Myocardial infarction early warning and first aid system

Abstract

The invention discloses a myocardial infarction early warning and emergency treatment system, and belongs to the technical field of medical treatment and emergency rescue, and the system comprises a data collection module which is used for collecting various physiological parameters of a user in real time through a wearable device; the analysis module is used for receiving and analyzing the physiological parameters and calculating a myocardial infarction risk score and a prediction time window of the user; the early warning module is used for giving out graded early warning and voice self-rescue guidance according to the myocardial infarction risk score; the alarm module is used for automatically triggering an alarm process when an emergency situation is detected; the AI emergency guidance module is used for providing augmented reality visual emergency guidance for field rescuers; and the cloud platform is used for data storage, model optimization and security management. According to the invention, risk quantitative scoring and time window prediction are realized, an early warning gateway is greatly brought forward, and precious time is won for patient intervention and medical treatment.

Description

Technical Field

[0001] This invention mainly relates to the technical field of medical and emergency rescue, specifically an early warning and emergency rescue system for myocardial infarction. Background Technology

[0002] With the continuous rise in the incidence of cardiovascular diseases, myocardial infarction has become one of the major diseases threatening human health. Myocardial infarction is characterized by its rapid onset, severe condition, and high mortality rate. Especially in emergency situations such as sudden cardiac arrest, if timely and effective rescue is not provided, the patient's survival rate will be significantly reduced.

[0003] Currently, early warning and emergency rescue systems for sudden cardiac death have seen some development. For example, Chinese patent CN115620483A discloses an interconnected early warning and emergency rescue system for sudden cardiac death. This system includes a hospital chest pain center terminal and patient-worn terminals, medical or emergency personnel terminals, and street or property management personnel terminals, all connected to the hospital chest pain center terminal. The patient-worn terminal includes a monitoring module, an alarm module, a positioning module, a data transmission module, and an identification module, used to monitor the patient's heart rate and vital signs in real time. When a patient experiences cardiac arrest or sudden death, the system can automatically trigger an alarm and transmit the patient's location information to the nearest hospital chest pain center terminal and emergency personnel.

[0004] In addition, Chinese patent CN115349868A discloses a pre-hospital emergency care method and system based on a smart emergency wristwatch. This system consists of a smart wristwatch, a smart wristwatch-compatible app, a remote smart wristwatch monitoring and management system, and a supporting system for 120 emergency centers. The system supports intelligent analysis of patient electrocardiogram (ECG) data, monitoring of device operating status, and automatic early warning functions. It provides real-time monitoring of the ECG data of users bound to the smart wristwatch, offering timely and efficient emergency care services.

[0005] In the area of ​​early warning for cardiovascular and cerebrovascular diseases, Chinese patent CN117064337A discloses an integrated emergency early warning system and method for cardiovascular and cerebrovascular diseases. This system includes wearable devices, a server, smart terminal A, smart terminal B, and a management platform. The system utilizes the characteristics of "blockchain" to form a management model, and by modularizing the original emergency procedures, it achieves full-process monitoring and management of user prevention, emergency calls, and emergency treatment.

[0006] In the field of remote emergency medical guidance, Chinese patent publication number CN109087488B discloses a remote video emergency medical system and emergency medical method. The system includes a 120 emergency center system, an APP video alarm system, and an ambulance APP terminal, which realizes remote emergency medical guidance through video calls and improves rescue efficiency.

[0007] Furthermore, Chinese patent CN119385510A discloses an emergency rescue system and method for cardiovascular diseases. This system includes a data collection and fusion module, a prediction module, a drone rescue module, and a storage module. The system predicts the patient's condition in real time by constructing a convolutional neural network (CNN) model and uses AR glasses and VR technology to guide emergency responders in their rescue efforts.

[0008] However, existing myocardial infarction early warning and emergency response systems still have the following technical problems:

[0009] 1. Lack of early warning capability for myocardial infarction symptoms. Existing systems mainly provide alarms and rescue for emergencies such as cardiac arrest that have already occurred, but cannot predict in advance when a patient may have a myocardial infarction, thus missing the best opportunity for early intervention.

