Heartbeat pulse monitoring system for critical care

Abstract

The invention discloses a heartbeat and pulse monitoring system for critical nursing, and relates to the technical field of heartbeat and pulse monitoring, the system comprises the following components: a non-contact sensing module which is used for capturing a heart micro-motion signal of a patient by integrating advanced radar and infrared sensors, converting the heart micro-motion signal into an electronic signal for filtering and feature extraction, and sending the electronic signal to a central processing module; according to the method, the discomfort caused by the fact that a sensor makes contact with the skin is avoided, non-contact monitoring of the heartbeat pulse of a patient is achieved, the stability and accuracy of data can still be kept when the patient slightly moves and moves, filtering, feature extraction and anomaly detection are conducted on captured heart micro-motion signals of the patient in combination with a deep learning algorithm, and the accuracy of the heart micro-motion signals is improved. The high-precision recognition module ensures that the system can still keep high accuracy even under the conditions that the patient moves and external noise interferes, medical staff can find abnormal changes of heartbeat pulses of the patient in time, and powerful support is provided for timely diagnosis and treatment.

Description

Technical Field

[0001] The present invention relates to the technical field of heartbeat and pulse monitoring, in particular to a heartbeat and pulse monitoring system for critical care. Background Art

[0002] In the modern medical field, intensive care is an important part of ensuring patient safety and promoting recovery. Especially when facing patients with complex conditions and unstable vital signs, real-time and accurate monitoring of their vital signs such as heartbeat and pulse is particularly critical. Heartbeat and pulse data are important indicators for assessing the health status of patients' cardiovascular system. They are of great significance for timely detection of abnormalities, prevention of complications and guidance of treatment. With the continuous advancement of medical technology, various new monitoring equipment and systems are constantly emerging, providing more possibilities for the monitoring of critically ill patients.

[0003] Traditional heartbeat and pulse monitoring technology is a monitoring method based on electrode patches, which has significant limitations. Contact monitoring requires patients to wear electrodes for a long time, which may not only cause skin allergies and pressure sores, but also restrict the patient's freedom of movement, affecting their comfort and recovery process. In addition, this type of monitoring equipment is easily affected by external interference, such as electromagnetic waves and patient body movements, which leads to deviations in monitoring data and affects diagnostic accuracy. It often relies on simple algorithms for data analysis and has limited ability to identify complex arrhythmias. It is difficult to meet the needs of high-precision, real-time monitoring data in intensive care environments. The lack of personalized adjustment and intelligent early warning mechanisms requires medical staff to frequently manually review data, increasing their workload and reducing response speed.

[0004] However, the use of traditional contact sensors increases the risk of cross-infection, especially in highly concentrated infection-sensitive areas such as the ICU. This problem is particularly prominent. On the other hand, data instability and analysis limitations may lead to misdiagnosis and missed diagnosis, delaying treatment, especially in the face of acute cardiac events and sudden changes in the disease. Such delays may have irreversible consequences. In addition, the lack of efficient data management and remote monitoring capabilities makes it difficult for medical institutions to achieve optimal resource allocation when responding to large-scale medical needs or emergencies, limiting the accessibility and flexibility of medical services. Therefore, the development of a new generation of heartbeat and pulse monitoring system that can not only ensure monitoring accuracy, but also overcome the shortcomings of traditional technology and incorporate intelligent management is of great significance to improving the quality of critical care and promoting patient safety. Summary of the Invention

[0005] The purpose of the present invention is to make up for the shortcomings of the existing technology and provide a heartbeat and pulse monitoring system for critical care. It can realize non-contact monitoring of the patient's heartbeat and pulse by integrating advanced radar and infrared sensors. This method avoids the discomfort that may be caused by traditional contact sensors, allowing patients to be monitored in a more natural and comfortable state. At the same time, non-contact monitoring also allows patients to maintain data stability and accuracy during slight activities and movements, thereby improving the flexibility and adaptability of monitoring.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a heartbeat and pulse monitoring system for critical care, characterized in that the system includes the following components:

[0007] Non-contact sensing module: By integrating advanced radar and infrared sensors, it captures the patient's heart micro-vibration signals and converts them into electronic signals for filtering and feature extraction. This avoids the discomfort caused by the sensor contacting the skin, enables non-contact monitoring of the patient's heartbeat and pulse, and maintains data stability and accuracy even when the patient is slightly active or moving.

