Intelligent watch device and system for blood pressure detection and analysis based on large model

Abstract

The invention discloses a smart watch device and system for blood pressure detection and analysis based on a large model, and the device comprises a smart watch terminal which comprises a wrist strap air bag pressure control module and a pulse wave monitoring module. The output end of the wrist strap air bag pressure control module and the output end of the pulse wave monitoring module are both connected with the blood pressure management module, the blood pressure management module is arranged in the intelligent watch terminal, the blood pressure management module is further connected with the cloud large model server, and the blood pressure detection and analysis system based on the large model comprises the blood pressure analysis module. The blood pressure analysis module is arranged in the cloud large model server, the blood pressure analysis module is in interactive connection with the blood pressure management module through wireless signals, the accuracy of user blood pressure calculation is greatly improved through mechanical iterative calculation of the blood pressure analysis large model, the blood pressure analysis large model is combined with historical measurement data of a user, and the user experience is improved. The blood pressure change trend is calculated, and the blood pressure prediction accuracy is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of blood pressure detection, and particularly relates to an intelligent watch device and system for blood pressure detection and analysis based on a large model. Background Art

[0002] When an intelligent watch monitors the blood pressure data of a user, there are mainly two technical solutions: the oscillometric method and the photoelectric detection method. Among them, the oscillometric method has a higher accuracy than the photoelectric monitoring method. Some manufacturers can achieve medical-grade accuracy, but there is still a certain measurement error compared with professional medical sphygmomanometers. Currently, the development trend in the industry is to evolve based on the oscillometric method, and the intelligent watch can also transmit the detected blood pressure.

[0003] For example, a cuffless continuous blood pressure estimation system based on online learning disclosed in the invention with the authorization announcement number CN115486823B belongs to the fields of data processing and deep learning. This solution can effectively solve the problem that the blood pressure estimation model obtained by training under a large amount of data has poor universality, so as to realize continuous real-time monitoring of individual blood pressure, reflect the real continuous blood pressure changes of patients, effectively avoid the randomness of clinic blood pressure measurement, exclude white coat hypertension, and detect blood pressure changes in a timely manner. It can comprehensively analyze the changes of in-vivo physiological activities or biochemical indicators that increase or decrease with blood pressure fluctuations.

[0004] The above solution has the following problems: it can detect blood pressure changes in a timely manner, but it cannot correct abnormal data measured in various life scenarios and individual differences, which greatly reduces the accuracy of blood pressure measurement and subsequent prediction, and is not convenient for judging the development direction of blood pressure. Therefore, we propose an intelligent watch device and system for blood pressure detection and analysis based on a large model. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent watch device and system for blood pressure detection and analysis based on a large model, so as to solve the problem that abnormal data measured in various life scenarios and individual differences cannot be corrected, which greatly reduces the accuracy of blood pressure measurement and subsequent prediction, and is not convenient for judging the development direction of blood pressure proposed in the above background art.

[0006] To achieve the above purpose, the present invention provides the following technical solution: an intelligent watch device for blood pressure detection and analysis based on a large model, including an intelligent watch terminal. The intelligent watch terminal includes a wristband airbag pressure control module and a pulse wave monitoring module. The output ends of the wristband airbag pressure control module and the pulse wave monitoring module are both connected to a blood pressure management module. The blood pressure management module is arranged in the intelligent watch terminal, and the blood pressure management module is also connected to a cloud large model server.

[0007] Preferably, the blood pressure management module is also connected to a display screen, and the display is arranged on one side surface of the smartwatch terminal.

[0008] A blood pressure detection and analysis system based on a large model includes a blood pressure analysis module. The blood pressure analysis module is arranged in a cloud large model server. The blood pressure analysis module is connected to a blood pressure management module through a wireless signal. The blood pressure management module is connected to a wristband airbag pressure control module and a pulse wave monitoring module interactively.

[0009] Preferably, the blood pressure management module is used to transmit the collected pulse wave data and airbag pressure data to the blood pressure model module.