[0010] 2. The existing emergency medical system's automatic alarm mechanism is not perfect in emergency situations such as cardiac arrest. It often requires the patient or people around to manually trigger the alarm, which means that the patient cannot get timely help when alone or unattended.

[0011] 3. On-site first aid guidance is not intuitive enough. Although there is a remote video guidance system, the lack of real-time visual guidance means that non-professionals cannot correctly implement first aid measures, which reduces the success rate of first aid.

[0012] 4. There is a lack of intelligent identification and timely intervention mechanisms for early myocardial infarction symptoms in patients. Existing systems mainly focus on the emergency process rather than prevention and early intervention, and cannot provide patients with self-help guidance when early symptoms appear.

[0013] Therefore, there is an urgent need to develop a comprehensive system that integrates early warning of myocardial infarction, automatic alarm, real-time visual emergency guidance, and intelligent intervention to improve the survival rate and quality of life of myocardial infarction patients. Summary of the Invention

[0014] The present invention addresses the problem that existing technical solutions are too simplistic and provides a solution that is significantly different from existing technologies. Specifically, the present invention mainly provides an early warning and emergency treatment system for myocardial infarction to solve the technical problems mentioned in the background.

[0015] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows:

[0016] An early warning and emergency treatment system for myocardial infarction, comprising:

[0017] The data acquisition module is used to collect various physiological parameters of the user in real time through wearable devices;

[0018] The analysis module is used to receive and analyze the physiological parameters, calculate the user's myocardial infarction risk score and prediction time window;

[0019] The early warning module is used to issue graded early warnings and voice self-rescue guidance based on the myocardial infarction risk score;

[0020] The alarm module is used to automatically trigger the alarm process when an emergency is detected.

[0021] The AI-powered first aid guidance module provides augmented reality visualization first aid guidance to on-site rescuers.

[0022] Cloud platforms are used for data storage, model optimization, and security management.

[0023] Preferably, the wearable device in the data acquisition module integrates at least one of an electrocardiogram sensor, a blood oxygen saturation sensor, a blood pressure sensor, a body temperature sensor, and a motion status sensor, for collecting heart rate, blood pressure, blood oxygen saturation, electrocardiogram signals, body temperature, and motion posture data; the wearable device is equipped with a low-power Bluetooth communication module for transmitting the collected data to a user terminal or a cloud server.

[0024] Preferably, the analysis module includes:

[0025] The local analysis unit, deployed on the user terminal, is used for real-time preprocessing and anomaly detection of physiological data;

[0026] The cloud-based analytics platform, deployed on a server, employs a hybrid architecture of convolutional neural networks and long short-term memory networks to comprehensively analyze and predict risks from users' historical and real-time physiological data.

[0027] Preferably, the workflow of the analysis module includes:

[0028] Outlier removal: The 3σ criterion is used to identify outliers in the physiological data, and the mean of the first few normal data points is used to replace them to maintain the continuity of the data sequence.

[0029] Feature fusion: ST segment offset and heart rate variability index are extracted from electrocardiogram signals, coefficient of variation is extracted from heart rate data, and main peak amplitude is extracted from eosinogram signals to form a multi-dimensional physiological feature set; the gradient boosting tree algorithm is used to calculate the dynamic weight of each feature, and a multimodal feature vector is generated by weighted summation;

[0030] Risk prediction: Using the multimodal feature vector sequence of the past hour as input, the deep learning model outputs a myocardial infarction risk score R ranging from 0 to 100, and outputs a prediction time window T based on the absolute value and trend of R. pre .

[0031] Preferably, the tiered early warning strategy of the early warning module includes:

[0032] Level 1 warning: When 20≤R<50, a health reminder is issued to the user through vibration and voice of the wearable device;

[0033] Level 2 alarm: When 50≤R<80, accompanied by persistent ST segment abnormalities or symptoms actively reported by the user, the wearable device’s audible and visual alarm will be activated and the emergency contact person will be notified.

[0034] Level 3 alarm: When R≥80, or when no effective QRS wave is detected and blood oxygen is below the threshold, the system automatically sends location and physiological data to the emergency center and activates the highest level audible and visual alarm on the wearable device.

