A dynamic heart rate monitoring method and system based on smart watch

Abstract

The present invention relates to the field of dynamic heart rate monitoring technology, and specifically discloses a dynamic heart rate monitoring method and system based on a smart watch. The user's heart rate data is collected in real time through the smart watch, and a bandpass filter is used for noise filtering. The signal processing is optimized in combination with a dynamic threshold adjustment method to ensure that the data is accurate and reliable. The pre-processed heart rate signal is time-series aligned and segmented for analysis, and time domain and frequency domain features are extracted, including the heart rate fluctuation amplitude and the instantaneous heart rate change rate. According to the analysis results, the user's heart rate abnormality is evaluated, and the abnormalities are divided into normal, mild abnormalities, and severe abnormalities. For severe abnormalities, the system issues an early warning in a timely manner; for mild abnormalities, a gradient boosting decision tree model is used to predict the heart rate change trend and potential abnormalities in the future monitoring period, and a personalized health report containing a heart rate fluctuation trend graph and abnormality analysis is generated to provide health guidance for the user.

Description

Technical Field

[0001] The present invention relates to the technical field of dynamic heart rate monitoring, and in particular to a dynamic heart rate monitoring method based on a smart watch. Background Art

[0002] With socioeconomic development and improved living standards, people are increasingly concerned about health management, especially the demand for real-time monitoring of physiological parameters such as heart rate. Smart wearable devices, such as smartwatches, have become important tools for personal health management due to their portability and diverse functionality. Heart rate monitoring, as a key indicator of health monitoring, can be used to assess the functional status of the cardiovascular system, detect physical load during exercise, and predict potential health risks. However, traditional heart rate monitoring methods, which mainly rely on medical equipment, are complex to operate and have poor real-time performance, making them unsuitable for long-term monitoring in daily life. Therefore, using smart wearable devices for high-precision, real-time dynamic heart rate monitoring has gradually become a hot topic in research and application.

[0003] The existing technology has the following deficiencies:

[0004] Although existing smartwatches can monitor heart rate in real time, their data processing methods have certain limitations, resulting in monitoring accuracy and reliability that cannot meet the higher demands of health management. Most existing devices only provide real-time heart rate values, lacking in-depth analysis of the dynamic characteristics of heart rate changes, and are unable to accurately assess heart rate abnormalities or provide forward-looking health predictions. Especially in the case of minor abnormalities, users often have difficulty receiving targeted health guidance, thus missing the optimal time for intervention. Therefore, how to improve the accuracy and robustness of heart rate monitoring and combine dynamic analysis and prediction models to achieve intelligent and personalized health management is an urgent problem that existing technologies need to solve. Summary of the Invention

[0005] The purpose of the present invention is to provide a dynamic heart rate monitoring method and system based on a smart watch to solve the problems in the above background.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A dynamic heart rate monitoring method based on a smart watch includes the following steps:

[0008] S1: Collects the user's heart rate data in real time through a smartwatch, and performs noise filtering and time alignment on the collected signal to ensure data accuracy and reliability;

[0009] S2: Perform segmented analysis on the processed heart rate signal to extract dynamic features in the time domain and frequency domain, including the heart rate fluctuation amplitude and instantaneous heart rate change rate;

[0010] S3: Comprehensively analyzing the heart rate fluctuation amplitude and instantaneous heart rate change rate of the heart rate signal, and assessing the abnormality of the current user's heart rate based on the analysis results;

[0011] S4: Classifying the abnormality of the user's heart rate into normal, slightly abnormal, and severely abnormal according to the judgment result;

[0012] S5: Perform early warning processing based on serious anomalies. For minor anomalies, establish a user heart rate prediction model based on the user's heart rate fluctuation amplitude and instantaneous heart rate change rate to predict the user's heart rate abnormality level in the future;

[0013] S6: Based on the prediction results, the smartwatch generates a prediction report of future heart rate changes and provides feedback to the user.

[0014] As a further solution of the present invention: the segmented analysis of the processed heart rate signal specifically includes:

[0015] The processed heart rate signal is segmented, and the length of each segment is determined by a fixed time window;

[0016] For each time period, the heart rate fluctuation amplitude is obtained by calculating the difference between the maximum heart rate value and the minimum heart rate value in the signal within each time period;

[0017] In each time period, the instantaneous heart rate change rate was obtained by calculating the heart rate difference between adjacent time points and dividing it by the time interval.

