Personalized health monitoring method based on multi-instrument data fusion

Abstract

The invention provides a personalized health monitoring method based on multi-instrument data fusion, and aims to solve the problems of non-uniform data standards and poor interoperability of existing multiple health monitoring devices. According to the method, a data standardization protocol is automatically selected and optimized according to data characteristics (such as sampling frequency, data precision and the like) of different devices (such as oximeters, sleep monitors and the like) through a machine learning algorithm, and seamless fusion of data between the devices is ensured. Different from a traditional static data processing mode, the scheme not only improves the collaboration efficiency among the devices, but also effectively solves the problems of inconsistent data formats and information loss. For different types of data generated by multiple devices, a data fusion strategy is automatically identified and optimized by adopting an intelligent fusion algorithm based on deep learning, and information conflict or delay caused by mismatching of data features is avoided.

Description

Technical Field

[0001] The present application relates to the technical field of document format detection, and in particular to a personalized health monitoring method based on multi-instrument data fusion. Background Art

[0002] With the widespread use of health monitoring equipment, the need for multiple devices (such as blood oximeters, heart rate monitors, sleep monitors, etc.) to work together has gradually emerged in hospital and home scenarios. However, these devices are often produced by different manufacturers, and their data formats, sampling frequencies, and transmission protocols vary, making it impossible to directly integrate and analyze the data. For example, in actual scenarios, some health management platforms need to process high-frequency updated blood oxygen data and low-frequency sampled sleep monitoring data at the same time, but the format conflicts and time alignment problems of the data in traditional methods greatly reduce the accuracy and real-time performance of the monitoring results. In addition, the collaborative work of multiple devices is often delayed due to inconsistent data standards, and may even result in important health events (such as abnormal heart rate or sudden drop in blood oxygen) not being warned in a timely manner.

[0003] In the prior art, in order to achieve data sharing and collaboration among multiple health monitoring devices, a simple unified format conversion or protocol compatibility approach is often used. However, this approach has the following problems:

[0004] 1. The data formats and collection methods of various devices are different. Direct conversion may lose important data information or cause data distortion.

[0005] 2. Existing data fusion methods usually rely on static processing rules, which are difficult to adapt to the data characteristics of different devices, resulting in the lack of accuracy of the fused data.

[0006] 3. Most existing methods do not take into account the real-time collaboration between devices and cannot give full play to the advantages of multi-device collaborative work.

[0007] To solve these problems, the industry has tried to adopt dynamic data standardization and fusion methods based on machine learning algorithms, hoping to automatically adjust the data processing process according to device characteristics. However, such technologies still face problems such as difficult data synchronization and high computing load, resulting in slow system response speed and affecting actual application effects.

[0008] Therefore, how to achieve seamless coordination and optimized integration of data among multiple devices while ensuring data accuracy and real-time performance has become a technical problem to be solved by the present invention. Summary of the invention

[0009] The present application provides a personalized health monitoring method based on multi-instrument data fusion to solve the problems existing in the prior art.

[0010] The present application provides a personalized health monitoring method based on multi-instrument data fusion, comprising the following steps: S1. Receive data from a variety of health monitoring devices, including but not limited to a blood oximeter, a heart rate monitor, and a sleep monitor; S2. extracting a feature vector of each health monitoring device data, wherein the feature vector includes data frequency, data volume, data accuracy and data format; S3. Based on the extracted feature vectors, a pre-trained classification model is used to select a standardized protocol that matches the data format of each device. The specific method is as follows: Construction and training of classification models: Based on the historical data of various health monitoring devices, the classification model is trained using a supervised learning algorithm to ensure that the classification model can accurately identify the format characteristics of data from different devices; Dynamic matching protocol: The system extracts the feature vector of device data in real time and inputs it into the classification model. The model outputs the optimal standardized protocol based on the input feature vector and verifies the adaptability of the protocol in real time. Specific content of the standardization protocol: including data format conversion rules, encoding requirements, and consistency calibration of data accuracy, to ensure that data from different devices can be losslessly converted into a unified standard format; S4. Convert and encode data in different formats to meet the requirements of a unified format and generate standardized data; specifically, the following steps are included: Data format conversion: Perform unified processing of data formats according to the standardized protocol selected in S3, including but not limited to converting JSON format data into XML format, or reorganizing segmented stored data into continuous structured data; Data encoding: Encode data according to the encoding rules defined by the protocol, including character set conversion to UTF-8 and standardization of data structure; Verification and generation: After completing format conversion and encoding, the data integrity and consistency are checked using the verification rules specified by the protocol. After passing the verification, standardized data that meets the requirements of the standardized protocol is generated; S5. Performing weighted fusion processing on the standardized data, including the following steps: Application of deep learning algorithms: weighted fusion of data from different devices through deep learning algorithms; Calculation and adjustment of weighting coefficients: a. The weighting coefficients are dynamically adjusted based on the reliability and relevance of the equipment data. The reliability is calculated by the error variance of the equipment data, and the error variance is calculated by the error distribution of historical data or the outlier elimination method of the equipment data; b. The correlation analysis method is used to calculate the correlation between different equipment data. The correlation analysis results are used as auxiliary parameters for weight adjustment. When the data correlation between the equipment is high, the joint weight is increased accordingly; c. The calculation formula for the final weighting coefficient is:

[0011] in, For the The weighting factor of each device, is the error variance of the equipment data; Anomaly detection and data verification: Perform anomaly detection on the fused data, recalculate the weighting coefficient by eliminating abnormal data points, and ensure the accuracy of the fused data; Real-time verification: By comparing the feature consistency of fused data with the original data, the accuracy and completeness of the data are verified to ensure that the fusion result meets the requirements; S6. Ensure the security and transparency of data transmission. The device data and its identifiers are recorded on the blockchain through blockchain technology. The blockchain records ensure that the data cannot be tampered with, and the integrity and consistency of the data transmission process are verified through smart contracts. S7. Dynamically adjust the parameters of data input, storage and output according to the device type and application scenario, including data update frequency, data transmission method and data storage format to optimize system performance and ensure real-time and high efficiency.

