular velocity timing information through multiple motion intensity level assessment networks to obtain multiple motion intensity level assessment results, and then calculate the average of the multiple motion intensity level assessment results to obtain the motion intensity level, which is set as the motion information.

[0026] Furthermore, the model parameters of multiple exercise intensity level assessment networks are the same. Any exercise intensity level assessment network is obtained through multiple sets of data based on neural network training. Any set of multiple sets of data includes: wrist movement position timing information, wrist acceleration timing information, wrist angular velocity timing information and labels for marking exercise intensity levels. The detailed training process is as follows: Construct the training loss function: ,in, Represents the loss value of the exercise intensity level assessment network after U consecutive trainings, Characterizes the preset loss function to calculate the number of training times, represents the output value of the intensity level of exercise in the i-th training, The label representing the marked exercise intensity level of the ith training, including multiple groups of data including wrist movement position timing information, wrist acceleration timing information, wrist angular velocity timing information and labels for marking exercise intensity levels, are divided into training data set, verification data set and test data set in an 8:1:1 ratio. In specific training, the training data set is first used, and the wrist movement position timing information, wrist acceleration timing information and wrist angular velocity timing information are used as input data for the long short-term memory neural network. When the loss function is less than or equal to the loss function threshold for consecutive preset times, the verification data set is retrieved to verify the long short-term memory neural network. When the loss function is less than or equal to the loss function threshold for consecutive preset times again, the test data set is retrieved to test the long short-term memory neural network. When the loss function is less than or equal to the loss function threshold for consecutive preset times again, it is deemed that the exercise intensity level assessment network has converged. Among them, the labels for marking the exercise intensity levels are all labeled by professionals.

[0027] Furthermore, as the motion data becomes more complex, the amount of calculation required by the motion intensity level assessment network becomes greater. If a single motion intensity level assessment network is always used to assess the motion intensity level, disturbances may occur. This can be solved by integrating multiple motion intensity level assessment networks. However, if the number of integrated motion intensity level assessment networks is large, the required computing power is large and the computing efficiency is low. In order to balance computing power and accuracy, this application proposes a solution for dynamically adjusting the number of integrated motion intensity level assessment networks according to the complexity of the motion data, as follows: The function for constructing the network integration quantity configuration is: The complexity of the movement can be characterized by the number of wrist turns. The greater the complexity of the movement, the greater the number of wrist turns. Preferably, the number of wrist turns can be determined by the angular velocity timing information, and then the number of wrist turns is input into the network integration number configuration function to determine the number of integrations of the required movement intensity level evaluation network. The specific judgment method of the number of wrist turns is as follows: Direction change: When an object turns, the direction of its angular velocity will change. For example, from clockwise rotation to counterclockwise rotation, the angular velocity value will change from positive to negative, or vice versa; Peak detection: During the turning process, the angular velocity may have a peak value. This is because when the object starts to turn, the angular velocity will increase rapidly, reach a peak value, and then gradually decrease until the turn ends; Zero crossing: When the object enters the turning state from a straight state, the angular velocity will increase from zero, and may return to zero again after a period of time. Therefore, the sum of the number of direction changes, peak values, and zero crossings can be counted to determine the number of turns.

[0028] Furthermore, the default =16, =256, the output accuracy data of the model with different turn number integration configurations are as follows: Sample No. Number of turns N Exercise intensity level Number of channels C Prediction accuracy (%) 1 0 1 16 85 2 5 2 32 88 3 10 3 54 90 4 15 4 72 92 5 20 5 89 94 6 25 6 104 95 7 30 7 117 96 8 35 8 129 97 9 40 9 140 98 10 45 10 150 99 It can be seen from the experimental results that as the number of turns increases, the number of channels gradually increases and the prediction accuracy also improves accordingly. This verifies that the formula for dynamically adjusting the number of model channels can improve the accuracy of motion intensity level prediction to a certain extent.