[0008] High-precision recognition module: The system combines deep learning algorithms to detect anomalies in captured cardiac micro-motion signals. It can maintain high accuracy and reliability even in the presence of patient movement and external noise interference, enabling precise processing and analysis of the patient's heartbeat and pulse data.

[0009] Intelligent Sensing Analysis and Alert Module: The system's built-in intelligent algorithm can analyze heart rate variability and identify arrhythmia patterns in real time. It also applies machine learning models to automatically trigger alerts based on preset parameters. It conducts in-depth analysis of continuously monitored heartbeat and pulse data and instantly notifies medical staff via wired and wireless methods.

[0010] Personalized monitoring plan module: The system supports adaptive adjustment of monitoring strategies based on the patient's age, gender, and underlying medical conditions. Combined with electronic medical records, it comprehensively analyzes the patient's recent medication use and changes in physiological indicators to provide the most suitable monitoring plan for the patient. It also dynamically adjusts monitoring parameters and alert thresholds based on the patient's basic information and historical health data.

[0011] User interface and management system module: The user interface provides clear monitoring data display and interactive interface, making it easy for medical staff to view and manage patient data. The management system supports remote monitoring and data management functions. Medical staff can remotely access the system via the Internet to view patients' real-time monitoring data and historical records, and perform data analysis and report generation operations.

[0012] Furthermore, the signal collected by the contactless sensing module often contains a large amount of environmental noise and high-frequency interference. Low-pass filtering is used to remove the high-frequency noise in the signal and retain the low-frequency component in the signal, that is, the part close to the heartbeat frequency. The cutoff frequency of the filter is determined according to the expected frequency range and noise characteristics of the heartbeat signal, allowing low-frequency signals to pass while suppressing high-frequency noise and interference. The filter coefficient is applied to the original signal, and the IIR is recursively calculated sample by sample to obtain the filtered signal. The algorithm formula is: y[n] = αx[n] + (1-α)y[n-1], y[n] is the filtered output signal, x[n] is the input signal, and α is the coefficient that determines the cutoff frequency. Low-pass filtering can effectively purify the heartbeat signal for subsequent feature extraction and analysis.

[0013] Furthermore, the non-contact sensing module uses Fast Fourier Transform (FFT) to convert the time domain signal into a frequency domain signal, analyze the signal spectrum in the frequency domain, identify the frequency components related to the heartbeat and pulse, help analyze the frequency composition of the heartbeat signal, identify the fundamental frequency of the heartbeat and its harmonics, thereby verifying the accuracy of the heart rate calculation and detecting frequency anomalies. The algorithm formula is: X[k] is the discrete-time signal, x[n] is the kth sample point in the frequency domain, N is the signal length, and j is the imaginary unit.

[0014] Furthermore, the high-precision identification module receives filtered and transformed heartbeat pulse signal data through the convolutional neural network CNN in the deep learning algorithm, and performs resampling and normalization processing to form a one-dimensional time series data matrix. It learns and extracts features related to heart rate and arrhythmia from the original signal, and then performs classification and anomaly detection. Through a set of learnable filters, a convolution operation is performed on the input signal. Each filter slides on the data and outputs a feature map to capture local features, that is, specific shapes and frequency patterns in the heartbeat waveform. The algorithm formula is: i =f(b i +∑ j w ij *x j ), c i is the result of the convolution operation, f is the activation function, b i is the bias term, w ij is the filter weight, x j It is a part of the input signal. It reduces the data dimension by downsampling, improves the robustness and computational efficiency of the model, integrates the previously extracted features, and performs classification and regression prediction to identify normal heart rhythm and specific types of arrhythmias.

