A blood lipid monitoring method based on wearable device

By synchronously obtaining electrocardiogram and pulse wave signals on wearable devices, and using deep learning models for blood lipid monitoring, the problems of inconvenient, expensive and inability to monitor blood lipids in the prior art are solved, and portable, real-time and low-cost blood lipid detection.

CN115363545BActive Publication Date: 2024-11-29BEIJING XUEYANG TECH CO LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202211067033.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-09-01
Publication Date
2024-11-29
Estimated Expiration
2042-09-01

AI Technical Summary

Technical Problem

The existing blood lipid monitoring methods require blood extraction, the equipment is expensive and cannot be monitored in real time, and is not portable.

Method used

Using a wearable device-based blood lipid monitoring method, the electrocardiogram and pulse wave signals are obtained synchronously, their multi-scale characteristics are determined, and a pre-trained deep learning model is input to output blood lipid information. The equipment is simple, low-cost, and can be used in real time. Test.

Benefits of technology

The portability, real-time and low cost of blood lipid monitoring are achieved, and the operability and accuracy of detection are improved, and the equipment is expensive and inconvenient in traditional methods is avoided.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115363545B_ABST
    Figure CN115363545B_ABST
Patent Text Reader

Abstract

The present invention discloses a blood lipid monitoring method based on a wearable device, comprising: detecting the state of the wearer based on the wearable device, and when it is determined that the wearer is in a stationary state, synchronously acquiring the wearer's electrocardiogram signal and pulse wave signal in each preset time period; determining the first multi-scale feature of the electrocardiogram signal and the second multi-scale feature of the pulse wave signal; inputting the first multi-scale feature and the second multi-scale feature into a pre-trained blood lipid deep learning model, and outputting the wearer's blood lipid information. Blood lipid monitoring through wearable devices is simple, highly operable, low-cost, easy to carry, and can be detected in real time.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of blood lipid monitoring, and in particular to a blood lipid monitoring method based on a wearable device. Background Art

[0002] At present, the existing blood lipid monitoring method is to draw venous blood and put the blood into a biochemical detector for testing, which has the following technical problems: 1) It is inconvenient to draw blood. 2) It requires a biochemical detector, which is expensive. 3) It cannot monitor in real time. Summary of the invention

[0003] The present invention aims to solve at least one of the technical problems in the above-mentioned technology to a certain extent. To this end, the purpose of the present invention is to propose a blood lipid monitoring method based on a wearable device, which is simple, highly operable, low-cost, easy to carry, and can detect in real time.

[0004] To achieve the above object, an embodiment of the present invention proposes a blood lipid monitoring method based on a wearable device, comprising:

[0005] Based on the wearable device detecting the state of the wearer, when it is determined that the wearer is in a stationary state, synchronously acquiring the wearer's electrocardiogram signal and pulse wave signal in each preset time period;

[0006] Determine a first multi-scale feature of the electrocardiogram signal and a second multi-scale feature of the pulse wave signal;

[0007] The first multi-scale features and the second multi-scale features are input into a pre-trained blood lipid deep learning model to output the wearer's blood lipid information.

[0008] According to some embodiments of the present invention, before determining the first multi-scale feature of the electrocardiogram signal and the second multi-scale feature of the pulse wave signal, it also includes: filtering the electrocardiogram signal and the pulse wave signal respectively.

[0009] According to some embodiments of the present invention, the preset time period is 8 minutes.

[0010] According to some embodiments of the present invention, determining a first multi-scale feature of an electrocardiogram signal and a second multi-scale feature of a pulse wave signal includes:

[0011] Performing time domain analysis on the electrocardiogram signal to determine a first frequency domain feature as a first multi-scale feature; the first multi-scale feature includes: heartbeat interval, heart rate variability, and signal power spectrum density change;

[0012] The pulse wave signal is subjected to time domain analysis to determine a second frequency domain feature as a second multi-scale feature; the second multi-scale feature includes: changes in blood volume, changes in blood color, changes in radial artery pressure waves, changes in signal power spectral density, and dicrotic wave amplitude.

