Dynamic monitoring system for senile cardiac patients

Abstract

The dynamic monitoring system for the senile cardiac patient comprises a multi-mode sensor module, a data processing module, an analysis module and an alarm module. The multi-modal sensor module collects parameters to generate a multi-modal physiological signal; the data processing module generates a preprocessing signal; the analysis module performs real-time analysis on the pre-processed signal based on a preset elderly electrocardiogram characteristic model to generate a graded alarm instruction; the alarm module triggers corresponding local reminding, remote notification or first-aid linkage response according to the graded alarm instruction, wherein the first-aid linkage response automatically sends the patient position and key physiological data to a specified first-aid center through the first-aid communication module; wherein the self-adaptive adjustment of the multi-mode sensor module comprises the step of dynamically optimizing signal gain according to a skin contact state. According to the dynamic monitoring system for the senile cardiac patients, the problems that dynamic monitoring of the senile cardiac patients is poor in real-time performance, high in false alarm rate and delayed in first-aid response can be solved.

Description

Technical Field

[0001] The present invention relates to the field of medical electronics technology and the field of wearable health monitoring devices, and particularly to a dynamic monitoring system for elderly cardiac patients. Background Art

[0002] For elderly patients with heart diseases, due to the decline of physiological functions and complex comorbidities, traditional monitoring methods have significant limitations. Although Holter devices can record 24-hour electrocardiogram data, there are problems such as lag in data transmission back and inability to give real-time warnings. Moreover, their relatively large size causes discomfort when worn, affecting the compliance of elderly patients.

[0003] Although existing wearable devices (such as smart watches) can monitor some physiological parameters, a single sensor is easily interfered by movement and has a high false alarm rate. Especially in the elderly population, due to characteristics such as loose skin and frequent body movements, the signal quality deteriorates further. Remote monitoring systems rely on wireless transmission and manual operation, which require a relatively high technical usage ability for the elderly. Moreover, most systems only achieve data storage rather than intelligent analysis and cannot identify subtle abnormalities such as ST segment deviation. Although artificial intelligence-assisted diagnosis technology has been applied in arrhythmia classification, the general model is not optimized for the electrocardiogram characteristics of elderly patients (such as conduction block and atrial fibrillation waveform variation), resulting in insufficient specificity.

[0004] In the first aid process, existing systems lack an automatic docking mechanism with medical institutions. After an abnormal event occurs, it is necessary to manually contact the emergency center, delaying the golden time for rescue. In addition, the contradiction between the power consumption and comfort of multi-parameter monitoring devices is prominent. Traditional devices often use high-frequency sampling to ensure accuracy, which exacerbates battery consumption, while reducing the sampling rate may miss sudden abnormalities. In view of the above problems, there is an urgent need to develop a monitoring system dedicated to elderly cardiac patients, which can achieve an integrated closed-loop management of high-precision monitoring and first aid linkage through sensor fusion, adaptive algorithms, and intelligent response mechanisms while ensuring wearing comfort. Summary of the Invention

[0005] In view of the above disadvantages of the prior art, the purpose of the present invention is to provide a dynamic monitoring system for elderly cardiac patients, which is used to solve the problems of poor real-time performance, high false alarm rate, and delayed first aid response in the dynamic monitoring of elderly patients with heart diseases. The present invention fuses electrocardiogram, blood oxygen, and respiratory signals through multi-modal sensors, dynamically optimizes the signal gain in combination with the skin contact state, and eliminates motion artifacts; uses a lightweight model that combines edge computing and the cloud to parse physiological data in real time, identify arrhythmia events, and classify risk levels; constructs a hierarchical alarm mechanism, triggers local reminders, remote notifications, or first aid linkages according to the risk levels. When first aid is responded, the patient's medical records are automatically retrieved and a redundant communication link is enabled to transmit key data, realizing a full-process closed-loop management from abnormal detection to first aid decision-making, and improving the monitoring accuracy and treatment timeliness.