[0015] Furthermore, the specific steps of the intelligent sensing analysis and early warning module are:

[0016] (1) Feature data reception and preprocessing: Receive heartbeat pulse data from the contactless sensing module and extract key features of the heartbeat signal, including heart rate, heart rate variability, and waveform morphology;

[0017] (2) Intelligent analysis and classification: Using machine learning algorithms to perform deep learning processing on the extracted features for classification tasks, distinguishing normal heart rhythm from various types of arrhythmias;

[0018] (3) Anomaly detection: Threshold-based detection identifies heartbeat patterns that deviate from the normal range and promptly detects potential anomalies;

[0019] (4) Warning trigger: Based on the model prediction results, once abnormal heart rhythm and heart rate variability are identified, a warning signal is immediately sent to medical staff through wired and wireless communication technologies, and a warning message is displayed on the user interface.

[0020] Furthermore, the intelligent sensing analysis and warning module extracts key features from the preprocessed data, including time-domain and frequency-domain parameters of heart rate and heart rate variability. These features represent specific properties of the heartbeat signal. A labeled dataset, i.e., known normal and abnormal heartbeat samples, is used to train the SVM model. The labeled data contains a variety of typical and atypical heartbeat patterns to ensure the generalization ability of the model. The goal of training the SVM is to find a decision hyperplane that correctly separates the two types of normal and abnormal data, namely: At the same time, the distance from the hyperplane to the nearest sample in the two types of data is maximized, that is, the interval is maximized: y i (w T x i +b)≥1-ξ i , i=1,...,n, w is the weight vector, b is the bias term, C is the penalty coefficient, ξ i is the slack variable, y i is the class label, x i is the eigenvector. For nonlinear problems, the original data is mapped to a high-dimensional space through the kernel function, and then a linearly separable hyperplane is found in the high-dimensional space. The RBF formula is: The trained SVM model is applied to the new heartbeat data for classification and prediction. If the model predicts that a certain heartbeat belongs to the abnormal category, the system will trigger an early warning and notify medical staff.

[0021] Furthermore, the intelligent sensing analysis and warning module measures the prediction model error index by the mean absolute error to evaluate the deviation between the model prediction value and the actual observation value. The algorithm formula is:

[0022] n is the number of samples, y iis the actual observed value of the i-th sample, It is the model's predicted value for the sample. By minimizing MAE, optimizing the model's parameters, improving the model's prediction accuracy, adjusting the model's hyperparameters, and using different model structures to reduce MAE, the model's performance can be improved.

[0023] Furthermore, the specific steps of the personalized monitoring solution module are:

[0024] (1) Patient information collection: Collect the patient's personal information, including age, gender, and underlying disease conditions, which is automatically obtained through the electronic medical record system and entered by medical staff;

[0025] (2) Analysis of physiological indicators and medication history: The system analyzes the cardiovascular status of patients based on their recent physiological indicators, namely blood pressure, blood oxygen saturation, and medication use;

[0026] (3) Monitoring strategy customization: using linear regression technology to adaptively adjust the monitoring strategy and set personalized monitoring parameters, including the normal range of heart rate, monitoring frequency, and alarm threshold;

[0027] (4) Dynamic adjustment: During continuous monitoring, the system dynamically adjusts monitoring parameters and warning thresholds based on the patient's real-time health data feedback to better adapt to changes in the patient's condition;

[0028] (5) Comprehensive evaluation and reporting: Regularly generate reports containing personalized monitoring results, trend analysis, and recommendations for medical staff to refer to in order to formulate or adjust treatment plans.

[0029] Furthermore, the personalized monitoring solution module collects the patient's historical monitoring data, known health status, historical records of physiological indicators, previously set monitoring parameters and actual monitored heartbeat pulse data through linear regression technology, constructs useful features from the original data, and adjusts the ideal heart rate alarm threshold according to the patient's specific situation. The linear regression technology is used, that is: y = β0 + β1x1 + β2x2 + ... + ∈, β0 is the intercept, β1, β2, β3 are coefficients related to each feature, and ∈ is the error term, representing unpredictable random fluctuations. By evaluating the generalization ability of the model, for new patients, their age, gender and disease status information are input into the regression formula to calculate the personalized alarm threshold. According to the patient's current physiological indicators and characteristics, their future physiological indicator changes are predicted, and based on the prediction results, the monitoring strategy, that is, the monitoring frequency and warning threshold, is adjusted and optimized to achieve personalized monitoring.

