Cardio-pulmonary resuscitation system, method and equipment based on artificial intelligence and big data

Abstract

The invention relates to the field of intelligent medical equipment, and discloses a cardio-pulmonary resuscitation system, method and equipment based on artificial intelligence and big data. According to the method, individual data, including body parameters or physiological data, of a patient is obtained, multi-modal feature extraction and fusion are carried out based on a cardio-pulmonary resuscitation pressing prediction model, individual pressing depth, pressure and frequency parameters are generated, and a pressing execution mechanism is controlled through a closed-loop control algorithm to achieve accurate pressing. Meanwhile, a fracture risk real-time monitoring function is integrated, and pressing parameters are dynamically adjusted based on pressure sensor data analysis. According to the system, individualization, intelligence and precision of cardio-pulmonary resuscitation can be achieved, the rescue success rate is increased, and the iatrogenic injury risk is reduced.

Description

Technical Field

[0001] This invention relates to the field of smart medical devices, and more specifically, to cardiopulmonary resuscitation systems, methods, and devices based on artificial intelligence and big data. Background Technology

[0002] Cardiopulmonary resuscitation (CPR) is a crucial technique for rescuing patients experiencing cardiac arrest. Current CPR guidelines recommend standard chest compression parameters that are universally applicable but do not adequately consider individual patient differences, such as chest anteroposterior diameter, age, and weight. For patients of different body types, a uniform compression depth may result in compressions that are either too shallow or too deep (causing iatrogenic injuries such as rib fractures or organ damage), thus affecting the success rate of resuscitation.

[0003] For example, in obese patients with a large anteroposterior diameter of the chest, the standard compression depth may not be sufficient to generate enough thoracic pressure to ensure effective blood flow; while in elderly or thin patients, excessive compression may lead to secondary injuries such as rib fractures.

[0004] Some existing CPR feedback devices monitor compression depth and frequency using sensors and provide real-time prompts to ensure compliance with guidelines. However, these devices remain based on fixed standards and cannot achieve true individualized optimization. Furthermore, some advanced automated cardiopulmonary resuscitation (APR) machines have mechanical adjustment capabilities, allowing healthcare professionals to set parameters (compression depth and frequency), but their "individualization" is limited to the physical fit of the robotic arm, rather than intelligent decision-making driven by patient physiological characteristics and prognostic data. While some studies have explored the relationship between compression parameters and body size, a practical system and device are lacking that can instantly integrate real-time feedback of patient physical characteristics and vital signs (parameters acquired from monitors) and continuously optimize compression protocols based on large-scale clinical outcome data.

[0005] Therefore, there is an urgent need in this field for a system and device that can dynamically generate and guide the execution of the optimal compression strategy based on the patient's specific physiological parameters through data-driven intelligent decision-making, in order to achieve truly precise cardiopulmonary resuscitation. Summary of the Invention

[0006] This invention provides a cardiopulmonary resuscitation system, method, and device based on artificial intelligence and big data, which solves the technical problems of lack of individualized parameters and fracture risk monitoring technology in related technologies.

[0007] This invention provides a cardiopulmonary resuscitation method based on artificial intelligence and big data, comprising the following steps:

[0008] S1, Obtain the patient's individualized data, wherein the patient's individualized data includes at least one of the following: physical parameters or physiological data;

[0009] S2, Based on the patient's individualized data, the individualized cardiopulmonary resuscitation (CPR) compression prediction model is used to generate individualized CPR compression parameters;

[0010] S3, based on the individualized cardiopulmonary resuscitation compression parameters, control the compression actuator to perform chest compressions on the patient.

[0011] In a preferred embodiment, the step of generating individualized cardiopulmonary resuscitation (CPR) compression parameters by processing the data through a CPR compression prediction model includes:

[0012] Feature extraction and fusion are performed on the individualized data of the patients to obtain a fused feature vector;

[0013] The fused feature vector is input into the cardiopulmonary resuscitation (CPR) compression prediction model, which then outputs the individualized CPR compression parameters.

[0014] In a preferred embodiment, the cardiopulmonary resuscitation method based on artificial intelligence and big data further includes:

[0015] During the compression process, the risk signal of fracture is monitored in real time through the feedback signal of the sensor;

[0016] Fracture risk assessment is performed based on the fracture risk signals, and the individualized cardiopulmonary resuscitation compression parameters are adjusted in real time according to the assessment results.