[0035] Preferably, the preset emergency conditions of the alarm module are: the electrocardiogram signal has no effective QRS waveform, the blood oxygen saturation is below 80%, and the user is in a static state.

[0036] Preferably, the AI ​​emergency rescue guidance module includes:

[0037] AR glasses hardware used to capture video and audio of the surrounding environment;

[0038] The emergency rescue guidance software intelligently identifies the patient's chest and marks the correct compression position, tracks and analyzes compression depth and frequency in real time and provides audio-visual feedback, autonomously locates the AED and guides its use, provides differentiated guidance based on the rescuer's ability, and supports multi-person collaborative rescue.

[0039] Preferably, the AI ​​emergency guidance module provides real-time feedback on the quality of chest compressions performed by medical personnel on patients based on the compression depth and frequency.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0041] (1) By integrating multimodal physiological parameters (electrocardiogram, blood oxygen, blood pressure, body movement, etc.) and combining them with the CNN-GRU hybrid deep learning model, the system can extract early features of myocardial infarction from weak and continuous physiological changes, realize risk quantification scoring and time window prediction, significantly advance the warning threshold, and buy valuable time for patient intervention and medical treatment.

[0042] (2) The system of this invention adopts a multi-criteria triggering strategy of "physiological indicators + risk score + user feedback", which significantly reduces the false alarm and false alarm rates. Especially when the user is unconscious or alone, it can automatically identify extreme situations such as cardiac arrest and directly link with the emergency center, realizing a truly "unattended" automatic rescue and greatly shortening the emergency response time.

[0043] (3) This invention uses AI glasses and computer vision technology to transform complex emergency procedures (such as CPR compression location, depth, frequency, and AED use) into immersive visual annotation and real-time voice feedback, which significantly reduces the rescue threshold for non-professionals and effectively improves the standardization and success rate of on-site emergency rescue.

[0044] (4) The system of this invention achieves full-chain coverage of myocardial infarction prevention and control, from continuous monitoring, intelligent analysis, hierarchical early warning, and automatic alarm to on-site AR guidance and remote expert intervention. Data is interconnected and logically consistent in each link, constructing an efficient, reliable, and user-friendly intelligent health management closed loop.

[0045] The present invention will be explained in detail below with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0046] Figure 1 This is a diagram of the overall system architecture of the present invention;

[0047] Figure 2 This is a flowchart of the data processing and analysis process of the present invention;

[0048] Figure 3 This invention relates to a hierarchical early warning decision tree;

[0049] Figure 4 This is a closed-loop diagram of the entire process of the system of the present invention;

[0050] Figure 5 This illustrates the relationship between risk scores and time windows in this invention. Detailed Implementation

[0051] To facilitate understanding of the present invention, a more comprehensive description of the present invention will be given below with reference to the accompanying drawings, which illustrate several embodiments of the present invention. However, the present invention can be implemented in different forms and is not limited to the embodiments described in the text. Rather, these embodiments are provided to make the disclosure of the present invention more thorough and complete.

[0052] It should be noted that when an element is referred to as being "fixed to" another element, it can be directly on the other element or there may be an intervening element. When an element is referred to as being "connected to" another element, it can be directly connected to the other element or there may be an intervening element. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0053] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly associated with those skilled in the art to which this invention pertains. The terminology used herein in the specification of this invention is for the purpose of describing particular embodiments and is not intended to limit the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0054] Example 1:

[0055] Please refer to the appendix carefully. Figure 1-5 As shown, the data acquisition module adopts a "wristwatch-style wearable device," integrating multiple types of sensors and a low-power communication module, as detailed below:

[0056] Electrocardiogram (ECG) sensor: dry electrode technology, sampling frequency 500Hz, accuracy 16-bit, acquiring single-lead ECG signals (including QRS wave and ST segment features);

[0057] Blood oxygen saturation (SpO2) sensor: photoplethysmography, which measures the ratio of oxyhemoglobin to total hemoglobin in blood by emitting red and infrared light of specific wavelengths, with a measurement range of 70%-100% and an accuracy of ±2%.