[0030] Furthermore, the specific steps of the user interface and management system module are:

[0031] (1) Data visualization: Design a user interface to display the currently monitored patient heartbeat and pulse data, including real-time heart rate, heart rate variability, and any abnormal heart rate warning information, to ensure that medical staff can quickly understand the patient's status;

[0032] (2) Interactive operation design: The user interface allows medical staff to easily switch to view the data of different patients, adjust display parameters, such as time range, heart rate range, and select to view the historical trend of specific physiological indicators through touch screen and mouse operation;

[0033] (3) Personalized monitoring parameter settings: options are provided on the interface to allow medical staff to adjust monitoring parameters according to the patient's age, gender, underlying diseases and current condition, set personalized warning thresholds, and select monitoring modes;

[0034] (4) Remote monitoring function: Through Internet connection, medical staff can log in to the management system at different locations and remotely view the patient's real-time monitoring data, realizing cross-departmental and cross-regional collaboration and monitoring;

[0035] (5) Data analysis and report generation: Provide data analysis tools that allow medical staff to perform statistical analysis on monitoring data, compare heart rate changes in different time periods, identify potential health risk trends, and automatically generate monitoring reports, including key data summaries, trend analysis, and abnormal event records, which are convenient for disease discussion, case research, and archiving.

[0036] Compared with the existing technology, this critical care heartbeat and pulse monitoring system has the following beneficial effects:

[0037] 1. The present invention combines deep learning algorithms to filter, extract features, and detect anomalies in captured cardiac micro-motion signals. This high-precision recognition module ensures that the system can maintain a high degree of accuracy and reliability even when the patient moves or is disturbed by external noise, helping medical staff to promptly detect abnormal changes in the patient's heartbeat and pulse, providing strong support for timely diagnosis and treatment.

[0038] 2. The built-in intelligent algorithm of the present invention can analyze heart rate variability in real time, identify arrhythmia patterns, and automatically trigger early warnings based on preset parameters. At the same time, the system also supports adaptive adjustment of monitoring strategies based on the patient's age, gender, and underlying disease conditions to provide patients with the most suitable monitoring plan. This intelligent and personalized monitoring method enables the system to better meet the needs of different patients and improve the efficiency and quality of medical services.

[0039] Other advantages, objects and features of the present invention will be described in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be learned from the practice of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0041] Figure 1 This is a process operation diagram of a heartbeat and pulse monitoring system for critical care.

[0042] Figure 2 The figure is a flow chart of a heartbeat pulse monitoring system for critical care. DETAILED DESCRIPTION

[0043] The following is a clear and complete description of the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0044] A heartbeat and pulse monitoring system for critical care, the system comprising the following components:

[0045] Non-contact sensing module: By integrating advanced radar and infrared sensors, it captures the patient's heart micro-vibration signals and converts them into electronic signals for filtering and feature extraction. This avoids the discomfort caused by the sensor contacting the skin, enables non-contact monitoring of the patient's heartbeat and pulse, and maintains data stability and accuracy even when the patient is slightly active or moving.

[0046] High-precision recognition module: The system combines deep learning algorithms to detect anomalies in captured cardiac micro-motion signals. It can maintain high accuracy and reliability even in the presence of patient movement and external noise interference, enabling precise processing and analysis of the patient's heartbeat and pulse data.

[0047] Intelligent Sensing Analysis and Alert Module: The system's built-in intelligent algorithm can analyze heart rate variability and identify arrhythmia patterns in real time. It also applies machine learning models to automatically trigger alerts based on preset parameters. It conducts in-depth analysis of continuously monitored heartbeat and pulse data and instantly notifies medical staff via wired and wireless methods.

[0048] Personalized monitoring plan module: The system supports adaptive adjustment of monitoring strategies based on the patient's age, gender, and underlying medical conditions. Combined with electronic medical records, it comprehensively analyzes the patient's recent medication use and changes in physiological indicators to provide the most suitable monitoring plan for the patient. It also dynamically adjusts monitoring parameters and alert thresholds based on the patient's basic information and historical health data.

[0049] User interface and management system module: The user interface provides clear monitoring data display and interactive interface, making it easy for medical staff to view and manage patient data. The management system supports remote monitoring and data management functions. Medical staff can remotely access the system via the Internet to view patients' real-time monitoring data and historical records, and perform data analysis and report generation operations.