[0006] The present invention provides a dynamic monitoring system for elderly cardiac patients, comprising: A multimodal sensor module that collects electrocardiogram signals, blood oxygen signals, and respiratory signals of elderly patients in real time. After adaptively adjusting the signal acquisition parameters to reduce environmental interference, it generates multimodal physiological signals; A data processing module that performs fusion processing on the multimodal physiological signals, eliminates motion artifacts, and extracts characteristic waveforms to generate preprocessed signals; An analysis module that performs real-time analysis on the preprocessed signals based on a preset elderly electrocardiogram feature model, identifies abnormal electrocardiogram events and their risk levels, and generates hierarchical alarm instructions; An alarm module that triggers corresponding local reminders, remote notifications, or first-aid linkage responses according to the hierarchical alarm instructions. Among them, the first-aid linkage response automatically sends the patient's location and key physiological data to a designated first-aid center through a first-aid communication module, and synchronously retrieves the patient's historical medical records to assist in first-aid decision-making; Wherein, the adaptive adjustment of the multimodal sensor module includes dynamically optimizing the signal gain according to the skin contact state, and the risk level determination of the analysis module combines the patient's individual threshold and the deviation degree of real-time physiological parameters.

[0007] In an embodiment of the present invention, the adaptive adjustment of the multimodal sensor module further includes dynamically adjusting the signal acquisition frequency by monitoring the impedance change at the skin contact interface of the sensor. When the detected contact impedance exceeds the preset range, it automatically switches to a backup electrode array and compensates for signal attenuation. At the same time, it performs closed-loop adjustment on the emission power of the optoelectronic sensor according to the environmental temperature fluctuation to maintain the signal-to-noise ratio of the blood oxygen signal within an analyzable range. The compensation parameters generated during the signal gain optimization process are fed back to the data processing module through the data bus for correcting the baseline drift of the multimodal physiological signals.

[0008] In an embodiment of the present invention, the fusion processing of the data processing module includes a dynamic filtering strategy based on motion state recognition. It constructs a motion interference model through the body motion signals collected by an inertial sensor, uses a time-frequency domain joint analysis method to separate the myoelectric noise components in the electrocardiogram signal, and incorporates the phase relationship between the respiratory signal and the blood oxygen signal into the motion artifact correction algorithm to generate a multimodal waveform data set with time alignment characteristics. This data set is stored in a cache queue bound to the patient identity identifier after being standardized for the analysis module to call according to the priority order.

[0009] In an embodiment of the present invention, the elderly electrocardiogram feature model is obtained by training with machine learning methods. Its input layer receives the preprocessed signal segments segmented by time windows, the middle layer uses an attention mechanism to weightedly fuse multi-modal feature vectors, and the output layer generates a probability distribution according to the electrophysiological characteristics of common arrhythmia types in elderly patients. The establishment of the individualized threshold is based on constructing a dynamic baseline from the patient's historical monitoring data. When the deviation degree of the real-time physiological parameters exceeds the baseline fluctuation range, the online fine-tuning of the model parameters is automatically triggered to reduce the misjudgment rate.

[0010] In an embodiment of the present invention, the generation of the hierarchical alarm instruction includes a three-level decision-making logic. The first level starts a local vibration reminder for occasional premature beat events and records the event characteristics. The second level superimposes an audible and visual alarm on persistent tachycardia events and sends an encrypted warning message to a preset guardian terminal. The third level immediately activates the full-duplex transmission channel of the first aid communication module when a life-threatening arrhythmia pattern is detected. The selection criteria for key physiological data include abnormal waveform segments, the change trend of hemodynamic parameters, and the current drug use status.

[0011] In an embodiment of the present invention, the first aid communication module is configured with a redundant communication link switching mechanism. When the transmission quality of the primary wireless channel drops below the threshold, it automatically enables the standby frequency band to establish a parallel data transmission path, and uses a differential encryption algorithm to process the patient's location information during the transmission process. The retrieval process of the historical medical record is realized through a medical data middleware to perform protocol conversion of heterogeneous systems, ensuring that the electronic medical record received by the first aid center includes the conclusions of the last three outpatient diagnoses and the list of allergic drugs.

[0012] In an embodiment of the present invention, the analysis module is built-in with a temporal correlation analysis unit for detecting the rhythm coupling relationship between the electrocardiogram signal and the respiratory signal. When it is found that the characteristics of respiratory sinus arrhythmia disappear and are accompanied by a stepwise decrease in blood oxygen saturation, the weight of the risk level determination is automatically increased. The update period of the individualized threshold is adaptively adjusted according to the patient's activity intensity, and the baseline calculation window is extended during the night rest stage to improve the stability of the criterion.

[0013] In an embodiment of the present invention, a dynamic priority scheduler is provided between the data processing module and the analysis module. When cross-parameter abnormal correlation features appear in the multi-modal physiological signals, the processing priority of the corresponding data stream is temporarily increased, and additional computing resources are allocated for multi-dimensional feature cross-verification. The generation process of the preprocessed signal includes timestamp calibration of the feature waveform to ensure that the data collected by different sensors are strictly synchronized in the time domain.