[0012] In some embodiments, the step of extracting feature vectors includes: preprocessing the data of each health monitoring device, wherein the preprocessing includes data cleaning, denoising and normalization; and extracting feature vectors from the preprocessed data using a feature extraction algorithm, wherein the feature extraction algorithm includes principal component analysis PCA or linear discriminant analysis LDA.

[0013] In some embodiments, the training step of the classification model includes: collecting historical data of multiple health monitoring devices, wherein the historical data includes device type, data format and data quality; based on the historical data, using a supervised learning algorithm to train the classification model, wherein the supervised learning algorithm includes a support vector machine SVM or a decision tree.

[0014] In some embodiments, the weighted fusion step also includes: performing anomaly detection on the fused data, identifying and eliminating abnormal data points; recalculating the weighting coefficients to optimize the accuracy of the fused data based on the results after eliminating the abnormal data; the historical data in step S3 includes device type, data format, sampling frequency, data accuracy and data volume; supervised learning algorithms such as support vector machines or decision trees.

[0015] In some embodiments, blockchain technology is used to ensure the security and transparency of data transmission. The blockchain records ensure that the data cannot be tampered with, and the integrity and consistency of the data transmission process are verified through smart contracts.

[0016] In some embodiments, the parameters of data input, storage and output are dynamically adjusted according to the device type and application scenario, wherein the parameters include data update frequency, data transmission method and data storage format.

[0017] In some embodiments, the weighted fusion step further includes: analyzing the correlation between data from different devices by a correlation analysis method, and adjusting the calculation method of the weighting coefficient according to the analyzed correlation to enhance the representativeness of the fused data.

[0018] In some embodiments, the data verification step includes: performing a feature consistency check on the fused data to check whether the data frequency, data volume and data accuracy are consistent; if the consistency is lower than a preset threshold, triggering the data re-fusion process.

[0019] In some embodiments, the dynamic adjustment step includes: identifying the current application scenario, which includes real-time monitoring, long-term analysis, and emergency warning, and adjusting the data update frequency and transmission method according to the identified application scenario.

[0020] In some embodiments, smart contracts are used to automatically verify the integrity and consistency of data, and trigger a data re-collection or verification process when verification fails.

[0021] This application designs a dynamic and adaptive data standardization engine, which can automatically select and optimize data standardization protocols according to the characteristics of different health monitoring devices (such as blood oximeters, sleep monitors, etc.) in the conversion between hospital and home scenarios. Unlike the traditional static data processing method, this solution uses a machine learning algorithm to automatically perform standardization processing according to the output data characteristics of each device (such as sampling frequency, data accuracy, etc.). It ensures the seamless integration of data between different devices, not only improves the collaborative work efficiency between devices, but also solves the problems of inconsistent data formats and information loss between devices in the prior art, greatly improving the comprehensiveness and accuracy of health monitoring; and for different types of data generated by multiple devices (such as blood oximeters, heart rate monitors, sleep monitors, etc.), the present invention proposes an intelligent fusion algorithm based on deep learning. This algorithm can automatically identify and analyze the characteristics of each device data, ensuring that when different devices provide data at the same time, the system can accurately identify the characteristics of various types of data and perform optimal fusion processing. Different from the simple data fusion in the existing technology, this solution intelligently adjusts the data fusion strategy through deep learning, optimizes data quality, and avoids information conflicts or processing delays caused by data feature mismatch. It not only improves the accuracy of data fusion, but also solves the coordination problem between devices, providing smarter and more comprehensive support for health monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0023] Figure 1 A flowchart of the steps of the personalized health monitoring method based on multi-instrument data fusion provided in this application; Figure 2 This is an architectural diagram of the personalized health monitoring method based on multi-instrument data fusion provided in this application.

[0024] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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 creative work are within the scope of protection of the present invention.

[0026] The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms "a", "an" and "the" used in the embodiments of the present application are also intended to include plural forms, unless the context clearly indicates other meanings.

[0027] It should be noted that when an element is referred to as being "fixed on" or "set on" another element, it can be directly on the other element or indirectly set on the other element; when an element is referred to as being "connected to" another element, it can be directly connected to the other element or indirectly connected to the other element.

[0028] It should be understood that the term "and / or" used in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0029] As used herein, the words "if" and "if" may be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrases "if it is determined" or "if (stated condition or event) is detected" may be interpreted as "when it is determined" or "in response to determining" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)", depending on the context.

[0030] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a product or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such a product or system. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the product or system including the elements.