[0018] As a further solution of the present invention, the comprehensive analysis of the heart rate fluctuation amplitude and instantaneous heart rate change rate of the heart rate signal specifically includes:

[0019] During a monitoring cycle, the heart rate fluctuation amplitude coefficient is calculated according to the degree of heart rate fluctuation in each time period, and the mean of the instantaneous heart rate change rate in each time period is calculated to obtain the instantaneous heart rate change rate in each time period. According to the degree of fluctuation of the instantaneous heart rate change rate in each time period, the instantaneous heart rate change rate abnormality coefficient during the monitoring cycle is calculated. The heart rate fluctuation amplitude coefficient and the instantaneous heart rate change rate abnormality coefficient are normalized and calculated to obtain the user's heart rate abnormality index during the current monitoring cycle.

[0020] As a further solution of the present invention: the process of obtaining the heart rate fluctuation amplitude coefficient is:

[0021] Applying wavelet transform to the denoised heart rate signal, the wavelet transform decomposes the signal into multiple frequency bands, each of which contains signal components within a different frequency range. Through wavelet transform, the signal coefficients of the multiple frequency bands are obtained;

[0022] The fluctuation amplitude is obtained by calculating the difference between the maximum and minimum values ​​of the signal of each frequency band in each time period. The heart rate fluctuation amplitude coefficient in each time period is obtained by taking the weighted average of the fluctuation amplitudes of all frequency bands. The heart rate fluctuation amplitude coefficient in the entire monitoring period is averaged to calculate the overall heart rate fluctuation amplitude coefficient.

[0023] As a further solution of the present invention: the process of obtaining the abnormal coefficient of the instantaneous heart rate change rate is:

[0024] Calculate the rate of change of the heart rate signal in each time period, and calculate the instantaneous rate of change of the heart rate signal by the differential formula. The calculation expression is: ;

[0025] Where, Indicates time The instantaneous heart rate change rate under Indicates the collection time point, Indicates the heart rate signal at time The value of Indicates a time interval;

[0026] For each instantaneous heart rate change rate , calculate its third-order spectrum , the calculation expression is: ;

[0027] in, For signal The third-order spectrum of , which represents the nonlinear interaction of different frequency components, is the frequency variable, Indicates time delay, represents the imaginary unit, Represents the logarithm of the base of a natural number;

[0028] By analyzing the third-order spectrum, abnormal fluctuations in the instantaneous heart rate change rate are detected, and the abnormal coefficient of the instantaneous change rate is calculated. The calculation expression is: ;

[0029] Where, For time The abnormal coefficient of instantaneous heart rate change rate under is the absolute value of the third-order spectrum, is the maximum value of the third-order spectrum;

[0030] By analyzing the abnormal coefficient of instantaneous heart rate change rate during the entire monitoring period Perform weighted averaging to obtain the overall instantaneous heart rate change rate abnormality coefficient .

[0031] As a further solution of the present invention, the abnormality of the user's heart rate is divided into normal, slightly abnormal, and severely abnormal according to the judgment result, specifically including:

[0032] Determine whether the user's heart rate abnormality index in the current monitoring period is greater than or equal to the first threshold. If so, it is recorded as a serious abnormality. If not, determine whether the user's heart rate abnormality index in the current monitoring period is less than the second threshold. If so, it is recorded as normal. If not, it is recorded as a slight abnormality.

[0033] As a further solution of the present invention: the establishment of the user heart rate prediction model specifically includes:

[0034] For users with slight abnormalities, the user's instantaneous heart rate change rate abnormality coefficient and heart rate fluctuation amplitude coefficient are obtained, and a comprehensive feature vector is constructed with the instantaneous heart rate change rate abnormality coefficient and the heart rate fluctuation amplitude coefficient. The vector is used as the input of the machine learning model, and the machine learning model is trained using historical heart rate data. The user's heart rate abnormality index in the future monitoring period is used as the output of the model. Based on the output of the model, the degree of heart rate abnormality of the user in the future monitoring period is predicted, and a final prediction report is generated and fed back to the user. The machine learning model is a gradient boosting decision tree model.

[0035] As a further solution of the present invention, the prediction of the abnormal degree of the user's heart rate in the future period specifically includes:

[0036] According to the prediction results of the model, the heart rate abnormality index for a period of time in the future is output, the heart rate abnormality index is compared with the first threshold and the second threshold, and each monitoring period is divided into abnormality levels.