[0014] In one embodiment of the present invention, the alarm module is integrated with an intelligent learning unit, which establishes a personalized feedback model by recording the patient's response behaviors to alarms at various levels. When it is detected that a high-risk alarm is ignored multiple times, the alarm intensity is automatically upgraded and the backup contact call chain is activated. The local reminder device adopts a gradually increasing stimulation strategy, with low-frequency vibration for reminder in the initial stage, and switches to a combined sound and light mode if no response is continuously received. At the same time, an attention signal is sent to the data processing module to trigger higher-precision physiological parameter sampling.

[0015] In one embodiment of the present invention, a data storage module is provided between the data processing module and the analysis module. The data storage module generates a dynamic storage strategy based on the characteristic waveform energy distribution and abnormal event markers of the preprocessed signal, stores the stable baseline segments in the multimodal physiological signals using compression coding, while retains the original sampling rate and attaches a timestamp index to the signal segments containing characteristic waveform mutations. The data storage module dynamically adjusts the storage period based on the risk level feedback by the analysis module. The data related to high-risk events is permanently stored and a logical association with the first aid linkage response is established. When the alarm module triggers the first aid communication module, multimodal data waveforms for a set time period before the incident are automatically extracted from the data storage module, and an auxiliary diagnosis report containing temporal evolution characteristics is generated and synchronously transmitted to the first aid center. The execution process of the storage strategy is realized through a circular buffer to achieve seamless connection between real-time data flow and historical data, ensuring that the complete physiological parameter change trajectory can be reconstructed when an abnormal event is traced back. A dynamic monitoring system for elderly cardiac patients provided by the present invention fuses electrocardiogram, blood oxygen and respiratory signals through multimodal sensors, dynamically optimizes the signal gain in combination with the skin contact state, and eliminates motion artifacts; uses a lightweight model with edge computing and cloud collaboration to real-time analyze physiological data, identify arrhythmia events and classify risk levels; constructs a hierarchical alarm mechanism, triggers local reminders, remote notifications or first aid linkages according to the risk level, automatically retrieves the patient's medical record and enables a redundant communication link to transmit key data during first aid response, realizing a full-process closed-loop management from abnormal detection to first aid decision-making, and improving the monitoring accuracy and treatment timeliness. Description of the Drawings

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings without creative efforts based on these drawings.

[0017] Figure 1 It is a system architecture diagram of a dynamic monitoring system for elderly cardiac patients. Detailed Embodiments

[0018] The following specific examples illustrate the embodiments of the present invention. Those skilled in the art can easily understand the other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0019] It should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner. Therefore, only the components related to the present invention are shown in the diagrams, rather than being drawn according to the number, shape, and size of the components in actual implementation. The type, quantity, and ratio of each component in its actual implementation can be arbitrarily changed, and the component layout type may also be more complex.

[0020] In the following description, a large number of details are explored to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0021] Please refer to Figure 1 , which shows a dynamic monitoring system for elderly heart patients of the present invention, including a multi-modal sensor module, a data processing module, an analysis module, and an alarm module. The multi-modal sensor module collects the electrocardiogram signal, blood oxygen signal, and respiratory signal of elderly patients in real time. After adaptively adjusting the signal acquisition parameters to reduce environmental interference, it generates multi-modal physiological signals. The data processing module performs fusion processing on the multi-modal physiological signals, eliminates motion artifacts, and extracts characteristic waveforms to generate preprocessed signals. The analysis module performs real-time analysis on the preprocessed signals based on a preset elderly electrocardiogram feature model, identifies abnormal electrocardiogram events and their risk levels, and generates graded alarm instructions. The alarm module triggers corresponding local reminders, remote notifications, or first-aid linkage responses according to the graded alarm instructions. Among them, the first-aid linkage response automatically sends the patient's location and key physiological data to the designated first-aid center through the first-aid communication module, and synchronously retrieves the patient's historical medical records to assist in first-aid decision-making. Among them, the adaptive adjustment of the multi-modal sensor module includes dynamically optimizing the signal gain according to the skin contact state, and the risk level determination of the analysis module combines the patient's individual threshold and the deviation degree of real-time physiological parameters.

