Intracardiac patient postoperative remote monitoring system based on biological feature recognition

Data quality and security issues are solved by using wearable devices and embedded signal processing, encryption technology, low-latency communication protocols and big data analysis in the postoperative remote monitoring system of intracardiac patients, and personalized rehabilitation plans and real-time feedback are achieved , improves patient monitoring efficiency and safety.

CN119069108AInactive Publication Date: 2024-12-03FUWAI CENTRAL CHINA CARDIOVASCULAR HOSPITAL
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202411105641.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-13
Publication Date
2024-12-03
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing postoperative remote monitoring system for intracardiac patients has shortcomings in data quality and integrity, individualized health status assessment, dynamic rehabilitation plan adjustment, real-time feedback mechanism and data security, especially the delay and security of data transmission.

Method used

Wearable devices collect biometric data in real time, pre-process and encrypt through embedded signal processing algorithms, transmit data using low-latency and high-bandwidth communication protocols, and establish distributed databases and big data processing technology on the cloud platform, combining The machine learning algorithm of the support vector machine performs real-time analysis and prediction, sets personalized thresholds and multi-level alarm mechanisms, and develops remote control modules and intelligent monitoring software.

Benefits of technology

Real-time and accurate monitoring and analysis of intracardiac patients' biometric data, timely feedback is provided, and rehabilitation plans are dynamically adjusted, which improves rehabilitation efficiency and patient safety, reduces emergency admissions, and ensures the safe storage and transmission of data.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119069108A_ABST
    Figure CN119069108A_ABST
Patent Text Reader

Abstract

The invention relates to an intracardiac patient postoperative remote monitoring system based on biological feature recognition. The system architecture comprises a front end, a rear end and a cloud platform. The front end is composed of a wearable device and a mobile application. The rear-end monitoring center is used for monitoring and analyzing data; the cloud platform is used for data storage and processing and model training; collecting biological characteristic data of the patient by using the wearable equipment; preprocessing the data through a signal processing algorithm; encrypting data by adopting an encryption standard AES, and performing data transmission by using a communication protocol; establishing a distributed database on the cloud platform; further, the biological characteristic data is processed and analyzed in real time by using a big data processing technology; a large-screen real-time monitoring system is arranged in the monitoring center, and biological characteristic data and the change trend of the patient are displayed; intelligent monitoring software is developed, the health condition of a patient is evaluated in real time through data analysis and model prediction, a threshold value of biological characteristic data is set, and when the data exceed the threshold value, an alarm is triggered.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a post-operative remote monitoring system for intracardiac patients, in particular to a post-operative remote monitoring system for intracardiac patients based on biometric identification. Background Art

[0002] The postoperative remote monitoring systems currently used for cardiac patients, including the digital postoperative remote monitoring and rehabilitation system for thoracic surgery patients described in Patent No.: 202211303650.4, have made positive contributions to improving rehabilitation efficiency, facilitating the recording and sharing of rehabilitation cases, but there are still some shortcomings and disadvantages. Although the system with Patent No. 202211303650.4 is able to collect rich physiological data through its data collection module, these systems often face problems with data quality and integrity. The collection of physiological data depends on the accuracy and stability of the sensor, and many sensors on the market are prone to data drift or loss during continuous long-term monitoring. In addition, the analysis and processing of biological signals such as electrocardiograms (ECGs) require highly accurate algorithms to ensure the correct interpretation of the data, and any algorithmic deficiencies lead to misdiagnosis or missed diagnosis.

[0003] Although the standard screening module is designed to screen and identify abnormal conditions by setting thresholds, the effectiveness of this module is limited by the balance between standardization and individualization. Each patient's physiological state and postoperative recovery are unique, and a single or overly general standard is not suitable for all individuals, which leads to insufficient assessment or excessive intervention of the health status of some patients. Although the group observation module mentioned in the patent can group patients according to different rehabilitation stages, this grouping often lacks sufficient dynamic adjustment mechanism. The patient's rehabilitation process will change due to a variety of factors, such as complications or unexpected situations during rehabilitation. Fixed grouping strategies cannot reflect these changes in a timely manner, thus affecting the real-time update and optimization of rehabilitation plans. Although the summary analysis module can provide valuable data insights and help doctors summarize experience, the accuracy of data interpretation depends on the statistical methods and analytical techniques used. Incorrect data interpretation leads to incorrect medical decisions. Especially in complex medical data analysis, how to ensure the accuracy and reliability of data analysis is still a major challenge. Although the doctor monitoring module in the patent system allows doctors to monitor the patient's rehabilitation status, it is insufficient in the real-time feedback mechanism. Doctors need to obtain critical alerts and updates in real time to make medical decisions quickly. System delays or insufficient alarm mechanisms result in untimely responses to patient emergencies. Although the use of cloud platforms promotes data sharing and collaboration and improves the efficiency of data processing, it also brings about issues of data security and privacy protection. Patients' medical data is extremely sensitive. How to ensure the secure storage and transmission of this data on the cloud platform and prevent unauthorized access and data leakage is an urgent problem that needs to be solved. Summary of the invention

[0004] The purpose of the present invention is to provide a postoperative remote monitoring system for intracardiac patients based on biometric identification, thereby solving some of the drawbacks and shortcomings pointed out in the background technology.

[0005] The present invention solves the above-mentioned technical problems by adopting the following technical solutions:

[0006] P1. The system architecture includes the front-end, back-end and cloud platform. The front-end consists of wearable devices and mobile applications, which are used to collect patients' biometric data in real time and perform preliminary processing. The back-end monitoring center is used to monitor and analyze data. The cloud platform is used for data storage, processing and model training.

[0007] P2. Use wearable devices to collect biometric data of patients including ECG, pulse, blood pressure and blood oxygen saturation; pre-process the data using signal processing algorithms, including filtering, denoising and normalization; encrypt the data using encryption standard AES, and use low-latency, high-bandwidth communication protocols including MQTT and HTTP / 2 for data transmission;

[0008] P3. Establish a distributed database on the cloud platform to store the patient's historical biometric data and use the InfluxDB time series database to manage continuous time data;

[0009] Furthermore, the biometric data is processed and analyzed in real time using technologies including Apache Spark big data processing, and the machine learning algorithm of support vector machine is used for model training to predict the health status of patients;

[0010] P4. Set up a large-screen real-time monitoring system in the monitoring center to display the patient's biometric data and its changing trends;

[0011] Furthermore, intelligent monitoring software is developed to assess the patient's health status in real time through data analysis and model prediction, and to set thresholds for biometric data. When the data exceeds the threshold, an alarm is triggered, and a multi-level alarm mechanism is adopted to ensure that corresponding measures are taken in a timely manner;

[0012] P5. Develop a remote control module to enable medical staff to remotely adjust and control the patient's treatment equipment through the system, and provide remote consultation functions; and establish a feedback mechanism to collect opinions and suggestions from patients and medical staff.