[0031] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0032] Figure 1 A flowchart of the steps of the personalized health monitoring method based on multi-instrument data fusion provided in this application; Figure 2The architecture diagram of the personalized health monitoring method based on multi-instrument data fusion provided in this application is as follows: Figure 1 , Figure 2 As shown, the present application provides a personalized health monitoring method based on multi-instrument data fusion, including: S1. Receive data from a variety of health monitoring devices, including but not limited to a blood oximeter, a heart rate monitor, and a sleep monitor; S2. extracting a feature vector of each health monitoring device data, wherein the feature vector includes data frequency, data volume, data accuracy and data format; S3. Based on the extracted feature vectors, a pre-trained classification model is used to select a standardized protocol that matches the data format of each device. The specific method is as follows: Construction and training of classification models: Based on the historical data of various health monitoring devices, the classification model is trained using a supervised learning algorithm to ensure that the classification model can accurately identify the format characteristics of data from different devices; Dynamic matching protocol: The system extracts the feature vector of device data in real time and inputs it into the classification model. The model outputs the optimal standardized protocol based on the input feature vector and verifies the adaptability of the protocol in real time. Specific content of the standardization protocol: including data format conversion rules, encoding requirements, and consistency calibration of data accuracy, to ensure that data from different devices can be losslessly converted into a unified standard format; S4. Convert and encode data in different formats to meet the requirements of a unified format and generate standardized data; specifically, the following steps are included: Data format conversion: Perform unified processing of data formats according to the standardized protocol selected in S3, including but not limited to converting JSON format data into XML format, or reorganizing segmented stored data into continuous structured data; Data encoding: Encode data according to the encoding rules defined by the protocol, including character set conversion to UTF-8 and standardization of data structure; Verification and generation: After completing format conversion and encoding, the data integrity and consistency are checked using the verification rules specified by the protocol. After passing the verification, standardized data that meets the requirements of the standardized protocol is generated; S5. Performing weighted fusion processing on the standardized data, including the following steps: Application of deep learning algorithms: weighted fusion of data from different devices through deep learning algorithms; Calculation and adjustment of weighting coefficients: a. The weighting coefficients are dynamically adjusted based on the reliability and relevance of the equipment data. The reliability is calculated by the error variance of the equipment data, and the error variance is calculated by the error distribution of historical data or the outlier elimination method of the equipment data; b. The correlation analysis method is used to calculate the correlation between different equipment data. The correlation analysis results are used as auxiliary parameters for weight adjustment. When the data correlation between the equipment is high, the joint weight is increased accordingly; c. The calculation formula for the final weighting coefficient is:

[0033] in, For the The weighting factor of each device, is the error variance of the equipment data; Anomaly detection and data verification: Perform anomaly detection on the fused data, recalculate the weighting coefficient by eliminating abnormal data points, and ensure the accuracy of the fused data; Real-time verification: By comparing the feature consistency of fused data with the original data, the accuracy and completeness of the data are verified to ensure that the fusion result meets the requirements; S6. Ensure the security and transparency of data transmission. The device data and its identifiers are recorded on the blockchain through blockchain technology. The blockchain records ensure that the data cannot be tampered with, and the integrity and consistency of the data transmission process are verified through smart contracts. S7. Dynamically adjust the parameters of data input, storage and output according to the device type and application scenario, including data update frequency, data transmission method and data storage format to optimize system performance and ensure real-time and high efficiency.

[0034] In some embodiments, the step of extracting feature vectors includes: preprocessing the data of each health monitoring device, wherein the preprocessing includes data cleaning, denoising and normalization; and extracting feature vectors from the preprocessed data using a feature extraction algorithm, wherein the feature extraction algorithm includes principal component analysis PCA or linear discriminant analysis LDA.

[0035] In some embodiments, the training step of the classification model includes: collecting historical data of multiple health monitoring devices, wherein the historical data includes device type, data format and data quality; based on the historical data, using a supervised learning algorithm to train the classification model, wherein the supervised learning algorithm includes a support vector machine SVM or a decision tree.

[0036] In some embodiments, the weighted fusion step also includes: performing anomaly detection on the fused data, identifying and eliminating abnormal data points; recalculating the weighting coefficients to optimize the accuracy of the fused data based on the results after eliminating the abnormal data; the historical data in step S3 includes device type, data format, sampling frequency, data accuracy and data volume; supervised learning algorithms such as support vector machines or decision trees.

[0037] In some embodiments, blockchain technology is used to ensure the security and transparency of data transmission. The blockchain records ensure that the data cannot be tampered with, and the integrity and consistency of the data transmission process are verified through smart contracts.

[0038] In some embodiments, the parameters of data input, storage and output are dynamically adjusted according to the device type and application scenario, wherein the parameters include data update frequency, data transmission method and data storage format.

[0039] In some embodiments, the weighted fusion step further includes: analyzing the correlation between data from different devices by a correlation analysis method, and adjusting the calculation method of the weighting coefficient according to the analyzed correlation to enhance the representativeness of the fused data.

[0040] In some embodiments, the data verification step includes: performing a feature consistency check on the fused data to check whether the data frequency, data volume and data accuracy are consistent; if the consistency is lower than a preset threshold, triggering the data re-fusion process.

[0041] In some embodiments, the dynamic adjustment step includes: identifying the current application scenario, which includes real-time monitoring, long-term analysis, and emergency warning, and adjusting the data update frequency and transmission method according to the identified application scenario.

[0042] In some embodiments, smart contracts are used to automatically verify the integrity and consistency of data, and trigger a data re-collection or verification process when verification fails.

[0043] Embodiment 1: In a comprehensive smart health management platform, users continuously collect health data through different types of health monitoring devices, such as blood oximeters, heart rate monitors, and sleep monitors. Traditional health monitoring systems often face problems such as inconsistent data formats, differences in standards, and poor interoperability among multiple devices. For example, a blood oximeter may collect high-precision data dozens of times per minute, while a sleep monitor records low-frequency data once a day, which is often not effectively integrated in traditional systems. The traditional approach is to rely on each device to output a standardized format, however, this approach not only has the risk of data loss, but also cannot flexibly adapt to the different operating frequencies and accuracy requirements of the equipment.