[0058] Blood pressure sensor: Based on pulse wave transit time technology and combined with user-specific calibration parameters, it can provide systolic and diastolic blood pressure data, with a measurement range of 60-180 mmHg (systolic blood pressure) and 40-120 mmHg (diastolic blood pressure), and an accuracy of ±5 mmHg;

[0059] Body temperature sensor: thermistor, measurement range 35-42℃, accuracy ±0.1℃;

[0060] Motion status sensors: a three-axis accelerometer and a three-axis gyroscope to detect changes in posture and falls;

[0061] Communication module: Low-power Bluetooth 5.0 + NB-IoT dual-mode, normally transmits data to the user terminal via Bluetooth, and connects directly to the cloud platform in emergencies.

[0062] Device features: Built-in 450mAh lithium polymer battery, normal monitoring battery life ≥72 hours, emergency mode battery life ≥24 hours; the shell is made of medical-grade silicone + stainless steel, waterproof rating IP68, suitable for long-term wear.

[0063] Example 2:

[0064] Please refer to the appendix carefully. Figure 1-5As shown, the analysis module is used to receive and analyze the physiological parameters, calculate the user's myocardial infarction risk score and prediction time window; the analysis module includes a local analysis unit and a cloud analysis platform; wherein, the local analysis unit is deployed on the user terminal and is used to perform preliminary real-time processing and anomaly detection on the physiological data; the cloud analysis platform is deployed on the server and uses a deep learning-based artificial intelligence model to perform comprehensive analysis and risk prediction on the user's historical and real-time physiological data, the artificial intelligence model being a hybrid model of convolutional neural network and long short-term memory network.

[0065] Outlier Removal: During physiological data acquisition, equipment noise and user actions (such as raising an arm or walking) may cause outliers (such as sudden increases in blood pressure or jumps in ECG signals). Failure to remove these outliers will severely affect the prediction accuracy of the AI ​​model. Therefore, the 3σ criterion (a robust outlier detection method derived from the characteristics of normal distribution) is used for processing, as detailed below:

[0066] Suppose the sample set of a certain physiological parameter is {x1, x2, ..., x...} n (n represents the sampled data from the past 5 minutes, with a sampling interval of 1 second, then n = 300)), calculate the sample mean. With respect to the sample standard deviation σ: like Then x i If the value is an outlier, it should be replaced with the mean of the top 5 normal samples for that parameter. (x k (The first 5 non-outlier values ​​are used to ensure that the data sequence is continuous and closely reflects the actual physiological trend.)

[0067] Feature fusion: Extracting multimodal core features, including ST segment offset ΔST (early ischemic marker of myocardial infarction, normal range ±0.1mV) and heart rate (HR) coefficient of variation from ECG (electrocardiogram). (Normal range is 2%-5%), SCG (cardiac angiography) main peak amplitude A SCG (Reflecting myocardial contractility); the importance of each feature is calculated using the Gradient Boosting Tree (GBDT) algorithm, and the dynamic weights ω are obtained by normalization. k ( m is the total number of features), which are ultimately fused into a multimodal feature vector. (f k The feature is standardized and mapped to the [0,1] interval.

[0068] Risk Prediction: The myocardial infarction risk prediction model architecture adopts a hybrid architecture of "Improved-CNN + Gated Recurrent Unit (GRU)";

[0069] Risk score calculation: The model input is a "multimodal feature vector sequence of the past hour" {F1, F2, ..., F...} T (Generate one feature vector F every minute) t Output a risk value of 0-100 R = Sigmoid(W2·GRU(Improved-CNN(F)) t ))+b2)×100;

[0070] Among them, Improved-CNN(F t ): For the feature vector F at each time step t Spatial feature extraction is performed, and a high-dimensional feature map is output; GRU(·): Temporal modeling is performed on the CNN output features at 60 time points, and the final hidden state is output (reflecting the temporal trend within 1 hour); W2, b2: Weight matrix and bias term of the fully connected layer (optimized through training with clinical data to ensure that the output is positively correlated with the risk of myocardial infarction); Sigmoid(·): Maps the output of the fully connected layer to the [0,1] interval (to avoid unbounded output values), and then multiplies it by 100 to scale it to 0-100.