[0010] Preferably, the blood pressure analysis module is used to establish and train a blood pressure analysis large model, and then calculate the current blood pressure value through the blood pressure analysis large model and predict the blood pressure change trend.

[0011] Preferably, the establishment and training of the blood pressure analysis large model are specifically as follows:

[0012] Data collection: Collect the original blood pressure measurement data of people of different ages and genders through professional medical blood pressure measurement instruments, including the original pulse wave measurement data and the original airbag pressure test data.

[0013] Feature extraction: Extract time domain features, waveform features, and morphological features from the original pulse wave data, and extract the pressure change curve, peak value, and change rate from the original airbag pressure data as features. Use the pulse wave features and airbag pressure features as the training data of the large model.

[0014] Model training: Adopt the iterative training method to perform multiple trainings according to the extracted feature data; in each iteration process, use the feature data as the input, and through the internal algorithm and neural network structure, calculate the predicted blood pressure value and blood pressure change trend.

[0015] Preferably, the calculation method of the blood pressure analysis large model is specifically as follows:

[0016] The blood pressure analysis large model extracts time domain features, waveform features, and morphological features from the original pulse wave data, and extracts the pressure change curve, peak value, and change rate from the original airbag pressure data as feature data. Use the pulse wave features and airbag pressure features as the analysis input data of the large model. The blood pressure analysis large model performs reasoning analysis according to the input feature data, calculates and outputs the optimal blood pressure measurement value, and combines the user's historical measurement data to calculate the blood pressure change trend.

[0017] Preferably, the wristband airbag pressure control module is used to collect the user's airbag pressure change curve, airbag pressure peak value, and airbag pressure change rate.

[0018] Preferably, the pulse wave monitoring module is used to collect the time domain features, frequency domain features, and morphological features in the user's pulse wave.

[0019] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0020] (1) By collecting the user's pulse wave, the airbag pressure change curve, peak value, and change rate, and then performing mechanical iterative calculations through the blood pressure analysis large model, the accuracy of calculating the user's blood pressure is greatly improved.

[0021] (2) This application can also extract time domain features, waveform features, and morphological features from the original data of the pulse wave, and extract the pressure change curve, peak value, and change rate from the original data of the airbag pressure as feature data. The pulse wave features and airbag pressure features are used as the analysis input data of the large model. The blood pressure analysis large model performs reasoning analysis based on the input feature data, calculates and outputs the optimal blood pressure measurement value, and combines the user's historical measurement data to calculate the blood pressure change trend to obtain a more accurate prediction and analysis result. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic diagram of the system structure of the first embodiment in the present invention;

[0023] Figure 2 It is a schematic diagram of the system structure of the second embodiment in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0025] Embodiment 1

[0026] Please refer to Figure 1 , the present invention provides a technical solution: an intelligent watch device for blood pressure detection and analysis based on a large model, including an intelligent watch terminal. The intelligent watch terminal includes a wristband airbag pressure control module and a pulse wave monitoring module. The output ends of the wristband airbag pressure control module and the pulse wave monitoring module are both connected to the blood pressure management module. The blood pressure management module is arranged in the intelligent watch terminal, and the blood pressure management module is also connected to the cloud large model server. A wearing wristband is also arranged on the intelligent watch terminal to facilitate wearing on the user's wrist.

[0027] Further, the blood pressure management module is also connected to a display screen, which is set on one side surface of the smartwatch terminal.

[0028] A blood pressure detection and analysis system based on a large model includes a blood pressure analysis module, which is set in the cloud large model server. The blood pressure analysis module is interconnected with the blood pressure management module through wireless signals. The blood pressure management module is interconnected with the wristband airbag pressure control module and the pulse wave monitoring module.

[0029] Further, the blood pressure management module is used to transmit the collected pulse wave data and airbag pressure data to the blood pressure model module.

[0030] Further, the blood pressure analysis module is used to establish and train a large blood pressure analysis model, and then calculate the current blood pressure value through the large blood pressure analysis model and predict the blood pressure change trend.