[0044] In order to solve the above problems, this embodiment combines a dynamically adaptive data standardization engine and an intelligent data fusion algorithm based on deep learning. In this implementation process, the data standardization engine automatically selects the most suitable standardized protocol according to the characteristics of the device, ensuring that the data from the blood oximeter to the sleep monitor can be seamlessly integrated between platforms. For example, the high-frequency data of the blood oximeter dynamically adjusts its sampling frequency through the intelligent protocol to adapt it to the unified format of the system, while the low-frequency data of the sleep monitor is adapted to the same platform according to its sampling period.

[0045] At the same time, the standardized data is fused through an intelligent algorithm based on deep learning. This algorithm can not only intelligently identify the characteristics of device data, but also dynamically adjust the weighting coefficient according to the actual situation of the data. For example, when the system processes the data of the blood oximeter and the heart rate monitor, it will recognize the close relationship between the two, and dynamically adjust the priority and method of data fusion through the deep learning model, avoiding information conflicts or delays caused by data mismatch. The introduction of the dynamic data standardization engine effectively solves the problem of inconsistent data formats between different devices in traditional methods, so that data from various devices can be shared in real time and losslessly fused. The intelligent data fusion algorithm based on deep learning can automatically identify and optimize the data fusion process, ensuring that all data can be accurately processed regardless of the source, avoiding errors caused by data conflicts or mismatches in traditional systems. Most importantly, the system's real-time processing capability ensures rapid data updates and instant feedback, greatly improving the accuracy and timeliness of health monitoring.

[0046] In actual applications, the system can monitor the health data of multiple devices in real time and conduct unified analysis. Whether it is a blood oximeter, heart rate monitor or sleep monitor, it can provide high-quality data output through a unified standardized protocol and intelligent fusion algorithm. This solution not only improves the efficiency of multi-device collaboration, but also ensures the integrity and continuity of the health management system, avoids the inefficiency of inter-device collaboration and data loss in traditional methods, and provides users with a more accurate and comprehensive health monitoring experience.

[0047] Embodiment 2: In an actual application scenario, in order to solve the problems of inconsistent data sources, fuzzy selection of standardized protocols, and incomplete logic of weighted coefficient calculation that may be encountered in the process of multi-device data fusion, this embodiment proposes an improved technical implementation method, which further improves the operability and practicality of the system by detailed design of data processing steps and algorithm logic.

[0048] First, this embodiment adds a dynamic verification module to the data processing process. This module aims to generate a multidimensional data characteristic matrix through dynamic analysis of historical data based on the differences in data formats of different devices. This matrix records the core parameters such as the sampling frequency, accuracy range and noise characteristics of the device data, and is used to guide the adaptive selection of data standardization protocols. The specific process includes:

[0049] Data characteristic matrix generation: Clean and preprocess historical data, use principal component analysis (PCA) to extract key characteristic variables, including equipment noise level, sampling period, etc., and store these variables as a multidimensional matrix for subsequent dynamic analysis.

[0050] Optimal selection of standardized protocols: Based on the matching calculation of real-time collected data and historical matrix, the support vector machine (SVM) model based on supervised learning is used to dynamically predict the optimal standardized protocol, and its execution parameters are adjusted according to the actual working status of the equipment.

[0051] Secondly, in order to avoid the inaccuracy caused by the single weight coefficient calculation logic in the fusion algorithm, the embodiment introduces a multi-level weight verification mechanism, which specifically includes the following steps: Weight calculation based on error distribution: historical error variance of device output data Based on the formula

[0052] Calculate the weighting coefficient and use the anomaly detection algorithm to remove outliers to ensure the accuracy of error distribution; Dynamic correlation analysis: Calculate the degree of correlation between data from different devices through the Pearson correlation coefficient, and use the correlation result as an auxiliary parameter for weight adjustment. For example, if the data of device A and device B are highly correlated, their joint weight will be increased in the fusion.

[0053] Verification feedback: The fusion results are verified in real time by calculating the matching degree between the fused data and the original data in terms of feature consistency. If the matching degree is lower than the set threshold (such as 95%), the re-fusion process is triggered.

[0054] In addition, this embodiment adds the dynamic adjustment capability of smart contracts to the application of blockchain technology to solve the consistency verification problem during data transmission. The specific implementation is as follows:

[0055] For the transmission of different data packets, the smart contract will generate corresponding verification rules based on the unique identifier of the data packet. The verification rules include data integrity verification (such as hash value comparison) and transmission order verification.

[0056] When verification fails, the system will automatically trigger the re-collection process, regenerate the corresponding data packet and verify it again to ensure the security and accuracy of the data; the introduction of the multi-dimensional data characteristic matrix realizes the precise optimization of the standardized protocol; the multi-level weight verification mechanism improves the accuracy and robustness of the data fusion results; the dynamic smart contract verification process ensures the transparency and consistency of data transmission.

[0057] Embodiment 3: This embodiment provides an improved personalized health monitoring method based on multi-instrument data fusion, aiming to introduce more detailed data standardization and fusion steps.

[0058] Step 1: Data feature extraction and standardization protocol selection; such as data feature extraction: During the implementation process, first extract data features from various health monitoring devices (such as blood oximeters, heart rate monitors, and sleep monitors). The feature vector contains the sampling frequency, data volume, data accuracy, and data format of the device.

[0059] Sampling frequency : Indicates the frequency of data collection per unit time by the device, in Hz.

[0060] Data volume : The total amount of data collected by the device within the specified time, in bytes.