[0037] As a further solution of the present invention: based on the prediction results, the smartwatch generates a prediction report of future heart rate changes and provides feedback to the user, specifically including:

[0038] Based on the prediction results of the gradient boosting decision tree model, a heart rate fluctuation trend chart is generated for the user's future monitoring period. The heart rate fluctuation trend chart uses time as the horizontal axis and the predicted heart rate value as the vertical axis to intuitively reflect the dynamic changes in heart rate.

[0039] A dynamic heart rate monitoring system based on a smart watch, comprising:

[0040] A data acquisition module collects the user's heart rate data in real time through a smartwatch, and performs noise filtering and time alignment on the collected signal to ensure the data is accurate and reliable;

[0041] A feature extraction module, which performs segmented analysis on the processed heart rate signal to extract dynamic features in the time domain and frequency domain, including the heart rate fluctuation amplitude and instantaneous heart rate change rate;

[0042] A heart rate abnormality assessment module, which comprehensively analyzes the heart rate fluctuation amplitude and instantaneous heart rate change rate of the heart rate signal and assesses the abnormality of the current user's heart rate based on the analysis results;

[0043] A heart rate abnormality classification module, which classifies the abnormality of the user's heart rate into normal, slightly abnormal, and severely abnormal according to the judgment result;

[0044] A heart rate prediction module, which performs early warning processing based on severe anomalies. For minor anomalies, it establishes a user heart rate prediction model based on the user's heart rate fluctuation amplitude and instantaneous heart rate change rate to predict the user's heart rate abnormality level in the future;

[0045] A heart rate report generation module generates a prediction report of future heart rate changes based on the prediction results, and provides feedback to the user.

[0046] Beneficial effects of the present invention:

[0047] (1) The present invention collects the user's heart rate data in real time through the built-in photoelectric pulse wave sensor of the smart watch, and applies multiple optimization technologies in the data processing link to improve the accuracy and reliability of heart rate monitoring. The heart rate signal is filtered by a bandpass filter to accurately eliminate high-frequency and low-frequency noise introduced by factors such as ambient light interference, wrist movement and unstable wearing position, thereby ensuring the purity of the data from the source. In order to further adapt to individual differences and dynamic environments, a dynamic threshold adjustment method is adopted to dynamically optimize the filter parameters according to the user's activity status and real-time scene, ensuring that the filtering process is both robust and does not lose key heart rate change information. The signal characteristics are analyzed by adaptive time window segmentation, and the signal is aligned and drift corrected in combination with cross-correlation analysis to accurately process potential errors in the time dimension. The overall processing flow not only achieves high-quality preprocessing of the heart rate signal, but also retains the dynamic change characteristics of the heart rate, providing high-precision basic data support for subsequent feature extraction, anomaly analysis and predictive modeling. This multi-level optimization strategy fully demonstrates the adaptability and technological advancement of the present invention in the heart rate monitoring scenario.

[0048] (2) The present invention comprehensively analyzes the dynamic characteristics of the user's heart rate by constructing a comprehensive feature model of the heart rate fluctuation amplitude coefficient and the instantaneous heart rate change rate abnormality coefficient, and deeply explores the potential abnormal patterns in the heart rate signal. Combined with the efficient learning ability of the gradient boosting decision tree model, the model is accurately trained using historical heart rate data, and can accurately predict the degree of heart rate abnormality in the future monitoring period. The prediction results clearly present the dynamic change trajectory of the heart rate through the heart rate fluctuation trend graph, and mark the specific time when the potential abnormality occurs, intuitively showing the changing trend of health status. Combined with personalized data analysis, a detailed health report is generated, covering key indicators such as heart rate fluctuation amplitude, instantaneous change rate, abnormal index, etc., providing a forward-looking health assessment. For users with serious abnormalities, the system can quickly issue an early warning signal to remind them to seek medical treatment in time; for users with minor abnormalities, the prediction results provide scientific and personalized health guidance to help users optimize their health management strategies. Through the organic combination of real-time monitoring and trend prediction, the present invention greatly improves the comprehensiveness and practicality of health management. It can not only actively prevent and control health risks, but also provide users with long-term health intervention support, reflecting the perfect integration of technological advancement and service innovation. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The present invention will be further described below with reference to the accompanying drawings.