[0031] Further, the establishment and training of the large blood pressure analysis model are specifically as follows:

[0032] Data collection: Collect the original blood pressure measurement data of people of different ages and genders through professional medical blood pressure measurement instruments, including the original pulse wave measurement data and the original airbag pressure test data.

[0033] Feature extraction: Extract time domain features, waveform features, and morphological features from the original pulse wave data, and extract the pressure change curve, peak value, change rate, etc. from the original airbag pressure data as features. The pulse wave features and airbag pressure features are used as the training data of the large model.

[0034] Model training: Adopt the iterative training method and conduct multiple trainings according to the extracted feature data. In each iteration process, the feature data is used as the input, and through the internal algorithm and neural network structure, the predicted blood pressure value and blood pressure change trend are calculated. To evaluate the accuracy of the model, the predicted blood pressure value and trend calculated in each iteration are compared with the actually measured blood pressure value. By calculating the error between the two, the prediction performance of the model is measured. If the error is large, it means that there is a deviation between the prediction result of the model and the actual situation, and the model needs to be corrected. According to the result of the error comparison, the calculation weight of the large model is adjusted. By optimizing the weight, the model can pay more attention to the features that have a greater impact on blood pressure prediction, improving the accuracy and stability of the model.

[0035] Further, the calculation method of the large blood pressure analysis model is specifically as follows:

[0036] The large blood pressure analysis model extracts time-domain features, waveform features, and morphological features from the original data of the pulse wave, and extracts the pressure change curve, peak value, and change rate from the original data of the airbag pressure as feature data. The pulse wave features and airbag pressure features are used as the analysis input data of the large model. The large blood pressure analysis model performs reasoning analysis based on the input feature data, calculates and outputs the optimal blood pressure measurement value, and combines the user's historical measurement data to calculate the blood pressure change trend.

[0037] Furthermore, the wristband airbag pressure control module is used to collect the user's airbag pressure change curve, airbag pressure peak value, and airbag pressure change rate.

[0038] Furthermore, the pulse wave monitoring module is used to collect the time-domain features, frequency-domain features, and morphological features in the user's pulse wave.

[0039] The time-domain features include the pulse wave period, wave peak height, wave valley height, rise time, fall time, and pulse wave area.

[0040] Frequency-domain features: Extract spectral features (such as main frequency and harmonic components) through Fourier transform.

[0041] Morphological features: The number of peaks, inflection point positions, waveform symmetry, etc. of the pulse wave.

[0042] Airbag pressure change curve: The curve of airbag pressure changing with time.

[0043] Airbag pressure peak value: The maximum value of the airbag pressure.

[0044] Airbag pressure change rate: The rising and falling rates of the airbag pressure.

[0045] The specific detection method is as follows:

[0046] Step 1: The wristband airbag pressure control module controls the pressure change in the airbag, and transmits the airbag pressure change curve with time, the maximum value of the airbag pressure, and the rising and falling rates of the airbag pressure to the blood pressure management module through its pressure sensor;

[0047] Step 2: The pulse wave monitoring module collects the time-domain features, frequency-domain features, and morphological features in the pulse wave, and transmits the time-domain features, frequency-domain features, and morphological features to the blood pressure management module;

[0048] Step 3: The blood pressure management module transmits the airbag pressure change curve with time, the maximum value of the airbag pressure, and the rising, falling rates of the airbag pressure, time-domain features, frequency-domain features, and morphological features to the cloud large model server through wireless network;

[0049] Step 4: The blood pressure analysis large model in the cloud large model server performs inference analysis based on the input feature data, calculates and outputs the optimal blood pressure measurement value;