[0013] According to some embodiments of the present invention, a method for obtaining a pre-trained blood lipid deep learning model includes:

[0014] Acquire multiple sets of data from the user, each set of data including a sample electrocardiogram signal and a sample pulse wave signal;

[0015] Analyze each set of data to determine the sample blood lipid information corresponding to each set of data;

[0016] Input a set of data to train the constructed neural network deep learning model and output the actual blood lipid information;

[0017] Compare the actual blood lipid information with the sample blood lipid information, determine the adjustment parameters of the neural network deep learning model according to the difference parameters, and make adjustments;

[0018] The next set of data is input into the adjusted neural network deep learning model for continued training until it is determined that the difference parameter is less than the preset difference parameter, thereby obtaining a pre-trained blood lipid deep learning model.

[0019] According to some embodiments of the present invention, the method further includes: the wearable device transmitting the blood lipid information to a server for storage.

[0020] According to some embodiments of the present invention, before the wearable device transmits the blood lipid information to the server, the method further includes:

[0021] The wearable device and the server mutually verify each other's reliability based on an asymmetric encryption algorithm.

[0022] According to some embodiments of the present invention, the wearable device and the server mutually verify each other's reliability based on an asymmetric encryption algorithm, including:

[0023] Acquire a first public key and a first private key of the wearable device; acquire a second public key and a second private key of the server; the server holds the first public key, and the wearable device holds the second public key;

[0024] The wearable device generates a verification request;

[0025] The wearable device calculates a hash value of the verification request to obtain a first hash value;

[0026] The wearable device encrypts the first hash value according to the first private key to obtain a digital signature;

[0027] The wearable device sends the digital signature and the verification request to the server;

[0028] The server calculates a hash value of the verification request to obtain a second hash value;

[0029] The server decrypts the digital signature according to the first public key to obtain a decrypted hash value;

[0030] The server compares the second hash value with the decrypted hash value;

[0031] When the server determines that the second hash value is the same as the decrypted hash value, the wearable device passes the reliability verification; otherwise, the wearable device fails the reliability verification; similarly, the reliability of the server is verified.

[0032] According to some embodiments of the present invention, detecting a state of a wearer based on a wearable device includes:

[0033] Determine the angular velocity and acceleration of the wearable device based on an inertial measurement module included in the wearable device;

[0034] Determine a first antisymmetric matrix of the angular velocity and a second antisymmetric matrix of the acceleration;

[0035] Determine when the wearer is in a stationary state based on the stationary angular velocity and stationary acceleration measured by the inertial measurement module;

[0036] Determine a third antisymmetric matrix of the static angular velocity and a fourth antisymmetric matrix of the static acceleration;

[0037] Performing a subtraction operation on the first antisymmetric matrix and the third antisymmetric matrix to obtain a fifth antisymmetric matrix, calculating three eigenvalues ​​of the fifth antisymmetric matrix, selecting the largest first eigenvalue, and determining whether it is less than 0;

[0038] Performing a subtraction operation on the second antisymmetric matrix and the fourth antisymmetric matrix to obtain a sixth antisymmetric matrix, calculating three eigenvalues ​​of the sixth antisymmetric matrix, selecting the largest second eigenvalue, and determining whether it is less than 0;

[0039] When it is determined that the first eigenvalue is less than 0 and the second eigenvalue is less than 0, it indicates that the wearer is in a stationary state; otherwise, the wearer is in a moving state.

[0040] According to some embodiments of the present invention, before the wearable device transmits the blood lipid information to the server, the method further includes:

[0041] The probability that the wearable device maintains connection with the server within time t is calculated, and it is determined whether it is greater than a preset probability. When it is determined that the probability is greater than the preset probability, a transmission instruction is generated.

[0042] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0043] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0045] Figure 1 is a flow chart of a blood lipid monitoring method based on a wearable device according to an embodiment of the present invention;

[0046] Figure 2 is a flow chart of a blood lipid monitoring method based on a wearable device according to another embodiment of the present invention;

[0047] Figure 3 is a flowchart of detecting a wearer's state based on a wearable device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0049] like Figure 1 As shown, the embodiment of the present invention proposes a blood lipid monitoring method based on a wearable device, including steps S1-S3:

[0050] S1. Detecting the state of the wearer based on the wearable device, when it is determined that the wearer is in a stationary state, synchronously acquiring the wearer's electrocardiogram signal and pulse wave signal in each preset time period;

[0051] S2, determining a first multi-scale feature of the electrocardiogram signal and a second multi-scale feature of the pulse wave signal;

[0052] S3. Input the first multi-scale feature and the second multi-scale feature into a pre-trained blood lipid deep learning model, and output the wearer's blood lipid information.