[0071] Meaning of risk value:

[0072] R < 20: No risk (normal physiological state, no intervention required);

[0073] 20≤R<50: Potential risk (slight abnormalities in physiological indicators, requiring enhanced monitoring);

[0074] 50≤R<80: High risk (significantly increased likelihood of myocardial infarction, requiring emergency intervention);

[0075] R≥80: Extremely high risk (high probability of short-term myocardial infarction, immediate emergency treatment required).

[0076] Prediction time window T pre Output: Combining the "absolute value" and "trend" of the risk value R (e.g., if R rises from 30 to 60 within 10 minutes, it indicates a rapid increase in risk), four time windows are defined to provide clear time basis for intervention:

[0077]

[0078] Example 3:

[0079] Please refer to the appendix carefully. Figure 1-5 As shown, the early warning module is used to issue graded early warnings and voice self-rescue guidance based on the myocardial infarction risk score; the alarm module is used to automatically trigger the alarm process when an emergency is detected.

[0080] Level 1 Warning (Potential Risk): The risk value is in the "attention required but no emergency treatment needed" range, and only a health reminder needs to be sent.

[0081] Triggering condition: 20≤R<50 (no other abnormal physiological signals); Note: During this stage, physiological indicators are slightly abnormal (such as a slight decrease in HRV and a slight decrease in SCG amplitude), which has not yet reached the level of "emergency intervention". Avoid excessive alarms that may cause user fatigue.

[0082] The wrist-worn wearable device vibrates at a frequency of 100Hz (a frequency sensitive to the human ear, easily noticeable) and an intensity of 500mG (approximately equivalent to a moderate vibration intensity on a mobile phone), lasting for 2 seconds before stopping. A voice alert is displayed: "Potential risk of myocardial infarction detected. Rest and avoid strenuous activity are recommended. If chest pain occurs, please press and hold the device button to trigger an emergency alarm." Simultaneously, the mobile app pushes a "risk report" (including interpretations of abnormal indicators, such as "HRV is slightly below the normal range, possibly related to recent fatigue") and provides a "one-click doctor consultation" entry.

[0083] Level 2 Alert (Emergency Intervention): The risk value is significantly elevated and accompanied by "early symptoms of myocardial infarction or abnormal key indicators", and contact person notification needs to be initiated;

[0084] Triggering conditions: 50≤R<80, and meet any of the following conditions: ST segment elevation ≥0.1mV (direct signal of myocardial ischemia) is detected in ECG for 3 consecutive sampling cycles (3 minutes); the user manually reports "chest pain, left arm numbness, shortness of breath" (typical symptoms of myocardial infarction) through the wearable device; Note: "Dual condition judgment" (risk value + symptoms / indicators) is introduced to avoid invalid alarms caused by misjudgment of risk value alone (e.g., short-term R increase after exercise but no ST segment abnormality, no secondary alarm is triggered).

[0085] At this moment, the wearable device on the wrist calls the emergency contact and emits an 85-decibel beep to attract the attention of people nearby; at the same time, the display screen of the wearable device on the wrist highlights "Emergency risk, contacting emergency contact" and plays the voice message "Your heart indicators are abnormal, please remain calm, emergency contact has been notified" repeatedly.

[0086] Level 3 Alarm (Risk of Cardiac Arrest): The risk value is extremely high or "life-threatening emergency signals" have appeared, requiring immediate professional rescue.

[0087] Triggering conditions: Meet any of the following conditions: R ≥ 80 (probability of short-term myocardial infarction ≥ 80%); ECG shows no valid QRS waves for 8 consecutive seconds (heart rate < 20 bpm, indicating impending cardiac arrest) and blood oxygen saturation (SpO2) < 80% (hypoxia signal). Note: Cardiac arrest is determined using a triple confirmation of "ECG + blood oxygen + body movement" (if only ECG shows no QRS waves but SpO2 is normal, it may be due to device detachment and will not trigger an alarm) to avoid false alarms; simultaneously, R ≥ 80 triggers directly, covering the "extremely high risk state before cardiac arrest".

[0088] The wearable device on the wrist emitted a beeping sound at 95 decibels, while a red LED flashed and the display screen displayed "Urgent emergency care needed, location: near No. XX, XX Road" in a loop.