[0022] Figure 1As shown, the system includes four core modules: a multimodal sensor module, a data processing module, an analysis module, and an alarm module. Each module is tightly coupled through signal flow and functional logic to form a complete monitoring closed-loop. The multimodal sensor module, as the front-end for data acquisition, integrates flexible electrodes, optoelectronic sensors, and impedance sensors, corresponding to the synchronous capture of electrocardiogram (ECG) signals, blood oxygen signals, and respiratory signals respectively. The core innovation of this module lies in the adaptive adjustment mechanism, which can dynamically optimize the signal gain according to the skin contact state of elderly patients. For example, when detecting changes in contact impedance caused by skin relaxation or sweat, it automatically adjusts the input impedance matching parameters of the electrodes to reduce signal distortion. For the acquisition of blood oxygen signals, the module is equipped with a closed-loop regulation system that compensates the emission power of the optoelectronic sensor by real-time monitoring of environmental temperature fluctuations to ensure signal stability under different wearing conditions. The respiratory signal is obtained using the thoracic impedance method, combined with a motion state recognition algorithm to filter out noise interference caused by limb movement. All sensor data are preliminarily processed to generate multimodal physiological signals, which are sent to the data processing module through a low-power wireless transmission protocol. The data processing module undertakes the functions of signal fusion and preprocessing, and its core technology lies in motion artifact elimination and characteristic waveform extraction. After receiving the multimodal physiological signals, this module first corrects the baseline drift of the ECG signals and uses a filtering algorithm based on wavelet transform to separate high-frequency electromyogram noise and low-frequency respiratory interference. Regarding the time-domain correlation between blood oxygen signals and respiratory signals, the module aligns the sampling time axes of the two through phase synchronization technology to eliminate data misalignment caused by sensor response delay. The elimination of motion artifacts adopts a dynamic filtering strategy. By using built-in inertial sensors (such as accelerometers) to capture the patient's body movement data in real time, a three-dimensional motion model is constructed and its coupling relationship with physiological signals is calculated. Finally, the motion interference components are inversely eliminated through an adaptive filter. Key characteristic waveforms are extracted from the processed signals, such as the QRS complex morphology and ST segment slope of the ECG signals, and the pulsation waveform period of the blood oxygen signals. These features are encoded after standardization to generate preprocessed signals, which are transmitted to the analysis module for in-depth analysis.

[0023] Furthermore, the core of the analysis module lies in a pre-set elderly electrocardiogram feature model, which is trained and optimized specifically for the electrocardiogram characteristics of elderly patients through machine learning methods. The input layer of the model receives the pre-processed signal segments segmented by time windows, and uses a convolutional neural network to extract the spatio-temporal features of multi-modal signals, such as the rhythm coupling relationship between electrocardiogram signals and respiratory signals, and the correlation between the oxygen saturation decline rate and ST segment deviation. The attention mechanism is introduced in the middle layer to dynamically adjust the weight distribution of different feature vectors according to the common pathological characteristics of elderly patients (such as atrial fibrillation waveform variation, conduction block pattern). The output layer combines the individualized threshold of the patient to determine the risk level. This threshold is not a fixed value, but a dynamic baseline constructed based on the patient's historical monitoring data. For example, the heart rate variability coefficient in the past 72 hours is used as a reference benchmark. When the real-time detected heart rate deviation exceeds the baseline fluctuation range, the online fine-tuning of the model parameters is automatically triggered to reduce the misjudgment rate. For the identified abnormal events, the module classifies them into three levels of risk according to the pre-set classification rules: level 1 risk corresponds to minor abnormalities such as occasional premature beats, level 2 risk targets moderate symptoms such as persistent tachycardia, and level 3 risk covers critical conditions such as ST segment elevation myocardial infarction. The risk determination result is converted into a graded alarm instruction and transmitted to the alarm module through an encrypted data packet. The alarm module realizes the transformation from data analysis to actual intervention, and its response mechanism strictly corresponds to the risk level. For level 1 risk, the module activates the local reminder device and adopts a gradually increasing stimulation strategy. In the initial stage, the patient is reminded by low-frequency vibration. If there is no response for a long time, it switches to a combined sound and light alarm, and at the same time sends a simple notice to the family terminal through short-range wireless communication. When level 2 risk is triggered, in addition to enhancing the local alarm intensity, the module also sends an encrypted early warning message to the pre-set guardian through the cellular network. The message contains the type of abnormal event, the occurrence time, and the brief waveform characteristics. When a level 3 high-risk event is detected, the module immediately activates the full-duplex transmission channel of the emergency communication module. This channel adopts redundant communication design. The primary channel preferentially selects the 5G network to transmit high-priority data packets, and the backup channel ensures the transmission of basic information through LoRa technology. During the emergency response process, the module automatically retrieves the key information in the patient's electronic medical record (including recent medication records, allergy history, and previous diagnosis conclusions), encapsulates it with the real-time physiological data into an emergency auxiliary decision-making report, converts it into a format conforming to the HL7 standard through the medical data middleware, and directly pushes it to the emergency center dispatching system. To ensure the reliability of data transmission, the patient's location information is processed by a differential encryption algorithm to achieve precise positioning on the premise of ensuring privacy and security.