[0053] The working principle of the above technical solution: Synchronous acquisition refers to the simultaneous collection of the user's wrist electrocardiogram (ECG) and pulse wave signal (PPG) in the same period of time; the pulse wave signal refers to the red light and infrared light signal. First, the wearer's state is determined based on the wearable device. When the wearer is in a stationary state, the wearer's ECG signal and pulse wave signal are collected to facilitate the accurate detection of blood lipids. Based on the first multi-scale feature of the ECG signal and the second multi-scale feature of the pulse wave signal, a pre-trained blood lipid deep learning model is input to output the wearer's blood lipid information.

[0054] The beneficial effects of the above technical solution are: blood lipid monitoring is carried out through wearable devices, the equipment is simple, easy to operate, low cost, easy to carry, and can be detected in real time.

[0055] According to some embodiments of the present invention, before determining the first multi-scale feature of the electrocardiogram signal and the second multi-scale feature of the pulse wave signal, it also includes: filtering the electrocardiogram signal and the pulse wave signal respectively.

[0056] The beneficial effect of the above technical solution is to remove noise and redundant components in the signal. The physiological signal of the human body is basically between 1-5 Hz. This filtering can not only remove conventional noise such as power frequency interference and baseline drift, but also add a method to remove motion artifacts. Therefore, various noises and motion interference can be effectively eliminated.

[0057] According to some embodiments of the present invention, the preset time period is 8 minutes.

[0058] According to some embodiments of the present invention, determining a first multi-scale feature of an electrocardiogram signal and a second multi-scale feature of a pulse wave signal includes:

[0059] Performing time domain analysis on the electrocardiogram signal to determine a first frequency domain feature as a first multi-scale feature; the first multi-scale feature includes: heartbeat interval, heart rate variability, and signal power spectrum density change;

[0060] The pulse wave signal is subjected to time domain analysis to determine a second frequency domain feature as a second multi-scale feature; the second multi-scale feature includes: changes in blood volume, changes in blood color, changes in radial artery pressure waves, changes in signal power spectral density, and dicrotic wave amplitude.

[0061] The beneficial effect of the above technical solution is that it is easy to accurately determine the first multi-scale feature of the electrocardiogram signal and the second multi-scale feature of the pulse wave signal.

[0062] According to some embodiments of the present invention, a method for obtaining a pre-trained blood lipid deep learning model includes:

[0063] Acquire multiple sets of data from the user, each set of data including a sample electrocardiogram signal and a sample pulse wave signal;

[0064] Analyze each set of data to determine the sample blood lipid information corresponding to each set of data;

[0065] Input a set of data to train the constructed neural network deep learning model and output the actual blood lipid information;

[0066] Compare the actual blood lipid information with the sample blood lipid information, determine the adjustment parameters of the neural network deep learning model according to the difference parameters, and make adjustments;

[0067] The next set of data is input into the adjusted neural network deep learning model for continued training until it is determined that the difference parameter is less than the preset difference parameter, thereby obtaining a pre-trained blood lipid deep learning model.

[0068] The working principle and beneficial effects of the above technical solution are as follows: the neural network deep learning model is trained based on sample data. During the training process, the actual blood lipid information is compared with the sample blood lipid information, and the adjustment parameters of the neural network deep learning model are determined according to the difference parameters, and then adjusted; the next set of data is input into the adjusted neural network deep learning model for continued training, until it is determined that the difference parameter is less than the preset difference parameter, and a pre-trained blood lipid deep learning model is obtained. The accuracy of determining the blood lipid deep learning model is improved.

[0069] like Figure 2 As shown, according to some embodiments of the present invention, it also includes: the wearable device transmits the blood lipid information to a server for storage.

[0070] The beneficial effects of the above technical solution are: it facilitates the server to display and store blood lipid information, improves the traceability of data, and avoids the wearable device from storing too much data, resulting in slow response speed.

[0071] According to some embodiments of the present invention, before the wearable device transmits the blood lipid information to the server, the method further includes:

[0072] The wearable device and the server mutually verify each other's reliability based on an asymmetric encryption algorithm.

[0073] The beneficial effects of the above technical solution are: it is easy to ensure the reliability of the wearable device and the server, and improves the security and reliability of data transmission.