[0050] Step 5: The blood pressure analysis large model adopts an iterative training method and conducts multiple trainings based on the extracted feature data. In each iteration process, the feature data is used as input, and through the internal algorithms and neural network structure, the predicted blood pressure value and the blood pressure change trend are calculated. To evaluate the accuracy of the model, the predicted blood pressure value and trend obtained from each iteration calculation are compared with the actually measured blood pressure value. By calculating the error between the two, the prediction performance of the model is measured. If the error is large, it indicates that there is a deviation between the prediction result of the model and the actual situation, and the model needs to be corrected. According to the result of the error comparison, the calculation weights of the large model are adjusted. By optimizing the weights, the model can pay more attention to the features that have a greater impact on blood pressure prediction, improving the accuracy and stability of the model;

[0051] Step 6: The cloud large model server transmits the current blood pressure value calculated by the blood pressure analysis large model and the predicted blood pressure change trend data to the blood pressure management module;

[0052] Step 7: The blood pressure management module sends the current blood pressure value and the predicted blood pressure change trend to the display screen, and the display screen shows the current blood pressure value and the predicted blood pressure change trend.

[0053] The smartwatch terminal in the present invention has a communication function and can be directly wirelessly connected to the cloud large model server.

[0054] Embodiment 2

[0055] Compared with Embodiment 1, the smartwatch terminal in this embodiment does not have a remote communication function.

[0056] An intelligent watch device for blood pressure detection and analysis based on a large model, including a smartwatch terminal. The smartwatch terminal includes a wristband airbag pressure control module and a pulse wave monitoring module. The output ends of the wristband airbag pressure control module and the pulse wave monitoring module are both connected to the blood pressure management module. The blood pressure management module is arranged in the smartwatch terminal, and the blood pressure management module is also interactively connected to the blood pressure management APP of the user's mobile phone terminal. A wearing wristband is also arranged on the smartwatch terminal for convenient wearing on the user's wrist.

[0057] A blood pressure detection and analysis system based on a large model, including a blood pressure analysis module. The blood pressure analysis module is set in the cloud large model server. The blood pressure analysis module is interactively connected to the blood pressure management APP of the user's mobile phone terminal through Bluetooth. The blood pressure management APP of the user's mobile phone terminal is connected to the cloud large model server through a wireless network. The blood pressure management module is interactively connected to the wristband airbag pressure control module and the pulse wave monitoring module.

[0058] The specific detection method is as follows:

[0059] Step 1: The wristband airbag pressure control module controls the pressure change in the airbag and transmits the curve of the airbag pressure changing with time, the maximum value of the airbag pressure, and the rising and falling rates of the airbag pressure to the blood pressure management module through its pressure sensor;

[0060] Step 2: The pulse wave monitoring module collects the time-domain features, frequency-domain features, and morphological features in the pulse wave and transports the time-domain features, frequency-domain features, and morphological features to the blood pressure management module;

[0061] Step 3: The blood pressure management module transmits the curve of the airbag pressure changing with time, the maximum value of the airbag pressure, and the rising, falling rates, time-domain features, frequency-domain features, and morphological features of the airbag pressure to the blood pressure management APP of the user's mobile phone terminal through a wireless network;

[0062] Step 4: The blood pressure management APP of the user's mobile phone terminal transports the curve of the airbag pressure changing with time, the maximum value of the airbag pressure, and the rising, falling rates, time-domain features, frequency-domain features, and morphological features of the airbag pressure to the cloud large model server through a wireless network;

[0063] Step 5: Through the blood pressure analysis large model in the cloud large model server, based on the input feature data, perform inference analysis, calculate and output the optimal blood pressure measurement value;

[0064] Step 6: The blood pressure analysis large model adopts an iterative training method and conducts multiple trainings based on the extracted feature data. In each iteration process, the feature data is used as input, and through the internal algorithm and neural network structure, the predicted blood pressure value and blood pressure change trend are calculated. To evaluate the accuracy of the model, the predicted blood pressure value and trend calculated in each iteration are compared with the actually measured blood pressure value. By calculating the error between the two, the prediction performance of the model is measured. If the error is large, it indicates that there is a deviation between the prediction result of the model and the actual situation, and the model needs to be corrected. According to the result of the error comparison, the calculation weight of the large model is adjusted. By optimizing the weight, the model can pay more attention to the features that have a greater impact on blood pressure prediction, improving the accuracy and stability of the model;