[0013] Furthermore, the biometric data processing process includes:

[0014] S1. First, use a wearable device equipped with high-precision sensors to continuously collect ECG, pulse, blood pressure and blood oxygen saturation data, and transmit them to a mobile application in real time via Bluetooth or Wi-Fi or upload them directly to a cloud platform.

[0015] S2. At the same time, embedded signal processing algorithms are used, including bandpass filters to remove high-frequency noise and baseline drift in the electrocardiogram, adaptive filtering and wavelet transform technology to remove noise in pulse, blood pressure and blood oxygen saturation data, and all data are normalized to ensure data accuracy and stability;

[0016] S3. Before data transmission, the data is encrypted using the encryption standard AES, and the encryption key is regularly updated in combination with the dynamic key management mechanism;

[0017] S4, using low-latency, high-bandwidth communication protocols MQTT and HTTP / 2 to ensure real-time and reliability of data transmission.

[0018] Furthermore, the process of collecting electrocardiogram (ECG), pulse, blood pressure and blood oxygen saturation data includes:

[0019] First, wearable devices equipped with high-precision sensors are used to continuously collect ECG, pulse, blood pressure, and blood oxygen saturation data through algorithms:

[0020]

[0021] Where s is the integration variable, t represents the acquisition time, and α and β are parameters for adjusting the filter strength and frequency;

[0022] Transmit data in real time via Bluetooth or Wi-Fi, using data stream compression function:

[0023]

[0024] Where x represents the data point, λ, k, and c are parameters for adjusting the efficiency and level of data compression. λ is the adjustment parameter that controls the shift of the function, k is the gain parameter that affects the speed and sensitivity of the exponential term, and c is the center point that represents the baseline value of the input data.

[0025] Furthermore, the embedded signal processing algorithm construction method is as follows:

[0026] First, a bandpass filter is used to remove high-frequency noise and baseline drift in the electrocardiogram (ECG) signal using the formula:

[0027]

[0028] Where t represents the time variable, f represents the frequency, α and β are parameters for adjusting the filter characteristics, and f0 is the center frequency, which ensures the accuracy of the filter for a specific frequency band;

[0029] Then, use adaptive filtering:

[0030]

[0031] Where N represents the number of delayed samples, w n (t) represents the adaptive filter weight at time t, and x(tn) is the signal value;

[0032] Next, the wavelet transform process is applied, through the transform function:

[0033]

[0034] Where t represents time, t′ is the integral variable, ψ is the wavelet function, and a is the scale factor, which removes noise from the signal in a multi-scale manner;

[0035] Finally, all processed data are normalized, and the normalization function is:

[0036]

[0037] Where x represents the original value of the processed biometric data.

[0038] Furthermore, the data encryption implementation process is as follows:

[0039] S1. Encrypt the data using the encryption standard AES before data transmission, using the encryption function based on the time series:

[0040]

[0041] Where x is the integral variable, t represents the encryption time, and α and β are encryption strength adjustment parameters;

[0042] S2. At the same time, combined with the dynamic key management mechanism, the key update is based on the nonlinear transformation function:

[0043] K n =γ·tan -1 (n)+δ·ln(1+n 2 )

[0044] Where n represents the period of key update, γ and δ are the adjustment parameters of key update;

[0045] S3, key generation and distribution through random number generation function:

[0046]

[0047] Where x1 represents the random input when generating the key, and θ is the adjustment parameter.

[0048] Furthermore, the process of adopting a low-latency, high-bandwidth communication protocol includes first using the MQTT protocol and adopting a delay function:

[0049]

[0050] Where x is the integral variable, t represents the transmission time, α and β are adjustment parameters, and the quality of service QoS level and heartbeat mechanism are used to maintain the stability of the connection;

[0051] The HTTP / 2 protocol uses multiplexing functions:

[0052]

[0053] Where s is the data stream transmitted in parallel, c represents the number of parallel connections, and γ and δ are optimization parameters;

[0054] Combine intelligent switching and load balancing strategies to switch functions:

[0055]

[0056] Where n is the network state parameter and k is the adjustment coefficient.

[0057] Furthermore, the method for establishing a distributed database includes: using a distributed database architecture Google Cloud Spanner on a cloud platform to support storage and management of large-scale data, and designing a data model through a database optimization function:

[0058]

[0059] x represents the data writing frequency, y represents the optimal query frequency, and z represents the balance parameter between data storage and query;

[0060] InfluxDB is used to store and manage time-series biometric data such as electrocardiogram (ECG), pulse, blood pressure, and blood oxygen saturation. Data compression and query optimization are implemented through time-series data management formulas:

[0061]

[0062] Where t represents the timestamp of the data.

[0063] Furthermore, the application process of the big data processing technology is as follows:

[0064] First, we use Apache Spark big data processing technology for real-time processing and analysis, and apply traffic optimization functions:

[0065]

[0066] Where x represents the real-time density of the data stream, and t represents the integral variable;

[0067] Then the support vector machine machine learning algorithm is used for model training:

[0068]

[0069] Where t represents the integral variable, v represents the dimensional vector of biometric data, and α is the adjustment parameter.

[0070] Furthermore, the process of developing the intelligent monitoring software is as follows:

[0071] S1. Set personalized biometric data thresholds based on a dynamic adjustment function:

[0072]

[0073] Where t is time, γ is the adjustment coefficient, which determines the time integral part The weight of ds in the total threshold, α is the parameter in the integral term, which directly affects sin(αs 2 ) controls the volatility and complexity of the time integral, s is the integral variable, which represents the instantaneous value of time, and δ is the adjustment coefficient of the nonlinear logarithmic term, which controls the logarithmic function part ln(1+βt 2 ) on the threshold, β is the adjustment parameter in the logarithmic term, controlling the influence of the square of time t in the logarithmic function;

[0074] S2. When the biometric data exceeds the threshold, the system triggers an alarm. The alarm trigger function is:

[0075]

[0076] Where z is the integral variable, y is the deviation that exceeds the threshold, λ is the attenuation coefficient, which controls the decay speed of the deviation over time, and μ adjusts the oscillation frequency of the deviation;

[0077] S3, adopt multi-level alarm mechanism, through response function:

[0078]

[0079] Where z is the deviation time, η is the adjustment coefficient, and N represents the number of alarm levels.

[0080] Beneficial effects of the present invention:

[0081] By continuously monitoring the biometric data of cardiac patients, including electrocardiogram (ECG), pulse, blood pressure, and blood oxygen saturation, and processing this data in real time, the medical team can be informed of the patient's physiological status in a timely manner. The advanced algorithms in the system can instantly analyze this data and provide immediate feedback on the patient's current health status, allowing the medical team to respond quickly when necessary; and customize personalized rehabilitation plans based on the patient's specific biometric data. These plans are dynamically adjusted based on real-time data to ensure that rehabilitation activities match the patient's current health status and recovery progress, thereby improving rehabilitation efficiency.