[0017] In a preferred embodiment, the step of controlling the pressing actuator includes: using a closed-loop control algorithm to adaptively adjust at least one of the pressing depth, frequency, or force based on real-time feedback during the pressing process.

[0018] In a preferred embodiment, the patient's individualized data includes body parameters and physiological data; the body parameters include anteroposterior diameter of the chest, weight, height, chest circumference, and abdominal circumference; the physiological data includes age, gender, data acquired by the monitor, and emergency outcomes; the data acquired by the monitor includes the patient's electrocardiogram, heart rate, blood pressure, blood oxygen saturation, and end-tidal carbon dioxide; the emergency outcomes include the patient's survival status and treatment methods recorded by medical personnel.

[0019] In a preferred embodiment, the cardiopulmonary resuscitation compression prediction model includes:

[0020] The input layer receives the fused feature vector, normalizes it, checks the integrity of the feature vector, fills in missing features with the mean, and maps the processed feature vector to the model's internal representation space.

[0021] Hidden layers, all of which pass through fully connected layers; the fused features are input into the fully connected layers, processed by linear transformations and activation functions, and output as the final layer of features;

[0022] The output layer contains parallel fully connected branches; the press depth prediction branch takes the features from the last layer and inputs them into the fully connected layer, then maps the output to the press depth range through a linear transformation.

[0023] The press frequency prediction branch inputs the last layer of features into the fully connected layer, and maps the output to a press frequency range through a linear transformation, thus outputting an individualized press frequency.

[0024] In a preferred embodiment, an optimization method for a cardiopulmonary resuscitation (CPR) compression prediction model includes:

[0025] Acquire multiple cardiopulmonary resuscitation (CPR) case datasets, each of which includes individualized patient data, compression parameter data, and emergency outcome data.

[0026] The cardiopulmonary resuscitation (CPR) case dataset was used to train or optimize the CPR compression prediction model.

[0027] In a preferred embodiment, an AI- and big data-based cardiopulmonary resuscitation (CPR) system is used to perform the steps of the AI- and big data-based CPR method described above, including:

[0028] The data acquisition module is used to acquire individualized patient data;

[0029] The intelligent prediction module is used to generate individualized cardiopulmonary resuscitation (CPR) compression parameters based on the patient's individualized data through a CPR compression prediction model.

[0030] The execution control module is used to control the compression actuator based on the individualized cardiopulmonary resuscitation compression parameters.

[0031] In a preferred embodiment, the AI- and big data-based cardiopulmonary resuscitation (CPR) device, used to execute the modules in the aforementioned AI- and big data-based CPR system, includes:

[0032] The compression actuator is connected to the execution control module and is used to perform chest compressions on the patient;

[0033] At least one sensor is used to collect feedback signals during the pressing process.

[0034] In a preferred embodiment, the feedback signals collected by the sensor include at least one of the following: compression depth, compression frequency, compression force, and chest recoil; and are transmitted in real time to the execution control module for dynamic adjustment.

[0035] The beneficial effects of this invention are as follows:

[0036] By constructing a cardiopulmonary resuscitation (CPR) prediction model based on artificial intelligence and big data, a technological breakthrough has been achieved, moving from standardized, one-size-fits-all compressions to individualized, precise compressions. The system integrates multi-dimensional information such as the patient's physical parameters and physiological data, outputting optimal compression depth, pressure, and frequency parameters through the CPR compression prediction model. This effectively solves the problem of improper compressions caused by individual patient differences in traditional CPR. Compared to existing technologies, this invention can improve the success rate of CPR while significantly reducing the incidence of iatrogenic complications such as rib fractures and organ damage caused by excessive compression.

[0037] This invention integrates real-time fracture risk monitoring and dynamic parameter adjustment functions, achieving intelligent closed-loop control of the compression process through multi-sensor fusion technology. This ensures maximum patient safety while guaranteeing the effectiveness of resuscitation. This technical solution has high clinical applicability and promotional value, and can be widely applied in various emergency scenarios such as hospital emergency departments, ICUs, and ambulances. Attached Figure Description

[0038] Figure 1 This is a flowchart of the cardiopulmonary resuscitation method based on artificial intelligence and big data of the present invention;

[0039] Figure 2 This is a block diagram of the cardiopulmonary resuscitation system based on artificial intelligence and big data of the present invention. Detailed Implementation

[0040] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, some features described in the examples may be combined in other examples.