[0050] Example 1

[0051] This embodiment provides a heartbeat and pulse monitoring system for critical care. By integrating advanced radar and infrared sensors, the system achieves non-contact monitoring of the patient's heartbeat and pulse. This method avoids the discomfort that may be caused by traditional contact sensors, allowing patients to receive monitoring in a more natural and comfortable state. At the same time, non-contact monitoring also allows patients to maintain data stability and accuracy during slight activities and movements. Combined with deep learning algorithms, the captured patient heart micro-motion signals are filtered, feature extracted, and anomaly detected.

[0052] In its implementation, the system integrates radar and infrared sensors, which can capture the patient's weak physical movement signals without direct contact with the skin, thus avoiding the discomfort and skin damage caused by traditional electrode patches. After the sensor signal is converted into an electronic signal, it is low-pass filtered to remove high-frequency noise from the signal, retaining the low-frequency components in the signal, that is, the part close to the heartbeat frequency. The filter cutoff frequency is determined based on the expected frequency range and noise characteristics of the heartbeat signal, allowing low-frequency signals to pass while suppressing high-frequency noise and interference. The filter coefficients are applied to the original signal, and the IIR is recursively calculated sample by sample to obtain the filtered signal. The algorithm formula is: y[n] = αx[n] + (1-α)y[n-1]. The signal is converted from the time domain to the frequency domain using the fast Fourier transform (FFT). The signal spectrum is analyzed in the frequency domain to identify the frequency components related to the heartbeat and pulse, which helps to analyze the frequency composition of the heartbeat signal, identify the fundamental frequency and its harmonics, thereby verifying the accuracy of the heart rate calculation and detecting frequency anomalies. The algorithm formula is: It is easy to identify the frequency components related to the heartbeat, including the fundamental frequency and possible harmonics, to ensure the accuracy of heart rate calculation.

[0053] The high-precision recognition module uses convolutional neural networks (CNNs) in deep learning for signal processing and analysis. First, the heartbeat pulse signal processed by filtering and transformation is resampled and normalized to form a one-dimensional time series data matrix. CNN uses multiple learnable filters to perform convolution operations on this data. Each filter can capture a specific shape and frequency pattern and output a feature map, that is, a specific shape and frequency pattern in the heartbeat waveform. The algorithm formula is: i =f(b i +∑ j w ij *x j ), which helps to extract features related to heart rate and arrhythmia from the signal. Through training, the model can classify arrhythmias and detect anomalies, improving the accuracy and reliability of data processing.

[0054] The intelligent sensing analysis and warning module uses machine learning algorithms to analyze heart rate variability (HRV) and arrhythmia patterns in real time. The system automatically triggers warnings based on preset parameters. When an abnormal heart rate or arrhythmia is detected, medical staff will be notified via wired or wireless means. In addition, the module evaluates model performance through mean absolute error (MAE) and continuously optimizes the prediction model to ensure the timeliness and accuracy of warnings.

[0055] The personalized monitoring plan module can adaptively adjust the monitoring strategy based on the patient's personal information (such as age, gender, underlying disease conditions), physiological indicators and medication history. Using linear regression technology, the system can dynamically adjust monitoring parameters, such as the normal heart rate range, monitoring frequency and alarm threshold, to ensure that the monitoring strategy matches the patient's condition. The system will also generate reports containing personalized monitoring results and trend analysis to support clinical decision-making.

[0056] The user interface and management system module provides an intuitive graphical interface, allowing medical staff to easily view patients' real-time monitoring data, historical records and early warning information. The interface supports personalized settings, allowing medical staff to adjust monitoring parameters according to the patient's condition. The remote monitoring function allows medical staff to access the system via the Internet for cross-regional collaboration and data management.

[0057] Example 2

[0058] In this embodiment, a heartbeat and pulse monitoring system for critical care is particularly emphasized. The system uses a machine learning algorithm to train an SVM model. By performing deep learning processing on the extracted features, it is used for classification tasks to distinguish normal heart rhythms from various types of arrhythmias. Once abnormal heart rhythms and heart rate variability are identified, the system will immediately send an early warning signal to medical staff through wired and wireless communication technologies, and display a warning message on the user interface.