[0061] Data accuracy : The precision of data records, usually expressed in decimal places.

[0062] Data Format : The data storage format used by the device, such as JSON, XML, etc.

[0063] Standardized protocol selection: Based on the extracted feature vectors, the system dynamically selects the appropriate standardized protocol through the trained classification model to ensure that data from different devices can be converted into a unified format. This selection is based on the historical data characteristics of the device and the optimal standardized protocol is predicted through supervised learning methods (such as support vector machine SVM).

[0064] Step 2: Data standardization and weighted fusion processing; such as data standardization: In the standardization stage, the system selects the protocol that matches the data format and accuracy requirements of each device, and performs format conversion and encoding on the data. For example, assuming that the output data format of the blood oximeter is JSON, and the heart rate monitor uses CSV format, the system will automatically convert these data into a unified standard format for subsequent processing.

[0065] Weighted fusion algorithm: The core of the weighted fusion algorithm is to assign appropriate weighting coefficients to the input data of each device. The weighting coefficients are calculated as follows:

[0066] in: For the Weighting coefficients for each device; For equipment The error variance of the data is calculated based on the error distribution of the device's historical data or by removing outliers from the data set.

[0067] The error variance reflects the reliability of the device data. High-precision devices will have smaller error variances and thus receive greater weights. The specific error variance calculation formula is:

[0068] in: Indicates the device In the The data in the sub-measurement; For equipment The mean of the data; For equipment The total number of measurement data.

[0069] Dynamic adjustment of weighting coefficients will take into account the reliability and relevance of device data, ensuring that higher-precision data is given priority during fusion.

[0070] Step 3: Anomaly detection and real-time data verification; Anomaly detection: After weighted fusion, the system will verify the fused data through anomaly detection algorithms. Abnormal data usually manifests as data points that differ greatly from historical data or data features of other devices. The anomaly detection methods used include Z-Score or DBSCAN (density-based spatial clustering algorithm) to eliminate these abnormal points.

[0071] For example, if the blood oxygen data value output by the device deviates significantly from the normal range, the system will trigger an automatic verification mechanism to eliminate outliers by looking back at historical data.

[0072] Real-time data verification: After the fusion data is completed, the system will perform feature consistency verification with the original data to ensure the feature consistency between the fused data and the original data. If the consistency is lower than the preset threshold (such as 95%), the system will trigger the re-fusion process to optimize the fusion result.

[0073] Step 4: Blockchain technology application and data transmission guarantee; Blockchain records: To ensure the security and transparency of data, all health monitoring data and their identifiers will be recorded on the blockchain. Through blockchain technology, once the data is uploaded, it cannot be tampered with, ensuring the integrity and immutability of the data.

[0074] Smart contracts: To further ensure the integrity of the data transmission process, smart contracts are used to automatically verify the integrity and consistency of the data. Smart contracts verify each data packet and trigger a re-collection or re-verification process when it is found that the data packet is not transmitted in sequence or the data integrity is compromised.

[0075] For example, if the smart contract verification finds that a data packet is lost, the system will automatically send a request to re-collect the data and upload it again.

[0076] Step 5: Dynamic adjustment and system optimization; Dynamic adjustment parameters: According to the device type and application scenario (such as real-time monitoring, long-term analysis or emergency warning), the system will dynamically adjust the parameters of data input, storage and output. Specific adjustment parameters include data update frequency, data transmission method and data storage format. For example, in an emergency warning scenario, the system will increase the data update frequency and optimize the data transmission method to ensure timely processing of information.

[0077] System optimization: The system will automatically optimize the data transmission strategy and processing flow between devices through the feedback of historical data. For example, if it is found in some application scenarios that a device with low data frequency may affect the real-time performance of data fusion, the system will dynamically increase the data update frequency of the device to ensure the timeliness of the fusion results.

[0078] Through the optimization of the above steps, the system can effectively solve problems such as inconsistent data formats of multiple devices and low collaborative efficiency, and ensure seamless integration of data between different types of devices. The introduction of weighted fusion algorithms and dynamic adjustment strategies has greatly improved the accuracy and real-time performance of data processing; for example, in actual applications, after users continuously collect data through a variety of health monitoring devices, the system can automatically process and fuse the data of each device to provide high-quality, real-time health analysis results. For example, the data of the blood oximeter and the heart rate monitor will be optimized and integrated through intelligent algorithms to avoid information conflicts and delays, providing users with more accurate and comprehensive health monitoring services.

[0079] Embodiment 4: In order to enhance the system's adaptability to data from different devices, a dynamic data feature analysis module is added to the existing data standardization engine. This module generates a multi-dimensional data feature matrix through dynamic analysis of historical data, including the device's sampling frequency, data accuracy, noise level and other influencing factors. This matrix is ​​used to optimize the selection of standardized protocols and dynamically adjust the data processing process.

[0080] Specific steps include feature extraction and matrix construction: by preprocessing the historical data of each device, key characteristic variables such as data accuracy, noise level and sampling period are extracted and stored in a multidimensional matrix. This matrix provides a basis for the subsequent selection of dynamic data standardization protocols; standardization protocol optimization: based on the data characteristics collected in real time, the system matches the historical data matrix, and predicts the most suitable standardization protocol through supervised learning algorithms (such as support vector machine SVM) to ensure that data from different devices can be smoothly converted into a unified format.

[0081] Based on the above solution, this embodiment introduces a multi-level weighting mechanism to further improve the accuracy and robustness of data fusion. The calculation of the weighting coefficient not only considers the error variance, but also introduces the correlation analysis between the device data and dynamically adjusts the calculation method of the weighting coefficient.