[0041] At least one embodiment of the present invention discloses a cardiopulmonary resuscitation method based on artificial intelligence and big data, such as Figure 1 As shown, it includes the following steps:

[0042] S1, Obtain the patient's individualized data, wherein the patient's individualized data includes at least one of the following: physical parameters or physiological data;

[0043] Obtain individualized patient data, which includes at least physical parameters and physiological data.

[0044] Specifically, the body parameters include the anteroposterior diameter of the chest, weight, height, chest circumference, and abdominal circumference. The anteroposterior diameter of the chest is obtained through the distance sensor of the compression device, the weight data is obtained through the weight sensor of the bed base on which the patient lies, and the chest and abdominal circumferences are obtained through device scanning.

[0045] Physiological data includes age, gender, data acquired by the monitor, and emergency outcomes. Age and gender are entered by medical staff or via voice recognition; data acquired by the monitor includes the patient's electrocardiogram, blood pressure, blood oxygen saturation, and end-tidal carbon dioxide, acquired in real time; emergency outcomes include the patient's survival status and treatment methods entered by medical staff.

[0046] S2, Based on the patient's individualized data, the individualized cardiopulmonary resuscitation (CPR) compression prediction model is used to generate individualized CPR compression parameters;

[0047] Based on the patient's individualized data, individualized cardiopulmonary resuscitation (CPR) compression parameters are generated through a CPR compression prediction model.

[0048] Specifically, the following steps are included:

[0049] S21, Feature extraction and fusion: The individualized data of the patient are subjected to feature extraction and fusion to obtain a fused feature vector.

[0050] For body parameters, a deep neural network algorithm is used for feature extraction to process numerical information such as the anteroposterior diameter of the thorax and body weight.

[0051] For time-series signal data (electrocardiogram, heart rate, blood pressure, blood oxygen saturation, end-tidal carbon dioxide) in physiological data, a convolutional neural network algorithm is used for feature extraction.

[0052] Feature vectors from different modalities are fused to obtain fused features. First, the feature vectors from each modality are preprocessed by mapping body parameter feature vectors and physiological data feature vectors to a unified feature space through fully connected layers and performing L2 normalization. Then, a feature concatenation algorithm is used to concatenate the preprocessed feature vectors sequentially to form a fused feature vector.

[0053] S22, Cardiopulmonary resuscitation (CPR) compression prediction model processing; the fused feature vector is input into the CPR compression prediction model, and the individualized CPR compression parameters are output by the CPR compression prediction model.

[0054] The cardiopulmonary resuscitation compression prediction model includes a deep neural network architecture, comprising an input layer, multiple hidden layers, and an output layer.

[0055] The input layer receives the fused feature vector, normalizes it, checks the integrity of the feature vector, fills in missing features with the mean, and maps the processed feature vector to the model's internal representation space.

[0056] The hidden layer, through a combination of fully connected layers, batch normalization layers, activation function layers, and dropout layers, achieves layer-by-layer abstraction and non-linear transformation of the input fusion features. Specifically:

[0057] First hidden layer: The input features are linearly transformed through the first fully connected layer. Batch normalization is applied to stabilize the training process. Non-linearity is introduced through the ReLU activation function, and dropout regularization is applied to prevent overfitting. The output is the feature representation of the first layer. .

[0058] Second hidden layer: Output the first layer A linear transformation is performed through a second fully connected layer, followed by batch normalization. A non-linear transformation is then performed using the ReLU activation function, and dropout regularization is applied to output the feature representation of the second layer. .

[0059] Third hidden layer: Output the second layer A linear transformation is performed through a third fully connected layer, followed by batch normalization. A non-linear transformation is then performed using the ReLU activation function, and dropout regularization is applied to output the feature representation of the third layer. .

[0060] The output layer contains multiple parallel fully connected branches that predict press depth, press force, and press frequency, respectively.

[0061] Compression depth prediction branch: Representing the features of the third layer Input the fourth fully connected layer, apply the tanh activation function to restrict the output to the range of [-1,1], and map the output to a reasonable range of compression depth [3.0,7.0]cm through linear transformation, thus outputting an individualized compression depth.

[0062] Pressure prediction branch: Incorporating deep features The fifth fully connected layer is input, and the sigmoid activation function is applied to restrict the output to the range of [0,1]. The output is mapped to a reasonable range of compression pressure [20,60] kg through linear transformation. The output is then adaptively adjusted based on the patient's weight to output an individualized compression pressure.