[0029] Furthermore, the motion information, temperature and humidity information and the pre-stored disease number are used as modal constraints, including: Construct a modality similarity evaluation function based on motion information, temperature and humidity information, and pre-stored disease numbers: , , in, Characterizes the modal similarity evaluation value, Characterize the modal similarity of the kth attribute, which are motion distance, temperature distance, humidity distance and disease distance. Characterizes the k-th attribute distance threshold, The k-th attribute constraint value is represented by motion information, temperature and humidity information, and pre-stored disease number. is the value of the kth attribute to be analyzed; When the modal similarity evaluation value is greater than or equal to the modal similarity evaluation threshold, the modal constraint is considered to be satisfied; otherwise, the modal constraint is considered not to be satisfied.

[0030] In a preferred embodiment, the pulse signal recording value collection of healthy users in the same state is realized by quantifying the modal constraints. The quantization method of the modal constraints is as follows: Construct a modality similarity evaluation function based on motion information, temperature and humidity information, and pre-stored disease numbers: , Since the aforementioned motion information, temperature and humidity information and pre-stored disease number are all quantifiable and comparable parameters, the modal similarity between the motion record information, temperature and humidity record information and pre-stored disease record number of healthy users recorded in the big data and the motion information, temperature and humidity information and pre-stored disease number of the bracelet wearing user can be analyzed through the modal similarity evaluation function. When the modal similarity evaluation value is greater than or equal to the modal similarity evaluation threshold, it is deemed to meet the modal constraint, otherwise it is deemed to not meet the modal constraint. For example, if the modal similarity threshold is 0.8, then when the modal similarity evaluation value is greater than or equal to 0.8, it is deemed to be the same modality, otherwise it is deemed to have modal differences.

[0031] The modality similarity evaluation function can be used to retrieve healthy users of the same modality, and record the pulse signal record value sets of these users in the corresponding modality, so as to facilitate the subsequent centralized value processing of the pulse signal.

[0032] Furthermore, the pulse signal record value set of the same modality healthy user is collected online, including: Collecting the first-level pulse signal record value set, first-level motion information, first-level temperature and humidity information, and first-level pre-stored disease number of the first-level healthy user that meets the modal constraints through networking; Using the first-level motion information, the first-level temperature and humidity information, and the first-level pre-stored disease number as modal constraints, the second-level pulse signal record value set of the second-level healthy user is collected online; The primary pulse signal record value set and the secondary pulse signal record value set are added to the pulse signal record value set.

[0033] Specifically, since the amount of sample data that can be referenced when performing data statistics is a prerequisite for ensuring the accuracy of analysis, in order to avoid the small number of samples of healthy users, the embodiment of the present application adopts a secondary index strategy to increase the amount of sample data, as detailed below: First, the pulse signal recording value set, motion recording information set, temperature and humidity recording information set and pre-stored disease recording number set of a set of healthy users with the same modality as the motion information, temperature and humidity information and pre-stored disease number of the user wearing the bracelet are collected online, and the pulse signal recording value set is stored as a first-level pulse signal recording value set, the motion recording information set is stored as first-level motion information, the temperature and humidity recording information set is stored as first-level temperature and humidity information, the pre-stored disease recording number set is stored as first-level pre-stored disease number, and the healthy user set is stored as first-level healthy users.

[0034] That is, any healthy user has a set of pulse signal record values, a first-level motion information, a first-level temperature and humidity information, and a first-level pre-stored disease number. Furthermore, for each first-level healthy user, a second-level pulse signal record value set of a second-level healthy user with the same modality as the first-level motion information, the first-level temperature and humidity information, and the first-level pre-stored disease number is collected online, and then the first-level pulse signal record value set and the second-level pulse signal record value set are added to the pulse signal record value set. Through the second-level adjacent sample collection, the adequacy of the sample data volume can be guaranteed, and at the same time, the correlation with the motion information, temperature and humidity information and pre-stored disease number of the user wearing the bracelet is ensured.