[0082] Specific steps: Dynamic calculation of weighted coefficients: The data of each device is processed by error variance ( ) is calculated, and the preliminary weighting coefficient is obtained.

[0083] In order to improve the accuracy of the fusion results, a dynamic adjustment strategy based on the Pearson correlation coefficient is introduced: if the data between devices are highly correlated, the system will increase their joint weight accordingly, thereby ensuring the priority of high-precision devices in the fusion.

[0084] Abnormal data detection and elimination: To avoid data distortion caused by abnormal data, Example 7 adds an abnormal detection algorithm to the data fusion step, such as an abnormal value identification method based on Z-Score. By detecting and eliminating outliers, the reliability of the fused data is further improved.

[0085] This embodiment further enhances the role of blockchain technology in data transmission, especially in data consistency verification and transmission order management. By recording the unique identifier of each data packet on the blockchain and automatically verifying it in combination with smart contracts, the integrity and consistency of the data during transmission are ensured.

[0086] Specific steps include data packet verification: a unique identifier is generated for each data packet on the blockchain, and the transmission order and integrity of the data packet are written into the blockchain. Smart contracts automatically verify based on this information to ensure the transparency and security of data transmission; transmission abnormality response mechanism: if the smart contract finds that the data packet transmission is missing or the sequence is wrong, the system will automatically trigger the re-collection or verification process to ensure the stability of the system.

[0087] This embodiment introduces a real-time data consistency verification mechanism. After data fusion is completed, the system will compare the feature consistency with the original data to ensure the accuracy of the fused data. If the feature consistency of the fused data is lower than the preset threshold (for example, 95%), the re-fusion process is automatically triggered.

[0088] By comparing the feature consistency of the fused data with the original data, such as data frequency, accuracy, etc., the feature consistency metric is calculated. If the consistency is lower than the preset threshold, the data re-fusion process is initiated; and the system dynamically optimizes the data fusion process through a real-time feedback mechanism to ensure that the output after each data fusion is highly accurate and timely.

[0089] In terms of data input, storage and output, a scenario-driven dynamic adjustment strategy can be adopted. This strategy adjusts the frequency, transmission method and storage format of data processing according to different application scenarios (such as real-time monitoring, long-term analysis, emergency warning, etc.) to further improve the performance and responsiveness of the system; that is, the system automatically identifies and adjusts the parameters of data input, storage and output according to the current application scenario. For example, in an emergency warning scenario, the system will increase the frequency of data updates to ensure timely response to emergencies; and the system will dynamically optimize the data transmission and processing strategies between devices based on historical data and feedback. If the data update frequency of a device is too low, the system will automatically increase its update frequency to ensure the overall real-time performance of the system.

[0090] Embodiment 5: In this embodiment, key data features of each device are first extracted from different types of health monitoring devices (such as blood oximeters, heart rate monitors, and sleep monitors). Data features include but are not limited to sampling frequency, data volume, data accuracy, and data format. The data feature vector of each device is as follows:

[0091] Sampling frequency : Indicates the frequency of data collection per unit time by the device, in Hz.

[0092] Data volume : The total amount of data collected by the device within the specified time, in bytes.

[0093] Data accuracy : Indicates the accuracy of the data recorded by the device, usually expressed in decimal places.

[0094] Data Format : The data storage format used by the device, such as JSON, XML, etc.

[0095] Based on these feature vectors, the system uses a pre-trained classification model (such as support vector machine SVM or decision tree) to dynamically select an adaptive standardized protocol. This protocol ensures that data from different devices can be converted into a unified standard format, thus laying the foundation for subsequent fusion processing.

[0096] Data standardization: For devices with different data formats, the system selects appropriate standardized protocols to convert and encode the data to ensure that it meets the requirements of a unified format.

[0097] Weighted fusion: After data standardization, the data from different devices are weighted and fused through deep learning algorithms. Weighting coefficient Dynamically adjusted according to the reliability and relevance of device data, the calculation formula is:

[0098] in, For equipment The error variance of the data is calculated as follows:

[0099] in, For equipment In the The data in the measurement, For equipment The mean of the data, For equipment The total number of measurement data.

[0100] Through this formula, the system can assign appropriate weights to devices based on the error range of each device's historical data. Devices with smaller error variance will receive higher weights, giving priority to high-precision data in data fusion.

[0101] Dynamic adjustment: To avoid the weighting coefficient being too static during data fusion, this embodiment introduces a dynamic adjustment mechanism. The correlation between data between devices is calculated using the Pearson correlation coefficient. If the correlation between data between two devices is high, their joint weight is increased. The calculation formula is:

[0102] in, is the weighted coefficient after combination, is the comprehensive error variance based on the correlation between devices.

[0103] At the same time, in order to ensure the accuracy of data fusion, this embodiment introduces an anomaly detection module. The fused data is detected for anomalies through Z-Score or density-based spatial clustering algorithm (DBSCAN). If the data of a device is found to be significantly deviated from the normal range or inconsistent with the data of other devices, the system will automatically remove these outliers and recalculate the weighting coefficients.

[0104] The verification process after data fusion includes comparing the matching degree of feature consistency between the fused data and the original data. If the consistency is lower than the preset threshold (such as 95%), the re-fusion process is triggered to ensure the high accuracy of the final data.