[0063] Pressing frequency prediction branch: Incorporating deep features Input the sixth fully connected layer, apply the sigmoid activation function for normalization, and map the output to a reasonable range of pressing frequency [100, 120] times / minute through linear transformation, thus outputting an individualized pressing frequency.

[0064] S3, based on the individualized cardiopulmonary resuscitation compression parameters, control the compression actuator to perform chest compressions on the patient;

[0065] Based on the individualized cardiopulmonary resuscitation compression parameters, the compression actuator is controlled to perform chest compressions on the patient.

[0066] The steps for controlling the pressing actuator include: using a closed-loop control algorithm to adaptively adjust at least one of the pressing depth, frequency, or force based on real-time feedback during the pressing process.

[0067] Specifically, the pressing actuator uses a stepper motor drive system, which controls the rotation angle and speed of the stepper motor through pulse signals, thereby precisely controlling the pressing depth and frequency.

[0068] Pressing depth control: The required stepper motor rotation angle is calculated based on individualized pressing depth parameters, the angle is converted into the corresponding number of pulses, and pulse signals are sent to the stepper motor according to the set pulse frequency. The pressing depth is monitored in real time by a displacement sensor, and the sending of pulse signals stops when the target depth is reached.

[0069] Pressing frequency control: The pressing cycle time is calculated based on individualized pressing frequency parameters. The cycle time is divided into a pressing phase and a rebound phase. During the pressing phase, the stepper motor is controlled to rotate forward, and during the rebound phase, the stepper motor is controlled to rotate in reverse, and the pressing and rebounding actions are executed cyclically.

[0070] Pressure control: The pressure sensor monitors the pressure in real time and compares the actual pressure with the target pressure. When the actual pressure deviates from the target pressure, the motor drive current is adjusted through PWM modulation technology to achieve precise control of the pressure.

[0071] In one embodiment of the present invention, in order to achieve real-time monitoring and dynamic parameter adjustment of fracture risk during cardiopulmonary resuscitation (CPR) compressions, a method for fracture risk monitoring and parameter adjustment is also provided, specifically including:

[0072] Real-time monitoring of fracture risk signals; During the compression process, fracture risk signals are monitored in real time through feedback signals from sensors;

[0073] The feedback signals collected by the sensors include compression depth, compression frequency, compression force, and chest recoil, and are transmitted in real time to the execution control module for dynamic adjustment.

[0074] Fracture risk monitoring is based on pressure sensor data analysis technology, which uses a pressure sensor array to collect pressure signals in real time during the compression process;

[0075] Pressure sensor array: Arranged on the surface of the compression head, it monitors the pressure distribution and pressure gradient changes in real time during the compression process, and identifies abnormal chest stiffness through pressure distribution non-uniformity analysis;

[0076] Fracture risk assessment; fracture risk assessment is performed based on the fracture risk signals.

[0077] A fracture risk assessment model was established, employing a pressure signal feature analysis algorithm:

[0078] Pressure gradient analysis: Calculate the spatial and temporal gradients of pressure distribution during compression. When the pressure gradient exceeds a preset threshold, it is identified as a potential fracture risk.

[0079] Thoracic stiffness change monitoring: By analyzing the pressure-displacement relationship, changes in thoracic stiffness are detected. A sudden decrease in stiffness may indicate a possible fracture.

[0080] Risk level assessment: Based on the comprehensive stress signal analysis results, the fracture risk is divided into three levels: low risk, medium risk, and high risk.

[0081] Adjust compression parameters in real time; adjust the individualized CPR compression parameters in real time based on the assessment results.

[0082] Automatically adjust compression parameters based on fracture risk assessment results:

[0083] Low risk level: Maintain individualized compression parameters unchanged and continue to perform compressions according to the optimal parameters output by the predictive model;

[0084] Medium risk level: Reduce compression depth by 10%-15%, reduce compression force by 5%-10%, keep the compression frequency unchanged, and increase monitoring frequency;

[0085] High-risk level: Reduce compression depth by 20%-25%, reduce compression force by 15%-20%, and increase compression frequency by 5%-8% to compensate for the compression effect. Activate the early warning mechanism to notify medical staff.

[0086] The parameter adjustment adopts a gradual adjustment strategy to avoid sudden parameter changes from affecting the continuity of compressions, and to ensure that effective cardiopulmonary resuscitation is maintained while reducing the risk of fracture.