[0089] The wrist-worn wearable device sends an alarm data packet to the "nearest emergency center" through a cloud platform. The data packet includes: basic user information such as name, age, and medical history; key physiological data such as ECG waveform in the past 5 minutes, blood pressure / blood oxygen change trend, R value and rate of change; and real-time location with a positioning accuracy of ≤5 meters (integrating Beidou and base station positioning, and supplementing Bluetooth beacon positioning in indoor scenarios).

[0090] Assistance from nearby users: Push notifications to users within 50 meters of this system who have installed the accompanying app.

[0091] Example 4:

[0092] Please refer to the appendix carefully. Figure 1-5 As shown, the AI ​​first aid guidance module is used to provide augmented reality visualization first aid guidance for on-site rescuers. The AI ​​first aid guidance module includes: AR glasses hardware, used to capture video and audio of the on-site environment; first aid guidance software, which intelligently identifies the patient's chest and marks the correct compression position, tracks and analyzes the compression depth and frequency in real time and provides audio-visual feedback, autonomously locates the AED and guides its use, provides differentiated guidance based on the rescuer's ability, and supports multi-person collaborative rescue.

[0093] The AR glasses feature a lightweight design, weighing no more than 50 grams, and are equipped with a high-definition camera, microphone, speaker, and transparent display. The camera has a 1080p resolution and a 120-degree field of view, enabling it to clearly capture the scene. The transparent display uses waveguide technology, with a brightness of 500 nits and a contrast ratio of 1000:1, providing clear image display under various lighting conditions.

[0094] The first aid guidance software, based on computer vision and natural language processing technologies, can recognize the scene environment and personnel movements, providing real-time first aid guidance. When on-site personnel wear AI glasses, the system automatically identifies the patient's location and condition, and overlays first aid guidance information onto the AR glasses' display screen. The guidance information includes text prompts, image markers, and voice commands, guiding on-site personnel to correctly perform CPR.

[0095] The system analyzes the position, depth, and frequency of the personnel's presses based on the images captured by the camera and provides real-time feedback. For example, when the press position is incorrect, the system will mark the correct press position on the AR display screen; when the press depth is insufficient, the system will provide a voice prompt saying "Please increase the press depth"; when the press frequency is too slow, the system will display a metronome to help the personnel maintain the correct press frequency.

[0096] Meanwhile, the AI ​​glasses will transmit the scene footage in real time to remote medical experts, enabling them to observe the situation remotely and provide professional guidance. Experts can communicate directly with people on site through the speakers on the AR glasses, and can also mark key locations or draw guidance graphics on the AR display screen.

[0097] The AR camera recognizes the patient's chest contour (distinguishing between the chest, abdomen, and shoulders) and automatically calculates the "midpoint of the line connecting the two nipples" (the international standard compression position), overlaying a red circular marker on the AI ​​glasses. If the rescuer's compression position deviates by more than 2cm (e.g., to the left of the nipple), the marker turns yellow and flashes, while a voice prompt says, "The compression position is too far to the left, please adjust to the red marker." If the deviation is greater than 4cm, the marker turns red and emits a "beep" warning sound, forcibly reminding the rescuer to adjust.

[0098] Data Acquisition: The compression depth d (unit: cm) is calculated using the AR camera's "motion capture algorithm" (analyzing the rescuer's arm displacement during compression), and the compression frequency f (unit: compressions / minute, e.g., timing each complete compression captured at 30fps, counting f over one minute) is calculated using "frame rate matching"; Real-time Feedback: Based on the values ​​of d and f, corresponding prompts are output using a "conditional judgment formula":

[0099]

[0100] The AI ​​glasses display a real-time "compression depth bar graph" (green segment for 5-6cm, red segment for <5cm or >6cm) and "frequency numbers" (e.g., 108 times / min, green for normal, red for abnormal), while the bone conduction speaker plays "100 beats / min" (e.g., "dong-dong-dong") to guide the rescuer to maintain a stable frequency.

[0101] The present invention has been described by way of example in conjunction with the accompanying drawings. Obviously, the specific implementation of the present invention is not limited to the above-described manner. Any non-substantial improvement made by adopting the inventive concept and technical solution of the present invention, or the direct application of the inventive concept and technical solution of the present invention to other occasions without modification, shall be within the protection scope of the present invention.