[0082] By remotely monitoring patients’ biometrics, healthcare providers are able to reduce the number of times patients need to be readmitted to the hospital for urgent issues. This not only helps reduce healthcare costs, but also improves patients’ quality of life by allowing them to recover in the comfort of their own homes. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] Figure 1 The present invention is a flow chart of the postoperative remote monitoring system for intracardiac patients based on biometric identification.

[0084] Figure 2 This is a flow chart of biometric data processing according to the present invention.

[0085] Figure 3 A module diagram of a distributed database method is established for the present invention.

[0086] Figure 4 Develop intelligent monitoring software module diagram for the present invention. DETAILED DESCRIPTION

[0087] The specific implementation modes of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0088] Combined with Figure 1 The postoperative remote monitoring system for intracardiac patients based on biometric identification includes five parts: system architecture design, biometric identification technology, data storage and processing, real-time monitoring and alarm, and remote intervention and feedback. The system architecture includes the front-end, back-end, and cloud platform. The front-end consists of wearable devices and mobile applications, which are used to collect patients' biometric data in real time and perform preliminary processing. The wearable devices have built-in high-precision sensors that can continuously monitor and record biometric data such as electrocardiogram (ECG), pulse, blood pressure, and blood oxygen saturation.

[0089] The mobile application is responsible for receiving data from wearable devices and performing preliminary processing on them, including removing noise and outliers. The back-end monitoring center is responsible for monitoring and analyzing the data transmitted from the front end, evaluating the patient's health status in real time, and issuing early warnings for abnormal situations. The cloud platform is used to centrally store and process patients' biometric data, support storage management of large-scale data, and ensure high availability and fast access to data through distributed database technology. Wearable devices are used to collect patients' electrocardiogram (ECG), pulse, blood pressure, and blood oxygen saturation biometric data, and the data is pre-processed through embedded signal processing algorithms, including bandpass filters to remove high-frequency noise in the ECG, adaptive filtering and wavelet transform technology to remove noise in pulse, blood pressure, and blood oxygen saturation data, and all data are normalized to ensure the accuracy and stability of the data under a unified scale.

[0090] Data is encrypted using the Advanced Encryption Standard (AES) before transmission, and is transmitted through low-latency, high-bandwidth communication protocols MQTT and HTTP / 2 to ensure real-time and reliability of data transmission. A distributed database is established on the cloud platform to store patients' historical biometric data, and the InfluxDB time series database is used to manage these continuous time data. Furthermore, Apache Spark big data processing technology is used to process and analyze biometric data in real time, and its distributed computing capabilities are used to process large amounts of real-time data streams. The machine learning algorithm of the support vector machine (SVM) is combined to model and train the processed data, predict the patient's health status, and generate a risk assessment report. A large-screen real-time monitoring system is set up in the monitoring center to display the patient's biometric data and its changing trends, including real-time electrocardiograms, pulse waveforms, blood pressure curves, and blood oxygen saturation change charts.

[0091] Furthermore, develop intelligent monitoring software to evaluate the health status of patients in real time through data analysis and model prediction, and dynamically set the threshold of biometric data based on the patient's historical data and real-time data. When the system detects that the data exceeds the threshold, it immediately triggers an alarm, and adopts a multi-level alarm mechanism to trigger different levels of alarms according to the severity of the threshold, from minor warnings to emergency medical intervention, to ensure that appropriate measures are taken in time to avoid dangerous conditions for patients. Develop a remote control module to enable medical staff to remotely adjust and control the patient's treatment equipment through the system, including adjusting the drug delivery rate or modifying the treatment plan, and provide a remote consultation function to support medical staff to communicate with patients in real time through video or audio. At the same time, the system establishes a feedback mechanism to regularly collect the opinions and suggestions of patients and medical staff in order to optimize the system and improve its functions, so as to ensure the effectiveness of the system and the continuous improvement of user experience.

[0092] Embodiment 1:

[0093] Combined with Figure 2 In the biometric identification-based remote monitoring system for postoperative cardiac patients, a smart bracelet equipped with a high-precision sensor continuously monitors the patient's ECG, pulse, blood pressure and blood oxygen saturation data, and transmits the data to the medical monitoring center for analysis in real time. The data is processed by the following algorithm:

[0094]

[0095] s is the integral variable, which represents the instantaneous time point in the time interval, and t represents the actual acquisition time. In order to ensure the accuracy of the data and filter the noise, the parameter α is set in the range of 0.1 to 0.5, which can effectively suppress the interference of high-frequency noise and enable the filter to process the ECG data more smoothly. The parameter β is set between 0.01 and 0.1 to adjust the sampling frequency of the data so that the system can capture subtle changes in heart rhythm. The normal range of the f(t) value is 6.0 to 9.0.

[0096] The patient's wristband collects data once a second, and processes the ECG data from t = 0 seconds to t = 10 seconds, with parameters α and β set to 0.2 and 0.05, respectively.

[0097] First, perform numerical integration calculation. Considering the time t = 10 seconds, the calculation formula is:

[0098]

[0099] Approximate calculations are performed using numerical integration methods. First, the time interval (0,10) is divided into small time steps, with a step length of 0.1 seconds. The numerical calculation of the integral can be approximated by accumulation. The specific calculation process is as follows:

[0100] Calculate the function value at each time point:

[0101]

[0102] And so on, until s=10, the function value corresponding to each time point is calculated.

[0103] The integral value is approximated by summing the function value at each time step and multiplying it by the step size:

[0104] f(10)≈0.1×(1+0.999+0.997+…+0.352)≈7.82

[0105] The calculated result of the patient's electrocardiogram characteristic value f(10) was 7.82, which was within the normal range.

[0106] Then the embodiment transmits these processed data to the cloud platform and monitoring center in real time via Bluetooth or Wi-Fi, using the data stream compression function:

[0107]

[0108] Where x represents the data point, λ, k, and c are parameters for adjusting the efficiency and level of data compression. λ is the adjustment parameter that controls the shift of the function, k is the gain parameter that affects the speed and sensitivity of the exponential term, and c is the center point that represents the baseline value of the input data.

[0109] The patient's ECG data point x is calculated to be 7.82 by the aforementioned integral formula f(t). In order to reduce the bandwidth occupied during data transmission, this data point is compressed. Assume that the parameters λ are 1.0, k is 0.5, and c is 5.0.

[0110] First, substitute the formula to calculate:

[0111]

[0112] Next, calculate the exponential term:

[0113] e -0.5×2.82 ≈e -1.41 ≈0.244

[0114] Substitute the value of the exponential term and continue the calculation:

[0115]

[0116] After compression processing, the original data point 7.82 is compressed to 6.285 and then transmitted to the cloud platform in real time via Bluetooth or Wi-Fi.