[0065] Step 7: The cloud large model server transmits the current blood pressure value calculated by the blood pressure analysis large model and the predicted blood pressure change trend data to the blood pressure management APP on the user's mobile phone terminal;

[0066] Step 8: The blood pressure management APP on the user's mobile phone terminal transmits the current blood pressure value and the predicted blood pressure change trend to the blood pressure management module through a Bluetooth signal;

[0067] Step 9: The blood pressure management module transmits the current blood pressure value and the predicted blood pressure change trend to the display screen, and the display screen displays the current blood pressure value and the predicted blood pressure change trend.

[0068] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A smart watch device for blood pressure detection and analysis based on a large model, characterized in that: It includes a smart watch terminal, which includes a wristband airbag pressure control module and a pulse wave monitoring module. The output ends of the wristband airbag pressure control module and the pulse wave monitoring module are both connected to a blood pressure management module. The blood pressure management module is arranged in the smart watch terminal, and the blood pressure management module is also connected to a cloud-based large model server.

2. According to claim 1, a smart watch device for blood pressure detection and analysis based on a large model is characterized in that: The blood pressure management module is also connected to a display screen, and the display is arranged on a surface of one side of the smart watch terminal.

3. A blood pressure detection and analysis system based on a large model according to claim 1 or 2, characterized in that: It includes a blood pressure analysis module, which is arranged in a cloud-based large model server. The blood pressure analysis module is interactively connected to a blood pressure management module via a wireless signal. The blood pressure management module is interactively connected to a wristband airbag pressure control module and a pulse wave monitoring module.

4. A blood pressure detection and analysis system based on a large model according to claim 3, characterized in that: The blood pressure management module is used to transmit the collected pulse wave data and air bag pressure data to the blood pressure model module.

5. The blood pressure detection and analysis system based on a large model according to claim 4 is characterized in that: The blood pressure analysis module is used to establish and train a large blood pressure analysis model, and then calculate the current blood pressure value through the large blood pressure analysis model and predict the blood pressure change trend.

6. A blood pressure detection and analysis system based on a large model according to claim 5, characterized in that: The blood pressure analysis model is established and trained specifically as follows: Data collection: Through professional medical blood pressure measuring equipment, the original blood pressure measurement data of people of different ages and genders are collected, including pulse wave measurement original data and air bag pressure test original data. Feature extraction: extract time domain features, waveform features, and morphological features from the raw data of the pulse wave, extract the pressure change curve, peak value, and change rate as features from the raw data of the airbag pressure, and use the pulse wave features and airbag pressure features as training data for the large model. Model training: An iterative training method is used to perform multiple trainings based on the extracted feature data. During each iteration, the feature data is used as input, and the predicted blood pressure value and blood pressure change trend are calculated through the internal algorithm and neural network structure.

7. The blood pressure detection and analysis system based on a large model according to claim 5, characterized in that: The calculation method of the blood pressure analysis model is specifically as follows: The blood pressure analysis model extracts time domain features, waveform features, and morphological features from the original data of the pulse wave, and extracts the pressure change curve, peak value, and change rate as feature data from the original data of the balloon pressure. The pulse wave features and balloon pressure features are used as the analysis input data of the large model. The blood pressure analysis model performs inference analysis based on the input feature data, calculates and outputs the optimal blood pressure measurement value, and calculates the blood pressure change trend in combination with the user's historical measurement data.

8. The blood pressure detection and analysis system based on a large model according to claim 3 is characterized by: The wristband airbag pressure control module is used to collect the user's airbag pressure change curve, airbag pressure peak value and airbag pressure change rate.

9. The smart watch device and system for blood pressure detection and analysis based on a large model according to claim 3, characterized in that: The pulse wave monitoring module is used to collect time domain features, frequency domain features and morphological features in the user's pulse wave.

Images