[0035] Furthermore, the pulse signal concentration interval is counted, including: Obtain the secondary exercise information, secondary temperature and humidity information, and secondary pre-stored disease number of the secondary health user; Based on the secondary motion information, the secondary temperature and humidity information, and the secondary pre-stored disease number, a first modal similarity set with the motion information, the temperature and humidity information, and the pre-stored disease number is calculated, and the first modal similarity set corresponds to the secondary pulse signal record value set one by one; Obtaining a second modality similarity set of the primary pulse signal record value set; Adding the first modality similarity set and the second modality similarity set to obtain a modality similarity sum value; Traversing the first modality similarity set and the second modality similarity set, comparing with the modality similarity sum value, to obtain a statistical weight set; The weighted mean of the first-level pulse signal recording value set and the second-level pulse signal recording value set is calculated according to the statistical weight set, and the pulse signal concentration interval is configured with the weighted mean as the center value according to the preset deviation threshold, and the preset deviation threshold is the preset value of the medical end.

[0036] In one implementation, the detailed process of counting the pulse signal concentration interval is as follows: Obtain the secondary motion information, secondary temperature and humidity information, and secondary pre-stored disease number of the secondary healthy user; based on the secondary motion information, secondary temperature and humidity information, and secondary pre-stored disease number, use the aforementioned modality similarity evaluation function: , , calculate the first modal similarity set with the motion information, temperature and humidity information and the pre-stored disease number, the first modal similarity set corresponds to the secondary pulse signal record value set one by one; then obtain the first pulse signal record value set and the second modal similarity set with the motion information, temperature and humidity information and the pre-stored disease number; add the first modal similarity set and the second modal similarity set to obtain the modal similarity sum value; traverse the first modal similarity set and the second modal similarity set, and compare them with the modal similarity sum value respectively, and store the ratio as a statistical weight set; calculate the weighted mean of the first pulse signal record value set and the second pulse signal record value set of each attribute according to the statistical weight set, and further, according to the deviation threshold of each attribute preset by the medical end, take the weighted mean as the center value, and configure the pulse signal concentration interval.

[0037] Furthermore, the pulse signal concentration interval is counted, including: Calculate the distances between the pulse signal recording value sets and obtain the pulse signal recording value deviation distance set; The pulse signal recording value deviation distance set is stripped of outliers to obtain the pulse signal concentration interval.

[0038] In another implementation, the pairwise distances of the pulse signal recording value set are calculated to obtain the pulse signal recording value deviation distance set; the pulse signal recording value deviation distance set is stripped of outliers to obtain the pulse signal concentration interval, and the deviation distance set is processed using outlier analysis technology to identify and exclude outliers. Common outlier analysis methods include statistical-based methods (such as Z-score), distance-based methods (such as LOF algorithm), and density-based methods.

[0039] A specific example of outlier removal method is LOF (local outlier factor) outlier analysis: Assume that we have a set of pulse signal recording values ​​(unit: times / minute) as follows: 70,72,75,78,80,82,85,88,90,92,95,98,100,102,105,110,150 On the surface, most of the data are between 70 and 110, while 150 is obviously high and may be an outlier.

[0040] Step 1: Calculate the local reachability density To apply LOF, we first need to calculate the local reachability density for each point. The local reachability density is the inverse of the average distance from the point to its kth nearest neighbor. Suppose we choose k=5, then for an intermediate point (such as 80), its 5 nearest neighbors may be between 70 and 85. For example, for point 90 (value 90): find the 5 nearest neighbors: 88, 92, 85, 95, 80; calculate the distances to these points, for example, the Euclidean distance is the absolute difference: 2 (88 to 90), 2 (90 to 92), 5 (85 to 90), 5 (90 to 95), 10 (80 to 90); the local reachability distance (for each neighbor j) is max(distance to j, kth distance of point j); assuming k=5, the kth distance of each point is the distance of its 5th nearest neighbor; the local reachability density is the inverse of the average of these local reachability distances; assuming the average local reachability distance of point 90 is 5, then the local reachability density is 1 / 5=0.2.