[0117] Embodiment 2:

[0118] Embodiment In a postoperative remote monitoring system for intracardiac patients based on biometric recognition, an embedded signal processing algorithm is used to remove high-frequency noise and baseline drift in the data, and a bandpass filter is used to process the signal through the formula:

[0119]

[0120] Here t represents the time variable, f represents the frequency, α and β are parameters used to adjust the filtering characteristics, and f0 is the center frequency of the filter, which ensures that the filter can accurately process the specific frequency band of the ECG signal.

[0121] α ranges from 0.01 to 0.1, which adjusts the frequency selectivity of the filter. Assume α = 0.05, and β ranges from 0.1 to 1.0, which is used to adjust the phase response characteristics of the filter. Assume β = 0.5, and set f = 10Hz to calculate the output of the bandpass filter. Substitute into the formula:

[0122]

[0123] First calculate the part in the exponential term:

[0124]

[0125] Next we calculate the cosine term:

[0126] cos(0.5×10)=cos(5)≈0.2837

[0127] Then, the output of the bandpass filter can be approximately expressed as the result of the accumulation of integral terms. If t = 1 second is assumed and the numerical integration method trapezoidal method is used for calculation, the output result after filtering is:

[0128]

[0129] After the above-mentioned bandpass filter processing, the high-frequency noise in the patient's ECG signal is significantly suppressed, and important ECG signal components are retained, especially the key frequency range of cardiac electrical activity.

[0130] The embodiment further uses adaptive filtering technology to eliminate the remaining low-frequency noise and other interference. The key formula of the adaptive filter is:

[0131]

[0132] Where N represents the number of delayed samples, w n (t) represents the adaptive filter weight at time t, and x(tn) is the value of the collected ECG signal. Through this formula, the system can dynamically adjust the weight of the filter to adapt to the changes of the ECG signal in real time to ensure the accuracy and clarity of the signal.

[0133] Assume that the patient's ECG signal x(t) collects the following data points within t = 10 seconds: x(10) = 0.9, x(9) = 0.85, x(8) = 0.88, x(7) = 0.87, and so on. These data points contain some noise due to motion or other interference. In order to deal with this noise and obtain a more accurate ECG signal, the system processes this data through an adaptive filter.

[0134] First, set the number of delayed samples N to 3, that is, the filter will consider the data of the first 3 sampling points.n (t) needs to be dynamically adjusted according to actual signal changes. Assume that at t = 10 seconds, the system sets the weight w based on past signal conditions. n (10) is: w0(10)=0.5, w1(10)=0.3, w2(10)=0.2.

[0135] According to the adaptive filtering formula, the system calculates the filtered signal value A(10) at t = 10 seconds as:

[0136] A(10)=0.5·x(10)+0.3·x(9)+0.2·x(8)

[0137] Substitute specific values:

[0138] A(10)=0.5·0.9+0.3·0.85+0.2·0.88

[0139] Calculated:

[0140] A(10)=0.45+0.255+0.176=0.881

[0141] Through the processing of the adaptive filter, the original signal x(10)=0.9 is adjusted to A(10)=0.881. This value has been smoothed to remove some noise, making the signal closer to the actual cardiac electrical activity.

[0142] The embodiment then introduces the wavelet transform method to deeply process the signal in a multi-scale manner, thereby extracting more detailed biometric information. The core formula of the wavelet transform is:

[0143]

[0144] Among them, t represents the current time, t′ is the integral variable, x(t′) is the ECG signal after pre-filtering, ψ is the selected wavelet function, and a is the scale factor, which determines the degree of expansion and contraction of the wavelet function.

[0145] Assume that at t = 10 seconds, a patient's ECG signal segment x(t') has been acquired and processed, which contains signal data from t' = 9.5 seconds to t' = 10.5 seconds. The sampling frequency is 1000 Hz, so there are 1000 data points in this 1 second time window.

[0146] The wavelet function ψ is selected as the commonly used Daubechies4 (db4) wavelet because the db4 wavelet performs well in signal singularity detection and denoising. The scale factor a determines the frequency range that the wavelet transform focuses on. Smaller a values ​​correspond to high-frequency components, and larger a values ​​correspond to low-frequency components. In order to comprehensively analyze the different frequency components in the signal, multiple scale factors are set, a=2 j , where j is an integer ranging from 1 to 5, that is, a=2,4,8,16,32.

[0147] Taking a=4 as an example, calculate the wavelet coefficient W(10,4) at t=10 seconds. First, calculate Next, the signal w(t′) is compared with the scaled and translated wavelet function Since the signal is discrete, the integration can be achieved by summation approximation:

[0148]

[0149] Among them, t′ i =9.5+(i-1)×0.001 seconds, Δt=0.001 seconds.

[0150] Assume that within this time window, part of the data of the signal x(t′) is as follows (for simplicity, take 5 data points):

[0151] x(9.5)=0.8, x(9.6)=0.85, x(9.7)=0.9, x(9.8)=0.88, x(9.9)=0.86.

[0152] The corresponding wavelet function value The calculation is as follows:

[0153] For t′ i =9.5 seconds:

[0154]

[0155] For t′ i =9.6 seconds:

[0156]

[0157] For t′ i =9.7 seconds:

[0158]

[0159] For t′ i =9.8 seconds:

[0160]

[0161] For t′ i =9.9 seconds:

[0162]

[0163] Calculate the values ​​of the db4 wavelet function at these points and get the following:

[0164] ψ(-0.125)=0.2

[0165] ψ(-0.1)=0.25

[0166] ψ(-0.075)=0.3

[0167] ψ(-0.05)=0.35

[0168] ψ(-0.025)=0.4

[0169] Substituting these data into the summation formula:

[0170]

[0171] Compute the products:

[0172] 0.8×0.2=0.16

[0173] 0.85×0.25=0.2125

[0174] 0.9×0.3=0.27

[0175] 0.88×0.35=0.308

[0176] 0.86×0.4=0.344

[0177] Sum:

[0178] 0.16+0.2125+0.27+0.308+0.344=1.2945

[0179] Multiply and Δt:

[0180]

[0181] This is only an approximate calculation for 5 data points. The actual system will calculate all 1000 data points to obtain the exact wavelet coefficients.

[0182] The final embodiment normalizes all processed data and maps signal data of different ranges to a unified scale, ranging from (0, 1). The formula of the normalization function is:

[0183]

[0184] Where x represents the original value of the ECG signal after the aforementioned processing, min(x) and max(x) represent the minimum and maximum values ​​in the signal data set, respectively.

[0185] After preliminary processing, a set of data points of the patient's ECG signal is obtained:

[0186] x1=0.88, x2=0.92, x3=0.85, x4=0.91, x5=0.87.