[0050] Figure 1 This is a flowchart of the specific steps of a dynamic heart rate monitoring method based on a smart watch of the present invention;

[0051] Figure 2 This is a flow chart of a dynamic heart rate monitoring system based on a smart watch in the present invention. DETAILED DESCRIPTION

[0052] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0053] See also Figure 1 As shown, the present invention is a dynamic heart rate monitoring method based on a smart watch, comprising the following steps:

[0054] S1: Collects the user's heart rate data in real time through a smartwatch, and performs noise filtering and time alignment on the collected signal to ensure data accuracy and reliability;

[0055] S2: Perform segmented analysis on the processed heart rate signal to extract dynamic features in the time domain and frequency domain, including the heart rate fluctuation amplitude and instantaneous heart rate change rate;

[0056] S3: Comprehensively analyzing the heart rate fluctuation amplitude and instantaneous heart rate change rate of the heart rate signal, and assessing the abnormality of the current user's heart rate based on the analysis results;

[0057] S4: Classifying the abnormality of the user's heart rate into normal, slightly abnormal, and severely abnormal according to the judgment result;

[0058] S5: Perform early warning processing based on serious anomalies. For minor anomalies, establish a user heart rate prediction model based on the user's heart rate fluctuation amplitude and instantaneous heart rate change rate to predict the user's heart rate abnormality level in the future;

[0059] S6: Based on the prediction results, the smartwatch generates a prediction report of future heart rate changes and provides feedback to the user. The report includes the predicted heart rate fluctuation trend and the time of occurrence of potential abnormalities.

[0060] In S1, the user's heart rate data is collected in real time through a smartwatch. The collected signal is then subjected to noise filtering and time alignment to ensure data accuracy and reliability. Specifically, the following steps are performed:

[0061] The smartwatch's built-in photoplethysmography sensor collects the user's heart rate data in real time and transmits the raw signal to the data processing module. Within the data processing module, a bandpass filter is used to filter the heart rate signal, eliminating high- and low-frequency noise caused by ambient light interference, wrist movement, or unstable wearing position, ensuring signal purity. Furthermore, a dynamic threshold adjustment method is used to dynamically optimize the filtering parameters based on the user's real-time activity status, further improving signal accuracy.

[0062] Next, the noise-filtered heart rate signal undergoes time alignment, using adaptive time windows to segment and analyze signal features to ensure temporal consistency. Cross-correlation analysis corrects for any signal drift or time misalignment, generating an accurate time series data set. Finally, the processed heart rate data serves as input for subsequent heart rate feature extraction and prediction models.

[0063] In S2, the processed heart rate signal is segmented and analyzed to extract dynamic features in the time and frequency domains, including the heart rate fluctuation amplitude and instantaneous heart rate change rate, specifically:

[0064] When performing segmented analysis on the processed heart rate signal, the signal is first divided into multiple time windows to capture dynamic changes within different time periods. Within each time period, a time-domain feature extraction method is used to calculate the heart rate fluctuation amplitude—the difference between the maximum and minimum heart rate values ​​within that period—to reflect the heart rate fluctuation range. Simultaneously, the mean and standard deviation of each signal segment are calculated to further analyze the concentration and volatility of the heart rate within that time period. These time-domain features reveal the underlying fluctuation patterns of the heart rate, providing a data foundation for subsequent anomaly detection and trend prediction.

[0065] In frequency domain analysis, a fast Fourier transform is performed on each signal segment to extract the signal's spectral information. By analyzing the low- and high-frequency components in the spectrum, the spectral bandwidth and the ratio of high to low frequencies are calculated. This frequency domain feature helps analyze the frequency distribution of heart rate variability and further reveals the regulatory state of the autonomic nervous system. Furthermore, the instantaneous heart rate change rate reflects rapid changes in heart rate by calculating the ratio of the heart rate difference between adjacent time points to the time interval. Combining the analysis of time and frequency domain features allows for a comprehensive assessment of heart rate fluctuation patterns and changing trends.