Claims

1. An early warning and emergency treatment system for myocardial infarction, characterized in that, include: The data acquisition module is used to collect various physiological parameters of the user in real time through wearable devices; The analysis module is used to receive and analyze the physiological parameters, calculate the user's myocardial infarction risk score and prediction time window; The early warning module is used to issue graded early warnings and voice self-rescue guidance based on the myocardial infarction risk score; The alarm module is used to automatically trigger the alarm process when an emergency is detected. The AI-powered first aid guidance module provides augmented reality visualization first aid guidance to on-site rescuers. Cloud platforms are used for data storage, model optimization, and security management.

2. The early warning and emergency treatment system for myocardial infarction according to claim 1, characterized in that: The wearable device in the data acquisition module integrates at least one of an electrocardiogram sensor, a blood oxygen saturation sensor, a blood pressure sensor, a body temperature sensor, and a motion status sensor, for collecting heart rate, blood pressure, blood oxygen saturation, electrocardiogram signals, body temperature, and motion posture data; the wearable device is equipped with a low-power Bluetooth communication module for transmitting the collected data to a user terminal or a cloud server.

3. The early warning and emergency treatment system for myocardial infarction according to claim 1, characterized in that: The analysis module includes: The local analysis unit, deployed on the user terminal, is used for real-time preprocessing and anomaly detection of physiological data; The cloud-based analytics platform, deployed on a server, employs a hybrid architecture of convolutional neural networks and long short-term memory networks to comprehensively analyze and predict risks from users' historical and real-time physiological data.

4. The early warning and emergency treatment system for myocardial infarction according to claim 3, characterized in that: The workflow of the analysis module includes: Outlier removal: The 3σ criterion is used to identify outliers in the physiological data, and the mean of the first few normal data points is used to replace them to maintain the continuity of the data sequence. Feature fusion: ST segment offset and heart rate variability index are extracted from electrocardiogram signals, coefficient of variation is extracted from heart rate data, and main peak amplitude is extracted from eosinogram signals to form a multi-dimensional physiological feature set; the gradient boosting tree algorithm is used to calculate the dynamic weight of each feature, and a multimodal feature vector is generated by weighted summation; Risk prediction: Using the multimodal feature vector sequence of the past hour as input, the deep learning model outputs a myocardial infarction risk score R ranging from 0 to 100, and outputs a prediction time window T based on the absolute value and trend of R. pre .

5. The early warning and emergency treatment system for myocardial infarction according to claim 4, characterized in that: The tiered early warning strategy of the early warning module includes: Level 1 warning: When 20≤R<50, a health reminder is issued to the user through vibration and voice of the wearable device; Level 2 alarm: When 50≤R<80, accompanied by persistent ST segment abnormalities or symptoms actively reported by the user, the wearable device’s audible and visual alarm will be activated and the emergency contact person will be notified. Level 3 alarm: When R≥80, or when no effective QRS wave is detected and blood oxygen is below the threshold, the system automatically sends location and physiological data to the emergency center and activates the highest level audible and visual alarm on the wearable device.

6. The early warning and emergency treatment system for myocardial infarction according to claim 1, characterized in that: The preset emergency conditions of the alarm module are: the electrocardiogram signal has no effective QRS waveform, the blood oxygen saturation is below 80%, and the user is stationary.

7. The early warning and emergency treatment system for myocardial infarction according to claim 1, characterized in that: The AI ​​emergency rescue guidance module includes: AR glasses hardware used to capture video and audio of the surrounding environment; The emergency rescue guidance software intelligently identifies the patient's chest and marks the correct compression position, tracks and analyzes compression depth and frequency in real time and provides audio-visual feedback, autonomously locates the AED and guides its use, provides differentiated guidance based on the rescuer's ability, and supports multi-person collaborative rescue.

8. The early warning and emergency treatment system for myocardial infarction according to claim 7, characterized in that: The AI ​​emergency care guidance module provides real-time feedback on the quality of chest compressions performed by medical personnel on patients, based on the depth and frequency of compressions.

Images