[0105] During the data transmission process, this embodiment uses blockchain technology to record device data and its identifier. Each data packet verifies the integrity and transmission order of the data through a smart contract. If the verification fails, the system will trigger a data re-collection or verification process to ensure the immutability and consistency of the data during transmission. In addition, in order to improve system performance and real-time performance, this embodiment dynamically adjusts the parameters of data input, storage, and output according to different application scenarios (such as real-time monitoring, long-term analysis, and emergency warning). These parameters include but are not limited to data update frequency, data transmission method, and data storage format. Specifically, in the emergency warning scenario, the system will increase the data update frequency and optimize the data transmission method to ensure the timely processing of information; through the introduction of a multi-dimensional data characteristic matrix, the system can optimize the standardized protocol according to the characteristics of historical data and improve the accuracy and robustness of data fusion.

[0106] Embodiment 6: This embodiment improves the weighted fusion and dynamic adjustment steps in the personalized health monitoring method of multi-instrument data fusion, aiming to improve the accuracy, real-time and system operability of data fusion. During the implementation process, the calculation method of the weighted coefficient is optimized by introducing a deep learning model and error analysis to ensure that the system can dynamically adjust the processing parameters according to the characteristics of different devices, thereby solving problems such as data inconsistency and fusion delay caused by device differences.

[0107] In actual operation, the data provided by each health monitoring device (such as blood oximeter, heart rate monitor, sleep monitor) contains multiple features (such as sampling frequency, data accuracy, data volume and data format). First, the system extracts the feature vector of each device by preprocessing the data of each device (such as denoising and normalization):

[0108] Sampling frequency( ): Indicates the data collection frequency of the device per unit time, in Hz.

[0109] Data volume( ): The amount of data recorded by the device during the collection period, in bytes.

[0110] Data accuracy( ): The precision of recorded data, usually expressed in decimal places.

[0111] Data Format( ): The data format used by the device (such as JSON, XML).

[0112] Based on the extracted feature vectors, the system dynamically selects a standardized protocol that matches the device data format through a pre-trained classification model (such as a support vector machine SVM or a decision tree). The model is trained based on historical data and automatically selects the most suitable standardized protocol based on the data characteristics of each device, ensuring that the output data of all devices can be converted into a unified standard format; at the same time, in order to improve the accuracy of data fusion, this embodiment calculates the weighting coefficient of each device through the error variance of historical data. The calculation formula of the weighting coefficient is:

[0113] in: For the Weighting coefficients for each device; For equipment The error variance of the data is calculated as follows:

[0114] in: Indicates Device No. The data of the measurement; For equipment The mean of the data; For equipment The total number of measurement data.

[0115] The weighting coefficient calculated by this formula can dynamically adjust the importance of device data in the fusion process according to the historical error distribution of the device. Devices with smaller errors (i.e., high-precision devices) will receive greater weights, giving priority to their data in the fusion process.

[0116] In deep learning model optimization, for example, in the process of calculating weighted coefficients, the system can also introduce deep learning models to automatically optimize weighted coefficients. The model can automatically adjust the weighted coefficients according to the correlation between device data (such as calculating the correlation between device data through the Pearson correlation coefficient) to ensure the best effect in the data fusion process.

[0117] In terms of anomaly detection and real-time data verification, such as anomaly detection: In order to eliminate abnormal data that may affect the fusion results, the system introduces an anomaly detection algorithm based on Z-Score or density clustering (DBSCAN). After weighted fusion, the system performs anomaly detection on the fused data, identifies and eliminates data points with large differences in characteristics from historical data or other device data; after the fusion of data is completed, the system verifies the fused data through feature consistency check to ensure the feature consistency between the fused data and the original data. If the consistency is lower than the preset threshold (such as 95%), the system will trigger the re-fusion process and use the optimized weighting coefficient to re-fuse the data.

[0118] At the same time, this embodiment dynamically adjusts the input, storage and output parameters of the data by intelligently analyzing the characteristics of the device data and combining the application scenarios (such as real-time monitoring, long-term analysis, and emergency warning). These parameters include: data update frequency: dynamically adjust the collection frequency of device data according to the device type and real-time monitoring requirements. For example, in the real-time monitoring scenario, the system will increase the frequency of data collection to ensure that the system can quickly respond to changes in user health; data transmission method: according to different application scenarios, the system automatically selects the best data transmission method (such as wireless transmission, Bluetooth, Wi-Fi, etc.) to ensure the timeliness and stability of data transmission; data storage format: according to the needs of the device and application scenarios, the system dynamically adjusts the storage format of the data to ensure the efficiency and security of data storage; scene recognition and adaptive adjustment: the system can automatically identify different application scenarios and dynamically adjust parameters according to the needs of the scene. For example, in the emergency warning scenario, the system will quickly increase the frequency of data updates and optimize the data transmission method to ensure that information can be fed back to the user in real time; data chain and smart contracts: in order to ensure the security and transparency of the data, this embodiment records the health monitoring data and its identifier on the blockchain through blockchain technology. Once the data is uploaded to the blockchain, it cannot be tampered with, ensuring the integrity and immutability of the data; Smart contract verification: The integrity and consistency of the data transmission process are verified through smart contracts to ensure that if data is lost or the sequence is wrong during the transmission process, the re-collection or verification process can be triggered in time. The smart contract will generate verification rules based on the unique identifier of the data packet to verify the integrity and transmission order of the data.

[0119] This embodiment significantly improves the accuracy and real-time performance of multi-device data fusion by introducing deep learning optimized weighted fusion and dynamic adjustment mechanisms. Especially when faced with different data characteristics of multiple devices, the system can flexibly adjust the weighting coefficients and data processing strategies to ensure the high accuracy and efficiency of the data fusion results. In addition, through anomaly detection and real-time data verification, the system can effectively eliminate abnormal data and improve the reliability of fusion results. The introduction of blockchain and smart contracts ensures the security and transparency of data transmission, providing users with more accurate and reliable health monitoring services.