[0066] In S3, the heart rate fluctuation amplitude and instantaneous heart rate change rate of the heart rate signal are comprehensively analyzed. Based on the analysis results, the abnormality of the current user's heart rate is evaluated, specifically including:

[0067] During a monitoring cycle, the heart rate fluctuation amplitude coefficient is calculated based on the degree of heart rate fluctuation in each time period, the mean of the instantaneous heart rate change rate in each time period is calculated to obtain the instantaneous heart rate change rate in each time period, the instantaneous heart rate change rate abnormality coefficient within the monitoring cycle is calculated based on the degree of fluctuation of the instantaneous heart rate change rate in each time period, the heart rate fluctuation amplitude coefficient and the instantaneous heart rate change rate abnormality coefficient are normalized and calculated to obtain the user's heart rate abnormality index within the current monitoring cycle;

[0068] Determine whether the user's heart rate abnormality index in the current monitoring period is greater than or equal to a first threshold. If so, it is recorded as a severe abnormality. If not, determine whether the user's heart rate abnormality index in the current monitoring period is less than a second threshold. If so, it is recorded as normal. If not, it is recorded as a slight abnormality.

[0069] The process of obtaining the heart rate fluctuation amplitude coefficient is as follows:

[0070] De-noise the heart rate signal to remove high-frequency noise and obtain the pre-processed signal ,in, Indicates the collection time period;

[0071] Perform wavelet transform on the preprocessed heart rate signal and decompose it into multiple frequency band signals ,in Indicates the scale of the frequency band, and the coefficient of each frequency band is obtained by wavelet transform , represents the contribution of the signal in the corresponding frequency band, and the decomposed signals of each frequency band are obtained;

[0072] For each time period and each frequency band , calculate the fluctuation amplitude of the corresponding frequency band , the fluctuation amplitude is expressed by the difference between the maximum and minimum values ​​of the frequency band signal;

[0073] By weighted averaging the fluctuation amplitudes of all frequency bands, we can get the The heart rate fluctuation amplitude coefficient is calculated as follows: ;

[0074] Where, For time period The heart rate fluctuation amplitude coefficient, For the The weight of each frequency band represents the contribution of the corresponding frequency band to the total fluctuation amplitude. The calculation formula is: ;

[0075] Where, is the normalized band energy ratio, For the The value of the frequency band signal;

[0076] The overall heart rate fluctuation coefficient is obtained by averaging the heart rate fluctuation coefficients of all time periods in the entire monitoring period. ;

[0077] It should be noted that: through the calculation steps of the heart rate fluctuation amplitude coefficient, the present invention can accurately extract the dynamic fluctuation characteristics of the heart rate signal based on wavelet transform, thereby calculating the heart rate fluctuation amplitude coefficient with high precision. Different from the traditional methods of the prior art, it provides an innovative fluctuation amplitude calculation process with multi-scale decomposition.

[0078] The process of obtaining the abnormal coefficient of the instantaneous heart rate change rate is as follows:

[0079] Calculate the rate of change of the heart rate signal in each time period, and calculate the instantaneous rate of change of the heart rate signal using the differential formula:

[0080] The calculation expression is: ;

[0081] Where, Indicates time The instantaneous heart rate change rate under Indicates the collection time point, Indicates the heart rate signal at time The value of Indicates a time interval;

[0082] The third-order spectrum analysis is used to evaluate the high-order statistical characteristics of the heart rate change rate and identify the nonlinear characteristics in the signal; for each instantaneous heart rate change rate , calculate its third-order spectrum , the third-order spectrum reflects the interaction of different frequency components in the signal, and the calculation expression is: ;

[0083] in, For signal The third-order spectrum of , which represents the nonlinear interaction of different frequency components, is the frequency variable, Indicates time delay, represents the imaginary unit, Represents the logarithm of the base of a natural number;

[0084] By analyzing the third-order spectrum, abnormal fluctuations in the instantaneous heart rate change rate are detected; the abnormal coefficient of the instantaneous change rate is calculated, and the calculation expression is: ;

[0085] Where, For time The abnormal coefficient of instantaneous heart rate change rate under is the absolute value of the third-order spectrum, reflecting the nonlinear intensity of the frequency component, is the maximum value of the third-order spectrum, which serves as the reference standard;

[0086] By analyzing the abnormal coefficient of instantaneous heart rate change rate during the entire monitoring period Perform weighted averaging to obtain the overall instantaneous heart rate change rate abnormality coefficient .

[0087] It should be noted that: by obtaining the abnormal coefficient of instantaneous heart rate change rate, abnormal fluctuations in the heart rate change rate can be accurately calculated and detected, and high-order spectrum analysis can be used to capture nonlinear characteristics, providing an effective tool for abnormality detection.