[0041] Step 2: Calculate LOF value The LOF value is a measure of the degree of outlier of a point relative to its neighbors. The calculation method is as follows: for point p, calculate the ratio of its local reachable density relative to its direct neighbors, and then take the average; for example, the neighbors of point 150 may only have a few points (such as 110 and 105), and the local reachable density of these neighbors may be high (because they are in the normal area), and the local reachable density of point 150 may be low (because there are not many points around), resulting in a high LOF value; suppose the LOF value of point 150 is 3, while the LOF values ​​of other normal points are close to 1. If the threshold is set to 1.5, then points with LOF values ​​exceeding 1.5 are considered outliers.

[0042] Therefore, 150 will be stripped off, and the remaining data forms a concentrated interval of the pulse signal: [70, 110]. In this way, the concentrated interval can better represent the normal pulse range of healthy users.

[0043] Furthermore, it also includes: The storage layer is used to upload abnormal signal types and abnormal time to the cloud when the pulse signal is abnormal and vibrates to prompt.

[0044] Specifically, the storage layer of the pulse signal monitoring system is mainly used for data storage and management. When the abnormal warning layer triggers an abnormal signal, the storage layer will store and upload the relevant information of the abnormal signal. The types of abnormal signals may include high heart rate, low heart rate, arrhythmia and other situations. These types reflect the specific manifestations of the user's pulse signal abnormality. The abnormal time refers to the specific time point when the abnormal signal occurs. Accurately recording the time is crucial for subsequent analysis of the cause of the abnormality and tracing the abnormal process. The storage layer uploads this information to the cloud to achieve remote storage and centralized management of data, which facilitates subsequent data analysis, system optimization and user queries.

[0045] Furthermore, it also includes: The collaborative early warning layer is used to send an abnormal pulse signal warning and vibration prompt to the collaborative mobile terminal for execution of the early warning if no user feedback information is received within a preset period of time. The collaborative mobile terminal is deployed among users who trust it.

[0046] Specifically, the working process of the collaborative early warning layer is as follows: when an abnormal pulse signal is detected, the system first prompts the user through vibration; if the user does not confirm receipt of the early warning information through specific actions (such as touching the bracelet or pressing a button) within a preset time period (such as 1 minute, 5 minutes, etc.), the collaborative early warning layer will send the early warning information to the collaborative mobile terminal; after receiving the information, the collaborative mobile terminal will remind the user's trusted group through notifications, text messages, phone calls, etc.; after receiving the notification, the trusted group can take appropriate measures according to the situation, such as contacting the user to understand the situation, going to the user's location to provide help, or notifying medical institutions.

[0047] The role of the collaborative warning layer is to provide multiple security guarantees for the pulse signal monitoring system. Even if the user cannot respond in time, the trusted group can promptly detect abnormal situations and take measures to ensure the health and safety of the user. For example, when the user has an abnormal pulse signal during sleep and fails to notice the vibration prompt of the bracelet, the collaborative warning layer can send the warning information to the guardian set by the user after a preset period of time. After receiving the information, the guardian can immediately go to the user's location to check the situation, or contact a medical institution for emergency rescue, thereby minimizing health risks.

[0048] Furthermore, it also includes: A database, used for storing an array with movement information, temperature and humidity information and pre-stored disease number as index tags and pulse signal concentration interval as index target after the pulse signal concentration interval is generated; When the modal constraint is updated, it is determined whether the database has a pulse signal concentration interval that satisfies the modal constraint; If it exists, the pulse signal concentration interval is directly output; if it does not exist, the pulse signal record value set of healthy users of the same modality is collected online, and the pulse signal concentration interval is counted.