[0087] In one embodiment of the present invention, in order to improve the accuracy and generalization ability of the cardiopulmonary resuscitation (CPR) compression prediction model, an optimization method for the CPR compression prediction model is also provided, specifically including:

[0088] Acquire multiple CPR case datasets, each dataset including individualized patient data, compression parameter data, and emergency outcome data.

[0089] The cardiopulmonary resuscitation (CPR) case dataset was used to train or optimize the CPR compression prediction model.

[0090] Specifically, a supervised learning method was employed, using individualized patient data as input and optimal compression parameters and emergency outcomes as labels to train a deep neural network model. Cross-validation was used to evaluate model performance, mean squared error was used as the loss function, and the Adam optimizer was employed for parameter updates.

[0091] During model training, data augmentation techniques are used to expand the training samples, regularization techniques are used to prevent overfitting, and a learning rate decay strategy is used to improve convergence stability.

[0092] Model optimization employs an online learning algorithm, continuously refining the predictive model based on newly added CPR case data. The model is updated every 100 new cases collected, using incremental learning to avoid retraining the entire model and ensuring it adapts to the changing characteristics of different patient groups.

[0093] In one embodiment of the present invention, in order to realize the systematic application of the above-mentioned cardiopulmonary resuscitation method based on artificial intelligence and big data, a cardiopulmonary resuscitation system based on artificial intelligence and big data is also provided, such as... Figure 2 As shown, it specifically includes:

[0094] The data acquisition module is used to acquire individualized patient data. This module includes various sensors and data interfaces.

[0095] The system includes a distance sensor to measure the anteroposterior diameter of the patient's chest; a weight sensor integrated into the bed base to acquire the patient's weight; a monitor interface to acquire real-time data such as the patient's electrocardiogram, heart rate, blood pressure, blood oxygen saturation, and end-tidal carbon dioxide; and a human-machine interface for medical staff to input basic information such as the patient's age and gender, as well as emergency outcomes.

[0096] The intelligent prediction module is used to generate individualized cardiopulmonary resuscitation (CPR) compression parameters based on the patient's individualized data through a CPR compression prediction model.

[0097] This module includes a feature extraction unit, a feature fusion unit, and a prediction model unit: The feature extraction unit uses deep neural network and convolutional neural network algorithms to extract features from body parameters (anteroposterior diameter of the chest, weight, height, chest circumference, and abdominal circumference) and physiological data (age, gender, data obtained from the monitor, and emergency outcome), respectively; The feature fusion unit preprocesses and concatenates feature vectors from different modalities to generate a fused feature vector; The prediction model unit is based on a deep neural network architecture and outputs individualized compression depth, compression force, and compression frequency parameters.

[0098] The execution control module is used to control the compression actuator based on the individualized cardiopulmonary resuscitation compression parameters.

[0099] This module employs a closed-loop control algorithm and includes a stepper motor control unit, a sensor feedback unit, and a parameter adjustment unit.

[0100] The stepper motor control unit precisely controls the compression depth, frequency, and force through pulse signals; the sensor feedback unit collects feedback signals in real time during the compression process; and the parameter adjustment unit adjusts the compression parameters in real time based on the feedback signals and fracture risk assessment results.

[0101] In one embodiment of the present invention, in order to realize the hardware deployment and clinical application of the above-mentioned cardiopulmonary resuscitation system, a cardiopulmonary resuscitation device based on artificial intelligence and big data is also provided, specifically including:

[0102] The compression actuator is connected to the execution control module and is used to perform chest compressions on the patient.

[0103] The device uses a stepper motor drive system, including: a stepper motor, providing precise position and speed control; a transmission mechanism, converting motor rotation into linear compression motion; a compression head, which directly contacts the patient's chest to perform compression operations; and a guide device, ensuring the accuracy and stability of the compression direction.

[0104] A sensor system, including at least one sensor, for acquiring feedback signals during the pressing process.

[0105] The sensor system includes: a displacement sensor for monitoring compression depth; a pressure sensor array for monitoring compression intensity and pressure distribution; and an acceleration sensor for monitoring compression frequency and chest recoil.

[0106] The feedback signals collected by the sensors include compression depth, compression frequency, compression force, and chest recoil, and are transmitted in real time to the execution control module for dynamic adjustment.

[0107] The embodiments of the present invention have been described above. However, the embodiments are not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make more equivalent embodiments under the guidance of the present embodiments, and all of them are within the protection scope of the present embodiments.