[0024] In an embodiment of the present invention, the adaptive adjustment mechanism of the multimodal sensor module focuses on solving the problem of signal acquisition caused by the physiological characteristics of elderly patients. The impedance monitoring system at the sensor-skin contact interface samples the contact impedance value at a millisecond-level frequency. When it detects that the impedance value exceeds the preset safety range (for example, a sudden increase in impedance due to the detachment of the electrode patch), the module automatically switches to the backup electrode array. This array is designed with a distributed layout. It selects the optimal signal source by scanning the contact quality of each backup electrode and applies a digital signal reconstruction algorithm to compensate for the signal attenuation caused by the electrode switch. For the blood oxygen monitoring of the optoelectronic sensor, the module has a built-in temperature-light intensity compensation curve and dynamically adjusts the LED emission power according to the real-time collected ambient temperature value. For example, it appropriately increases the intensity of infrared light in a low-temperature environment to penetrate the thickened epidermal tissue, and reduces the intensity of green light in a high-temperature and sweaty situation to reduce the interference of epidermal reflection. The compensation parameters generated during the signal gain optimization process are fed back to the data processing module through a dedicated data bus to correct the baseline drift of the multimodal physiological signals. For example, when the blood oxygen sensor detects environmental light interference, this parameter will trigger the adjustment of the center frequency of the band-stop filter in the data processing module, forming a feedforward-feedback joint control loop. The fusion processing algorithm of the data processing module is extended to propose a dynamic filtering strategy based on motion state recognition. This strategy captures the patient's body motion data in real time through a six-axis inertial sensor (including a three-axis accelerometer and a three-axis gyroscope), fuses the multi-dimensional motion signals using a Kalman filter, and constructs a feature vector including motion amplitude, frequency, and direction. These vectors are input into a pre-trained motion interference classification model to identify typical activity patterns of elderly patients such as walking, turning over, and trembling, and generate corresponding motion artifact templates. In the time-frequency domain joint analysis stage, the module performs a short-time Fourier transform on the electrocardiogram signal, extracts the energy distribution characteristics of each frequency band, and performs a convolution operation with the motion artifact template to separate the pure electrocardiogram components. For the processing of the respiratory signal, the algorithm performs a correlation analysis on the basic respiratory waveform measured by the impedance method and the volume change signal caused by chest and abdominal movements, and eliminates the false respiratory fluctuations caused by postural changes. The finally generated preprocessed signal not only contains the multimodal waveform data aligned in the time domain but also additional metadata tags (such as motion state encoding, signal quality score) for subsequent analysis modules to preferentially process the data segments in high-risk periods. This fusion processing mechanism particularly optimizes the ability to eliminate small-amplitude high-frequency trembling interference common in elderly patients, improving the signal fidelity by approximately 40% compared to traditional filtering methods.

[0025] Such as Figure 1As shown, the construction and dynamic optimization mechanism of the elderly electrocardiogram feature model. This model is trained by supervised learning methods, and the training dataset specifically collects multimodal physiological signals of elderly heart disease patients, covering common arrhythmia types such as atrial fibrillation, ventricular premature beats, and ST segment abnormalities. The input layer receives the preprocessed signal segments segmented by time windows, and the length of each time window is adaptively adjusted according to the electrocardiogram signal characteristics. For example, when an increase in the RR interval variability is detected, the window is automatically extended to capture the complete arrhythmia cycle. The middle layer adopts a multi-head attention mechanism, which assigns different weights to the feature vectors of different physiological signals. For example, higher attention is given to features such as the absence of P waves or the widening of QRS complexes commonly seen in elderly patients. The output layer generates the probability distribution of each abnormal type based on the Softmax function and combines the individual thresholds of the patients for risk grading. The establishment of individual thresholds relies on historical monitoring data to construct a dynamic baseline. The specific method is as follows: The standard deviations of parameters such as heart rate variability and blood oxygen fluctuation range are statistically calculated with a sliding time window (such as 72 hours) as the reference benchmark for the current threshold. When the deviation degree of the real-time detected physiological parameters exceeds a preset multiple of the baseline fluctuation range, the system triggers online fine-tuning of the model parameters. The fine-tuning process adopts a transfer learning strategy, and only a small adjustment is made to the weights of the fully connected layer to retain the generalization ability of the pre-trained model. In addition, the model introduces an adversarial training mechanism, which enhances the training samples by generating noise data simulating the skin impedance changes of the elderly, and improves the classification robustness under low signal-to-noise ratio conditions.