[0059] First, the heartbeat pulse data from the contactless sensing module is received and intelligently analyzed and anomaly detected. The key features of the heartbeat signal are extracted, including the time and frequency domain parameters of the heart rate, heart rate variability, and waveform morphology. These features represent the specific properties of the heartbeat signal. A labeled data set, i.e., known normal and abnormal heartbeat samples, is used to train the SVM model. The labeled data contains a variety of typical and atypical heartbeat patterns to ensure the generalization ability of the model. The goal of training the SVM is to find a decision hyperplane that correctly separates the two types of normal and abnormal data, namely: At the same time, the distance from the hyperplane to the nearest sample in the two types of data is maximized, that is, the interval is maximized: y i (w T x i +b)≥1-ξ i , i=1,...,n, w is the weight vector, b is the bias term, C is the penalty coefficient, ξ i is the slack variable, y i is the class label, x i is the eigenvector. For nonlinear problems, the original data is mapped to a high-dimensional space through the kernel function, and then a linearly separable hyperplane is found in the high-dimensional space. The RBF formula is: The trained SVM model is applied to the new heartbeat data for classification and prediction. If the model predicts that a certain heartbeat belongs to the abnormal category, the system will trigger an early warning and notify medical staff.

[0060] Then, the forecast model error index is measured by the mean absolute error to evaluate the deviation between the model prediction value and the actual observation value. The algorithm formula is: n is the number of samples, y i is the actual observed value of the i-th sample, It is the model's predicted value for the sample. By minimizing MAE, optimizing the model's parameters, improving the model's prediction accuracy, adjusting the model's hyperparameters and using different model structures to reduce MAE, the model's performance can be improved. The monitoring strategy is adaptively adjusted based on the patient's age, gender and underlying disease conditions. By combining electronic medical records and the patient's recent drug use and physiological indicator changes, the most suitable monitoring plan is provided for the patient. At the same time, the system can also dynamically adjust monitoring parameters and warning thresholds based on the patient's basic information and historical health data to ensure the accuracy and effectiveness of monitoring.

[0061] Finally, the user interface provides clear monitoring data display and interactive interface, making it convenient for medical staff to view and manage patient data. The management system supports remote monitoring and data management functions. Medical staff can remotely access the system through the Internet to view the patient's real-time monitoring data and historical records, and perform data analysis and report generation operations.

[0062] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A heartbeat and pulse monitoring system for critical care, characterized in that: The system includes the following components: Non-contact sensing module: By integrating advanced radar and infrared sensors, it captures the patient's heart micro-vibration signals and converts them into electronic signals for filtering and feature extraction. This avoids the discomfort caused by the sensor contacting the skin, enables non-contact monitoring of the patient's heartbeat and pulse, and maintains data stability and accuracy even when the patient is slightly active or moving. High-precision recognition module: The system combines deep learning algorithms to detect anomalies in captured cardiac micro-motion signals. It can maintain high accuracy and reliability even in the presence of patient movement and external noise interference, enabling precise processing and analysis of the patient's heartbeat and pulse data. Intelligent Sensing Analysis and Alert Module: The system's built-in intelligent algorithm can analyze heart rate variability and identify arrhythmia patterns in real time. It also applies machine learning models to automatically trigger alerts based on preset parameters. It conducts in-depth analysis of continuously monitored heartbeat and pulse data and instantly notifies medical staff via wired and wireless methods. Personalized monitoring plan module: The system supports adaptive adjustment of monitoring strategies based on the patient's age, gender, and underlying medical conditions. Combined with electronic medical records, it comprehensively analyzes the patient's recent medication use and changes in physiological indicators to provide the most suitable monitoring plan for the patient. It also dynamically adjusts monitoring parameters and alert thresholds based on the patient's basic information and historical health data. User interface and management system module: The user interface provides clear monitoring data display and interactive interface, making it easy for medical staff to view and manage patient data. The management system supports remote monitoring and data management functions. Medical staff can remotely access the system via the Internet to view patients' real-time monitoring data and historical records, and perform data analysis and report generation operations.