[0187] The minimum value min(x) of these data points in the current time window is 0.85, and the maximum value max(x) is 0.92.

[0188] Next, the system calculates for each data point x i Perform normalization:

[0189]

[0190] After normalization, the data set becomes:

[0191] N(x1)≈0.4286

[0192] N(x2)=1.0

[0193] N(x3)=0.0

[0194] N(x4)≈0.8571

[0195] N(x5)≈0.2857

[0196] These normalized values ​​will be used for subsequent data analysis, trend determination, and training of machine learning models to ensure that data from different time periods and different patients can be compared and processed on the same scale.

[0197] Embodiment 3:

[0198] In the biometric identification-based remote monitoring system for postoperative cardiac patients, the encryption standard AES is combined with the encryption function based on time series to achieve data encryption. The encryption function is in the form of:

[0199]

[0200] Among them, x is the integral variable, t represents the encryption time, and α and β are the adjustment parameters of the encryption strength.

[0201] The parameters α=0.05 and β=0.1 are set, and the encryption time t is set to 5 seconds, which means that the system will complete the data encryption process within 5 seconds.

[0202] First, the value of the encryption function C(t) is calculated by the formula. Substitute the specific parameters α=0.05 and β=0.1, and use the numerical integration method to calculate the encryption function. The upper limit of the integration is t=5 seconds.

[0203]

[0204] The integration interval (0,5) is divided into several small intervals with a step length of 0.1 seconds. This means that at each step length, the and sin(0.1x), and then accumulate the results.

[0205] At x=0, calculate:

[0206]

[0207] At x = 0.1, calculate:

[0208]

[0209] At x = 0.2, calculate:

[0210]

[0211] And so on, until x = 5.0 seconds, these results are gradually accumulated.

[0212] After numerical integration, it is calculated that at t = 5 seconds, the value of the encryption function C(5)≈0.25. This value will be used in the key generation process of the AES encryption algorithm to ensure that each encryption result is unique and complex.

[0213] The embodiment further combines a dynamic key management mechanism to ensure that the key remains secure during transmission. The dynamic key management mechanism periodically updates the key through a nonlinear transformation function, and the formula is as follows:

[0214] K n =γ·tan -1 (n)+δ·ln(1+n 2 )

[0215] Where n represents the period of key update, γ and δ are parameters for adjusting key update, and the frequency and complexity of key update can be controlled by adjusting these parameters.

[0216] Set the key to be updated every hour, that is, the unit of n is hours. Set γ = 50 and δ = 20. Each time the key is updated, the current time period n is used to calculate the new key value K. n .

[0217] Generate a new encryption key in the 10th hour of postoperative monitoring of the patient and substitute it into the formula:

[0218] K 10 =50·tan -1 (10)+20·ln(1+10 2 )

[0219] Calculate tan -1 (10):

[0220] tan -1 (10)≈1.4711 radians

[0221] Calculate ln(1+10 2 )=ln(101):

[0222] ln(101)≈4.6151

[0223] Calculate the key value K 10 :

[0224] K 10 =50×1.4711+20×4.6151=73.555+92.302=165.857

[0225] Therefore, the key value K updated at the 10th hour 10 The key value will be used to generate a new AES encryption key to protect subsequent biometric data transmission.

[0226] The last embodiment adopts a key generation and distribution mechanism based on a random number generation function, so that the key is not only generated by time and parameter changes, but also combined with a random number generation function to increase the unpredictability of the key. The key generation function is:

[0227]

[0228] Among them, x1 represents the random input when the key is generated, and θ is the adjustment parameter.

[0229] The system generates a new key to encrypt the biometric data before each data transmission. To generate this key, the system first obtains a random input value x1 from the random number generator. Assume that the random input value generated by the system in a certain operation is x1=5, and θ=0.5 is selected as the adjustment parameter.

[0230] To generate the key, the system uses the above random number generation function R(x1) for calculation. Substituting parameters x1=5 and θ=0.5:

[0231]

[0232] Calculate sin(0.5 5):

[0233] sin(2.5)≈0.5985

[0234] Calculate the denominator 1+0.5 2 :

[0235] 1+0.25=1.25

[0236] Substituting these values ​​into the integral:

[0237]

[0238] The final calculated key value R(5)≈0.762 will be used to generate the key of the AES encryption algorithm. This key value is calculated based on random input to ensure the unpredictability of the generated key.

[0239] Embodiment 4:

[0240] In the biometric recognition-based remote monitoring system for postoperative cardiac patients, the low-latency, high-bandwidth communication protocol MQTT (Message Queuing Telemetry Transport) is adopted, and the delay function D(t) is introduced to evaluate and adjust the efficiency of data transmission. The delay function is in the form of:

[0241]

[0242] Where x is the integral variable representing a small change in time, t is the data transmission time, and α and β are parameters for adjusting the delay characteristics.

[0243] Assume that the transmission time t is set to 3 seconds, select α = 0.01 and β = 0.05 to optimize the delay performance, and use the delay function D(t) to calculate the cumulative delay within the 3-second transmission time. Substituting the specific parameters α = 0.01 and β = 0.05, the calculation is as follows:

[0244]

[0245] The integration interval (0,3) is divided into small intervals with a step length of 0.1 seconds, and the delay contribution of each time point is calculated step by step.

[0246] At x=0, calculate:

[0247]

[0248] At x = 0.1, calculate:

[0249]

[0250] At x = 0.2, calculate:

[0251]

[0252] And so on, until x = 3.0 seconds, these results are gradually accumulated.

[0253] Assume that the final result is obtained through numerical integration:

[0254] D(3)≈2.987

[0255] This means that during the 3-second data transmission process, the cumulative delay is about 2.987 seconds. This delay value is fed back to the system's QoS (Quality of Service) management module. The system can further optimize the transmission delay by adjusting the α and β parameters.

[0256] The embodiment further introduces the HTTP / 2 protocol to optimize data transmission. An important feature of the HTTP / 2 protocol is multiplexing, which allows multiple data streams to be transmitted in parallel on the same connection, thereby improving transmission efficiency and bandwidth utilization. The multiplexing function is used:

[0257]

[0258] Among them, s represents the data stream transmitted in parallel, c is the number of parallel connections, and γ and δ are parameters used to optimize transmission performance.

[0259] The embodiment transmits the patient's electrocardiogram data, pulse data, blood pressure data and blood oxygen saturation data in parallel through HTTP / 2 protocol. Use c=4 parallel connections to transmit these data streams, and allocate different transmission bandwidths (unit: Mbps) to each data stream, which are:

[0260] s1=10, s2=8, s3=12, s4=7.