[0088] The calculation expression of the abnormal heart rate index is: ;

[0089] Where, Indicates heart rate abnormality index, represents the heart rate fluctuation amplitude coefficient, Indicates the abnormal coefficient of instantaneous heart rate change rate, and represents the preset scale factor, and and Both are greater than 0.

[0090] In S5, an early warning process is performed based on severe anomalies. For minor anomalies, a user heart rate prediction model is established based on the user's heart rate fluctuation amplitude and instantaneous heart rate change rate to predict the user's heart rate abnormality level in the future. Specifically, the model includes:

[0091] If it is detected that the user's heart rate is in a seriously abnormal state during the monitoring period, an early warning signal will be immediately issued through the smart watch, and the user will need to go to the hospital for further heart rate testing.

[0092] For users with mild heart rate abnormalities, the user's instantaneous heart rate change rate abnormality coefficient and heart rate fluctuation amplitude coefficient are obtained, and a comprehensive feature vector is constructed from the instantaneous heart rate change rate abnormality coefficient and the heart rate fluctuation amplitude coefficient. This is used as the input of the machine learning model, and the machine learning model is trained using historical heart rate data. The user's heart rate abnormality index in the future monitoring period is used as the model output. Based on the model output, the user's heart rate abnormality level in the future monitoring period is predicted, and a final prediction report is generated and fed back to the user. The machine learning model is a gradient boosting decision tree model.

[0093] The training process of the gradient boosting decision tree model is:

[0094] Historical heart rate data is preprocessed to remove noise and ensure the quality of training data. A gradient boosting decision tree is constructed and each decision tree is iteratively trained to minimize the model's loss function. The weights and biases of each tree are adjusted using a gradient descent optimization algorithm to improve the model's prediction accuracy.

[0095] In S6, based on the prediction results, the smartwatch generates a forecast report of future heart rate changes and provides feedback to the user. The report includes the predicted heart rate fluctuation trend and the time of potential abnormalities, including:

[0096] Based on the prediction results of the gradient boosting decision tree model, a heart rate fluctuation trend chart for the user's future monitoring period is generated. The chart uses time as the horizontal axis and the predicted heart rate value as the vertical axis to intuitively reflect the dynamic changes in heart rate.

[0097] Identify the time period during the monitoring period when the heart rate anomaly index exceeds the set threshold, mark it as the time of potential anomaly occurrence, and highlight it in the report;

[0098] Provides detailed analysis data for each potential abnormal time period, including the heart rate fluctuation amplitude coefficient, the instantaneous heart rate change rate abnormal coefficient, and the specific value of the comprehensive heart rate abnormality index;

[0099] By analyzing heart rate fluctuation trends and potential abnormal data, personalized health guidance is generated, and users are fed back abnormal signals that may need attention.

[0100] See also Figure 2 As shown, a dynamic heart rate monitoring system based on a smart watch includes:

[0101] A data acquisition module collects the user's heart rate data in real time through a smartwatch, and performs noise filtering and time alignment on the collected signal to ensure the data is accurate and reliable;

[0102] A feature extraction module, which performs segmented analysis on the processed heart rate signal to extract dynamic features in the time domain and frequency domain, including the heart rate fluctuation amplitude and instantaneous heart rate change rate;

[0103] A heart rate abnormality assessment module, which comprehensively analyzes the heart rate fluctuation amplitude and instantaneous heart rate change rate of the heart rate signal and assesses the abnormality of the current user's heart rate based on the analysis results;

[0104] A heart rate abnormality classification module, which classifies the abnormality of the user's heart rate into normal, slightly abnormal, and severely abnormal according to the judgment result;

[0105] A heart rate prediction module, which performs early warning processing based on severe anomalies. For minor anomalies, it establishes a user heart rate prediction model based on the user's heart rate fluctuation amplitude and instantaneous heart rate change rate to predict the user's heart rate abnormality level in the future;

[0106] A heart rate report generation module generates a prediction report of future heart rate changes based on the prediction results, and provides feedback to the user.