2. A heartbeat and pulse monitoring system for critical care according to claim 1, characterized in that: The signal collected by the non-contact sensing module often contains a large amount of environmental noise and high-frequency interference. Low-pass filtering is used to remove the high-frequency noise in the signal and retain the low-frequency component in the signal, that is, the part close to the heartbeat frequency. The cutoff frequency of the filter is determined according to the expected frequency range and noise characteristics of the heartbeat signal, allowing low-frequency signals to pass while suppressing high-frequency noise and interference. The filter coefficient is applied to the original signal, and the IIR is recursively calculated sample by sample to obtain the filtered signal. The algorithm formula is as follows: y[n]=αx[n]+(1-α)y[n-1], where y[n] is the filtered output signal, x[n] is the input signal, and α is the coefficient that determines the cutoff frequency. Low-pass filtering can effectively purify the heartbeat signal for subsequent feature extraction and analysis.

3. A heartbeat and pulse monitoring system for critical care according to claim 1, characterized in that: The non-contact sensing module uses Fast Fourier Transform (FFT) to convert time domain signals into frequency domain signals, analyze the signal spectrum in the frequency domain, identify the frequency components related to heartbeat and pulse, help analyze the frequency composition of the heartbeat signal, identify the fundamental frequency of the heartbeat and its harmonics, thereby verifying the accuracy of heart rate calculation and detecting frequency anomalies. The algorithm formula is: X[k] is the discrete-time signal, x[n] is the kth sample point in the frequency domain, N is the signal length, and j is the imaginary unit.

4. A heartbeat and pulse monitoring system for critical care according to claim 3, characterized in that: The high-precision identification module receives filtered and transformed heartbeat pulse signal data through the convolutional neural network (CNN) in the deep learning algorithm, and performs resampling and normalization processing to form a one-dimensional time series data matrix. It learns and extracts features related to heart rate and arrhythmia from the original signal, and then performs classification and anomaly detection. Through a set of learnable filters, a convolution operation is performed on the input signal. Each filter slides on the data and outputs a feature map to capture local features, that is, specific shapes and frequency patterns in the heartbeat waveform. The algorithm formula is: i =f(b i +∑ j w ij *x j ), c i is the result of the convolution operation, f is the activation function, b i is the bias term, w ij is the filter weight, x j It is a part of the input signal. It reduces the data dimension by downsampling, improves the robustness and computational efficiency of the model, integrates the previously extracted features, and performs classification and regression prediction to identify normal heart rhythm and specific types of arrhythmias.

5. A heartbeat and pulse monitoring system for critical care according to claim 1, characterized in that: The specific steps of the intelligent sensing analysis and early warning module are: (1) Feature data reception and preprocessing: Receive heartbeat pulse data from the contactless sensing module and extract key features of the heartbeat signal, including heart rate, heart rate variability, and waveform morphology; (2) Intelligent analysis and classification: Using machine learning algorithms to perform deep learning processing on the extracted features for classification tasks, distinguishing normal heart rhythm from various types of arrhythmias; (3) Anomaly detection: Threshold-based detection identifies heartbeat patterns that deviate from the normal range and promptly detects potential anomalies; (4) Warning trigger: Based on the model prediction results, once abnormal heart rhythm and heart rate variability are identified, a warning signal is immediately sent to medical staff through wired and wireless communication technologies, and a warning message is displayed on the user interface.

6. A heartbeat and pulse monitoring system for critical care according to claim 5, characterized in that: The intelligent sensing analysis and warning module extracts key features from the preprocessed data, including time-domain and frequency-domain parameters of heart rate and heart rate variability. These features represent specific properties of the heartbeat signal. A labeled dataset, i.e., known normal and abnormal heartbeat samples, is used to train the SVM model. The labeled data contains a variety of typical and atypical heartbeat patterns to ensure the model's generalization ability. The goal of training the SVM is to find a decision hyperplane that correctly separates the two types of normal and abnormal data, namely: At the same time, the distance from the hyperplane to the nearest sample in the two types of data is maximized, that is, the interval is maximized: y i (w T x i +b)≥1-ξ i , i = 1, ..., n, w is the weight vector, b is the bias term, C is the penalty coefficient, ξ i is the slack variable, y i is the class label, x i is the eigenvector. For nonlinear problems, the original data is mapped to a high-dimensional space through the kernel function, and then a linearly separable hyperplane is found in the high-dimensional space. The RBF formula is: The trained SVM model is applied to the new heartbeat data for classification and prediction. If the model predicts that a certain heartbeat belongs to the abnormal category, the system will trigger an early warning and notify medical staff.