[0261] Set the parameters γ = 0.5 and δ = 2, use the multiplexing function to calculate the overall transmission efficiency, and substitute it into the formula for calculation:

[0262]

[0263] Calculate sin(γs i ):

[0264] sin(5)≈-0.9589

[0265] sin(4)≈-0.7568

[0266] sin(6)≈-0.2794

[0267] sin(3.5)≈0.3508

[0268] Calculate the denominator for each data stream

[0269] 10 2 +2=102

[0270] 8 2 +2=66

[0271] 12 2 +2=146

[0272] 7 2 +2=51

[0273] Calculate the itemized value for each data stream:

[0274]

[0275] Add all the sub-values ​​and divide by the number of parallel connections:

[0276]

[0277] The result M(4,c)≈-0.003975 indicates that during the current transmission process, the transmission efficiency is affected to a certain extent due to the negative sine value and the transmission speed of different data streams. The system will adjust the γ and δ parameters according to this result, or dynamically adjust the number of parallel connections c to optimize the overall transmission performance.

[0278] The last embodiment combines intelligent switching and load balancing strategies to ensure that data transmission remains stable and efficient when network conditions change, using a switching function:

[0279]

[0280] Wherein, n is a network status parameter, which indicates the quality or load of the current network, and k is an adjustment coefficient, which is used to control the response sensitivity of the system to changes in the network status.

[0281] Set the network status parameter n to 3, indicating that the current network has a certain load pressure. In order to adapt to this situation, the system needs to decide whether to switch the data transmission strategy or reallocate bandwidth resources through the switching function S(n). Assume that the adjustment coefficient k = 0.7, substitute it into the function calculation:

[0282]

[0283] Calculate ln(1+3 2 ):

[0284] ln(1+9)=ln(10)≈2.3026

[0285] Calculate the exponential term e -0.7×3 :

[0286] e -2.1 ≈0.1225

[0287] Calculate the denominator 1+e -2.1 :

[0288] 1+0.1225=1.1225

[0289] Calculate the second term

[0290]

[0291] Calculate the final switching function value:

[0292] S(3)=2.3026-2.6735≈-0.3709

[0293] The result S(3)≈-0.3709 indicates that under the current network state, the necessity of switching is high, and the system needs to adjust the transmission strategy, reduce the number of parallel transmissions, or transmit part of the data through other network paths to ensure the stability and efficiency of the overall transmission.

[0294] Embodiment 5:

[0295] View attached Figure 3 In the biometric recognition-based postoperative remote monitoring system for intracardiac patients, the Google Cloud Spanner distributed database architecture is used to support the storage and management of large-scale data, ensure the high availability and consistency of data, and use the database optimization function:

[0296]

[0297] Among them, x represents the frequency of data writing, y represents the optimal query frequency, and z is the balance parameter between data storage and query.

[0298] Set the data writing frequency x to 100 times per second, the doctor's query frequency y to 30 times per minute, which is converted to 0.5 times per second, set z = 10, and substitute it into the formula for calculation:

[0299]

[0300] calculate:

[0301] (xy) = 100 - 0.5 = 99.5

[0302] calculate:

[0303] (xy) 2 =99.5 2=9900.25

[0304] Calculate the denominator:

[0305] 2z 2 =2×10 2 =200

[0306] Calculate the exponential part:

[0307]

[0308] Calculate e -49.50125 :

[0309] e -49.50125 ≈2.91×10 -22

[0310] The final calculated optimization function value F(100,0.5,10):

[0311] F(100,0.5,10)=10×2.91×10 -22 ≈2.91×10 -21

[0312] From the results, we can see that there is a large mismatch between the current data writing frequency and the query frequency, which leads to a very low output value of the optimization function, indicating that the system needs to adjust the database configuration parameters z or strategies to better balance data storage and query.

[0313] In the next step of the embodiment, the InfluxDB database is used to store and manage these time series data, and the management formula is adopted:

[0314]

[0315] Among them, t represents the timestamp of the data, and t′ is the integral variable, which represents all data points from the starting time to the current time t.

[0316] Assume that during the patient monitoring process, the system needs to store the ECG data for one hour from 10 am to 11 am, with a sampling frequency of 100 times per second, that is, 100 data points are generated every second. In order to efficiently store this data, the formula G(t) is used to compress and manage the data for every second.

[0317] Assume that at a certain time t = 60 seconds, that is, the data storage has been going on for 1 minute, the system uses formula G(60) for integral calculation, and substitutes into the formula:

[0318]

[0319] Calculate every moment value:

[0320] When t′=0:

[0321]

[0322] When t′=1:

[0323]

[0324] When t′=4:

[0325]

[0326] When t′=9:

[0327]

[0328] Similar calculations continue until t′=60.

[0329] The values ​​of the above points are accumulated by the numerical integration method trapezoidal method, and the integral result from t′=0 to t′=60 is calculated, and G(60)≈37.84 is obtained. This value represents the result of all sampled data in 1 minute after being compressed by the time series management formula. The system stores this value in InfluxDB as a compressed time series data point.

[0330] Embodiment 6:

[0331] In the biometric recognition-based postoperative remote monitoring system for cardiac patients, Apache Spark big data processing technology is used to process and analyze large-scale real-time data streams, and a flow optimization function is introduced:

[0332]

[0333] Among them, x represents the real-time density of the data flow, and t is the integration variable.

[0334] Set the data flow density x to 50 data points / second and substitute it into the formula to calculate:

[0335]

[0336] First, calculate the denominator

[0337]

[0338] 2500+7.071=2507.071

[0339] Then calculate the integral part Since sin(t 2 ) has no analytical solution, so the trapezoidal method is used for numerical integration, and the integral result is:

[0340]

[0341] Calculate the final optimization function value:

[0342]

[0343] This result R(50)≈0.000348 shows the processing optimization effect under the current data flow density.

[0344] Next, the embodiment adopts the support vector machine SVM algorithm for model training. The core is to establish a model that can effectively distinguish between healthy and abnormal states, and adjust and train it through an optimization function. The optimization function is:

[0345]

[0346] Where v represents the dimensional vector of biometric data, t is the integration variable, and α is a parameter used to adjust the model bias and variance.

[0347] Use historical data as input for training the SVM model, where each data point can be represented as a vector v. Assume that the system selects data vector v = 2, and selects the adjustment parameter α = 0.7 to balance the bias and variance of the model.

[0348] Substitute into the formula to calculate:

[0349]

[0350] First calculate the exponential term

[0351] e -4 ≈0.0183

[0352] 0.7×0.0183≈0.01281

[0353] Then calculate the integral term Since cos(t 2 ) has no analytical solution, so the system uses the numerical integration method trapezoidal method to calculate:

[0354] approximation

[0355] 0.3×1.382≈0.4146

[0356] Finally, the optimization function value is calculated:

[0357] Q(2,0.7)=0.01281+0.4146≈0.42741

[0358] This result Q(2,0.7)≈0.42741 represents the performance of the model after adjustment by the optimization function under the current specific biometric dimension.