[0049] Specifically, the database in the pulse signal monitoring system is a key component used to store and manage pulse signal concentration intervals and related information. Its main function is to store arrays with motion information, temperature and humidity information, and pre-stored disease numbers as index tags, and pulse signal concentration intervals as index targets. These index tags are like file directories, helping the system to quickly locate and retrieve pulse signal concentration intervals. For example, motion information may include the user's exercise intensity level, temperature and humidity information includes the temperature and humidity of the user's environment, and pre-stored disease numbers can identify the user's disease type (such as hypertension, diabetes, etc.). The pulse signal concentration interval is the normal range of the pulse signal of a healthy user calculated based on these index tags.

[0050] When the user uses the pulse signal monitoring system in a new environment, the system will search the database for a matching pulse signal concentration interval based on the current motion information, temperature and humidity information, and pre-stored disease numbers. If there is ready-made data in the database, the system will immediately use this data for abnormal analysis; if there is no relevant data in the database, the system will prompt the user to collect data online to update the information in the database.

[0051] The pulse signal monitoring system based on a wearable pulse diagnosis bracelet provided by the embodiment of the present invention has at least the following technical effects: By deploying multiple types of sensors on the pulse diagnosis bracelet to monitor the user's movement information and temperature and humidity information in real time, and then combining the user's disease portrait, the pulse signal concentration interval of healthy users in the same state is counted, and then it is determined whether the pulse monitoring signal is abnormal before executing the alarm. The dynamic adjustment of the pulse signal concentration interval ensures the alarm stability of the pulse signal monitoring, achieving the technical effect of reducing the false alarm rate.

[0052] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0053] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as systems, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0054] The present invention is described with reference to flowcharts and / or block diagrams of systems, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0055] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0056] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0057] Although preferred embodiments of the present invention have been described, additional changes and modifications may occur to these embodiments once those skilled in the art understand the basic inventive concepts.

[0058] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention belong to the scope of the present invention and its equivalent technologies, the present invention is also intended to include these changes and variations.

Claims

1. A pulse signal monitoring system based on a wearable pulse diagnosis bracelet, characterized in that: Embedded in a wearable pulse diagnosis bracelet, the wearable pulse diagnosis bracelet includes an acceleration sensor, a temperature and humidity sensor, and a pulse signal sensor, including: The perception layer is used to collect motion information through an acceleration sensor, collect temperature and humidity information through a temperature and humidity sensor, and collect pulse monitoring signals through a pulse signal sensor; The index layer is used to collect the pulse signal record value set of healthy users with the same modality through networking, using motion information, temperature and humidity information, and pre-stored disease numbers as modal constraints, and to count the pulse signal concentration intervals. The greater the deviation of the disease number, the lower the degree of disease similarity. The abnormality analysis layer is used to increase the abnormality factor by one when the pulse monitoring signal does not belong to the pulse signal concentration interval, and decrease the abnormality factor by one when the pulse monitoring signal belongs to the pulse signal concentration interval for a continuous preset time, and the initial value of the abnormality factor is equal to 0; The abnormal warning layer is used to generate a pulse signal abnormal warning and vibrate when the abnormal factor is greater than or equal to the abnormal factor threshold.

2. The system according to claim 1, characterized in that The wearable pulse diagnosis bracelet also includes a position sensor and a gyroscope, and collects motion information through an acceleration sensor, including: The position sensor is used to collect the motion position time series information, the acceleration sensor is used to collect the acceleration time series information, and the gyroscope is used to collect the angular velocity time series information; Determine the number of wrist position turns according to the angular velocity time series information, input the network integration quantity configuration function, and obtain the network integration quantity; Select multiple motion intensity level assessment networks equal to the number of network integrations, perform mean calculation configuration on the outputs, analyze the motion position timing information, acceleration timing information, and angular velocity timing information, and obtain the motion intensity level, which is set as motion information, wherein the motion intensity level assessment network is obtained through multiple sets of data based on neural network training, and any set of the multiple sets of data includes: wrist motion position timing information, wrist acceleration timing information, wrist angular velocity timing information, and a label indicating the motion intensity level.