[0074] According to some embodiments of the present invention, the wearable device and the server mutually verify each other's reliability based on an asymmetric encryption algorithm, including:

[0075] Acquire a first public key and a first private key of the wearable device; acquire a second public key and a second private key of the server; the server holds the first public key, and the wearable device holds the second public key;

[0076] The wearable device generates a verification request;

[0077] The wearable device calculates a hash value of the verification request to obtain a first hash value;

[0078] The wearable device encrypts the first hash value according to the first private key to obtain a digital signature;

[0079] The wearable device sends the digital signature and the verification request to the server;

[0080] The server calculates a hash value of the verification request to obtain a second hash value;

[0081] The server decrypts the digital signature according to the first public key to obtain a decrypted hash value;

[0082] The server compares the second hash value with the decrypted hash value;

[0083] When the server determines that the second hash value is the same as the decrypted hash value, the wearable device passes the reliability verification; otherwise, the wearable device fails the reliability verification; similarly, the reliability of the server is verified.

[0084] The working principle of the above technical solution is as follows: obtaining the first public key and the first private key of the wearable device; obtaining the second public key and the second private key of the server; the server holds the first public key, and the wearable device holds the second public key; the wearable device generates a verification request; the wearable device calculates the hash value of the verification request to obtain a first hash value; the wearable device encrypts the first hash value according to the first private key to obtain a digital signature; the wearable device sends the digital signature and the verification request to the server; the server calculates the hash value of the verification request to obtain a second hash value; the server decrypts the digital signature according to the first public key to obtain a decrypted hash value; the server compares the second hash value with the decrypted hash value; when the server determines that the second hash value is the same as the decrypted hash value, the wearable device passes the reliability verification; otherwise, the wearable device fails the reliability verification; similarly, verify the reliability of the server.

[0085] The beneficial effects of the above technical solution are: respectively verifying the reliability of the wearable device and the reliability of the server, improving the verification accuracy and verification rate, facilitating the judgment of whether the wearable device and the server are reliably connected, and improving the accuracy of data transmission.

[0086] like Figure 3 As shown, according to some embodiments of the present invention, detecting the state of the wearer based on the wearable device includes steps S11-S17:

[0087] S11. Determine the angular velocity and acceleration of the wearable device based on an inertial measurement module included in the wearable device;

[0088] S12, determining a first antisymmetric matrix of the angular velocity and a second antisymmetric matrix of the acceleration;

[0089] S13, when determining that the wearer is in a stationary state, the stationary angular velocity and stationary acceleration measured by the inertial measurement module;

[0090] S14, determining a third antisymmetric matrix of the static angular velocity and a fourth antisymmetric matrix of the static acceleration;

[0091] S15, performing a subtraction operation on the first antisymmetric matrix and the third antisymmetric matrix to obtain a fifth antisymmetric matrix, calculating three eigenvalues ​​of the fifth antisymmetric matrix, selecting the largest first eigenvalue, and determining whether it is less than 0;

[0092] S16, performing a subtraction operation on the second antisymmetric matrix and the fourth antisymmetric matrix to obtain a sixth antisymmetric matrix, calculating three eigenvalues ​​of the sixth antisymmetric matrix, selecting the largest second eigenvalue, and determining whether it is less than 0;

[0093] S17. When it is determined that the first eigenvalue is less than 0 and the second eigenvalue is less than 0, it indicates that the wearer is in a stationary state; otherwise, the wearer is in a moving state.

[0094] The working principle of the above technical solution: the angular velocity is a three-axis angular velocity; the acceleration is a three-axis acceleration. Based on the inertial measurement module included in the wearable device, the angular velocity and acceleration of the wearable device are determined; the first antisymmetric matrix of the angular velocity and the second antisymmetric matrix of the acceleration are determined; when the wearer is in a stationary state, the stationary angular velocity and the stationary acceleration measured by the inertial measurement module are determined; the stationary angular velocity is a three-axis stationary angular velocity; the stationary acceleration is a three-axis stationary acceleration. Determine the third antisymmetric matrix of the stationary angular velocity and the fourth antisymmetric matrix of the stationary acceleration; perform subtraction operations on the first antisymmetric matrix and the third antisymmetric matrix to obtain the fifth antisymmetric matrix, calculate the three eigenvalues ​​of the fifth antisymmetric matrix, and select the largest first eigenvalue to determine whether it is less than 0; because of the three-axis angular velocity and the three-axis stationary angular velocity, the fifth antisymmetric matrix has three eigenvalues. Determining the eigenvalues ​​of a matrix is ​​a conventional technical means, which will not be repeated here. A subtraction operation is performed on the second antisymmetric matrix and the fourth antisymmetric matrix to obtain a sixth antisymmetric matrix, three eigenvalues ​​of the sixth antisymmetric matrix are calculated, and the largest second eigenvalue is selected to determine whether it is less than 0; when it is determined that the first eigenvalue is less than 0 and the second eigenvalue is less than 0, it indicates that the wearer is in a stationary state; otherwise, the wearer is in a moving state.