[0359] Embodiment 7:

[0360] View attached Figure 4 , Embodiment In a postoperative remote monitoring system for intracardiac patients based on biometric identification, an intelligent monitoring software is developed to set personalized biometric data thresholds so as to detect any abnormalities in time and issue an alarm, and a threshold setting method based on a dynamic adjustment function is adopted:

[0361]

[0362] Among them, t represents time, and γ is the adjustment coefficient, which is used to determine the time integral part The weight in the total threshold; α is the parameter in the integral term, which directly affects sin(αs 2 ) is the oscillation frequency, which controls the volatility and complexity of the time integral; s is the integral variable, which represents the instantaneous value of time; δ is the adjustment coefficient of the nonlinear logarithmic term, which controls the logarithmic function part ln(1+βt 2 ) on the threshold; β is the adjustment parameter in the logarithmic term, which controls the influence of the square of time t in the logarithmic function.

[0363] Assume that the system monitors the patient's heart rate changes and sets thresholds during the period from 8 am to 12 pm. Assume that time t starts from 0 and is measured in hours. Select parameters γ = 2, α = 0.5, δ = 1.5 and β = 0.2 as adjustment coefficients.

[0364] Assume that the time t=3 hours at this time, and calculate the threshold value V(3) at this time point.

[0365] First calculate the time integral part

[0366] Since sin(0.5s 2 ) has no analytical solution, so the system uses numerical integration methods to calculate:

[0367] Assume that the numerical integration result is:

[0368]

[0369] Then calculate the logarithmic function part ln(1+0.2×3 2 ):

[0370] ln(1+0.2×9)=ln(1+1.8)=ln(2.8)≈1.0296

[0371] Finally, substitute the parameters γ = 2, δ = 1.5 to calculate:

[0372] V(3)=2×1.327+1.5×1.0296≈2.654+1.5444≈4.1984

[0373] This result V(3)≈4.1984 represents the heart rate monitoring threshold set by the system for the patient at the current time point.

[0374] In the next step of the embodiment, the patient's biometric data, including electrocardiogram (ECG), pulse, blood pressure and blood oxygen saturation, is monitored in real time through intelligent monitoring software. When any of these data exceeds the set personalized threshold, the system will trigger an alarm mechanism using the alarm trigger function:

[0375]

[0376] Among them, z is the integral variable, which represents the change of the deviation exceeding the threshold over time; y is the deviation of the current data exceeding the threshold; λ is the attenuation coefficient, which controls the decay speed of the deviation over time; μ adjusts the oscillation frequency of the deviation, reflecting the volatility of the deviation exceeding the threshold over time.

[0377] The current heart rate deviation y = 0.8 bpm. In order to determine whether to trigger an alarm, calculate the alarm trigger function A(y). Assume that the parameters are λ = 0.1 and μ = 2, and substitute them into the formula for calculation:

[0378]

[0379] Calculate the integral term

[0380] When z = 0:

[0381]

[0382] At z = 0.2:

[0383] e -0.1×0.04 cos(0.4)≈e -0.004 ×0.9211≈0.996×0.9211≈0.9181

[0384] At z = 0.4:

[0385] e -0.1×0.16 cos(0.8)≈e -0.016 ×0.6967≈0.984×0.6967≈0.6854

[0386] At z = 0.6:

[0387] e -0.1×0.36cos(1.2)≈e -0.036 ×0.3624≈0.965×0.3624≈0.3497

[0388] At z = 0.8:

[0389] e -0.1×0.64 cos(1.6)≈e -0.064 ×-0.0292≈0.938×-0.0292≈-0.0274

[0390] The values ​​of these integration points are accumulated using the trapezoidal method of numerical integration, and y=0.8 is taken as the upper limit of integration, and it is calculated that A(0.8)≈0.293.

[0391] The result A(0.8)≈0.293 indicates the alarm intensity triggered by the current deviation. If the alarm threshold T=0.25 is set, then A(0.8)=0.293 will trigger an alarm, notifying the doctor or nursing staff to check the patient's status immediately.

[0392] The last embodiment adopts a multi-level alarm mechanism, which gradually upgrades the alarm level according to the severity and duration of the abnormality, thereby ensuring that doctors and nurses can take appropriate measures according to the situation, and introduces a response function:

[0393]

[0394] Among them, z represents the deviation time, that is, the duration of the biometric data deviating from the normal value; η is the adjustment coefficient, which controls the sensitivity of the response function to time deviation; N represents the number of alarm levels, which is usually set to different values ​​according to different monitoring needs.

[0395] Assume that during the monitoring process, the system finds that the patient's heart rate deviates from the normal range, and this deviation has lasted for z = 5 minutes. The system sets a three-level alarm mechanism N = 3, and selects an adjustment coefficient η = 0.2 to balance the response speed and the accuracy of the alarm.

[0396] Substitute into the formula to calculate:

[0397]

[0398] Calculate the exponential term exp(-0.2×5 2 ):

[0399] exp(-0.2×25)=exp(-5)≈0.00674

[0400] Calculate each level of response in the three-level alarm mechanism separately:

[0401] When n=1, the response value is

[0402] When n=2, the response value is

[0403] When n=3, the response value is

[0404] Finally, add up all the response values ​​to get the total response function value:

[0405] R(5)=0.00674+0.00337+0.00225≈0.01236

[0406] The result R(5)≈0.01236 indicates the alarm response strength under the current deviation time. Based on this result, the system can decide whether to upgrade the alarm level to a higher level or maintain the current level and continue monitoring.

[0407] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. The postoperative remote monitoring system for cardiac patients based on biometric identification includes five parts: system architecture design, biometric identification technology, data storage and processing, real-time monitoring and alarm, and remote intervention and feedback. The system features: P1. The system architecture includes the front-end, back-end and cloud platform. The front-end consists of wearable devices and mobile applications, which are used to collect patients' biometric data in real time and perform preliminary processing. The back-end monitoring center is used to monitor and analyze data. The cloud platform is used for data storage, processing and model training. P2. Use wearable devices to collect biometric data of patients including ECG, pulse, blood pressure and blood oxygen saturation; pre-process the data using signal processing algorithms, including filtering, denoising and normalization; encrypt the data using encryption standard AES, and use low-latency, high-bandwidth communication protocols including MQTT and HTTP / 2 for data transmission; P3. Establish a distributed database on the cloud platform to store the patient's historical biometric data and use the InfluxDB time series database to manage continuous time data; Furthermore, Apache Spark big data processing technology is used to process and analyze biometric data in real time, and the machine learning algorithm of support vector machine is used to train the model and predict the health status of patients; P4. Set up a large-screen real-time monitoring system in the monitoring center to display the patient's biometric data and its changing trends; Furthermore, intelligent monitoring software is developed to assess the patient's health status in real time through data analysis and model prediction, and to set thresholds for biometric data. When the data exceeds the threshold, an alarm is triggered, and a multi-level alarm mechanism is adopted to ensure that corresponding measures are taken in a timely manner; P5. Develop a remote control module to enable medical staff to remotely adjust and control the patient's treatment equipment through the system, and provide remote consultation functions; and establish a feedback mechanism to collect opinions and suggestions from patients and medical staff.