[0026] Such as Figure 1As shown, the three-level judgment logic and execution process of the hierarchical alarm instruction. The first-level alarm targets occasional physiological abnormalities, such as single atrial premature beats or transient oxygen saturation drops. The system activates the local vibration reminder device, which adopts a progressive stimulation strategy: in the initial stage, it prompts with low-frequency intermittent vibration. If the patient does not confirm receipt through the physical button within the set time, the vibration frequency is gradually increased and a soft light effect is superimposed to avoid scaring elderly patients. The second-level alarm responds to persistent abnormal events. For example, when sinus tachycardia accompanied by an increased respiratory rate occurs continuously for ten minutes, the system sends a warning message to the preset guardian terminal through an encrypted communication protocol. The message body is encapsulated in JSON format and includes the abnormal type code, occurrence timestamp, current physiological parameter summary, and recommended treatment measures. The encryption process uses the national cipher SM4 algorithm to ensure the security of data transmission. The third-level alarm is immediately activated when life-threatening symptoms are identified (such as continuous ST segment elevation exceeding the threshold or ventricular fibrillation waveform). The system executes three tasks in parallel: First, the emergency communication module enables redundant channels to synchronously transmit the patient's geographical location, real-time vital signs, and high-resolution physiological waveforms in the last 15 minutes; Second, the medical data middleware automatically retrieves the patient's electronic medical record from the hospital information system, extracts key information including recent medication records, allergy history, previous surgical records, and the latest inspection reports, and converts them into a structured document that conforms to the emergency center data standard; Finally, the system activates the voice guidance module to guide the patient or on-site personnel to implement preliminary first aid measures (such as taking nitroglycerin or maintaining a specific position) with clear and short voice instructions. The selection of key physiological data follows the clinical first aid guidelines and preferably includes abnormal electrocardiogram segments, continuous blood pressure trend charts, oxygen saturation change curves, and the current activity status (judging whether the patient has fallen or is stationary through inertial sensor data).

[0027] Furthermore, the transmission reliability and data security of the first-aid communication module. The redundant communication link switching mechanism is based on real-time channel quality assessment. The primary channel preferentially adopts the eMBB service type of the 5G network to ensure the data transmission rate under high bandwidth requirements; the backup channel is configured for NB-IoT and LoRa dual-mode standby to ensure basic data transmission when the 5G signal coverage is insufficient. The channel switching decision algorithm comprehensively evaluates three indicators: signal strength, bit error rate, and network latency. When any of the indicators exceeds the preset threshold, the switch is automatically triggered. The switching process uses seamless connection technology to avoid information loss through packet sequence number synchronization and cache retransmission mechanisms. The patient's location information is processed using differential privacy protection technology. The specific method is as follows: Add random noise conforming to the Laplace distribution to the original GPS coordinates, so that the location information obtained by the first-aid center is valid within a range of 100-meter accuracy, but cannot be traced back to the specific street or house number. The medical data middleware is built with a protocol conversion engine that supports two-way conversion of multiple medical data standards such as HL7, FHIR, and DICOM. When retrieving historical medical records, it automatically matches the keyword fields required for first aid. For example, it integrates scattered test results, imaging reports, and medication records into a diagnosis and treatment summary with a unified time axis. In addition, the middleware implements data desensitization processing, masking and replacing sensitive information such as the patient's ID number and contact information, while retaining the diagnosis conclusion and treatment record for first-aid reference.