Claims

1. A cardiopulmonary resuscitation method based on artificial intelligence and big data, characterized in that, Includes the following steps: S1, Obtain the patient's individualized data, wherein the patient's individualized data includes at least one of the following: physical parameters or physiological data; S2, Based on the patient's individualized data, the individualized cardiopulmonary resuscitation (CPR) compression prediction model is used to generate individualized CPR compression parameters; S3, based on the individualized cardiopulmonary resuscitation compression parameters, control the compression actuator to perform chest compressions on the patient.

2. The cardiopulmonary resuscitation method based on artificial intelligence and big data according to claim 1, characterized in that, The step of generating individualized cardiopulmonary resuscitation (CPR) compression parameters by processing the data through a CPR compression prediction model includes: Feature extraction and fusion are performed on the individualized data of the patients to obtain a fused feature vector; The fused feature vector is input into the cardiopulmonary resuscitation (CPR) compression prediction model, which then outputs the individualized CPR compression parameters.

3. The cardiopulmonary resuscitation method based on artificial intelligence and big data according to claim 1, characterized in that, Also includes: During the compression process, the risk signal of fracture is monitored in real time through the feedback signal of the sensor; Fracture risk assessment is performed based on the fracture risk signals, and the individualized cardiopulmonary resuscitation compression parameters are adjusted in real time according to the assessment results.

4. The cardiopulmonary resuscitation method based on artificial intelligence and big data according to claim 1, characterized in that, The steps of controlling the pressing actuator include: using a closed-loop control algorithm to adaptively adjust at least one of the pressing depth, frequency, or force based on real-time feedback during the pressing process.

5. The cardiopulmonary resuscitation method based on artificial intelligence and big data according to claim 1, characterized in that, The individualized patient data includes body parameters and physiological data; the body parameters include the anteroposterior diameter of the chest, weight, height, chest circumference, and abdominal circumference; the physiological data includes age, gender, data acquired by the monitor, and emergency outcomes; the data acquired by the monitor includes the patient's electrocardiogram, heart rate, blood pressure, blood oxygen saturation, and end-tidal carbon dioxide; the emergency outcomes include the patient's survival status and treatment methods recorded by medical staff.

6. The cardiopulmonary resuscitation method based on artificial intelligence and big data according to claim 1, characterized in that, Cardiopulmonary resuscitation compression prediction models include: The input layer receives the fused feature vector, normalizes it, checks the integrity of the feature vector, fills in missing features with the mean, and maps the processed feature vector to the model's internal representation space. Hidden layers, all of which pass through fully connected layers; the fused features are input into the fully connected layers, processed by linear transformations and activation functions, and output as the final layer of features; The output layer contains parallel fully connected branches; the press depth prediction branch takes the features from the last layer and inputs them into the fully connected layer, then maps the output to the press depth range through a linear transformation. The press frequency prediction branch inputs the last layer of features into the fully connected layer, and maps the output to a press frequency range through a linear transformation, thus outputting an individualized press frequency.

7. An optimization method for a cardiopulmonary resuscitation (CPR) compression prediction model, characterized in that, include: Acquire multiple cardiopulmonary resuscitation (CPR) case datasets, each of which includes individualized patient data, compression parameter data, and emergency outcome data. The cardiopulmonary resuscitation (CPR) case dataset was used to train or optimize the CPR compression prediction model.

8. A cardiopulmonary resuscitation (CPR) system based on artificial intelligence and big data, used to perform the steps of the CPR method based on artificial intelligence and big data as described in any one of claims 1-6, characterized in that, include: The data acquisition module is used to acquire individualized patient data; The intelligent prediction module is used to generate individualized cardiopulmonary resuscitation (CPR) compression parameters based on the patient's individualized data through a CPR compression prediction model. The execution control module is used to control the compression actuator based on the individualized cardiopulmonary resuscitation compression parameters.

9. A cardiopulmonary resuscitation (CPR) device based on artificial intelligence and big data, used to execute a module in the CPR system based on artificial intelligence and big data as described in claim 8, characterized in that, include: The compression actuator is connected to the execution control module and is used to perform chest compressions on the patient; At least one sensor is used to collect feedback signals during the pressing process.

10. The cardiopulmonary resuscitation device based on artificial intelligence and big data according to claim 9, characterized in that, The feedback signals collected by the sensors include at least one of the following: compression depth, compression frequency, compression force, and chest recoil; and are transmitted in real time to the execution control module for dynamic adjustment.

Images