[0120] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the application disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the claims.

[0121] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A personalized health monitoring method based on multi-instrument data fusion, characterized in that: The following steps are involved: S1. Receive data from a variety of health monitoring devices, including but not limited to a blood oximeter, a heart rate monitor, and a sleep monitor; S2. extracting a feature vector of each health monitoring device data, wherein the feature vector includes data frequency, data volume, data accuracy and data format; S3. Based on the extracted feature vectors, a pre-trained classification model is used to select a standardized protocol that matches the data format of each device. The specific method is as follows: Construction and training of classification models: Based on historical data from a variety of health monitoring devices, the classification model is trained using a supervised learning algorithm; Dynamic matching protocol: The system extracts the feature vector of device data in real time and inputs it into the classification model. The model outputs the optimal standardized protocol based on the input feature vector and verifies the adaptability of the protocol in real time. Specific content of the standardization protocol: including data format conversion rules, encoding requirements, and consistency calibration of data accuracy, to ensure that data from different devices can be losslessly converted into a unified standard format; S4. Convert and encode data in different formats to meet the requirements of a unified format and generate standardized data; The specific steps include: Data format conversion: Perform unified processing of data formats according to the standardized protocol selected in S3, including but not limited to converting JSON format data into XML format, or reorganizing segmented stored data into continuous structured data; Data encoding: Encode data according to the encoding rules defined by the protocol, including character set conversion to UTF-8 and standardization of data structure; Verification and generation: After completing format conversion and encoding, the data integrity and consistency are checked using the verification rules specified by the protocol. After passing the verification, standardized data that meets the requirements of the standardized protocol is generated; S5. Performing weighted fusion processing on the standardized data, including the following steps: Application of deep learning algorithms: weighted fusion of data from different devices through deep learning algorithms; Calculation and adjustment of weighting coefficients: a. The weighting coefficients are dynamically adjusted based on the reliability and relevance of the equipment data. The reliability is calculated by the error variance of the equipment data, and the error variance is calculated by the error distribution of historical data or the outlier elimination method of the equipment data; b. The correlation analysis method is used to calculate the correlation between different equipment data. The correlation analysis results are used as auxiliary parameters for weight adjustment. When the data correlation between the equipment is high, the joint weight is increased accordingly; c. The calculation formula for the final weighting coefficient is: in, For the The weighting factor of each device, is the error variance of the equipment data; Anomaly detection and data verification: perform anomaly detection on the fused data and recalculate the weighting coefficient by removing abnormal data points; Real-time verification: Verify the accuracy and completeness of data by comparing the feature consistency of fused data with the original data; S6. Ensure the security and transparency of data transmission. The device data and its identifiers are recorded on the blockchain through blockchain technology. The blockchain records ensure that the data cannot be tampered with, and the integrity and consistency of the data transmission process are verified through smart contracts. S7. Dynamically adjust the parameters of data input, storage and output according to the device type and application scenario, including data update frequency, data transmission method and data storage format.

2. The personalized health monitoring method based on multi-instrument data fusion according to claim 1 is characterized in that: The step of extracting feature vectors includes: preprocessing the data of each health monitoring device, wherein the preprocessing includes data cleaning, denoising and normalization; and extracting feature vectors from the preprocessed data using a feature extraction algorithm, wherein the feature extraction algorithm includes principal component analysis PCA or linear discriminant analysis LDA.

3. The personalized health monitoring method based on multi-instrument data fusion according to claim 1 is characterized in that: The training steps of the classification model include: collecting historical data of various health monitoring devices, wherein the historical data includes device type, data format and data quality; based on the historical data, using a supervised learning algorithm to train the classification model, wherein the supervised learning algorithm includes a support vector machine (SVM) or a decision tree.

4. The personalized health monitoring method based on multi-instrument data fusion according to claim 1 is characterized in that: The weighted fusion step also includes: performing anomaly detection on the fused data, identifying and eliminating abnormal data points; recalculating the weighting coefficients to optimize the accuracy of the fused data based on the results after eliminating the abnormal data; the historical data in step S3 includes device type, data format, sampling frequency, data accuracy and data volume; supervised learning algorithms such as support vector machines or decision trees.

5. The personalized health monitoring method based on multi-instrument data fusion according to claim 1 is characterized in that: The parameters of data input, storage and output are dynamically adjusted according to the device type and application scenario, wherein the parameters include data update frequency, data transmission method and data storage format.

6. The personalized health monitoring method based on multi-instrument data fusion according to claim 1 is characterized in that: The weighted fusion step also includes: analyzing the correlation between data from different devices by a correlation analysis method.

7. The personalized health monitoring method based on multi-instrument data fusion according to claim 1 is characterized in that: The data verification steps include: performing feature consistency check on the fused data to check whether the data frequency, data volume and data accuracy are consistent; if the consistency is lower than the preset threshold, the data re-fusion process is triggered.

8. The personalized health monitoring method based on multi-instrument data fusion according to claim 1 is characterized in that: The dynamic adjustment step includes: identifying the current application scenario, which includes real-time monitoring, long-term analysis and emergency warning, and adjusting the data update frequency and transmission method according to the identified application scenario.

9. The personalized health monitoring method based on multi-instrument data fusion according to claim 1 is characterized in that: Smart contracts are used to automatically verify the integrity and consistency of data, and trigger the data re-collection or verification process when verification fails.

Images