7. A heartbeat and pulse monitoring system for critical care according to claim 6, characterized in that: The intelligent sensing analysis and early warning module measures the prediction model error index by the mean absolute error to evaluate the deviation between the model prediction value and the actual observation value. The algorithm formula is: n is the number of samples, y i is the actual observed value of the i-th sample, It is the model's predicted value for the sample. By minimizing MAE, optimizing the model's parameters, improving the model's prediction accuracy, adjusting the model's hyperparameters, and using different model structures to reduce MAE, the model's performance can be improved.

8. A heartbeat and pulse monitoring system for critical care according to claim 1, characterized in that: The specific steps of the personalized monitoring solution module are: (1) Patient information collection: Collect the patient's personal information, including age, gender, and underlying disease conditions, which is automatically obtained through the electronic medical record system and entered by medical staff; (2) Analysis of physiological indicators and medication history: The system analyzes the cardiovascular status of patients based on their recent physiological indicators, namely blood pressure, blood oxygen saturation, and medication use; (3) Monitoring strategy customization: using linear regression technology to adaptively adjust the monitoring strategy and set personalized monitoring parameters, including the normal range of heart rate, monitoring frequency, and alarm threshold; (4) Dynamic adjustment: During continuous monitoring, the system dynamically adjusts monitoring parameters and warning thresholds based on the patient's real-time health data feedback to better adapt to changes in the patient's condition; (5) Comprehensive evaluation and reporting: Regularly generate reports containing personalized monitoring results, trend analysis, and recommendations for medical staff to refer to in order to formulate or adjust treatment plans.

9. A heartbeat and pulse monitoring system for critical care according to claim 1, characterized in that: The personalized monitoring solution module collects the patient's historical monitoring data, known health status, historical records of physiological indicators, previously set monitoring parameters and actual monitored heartbeat pulse data through linear regression technology, constructs useful features from the original data, and adjusts the ideal heart rate alarm threshold according to the patient's specific situation. The linear regression technology is used, that is: y = β0 + β1x1 + β2x2 + ... + ∈, β0 is the intercept, β1, β2, β3 are coefficients related to each feature, and ∈ is the error term, representing unpredictable random fluctuations. By evaluating the generalization ability of the model, for new patients, their age, gender and disease status information are input into the regression formula to calculate the personalized alarm threshold. According to the patient's current physiological indicators and characteristics, their future physiological indicator changes are predicted. Based on the prediction results, the monitoring strategy, that is, the monitoring frequency and warning threshold, is adjusted and optimized to achieve personalized monitoring.

10. A heartbeat and pulse monitoring system for critical care according to claim 1, characterized in that: The specific steps of the user interface and management system module are: (1) Data visualization: Design a user interface to display the currently monitored patient heartbeat and pulse data, including real-time heart rate, heart rate variability, and any abnormal heart rate warning information, to ensure that medical staff can quickly understand the patient's status; (2) Interactive operation design: The user interface allows medical staff to easily switch to view the data of different patients, adjust display parameters, such as time range, heart rate range, and select to view the historical trend of specific physiological indicators through touch screen and mouse operation; (3) Personalized monitoring parameter settings: options are provided on the interface to allow medical staff to adjust monitoring parameters according to the patient's age, gender, underlying diseases and current condition, set personalized warning thresholds, and select monitoring modes; (4) Remote monitoring function: Through Internet connection, medical staff can log in to the management system at different locations and remotely view the patient's real-time monitoring data, realizing cross-departmental and cross-regional collaboration and monitoring; (5) Data analysis and report generation: Provide data analysis tools that allow medical staff to perform statistical analysis on monitoring data, compare heart rate changes in different time periods, identify potential health risk trends, and automatically generate monitoring reports, including key data summaries, trend analysis, and abnormal event records, which are convenient for disease discussion, case research, and archiving.

Images