2. The biometric identification-based remote monitoring system for postoperative cardiac patients according to claim 1 is characterized in that The biometric data processing process includes: S1. First, use a wearable device equipped with high-precision sensors to continuously collect ECG, pulse, blood pressure and blood oxygen saturation data, and transmit them to a mobile application in real time via Bluetooth or Wi-Fi or upload them directly to a cloud platform. S2. At the same time, embedded signal processing algorithms are used, including bandpass filters to remove high-frequency noise and baseline drift in the electrocardiogram, adaptive filtering and wavelet transform technology to remove noise in pulse, blood pressure and blood oxygen saturation data, and all data are normalized to ensure data accuracy and stability; S3. Before data transmission, the data is encrypted using the encryption standard AES, and the encryption key is regularly updated in combination with the dynamic key management mechanism; S4, using low-latency, high-bandwidth communication protocols MQTT and HTTP / 2 to ensure real-time and reliability of data transmission.

3. The biometric recognition-based remote monitoring system for postoperative cardiac patients according to claim 2 is characterized in that The process of collecting ECG, pulse, blood pressure and blood oxygen saturation data includes: First, wearable devices equipped with high-precision sensors are used to continuously collect ECG, pulse, blood pressure, and blood oxygen saturation data through algorithms: Where s is the integration variable, t represents the acquisition time, and α and β are parameters for adjusting the filter strength and frequency; Transmit data in real time via Bluetooth or Wi-Fi, using data stream compression function: Where x represents the data point, λ, k, and c are parameters for adjusting the efficiency and level of data compression. λ is the adjustment parameter that controls the shift of the function, k is the gain parameter that affects the speed and sensitivity of the exponential term, and c is the center point that represents the baseline value of the input data.

4. The biometric identification-based postoperative remote monitoring system for cardiac patients according to claim 2 is characterized in that The embedded signal processing algorithm construction method is as follows: First, a bandpass filter is used to remove high-frequency noise and baseline drift in the electrocardiogram (ECG) signal using the formula: Where t represents the time variable, f represents the frequency, α and β are parameters for adjusting the filter characteristics, and f0 is the center frequency, which ensures the accuracy of the filter for a specific frequency band; Then, use adaptive filtering: Where N represents the number of delayed samples, w n (t) represents the adaptive filter weight at time t, and x(tn) is the signal value; Next, the wavelet transform process is applied, through the transform function: Where t represents time, t ′ is the integral variable, ψ is the wavelet function, and a is the scale factor, which removes noise from the signal in a multi-scale manner; Finally, all processed data are normalized, and the normalization function is: Where x represents the original value of the processed biometric data.

5. The biometric identification-based postoperative remote monitoring system for cardiac patients according to claim 2 is characterized in that The process of encrypting data is as follows: S1. Encrypt the data using the encryption standard AES before data transmission, using the encryption function based on the time series: Where x is the integral variable, t represents the encryption time, and α and β are encryption strength adjustment parameters; S2. At the same time, combined with the dynamic key management mechanism, the key update is based on the nonlinear transformation function: K n =γ·tan -1 (n)+δ·ln(1+n 2 ) Where n represents the period of key update, γ and δ are the adjustment parameters of key update; S3, key generation and distribution through random number generation function: Where x1 represents the random input when generating the key, and θ is the adjustment parameter.

6. The biometric identification-based postoperative remote monitoring system for cardiac patients according to claim 2 is characterized in that The process of using a low-latency, high-bandwidth communication protocol includes first using the MQTT protocol and using a delay function: Where x is the integral variable, t represents the transmission time, α and β are adjustment parameters, and the quality of service QoS level and heartbeat mechanism are used to maintain the stability of the connection; The HTTP / 2 protocol uses multiplexing functions: Where s is the data stream transmitted in parallel, c represents the number of parallel connections, and γ and δ are optimization parameters; Combine intelligent switching and load balancing strategies to switch functions: Where n is the network state parameter and k is the adjustment coefficient.

7. The biometric recognition-based remote monitoring system for postoperative cardiac patients according to claim 1 is characterized in that The method for establishing a distributed database includes: using the distributed database architecture Google Cloud Spanner on the cloud platform to support the storage and management of large-scale data, and the data model design is performed through the database optimization function: x represents the data writing frequency, y represents the optimal query frequency, and z represents the balance parameter between data storage and query; InfluxDB is used to store and manage time-series biometric data such as electrocardiogram (ECG), pulse, blood pressure, and blood oxygen saturation. Data compression and query optimization are implemented through time-series data management formulas: Where t represents the timestamp of the data.

8. The biometric identification-based postoperative remote monitoring system for cardiac patients according to claim 1 is characterized in that The application process of the big data processing technology is as follows: First, we use Apache Spark big data processing technology for real-time processing and analysis, and apply traffic optimization functions: Where x represents the real-time density of the data stream, and t represents the integral variable; Then the support vector machine machine learning algorithm is used for model training: Where t represents the integral variable, v represents the dimensional vector of biometric data, and α is the adjustment parameter.

9. The biometric identification-based postoperative remote monitoring system for cardiac patients according to claim 1 is characterized in that The process of developing the intelligent monitoring software is as follows: S1. Set personalized biometric data thresholds based on a dynamic adjustment function: V(t)=γ·∫0 t sin(αs 2 )ds+δ·ln(1+βt 2 ) Where t is time, γ is the adjustment coefficient, which determines the time integral part The weight in the total threshold, α is the parameter in the integral term, which directly affects sin(αs 2 ) controls the volatility and complexity of the time integral, s is the integral variable, which represents the instantaneous value of time, and δ is the adjustment coefficient of the nonlinear logarithmic term, which controls the logarithmic function part ln(1+βt 2 ) on the threshold, β is the adjustment parameter in the logarithmic term, controlling the influence of the square of time t in the logarithmic function; S2. When the biometric data exceeds the threshold, the system triggers an alarm. The alarm trigger function is: Where z is the integral variable, y is the deviation that exceeds the threshold, λ is the attenuation coefficient, which controls the decay speed of the deviation over time, and μ adjusts the oscillation frequency of the deviation; S3, adopt multi-level alarm mechanism, through response function: Where z is the deviation time, η is the adjustment coefficient, and N represents the number of alarm levels.

Citation Information

Patent Citations

  • Digitization-based postoperative remote monitoring and rehabilitation system for thoracic surgery patients

    CN115620890A