[0107] The present invention works by collecting the user's heart rate data in real time through a smartwatch, filtering the signal for noise and aligning the time series to ensure data accuracy and reliability. The processed heart rate signal is segmented and analyzed to extract dynamic features in the time and frequency domains, including heart rate fluctuation amplitude and instantaneous heart rate change rate. The extracted features are then comprehensively analyzed to calculate a heart rate abnormality index (HRI), assess the degree of abnormality in the current user's heart rate, and categorize the abnormality into normal, mild, or severe. For severe abnormalities, an immediate warning feedback is triggered, prompting the user to take medical measures. For mild abnormalities, a comprehensive feature vector is constructed by obtaining the heart rate fluctuation amplitude coefficient and the instantaneous heart rate change rate abnormality coefficient. The HRI is then predicted for the next monitoring period using a gradient boosting decision tree model. A prediction report is then generated, including a heart rate fluctuation trend chart and the time of potential abnormality occurrence, providing personalized health solutions. The system comprises a data acquisition module, a feature extraction module, a heart rate abnormality assessment module, a heart rate abnormality classification module, a heart rate prediction module, and a report generation module. It enables full-cycle monitoring of the user's heart rate, real-time analysis, and abnormality prediction, providing efficient health management services.

[0108] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0109] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0110] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0111] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0112] The above is a detailed description of an embodiment of the present invention. However, the content described is only a preferred embodiment of the present invention and should not be considered to limit the scope of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A dynamic heart rate monitoring method based on a smart watch, characterized in that: The following steps are involved: S1: Filters noise and aligns the user's heart rate signal to ensure data accuracy and reliability; S2: Perform segmented analysis on the processed heart rate signal to extract dynamic features in the time domain and frequency domain, including the heart rate fluctuation amplitude and instantaneous heart rate change rate; S3: Comprehensively analyze the heart rate fluctuation amplitude and instantaneous heart rate change rate of the heart rate signal, and evaluate the abnormality of the current user's heart rate based on the analysis results, specifically including: During a monitoring cycle, the heart rate fluctuation amplitude coefficient is calculated based on the degree of heart rate fluctuation in each time period, the mean of the instantaneous heart rate change rate in each time period is calculated to obtain the instantaneous heart rate change rate in each time period, the instantaneous heart rate change rate abnormality coefficient within the monitoring cycle is calculated based on the degree of fluctuation of the instantaneous heart rate change rate in each time period, the heart rate fluctuation amplitude coefficient and the instantaneous heart rate change rate abnormality coefficient are normalized and calculated to obtain the user's heart rate abnormality index within the current monitoring cycle; The process of obtaining the heart rate fluctuation amplitude coefficient is as follows: Applying wavelet transform to the denoised heart rate signal, the wavelet transform decomposes the signal into multiple frequency bands, each of which contains signal components within a different frequency range. Through wavelet transform, the signal coefficients of the multiple frequency bands are obtained; The fluctuation amplitude is obtained by calculating the difference between the maximum and minimum values ​​of each frequency band signal in each time period. The heart rate fluctuation amplitude coefficient in each time period is obtained by taking a weighted average of the fluctuation amplitudes of all frequency bands. The heart rate fluctuation amplitude coefficient in the entire monitoring period is averaged to calculate the overall heart rate fluctuation amplitude coefficient. The process of obtaining the abnormal coefficient of the instantaneous heart rate change rate is as follows: Calculate the rate of change of the heart rate signal in each time period, and calculate the instantaneous rate of change of the heart rate signal by the differential formula. The calculation expression is: ; Where, Indicates time The instantaneous heart rate change rate under Indicates the collection time point, Indicates the heart rate signal at time The value of Indicates a time interval; For each instantaneous heart rate change rate , calculate its third-order spectrum , the calculation expression is: , in, For signal The third-order spectrum of , which represents the nonlinear interaction of different frequency components, is the frequency variable, Indicates time delay, represents the imaginary unit, Represents the logarithm of the base of a natural number; By analyzing the third-order spectrum, abnormal fluctuations in the instantaneous heart rate change rate are detected, and the abnormal coefficient of the instantaneous change rate is calculated. The calculation expression is: ; Where, For time The abnormal coefficient of instantaneous heart rate change rate under is the absolute value of the third-order spectrum, is the maximum value of the third-order spectrum; By analyzing the abnormal coefficient of instantaneous heart rate change rate during the entire monitoring period Perform weighted averaging to obtain the overall instantaneous heart rate change rate abnormality coefficient ; S4: Classifying the abnormality of the user's heart rate into normal, slightly abnormal, and severely abnormal according to the judgment result; S5: Perform early warning processing based on serious anomalies. For minor anomalies, establish a user heart rate prediction model based on the user's heart rate fluctuation amplitude and instantaneous heart rate change rate to predict the user's heart rate abnormality level in the future; S6: Based on the prediction results, the smartwatch generates a prediction report of future heart rate changes and provides feedback to the user.