3. The system according to claim 2, characterized in that The network integration quantity configuration function is: , in, Characterize the number of network integrations, Characterize the number of basic networks, ≥16, N represents the number of wrist position turns, Characterize the maximum number of channels, ≤256.

4. The system according to claim 1, characterized in that The modal constraints include motion information, temperature and humidity information, and pre-stored disease numbers, including: Construct a modality similarity evaluation function based on motion information, temperature and humidity information, and pre-stored disease numbers: ， ， in, Characterizes the modal similarity evaluation value, Characterize the k-th attribute modal similarity, which are motion distance, temperature distance, humidity distance and disease distance. Characterizes the k-th attribute distance threshold, The k-th attribute constraint value is represented by motion information, temperature and humidity information, and pre-stored disease number. is the value of the kth attribute to be analyzed; When the modal similarity evaluation value is greater than or equal to the modal similarity evaluation threshold, the modal constraint is considered to be satisfied; otherwise, the modal constraint is considered not to be satisfied.

5. The system according to claim 4, characterized in that The pulse signal record value set of the same modality healthy user is collected online, including: Collecting the first-level pulse signal record value set, first-level motion information, first-level temperature and humidity information, and first-level pre-stored disease number of the first-level healthy user that meets the modal constraints through networking; Using the first-level motion information, the first-level temperature and humidity information, and the first-level pre-stored disease number as modal constraints, the second-level pulse signal record value set of the second-level healthy user is collected online; The primary pulse signal record value set and the secondary pulse signal record value set are added to the pulse signal record value set.

6. The system according to claim 5, characterized in that Statistical pulse signal concentration interval, including: Obtain the secondary movement information, secondary temperature and humidity information, and secondary pre-stored disease number of the secondary health user; Based on the secondary motion information, the secondary temperature and humidity information, and the secondary pre-stored disease number, a first modal similarity set with the motion information, the temperature and humidity information, and the pre-stored disease number is calculated, and the first modal similarity set corresponds to the secondary pulse signal record value set one by one; Obtaining a second modality similarity set of the primary pulse signal record value set; Adding the first modality similarity set and the second modality similarity set to obtain a modality similarity sum value; Traversing the first modality similarity set and the second modality similarity set, comparing with the modality similarity sum value, to obtain a statistical weight set; The weighted mean of the first-level pulse signal recording value set and the second-level pulse signal recording value set is calculated according to the statistical weight set, and the pulse signal concentration interval is configured with the weighted mean as the center value according to the preset deviation threshold, and the preset deviation threshold is the preset value of the medical end.

7. The system according to claim 1, characterized in that Statistical pulse signal concentration interval, including: Calculate the distances between the pulse signal recording value sets and obtain the pulse signal recording value deviation distance set; The pulse signal recording value deviation distance set is stripped of outliers to obtain the pulse signal concentration interval.

8. The system of claim 1, wherein: Also includes: The storage layer is used to upload abnormal signal types and abnormal time to the cloud when the pulse signal is abnormal and vibrates to prompt.

9. The system according to claim 1, characterized in that Also includes: The collaborative early warning layer is used to send an abnormal pulse signal warning and vibration prompt to the collaborative mobile terminal for execution of the early warning if no user feedback information is received within a preset period of time. The collaborative mobile terminal is deployed among users who trust it.

10. The system according to claim 1, wherein: Also includes: A database, used for storing an array with movement information, temperature and humidity information and pre-stored disease number as index tags and pulse signal concentration interval as index target after the pulse signal concentration interval is generated; When the modal constraint is updated, it is determined whether the database has a pulse signal concentration interval that satisfies the modal constraint; If it exists, the pulse signal concentration interval is directly output; if it does not exist, the pulse signal record value set of healthy users of the same modality is collected online, and the pulse signal concentration interval is counted.

Images