[0095] The beneficial effect of the above technical solution is: instead of directly comparing and judging the angular velocity and acceleration with the corresponding static angular velocity and static acceleration, the maximum eigenvalue is obtained by the difference of their corresponding antisymmetric matrices and compared with 0. This can filter out the influence of the zero bias and offset of the acceleration and angular velocity measured by the inertial measurement module, and obtain more accurate data. The threshold is set to 0, and there is no need to adjust the threshold each time. It has good versatility and is convenient for accurately judging the wearer's status.

[0096] According to some embodiments of the present invention, before the wearable device transmits the blood lipid information to the server, the method further includes:

[0097] The probability that the wearable device maintains connection with the server within time t is calculated, and it is determined whether it is greater than a preset probability. When it is determined that the probability is greater than the preset probability, a transmission instruction is generated.

[0098] Calculating the probability that the wearable device maintains connection with the server within time t, including:

[0099] a n =2t(δ 2 n +μ 2 n ) / λ n

[0100]

[0101] Where P is the probability that the wearable device maintains connection with the server within time t; a n is the parameter of the Rayleigh distribution that the motion vector of the wearable device approximately obeys; δ 2 n is the variance, μ 2 n is the mean, V is the maximum moving speed of the wearable device; λ n is the moving position of the wearable device; e is a natural constant; R is the communication radius of the wearable device; D is the distance between the wearable device and the server; θ is the movement direction of the wearable device, which obeys the uniform distribution of (0,2π).

[0102] The working principle and beneficial effects of the above technical solution: Before the wearable device transmits the blood lipid information to the server, it also includes: calculating the probability that the wearable device and the server remain connected within time t, and judging whether it is greater than a preset probability, and generating a transmission instruction when it is determined that the probability is greater than the preset probability. When the probability is less than the preset probability, it indicates that the connection between the wearable device and the server is unstable, and the wearable device is stopped from transmitting the blood lipid information to the server. When the probability is greater than the preset probability and the transmission instruction is generated, it indicates that the connection between the wearable device and the server is stable, and stable data transmission can be achieved, thereby improving the accuracy of data transmission. The complex movement of the wearable device is approximated as the superposition of a series of uniform linear motions. When the maximum movement distance of the wearable device is less than the communication range of the server, the motion vector of the wearable device approximately obeys a n Based on the above formula, it is easy to accurately determine the probability that the wearable device and the server remain connected within time t, which improves the accuracy of the judgment probability and the preset probability size, and then it can accurately determine whether to perform data transmission between the wearable device and the server.

[0103] 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 fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A blood lipid monitoring method based on wearable devices, It is characterized in that include: Based on the wearable device detecting the state of the wearer, when it is determined that the wearer is in a stationary state, synchronously acquiring the wearer's electrocardiogram signal and pulse wave signal in each preset time period; Determine a first multi-scale feature of the electrocardiogram signal and a second multi-scale feature of the pulse wave signal; Inputting the first multi-scale feature and the second multi-scale feature into a pre-trained blood lipid deep learning model, and outputting the wearer's blood lipid information; It also includes: the wearable device transmitting the blood lipid information to a server for storage; Before the wearable device transmits the blood lipid information to the server, the method further includes: The wearable device and the server mutually verify each other's reliability based on an asymmetric encryption algorithm; The wearable device and the server mutually verify each other's reliability based on an asymmetric encryption algorithm, including: Acquire a first public key and a first private key of the wearable device; acquire a second public key and a second private key of the server; the server holds the first public key, and the wearable device holds the second public key; The wearable device generates a verification request; The wearable device calculates a hash value of the verification request to obtain a first hash value; The wearable device encrypts the first hash value according to the first private key to obtain a digital signature; The wearable device sends the digital signature and the verification request to the server; The server calculates a hash value of the verification request to obtain a second hash value; The server decrypts the digital signature according to the first public key to obtain a decrypted hash value; The server compares the second hash value with the decrypted hash value; When the server determines that the second hash value is the same as the decrypted hash value, the wearable device passes the reliability verification; otherwise, the wearable device fails the reliability verification; similarly, the reliability of the server is verified; Calculating the probability that the wearable device maintains connection with the server within time t, and determining whether the probability is greater than a preset probability, and generating a transmission instruction when it is determined that the probability is greater than the preset probability; Calculating the probability that the wearable device maintains connection with the server within time t, including: a n =2t(δ 2 n +m 2 n ) / min n Where P is the probability that the wearable device maintains connection with the server within time t; a n is the parameter of the Rayleigh distribution that the motion vector of the wearable device approximately obeys; δ 2 n is the variance, μ 2 n is the mean, V is the maximum moving speed of the wearable device; λ n is the moving position of the wearable device; e is a natural constant; R is the communication radius of the wearable device; D is the distance between the wearable device and the server; θ is the movement direction of the wearable device, which obeys the uniform distribution of (0,2π).