2. The method for dynamic heart rate monitoring based on a smart watch according to claim 1, characterized in that: The segmented analysis of the processed heart rate signal specifically includes: The processed heart rate signal is segmented, and the length of each segment is determined by a fixed time window; For each time period, the heart rate fluctuation amplitude is obtained by calculating the difference between the maximum heart rate value and the minimum heart rate value in the signal within each time period; In each time period, the instantaneous heart rate change rate was obtained by calculating the heart rate difference between adjacent time points and dividing it by the time interval.

3. The method for dynamic heart rate monitoring based on a smart watch according to claim 1, characterized in that: According to the judgment result, the abnormality of the user's heart rate is divided into normal, slightly abnormal and seriously abnormal, specifically including: Determine whether the user's heart rate abnormality index in the current monitoring period is greater than or equal to the first threshold. If so, it is recorded as a serious abnormality. If not, determine whether the user's heart rate abnormality index in the current monitoring period is less than the second threshold. If so, it is recorded as normal. If not, it is recorded as a slight abnormality.

4. The method for dynamic heart rate monitoring based on a smart watch according to claim 1, characterized in that: The step of establishing a user heart rate prediction model specifically includes: For users with slight abnormalities, the user's instantaneous heart rate change rate abnormality coefficient and heart rate fluctuation amplitude coefficient are obtained, and a comprehensive feature vector is constructed with the instantaneous heart rate change rate abnormality coefficient and the heart rate fluctuation amplitude coefficient. The vector is used as the input of the machine learning model, and the machine learning model is trained using historical heart rate data. The user's heart rate abnormality index in the future monitoring period is used as the output of the model. Based on the output of the model, the degree of heart rate abnormality of the user in the future monitoring period is predicted, and a final prediction report is generated and fed back to the user. The machine learning model is a gradient boosting decision tree model.

5. The method for dynamic heart rate monitoring based on a smart watch according to claim 1, characterized in that: The prediction of the abnormal degree of the user's heart rate in the future period specifically includes: According to the prediction results of the model, the heart rate abnormality index for a period of time in the future is output, the heart rate abnormality index is compared with the first threshold and the second threshold, and each monitoring period is divided into abnormality levels.

6. The method for dynamic heart rate monitoring based on a smart watch according to claim 1, characterized in that: Based on the prediction results, the smartwatch generates a prediction report of future heart rate changes and provides feedback to the user, specifically including: Based on the prediction results of the gradient boosting decision tree model, a heart rate fluctuation trend chart is generated for the user's future monitoring period. The heart rate fluctuation trend chart uses time as the horizontal axis and the predicted heart rate value as the vertical axis to intuitively reflect the dynamic changes in heart rate.

7. A dynamic heart rate monitoring system based on a smart watch, characterized in that: A method for dynamic heart rate monitoring based on a smartwatch as described in any one of claims 1 to 6, comprising: A data acquisition module collects the user's heart rate data in real time through a smartwatch, and performs noise filtering and time alignment on the collected signal to ensure the data is accurate and reliable; A feature extraction module, which performs segmented analysis on the processed heart rate signal to extract dynamic features in the time domain and frequency domain, including the heart rate fluctuation amplitude and instantaneous heart rate change rate; A heart rate abnormality assessment module, which comprehensively analyzes the heart rate fluctuation amplitude and instantaneous heart rate change rate of the heart rate signal and assesses the abnormality of the current user's heart rate based on the analysis results; A heart rate abnormality classification module, which classifies the abnormality of the user's heart rate into normal, slightly abnormal, and severely abnormal according to the judgment result; A heart rate prediction module, which performs early warning processing based on severe anomalies. For minor anomalies, it establishes a user heart rate prediction model based on the user's heart rate fluctuation amplitude and instantaneous heart rate change rate to predict the user's heart rate abnormality level in the future; A heart rate report generation module generates a prediction report of future heart rate changes based on the prediction results, and provides feedback to the user.

Images