[0028] As Figure 1 shown, the design of the temporal correlation analysis unit, which mines potential pathological features through the time series association of multi-modal signals. The analysis unit first performs phase synchronization analysis on the RR interval sequence of the electrocardiogram signal and the inhalation-exhalation cycle of the respiratory signal, and calculates their mutual information entropy to evaluate the autonomic nerve regulation function. When it is detected that the characteristics of respiratory sinus arrhythmia are significantly weakened or disappeared (manifested as the correlation coefficient between the RR interval variation and the respiratory cycle being lower than the threshold), the system automatically increases the weight of the risk level determination. Especially in the case of a simultaneous stepwise decrease in blood oxygen saturation, this combined feature is marked as a precursor to respiratory failure. The individualized threshold update cycle introduces an activity intensity adaptive mechanism: Calculate the patient's metabolic equivalent of task (MET) through inertial sensor data. In the low activity intensity stage (such as during sleep), extend the baseline calculation window to 2 hours and use a moving average algorithm to smooth the fluctuations of physiological parameters; while in the high activity intensity stage (such as walking or climbing stairs), shorten the window to 30 minutes and increase the sampling frequency to capture the rapid change trend. In response to the common circadian rhythm differences in elderly patients, the system establishes two independent baseline models for daytime and nighttime, and performs a smooth transition of the baseline during mode switching to avoid false alarms caused by threshold mutations. In addition, the analysis unit integrates a pharmacokinetic model. After detecting that the patient takes cardiovascular drugs such as beta blockers, it automatically adjusts the time decay curve of the heart rate abnormality determination threshold to avoid misjudgment of physiological parameter changes caused by the drug taking effect.

[0029] A dynamic monitoring system for elderly cardiac patients of the present invention fuses electrocardiogram, blood oxygen and respiratory signals through multimodal sensors, dynamically optimizes signal gain in combination with skin contact status, and eliminates motion artifacts; uses a lightweight model with edge computing and cloud collaboration to parse physiological data in real time, identify arrhythmia events and classify risk levels; constructs a hierarchical alarm mechanism, triggers local reminders, remote notifications or emergency linkages according to risk levels, automatically retrieves the patient's medical records during emergency response and enables redundant communication links to transmit key data, realizing the full-process closed-loop management from anomaly detection to emergency decision-making, and improving the monitoring accuracy and treatment timeliness.

[0030] Therefore, through a dynamic monitoring system for elderly cardiac patients of the present invention, the problems of poor real-time performance, high false alarm rate and delayed emergency response in the dynamic monitoring of elderly heart disease patients can be solved.

[0031] The above embodiments are only illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Any person familiar with this technology can modify or change the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or changes made by those with ordinary knowledge in the technical field without departing from the spirit and technical ideas disclosed by the present invention should still be covered by the claims of the present invention.

Claims

1. A dynamic monitoring system for elderly heart patients, characterized in that, Comprising: A multimodal sensor module that collects electrocardiogram signals, blood oxygen signals, and respiratory signals of elderly patients in real time. After adaptively adjusting the signal acquisition parameters to reduce environmental interference, it generates multimodal physiological signals; A data processing module that performs fusion processing on the multimodal physiological signals, eliminates motion artifacts, and extracts characteristic waveforms to generate preprocessed signals; An analysis module that performs real-time analysis on the preprocessed signals based on a preset elderly electrocardiogram feature model, identifies abnormal electrocardiogram events and their risk levels, and generates hierarchical alarm instructions; An alarm module that triggers corresponding local reminders, remote notifications, or first aid linkage responses according to the hierarchical alarm instructions. Among them, the first aid linkage response automatically sends the patient's location and key physiological data to a designated first aid center through a first aid communication module, and synchronously retrieves the patient's historical medical records to assist in first aid decision-making; Among them, the adaptive adjustment of the multimodal sensor module includes dynamically optimizing the signal gain according to the skin contact state, and the risk level determination of the analysis module combines the patient's individual threshold and the deviation degree of real-time physiological parameters.

2. The dynamic monitoring system for elderly cardiac patients according to claim 1, wherein, The adaptive adjustment of the multimodal sensor module further includes dynamically adjusting the signal acquisition frequency by monitoring the impedance change at the sensor-skin contact interface. When the detected contact impedance exceeds the preset range, it automatically switches to a backup electrode array and compensates for signal attenuation. At the same time, it performs closed-loop adjustment on the emission power of the optoelectronic sensor according to environmental temperature fluctuations to keep the signal-to-noise ratio of the blood oxygen signal within an analyzable range. The compensation parameters generated during the signal gain optimization process are fed back to the data processing module through a data bus to correct the baseline drift of the multimodal physiological signals.

3. The dynamic monitoring system for elderly heart patients according to claim 1, wherein The fusion processing of the data processing module includes a dynamic filtering strategy based on motion state recognition. A motion interference model is constructed using body motion signals collected by an inertial sensor. The motion interference model uses a time-frequency domain joint analysis method to separate the myoelectric noise components in the electrocardiogram signal, and incorporates the phase relationship between the respiratory signal and the blood oxygen signal into the motion artifact correction algorithm to generate a multimodal waveform dataset with time alignment characteristics. This dataset is normalized and bound to the patient identity identifier and stored in a cache queue for the analysis module to call according to the priority order.