2. The blood lipid monitoring method based on a wearable device as claimed in claim 1, It is characterized in that Before determining the first multi-scale feature of the electrocardiogram signal and the second multi-scale feature of the pulse wave signal, the method further includes: filtering the electrocardiogram signal and the pulse wave signal respectively.

3. The blood lipid monitoring method based on a wearable device as claimed in claim 1, It is characterized in that The preset time period is 8 minutes.

4. The blood lipid monitoring method based on a wearable device as claimed in claim 1, It is characterized in that Determining a first multi-scale feature of an electrocardiogram signal and a second multi-scale feature of a pulse wave signal includes: Performing time domain analysis on the electrocardiogram signal to determine a first frequency domain feature as a first multi-scale feature; the first multi-scale feature includes: heartbeat interval, heart rate variability, and signal power spectrum density change; The pulse wave signal is subjected to time domain analysis to determine a second frequency domain feature as a second multi-scale feature; the second multi-scale feature includes: changes in blood volume, changes in blood color, changes in radial artery pressure waves, changes in signal power spectral density, and dicrotic wave amplitude.

5. The blood lipid monitoring method based on a wearable device as claimed in claim 1, It is characterized in that Methods for obtaining a pre-trained blood lipid deep learning model include: Acquire multiple sets of data from the user, each set of data including a sample electrocardiogram signal and a sample pulse wave signal; Analyze each set of data to determine the sample blood lipid information corresponding to each set of data; Input a set of data to train the constructed neural network deep learning model and output the actual blood lipid information; Compare the actual blood lipid information with the sample blood lipid information, determine the adjustment parameters of the neural network deep learning model according to the difference parameters, and make adjustments; The next set of data is input into the adjusted neural network deep learning model for continued training until it is determined that the difference parameter is less than the preset difference parameter, thereby obtaining a pre-trained blood lipid deep learning model.

6. The blood lipid monitoring method based on a wearable device as claimed in claim 1, It is characterized in that Detect the wearer's status based on wearable devices, including: Determine the angular velocity and acceleration of the wearable device based on an inertial measurement module included in the wearable device; Determine a first antisymmetric matrix of the angular velocity and a second antisymmetric matrix of the acceleration; Determine when the wearer is in a stationary state based on the stationary angular velocity and stationary acceleration measured by the inertial measurement module; Determine a third antisymmetric matrix of the static angular velocity and a fourth antisymmetric matrix of the static acceleration; Performing a subtraction operation on the first antisymmetric matrix and the third antisymmetric matrix to obtain a fifth antisymmetric matrix, calculating three eigenvalues ​​of the fifth antisymmetric matrix, selecting the largest first eigenvalue, and determining whether it is less than 0; Performing a subtraction operation on the second antisymmetric matrix and the fourth antisymmetric matrix to obtain a sixth antisymmetric matrix, calculating three eigenvalues ​​of the sixth antisymmetric matrix, selecting the largest second eigenvalue, and determining whether it is less than 0; When it is determined that the first eigenvalue is less than 0 and the second eigenvalue is less than 0, it indicates that the wearer is in a stationary state; otherwise, the wearer is in a moving state.

Citation Information

Patent Citations

  • Low-delay node Internet of Things applied to intelligent building and communication method thereof

    CN110191173A

  • Method and device for judging static state of robot and robot

    CN110916577A

  • Wearable device supporting PPG + ECG function

    CN113208573A

  • Identity verification method and device, equipment and storage medium

    CN113347143A