4. The dynamic monitoring system for elderly heart patients according to claim 1, wherein The elderly electrocardiogram feature model is obtained through machine learning methods. Its input layer receives preprocessed signal segments segmented by time windows. The middle layer uses an attention mechanism to weightedly fuse multimodal feature vectors. The output layer generates a probability distribution according to the electrophysiological characteristics of common arrhythmia types in elderly patients. The establishment of the individual threshold is based on constructing a dynamic baseline from the patient's historical monitoring data. When the deviation degree of real-time physiological parameters exceeds the baseline fluctuation range, it automatically triggers online fine-tuning of the model parameters to reduce the misjudgment rate.

5. The dynamic monitoring system for elderly heart patients according to claim 1, wherein, The generation of the hierarchical alarm instruction includes a three-level decision logic. The first level activates the local vibration reminder for occasional premature beat events and records the event characteristics. The second level superimposes the sound and light alarm on the persistent tachycardia event and sends an encrypted warning message to the preset guardian terminal. The third level immediately activates the full-duplex transmission channel of the emergency communication module when a life-threatening arrhythmia pattern is detected. The selection criteria for the key physiological data include abnormal waveform segments, the change trend of hemodynamic parameters, and the current drug use status.

6. The dynamic monitoring system for elderly heart patients according to claim 1, characterized in that, The emergency communication module is configured with a redundant communication link switching mechanism. When the transmission quality of the primary wireless channel drops below the threshold, it automatically enables the backup frequency band to establish a parallel data transmission path and processes the patient's location information using a differential encryption algorithm during the transmission. The retrieval process of the historical medical record is realized through the medical data middleware to perform protocol conversion of heterogeneous systems, ensuring that the electronic medical record received by the emergency center includes the conclusions of the last three outpatient diagnoses and the list of allergic drugs.

7. The dynamic monitoring system for elderly heart patients according to claim 1, wherein, The analysis module is built-in with a temporal correlation analysis unit for detecting the rhythm coupling relationship between the electrocardiogram signal and the respiratory signal. When it is found that the characteristics of respiratory sinus arrhythmia disappear and are accompanied by a stepwise decrease in blood oxygen saturation, it automatically increases the weight of the risk level determination. The update period of the individualized threshold is adaptively adjusted according to the patient's activity intensity, and the baseline calculation window is extended during the night rest stage to improve the stability of the criterion.

8. The dynamic monitoring system for elderly heart patients according to claim 1, characterized in that, A dynamic priority scheduler is provided between the data processing module and the analysis module. When cross-parameter abnormal correlation characteristics appear in the multi-modal physiological signals, it temporarily increases the processing priority of the corresponding data stream and allocates additional computing resources for multi-dimensional feature cross-validation. The generation process of the preprocessed signal includes timestamp calibration of the feature waveform to ensure that the data collected by different sensors are strictly synchronized in the time domain.

9. The dynamic monitoring system for elderly heart patients according to claim 1, characterized in that The alarm module is integrated with an intelligent learning unit to establish a personalized feedback model by recording the patient's response behaviors to various levels of alarms. When it is monitored that the high-risk alarm is ignored multiple times, it automatically upgrades the alarm intensity and activates the backup contact call chain. The local reminder device adopts a gradually increasing stimulation strategy, using low-frequency vibration to prompt in the initial stage, and switching to a composite sound and light mode if there is no response continuously. At the same time, it sends an attention signal to the data processing module to trigger higher-precision physiological parameter sampling.

10. The dynamic monitoring system for elderly heart patients according to claim 1, wherein A data storage module is provided between the data processing module and the analysis module. The data storage module generates a dynamic storage strategy based on the characteristic waveform energy distribution and abnormal event markers of the preprocessed signals, stores the stable baseline segments in the multimodal physiological signals in a compressed coding manner, and retains the original sampling rate and attaches a timestamp index to the signal segments containing characteristic waveform mutations. The data storage module dynamically adjusts the storage period according to the risk level feedback by the analysis module. The data related to high-risk events is permanently stored and a logical association with the first aid linkage response is established. When the alarm module triggers the first aid communication module, the multimodal data waveforms of the set time period before the incident are automatically extracted from the data storage module, and an auxiliary diagnosis report containing the temporal evolution characteristics is generated and synchronously transmitted to the first aid center. The execution process of the storage strategy is realized through a circular buffer to achieve seamless connection between real-time data streams and historical data, ensuring that the complete physiological parameter change trajectory can be reconstructed when abnormal events are traced back.

Images