AED intelligent first-aid management system based on Internet of Things

Abstract

The invention discloses an AED intelligent first-aid management system based on the Internet of Things, and relates to the technical field of first-aid management, and the system comprises a data collection and transmission module, a first-aid decision support module, a resource scheduling and cooperation module, and a first-aid training and evaluation module. The first-aid decision support module processes acquired data by using big data analysis and an artificial intelligence algorithm, quickly judges the type and severity of cardiac arrest, generates a personalized first-aid scheme and pushes the personalized first-aid scheme to first-aid personnel equipment, and the resource scheduling and collaboration module tracks AED equipment and personnel information in real time, plans and deploys a path and an attendance route in combination with road conditions, and sends the path to the first-aid personnel equipment. The system is connected with a hospital system to realize efficient cooperation of first aid; through deep fusion of the Internet of Things technology, big data and artificial intelligence, an efficient and intelligent AED first-aid management system is constructed, the first-aid response speed and success rate are remarkably improved, and precious treatment time is won for patients.

Description

Technical Field

[0001] The present invention relates to the technical field of emergency management, and in particular to an AED intelligent emergency management system based on the Internet of Things. Background Art

[0002] In the traditional emergency system, there are many deficiencies in the use of AED (automated external defibrillator) equipment and emergency response. First, the distribution and use of AED equipment are often not transparent enough, and it is difficult for emergency personnel to quickly locate and obtain equipment in an emergency. Secondly, emergency personnel's decisions on the scene often rely on limited information and experience, lack of effective data support and intelligent analysis tools, resulting in insufficient personalization and accuracy of emergency plans. In addition, the dispatch of emergency resources is not efficient enough, and it is impossible to optimize and deploy them according to real-time road conditions and personnel distribution, which affects the response speed and success rate of emergency.

[0003] At present, data collection at the emergency scene mostly relies on the subjective judgment of emergency personnel and simple equipment records, and lacks comprehensive and real-time data collection methods. For example, key information such as patients' heart rate data, symptoms and environmental factors has not been fully collected and integrated. At the same time, there is a lack of effective analysis and processing methods for the collected data, and it is impossible to fully utilize big data and artificial intelligence technologies for in-depth mining, thus failing to provide scientific basis and precise guidance for emergency decision-making.

[0004] In order to solve the above defects, a technical solution is now provided. Summary of the invention

[0005] The purpose of the present invention is to solve the problems of difficult positioning of AED equipment, lack of data support for emergency decision-making, inefficient resource scheduling, and imperfect training and evaluation of emergency personnel in existing emergency systems, and to propose an AED intelligent emergency management system based on the Internet of Things.

[0006] The purpose of the present invention can be achieved through the following technical solutions: AED intelligent emergency management system based on the Internet of Things, including: Data collection and transmission module, which is used to integrate multiple sensors into AED equipment and field terminals, collect and transmit patient, equipment and environmental data to the cloud to provide a basis for subsequent decision-making; The emergency decision support module is used to process the collected data using big data and AI algorithms, determine the type and severity of cardiac arrest, and generate personalized emergency plans and push them to the emergency personnel's equipment; Resource scheduling and coordination module, used to track AED equipment and personnel information, plan dispatch routes and attendance routes based on road conditions, and connect with the hospital system to achieve efficient coordination of emergency rescue; The first aid training and assessment module is used to record and analyze the performance data of first aid personnel, screen personnel to be trained after quantitative assessment, and carry out targeted training according to categories.

[0007] Furthermore, the execution process of the emergency decision support module is as follows: After receiving data from AED devices and on-site collection terminals, it is processed using big data analysis and artificial intelligence algorithms; Based on different historical emergency cases and medical research data, an intelligent diagnosis model is established to determine the type and severity of a patient's cardiac arrest based on the patient's real-time heart rhythm data, symptoms, and environmental factors; Based on the diagnosis results, a personalized first aid plan is automatically generated, including AED defibrillation energy setting recommendations and CPR compression frequency and depth guidance, and pushed to the first aid personnel's smart devices through voice prompts and mobile applications, ensuring that first aid personnel can operate accurately and in a standardized manner during the tense rescue process.

[0008] Furthermore, the specific operation steps of the emergency decision support module for determining the type of cardiac arrest of the patient are as follows: First, the patient's real-time heart rhythm data presented in the form of a continuous electrocardiogram waveform is obtained from the AED device, and discrete data points are collected within a preset time interval. At the same time, the symptom data provided by the on-site collection terminal is quantified through a preset code. Each symptom has a corresponding numerical weight, and environmental factors record temperature and humidity. Then, the heart rhythm data was denoised using Daubechies wavelet to obtain a pure heart rhythm signal. The symptom data and environmental factor data were also normalized to the range of 0-1. Then, an architecture combining a deep convolutional neural network and a long short-term memory network is constructed. The deep convolutional neural network layer uses convolution kernels and step sizes to extract features of heart rhythm data and convert them into high-level abstract feature vectors. The long short-term memory network layer processes symptom manifestations and environmental factor data in time series to obtain feature vectors. The two are concatenated and fused, and then classified by a fully connected layer and a softmax classifier to calculate the probability that the patient belongs to different types of cardiac arrest. The type is accurately determined by continuously training the model and adjusting the weights.

[0009] Furthermore, the specific operation steps of the emergency decision support module for judging the severity are as follows: For severity judgment, the emergency decision support module constructs a multi-factor regression model, with the patient's cardiac arrest type, the sum of symptom weights, and the comprehensive index of environmental factors as input; the sum of symptom weights is calculated based on the number of symptoms and their corresponding weights, and the comprehensive index of environmental factors is obtained by weighted summation of temperature, humidity, etc. The model outputs a severity score, which is trained through a historical emergency case database and continuously adjusts model parameters using the stochastic gradient descent algorithm or its variants. The cross entropy loss function is used for classification tasks, and the mean square error loss function is used for regression tasks to minimize the loss between the prediction and the true label. The severity levels are divided into mild, moderate, and severe according to the score range.

[0010] Furthermore, the first aid decision support module includes the following suggestions for setting the AED's defibrillation energy, the compression frequency and depth of cardiopulmonary resuscitation: If the patient is in ventricular fibrillation, the severity of the condition will be classified into different levels; In mild cases, the initial defibrillation energy is increased by a preset ratio based on the basic defibrillation energy of the AED device, and the proportionality coefficient contains a random factor distributed through the training model; In moderate and severe cases, only the random factors of the proportional coefficients are different and have preset initial values; For pulseless ventricular tachycardia, the defibrillation energy is calculated using different formulas for sustained low-severity and high-severity ventricular tachycardia; In terms of cardiopulmonary resuscitation, the compression frequency is determined by the adjustment factor based on the severity of the patient and the type of cardiac arrest, combined with the basic compression frequency; The compression depth for adults is related to their weight, and different coefficients correspond to different degrees of severity. The compression depth for children and infants is adjusted according to age and weight, and each age group is analyzed using different calculation formulas.

[0011] Furthermore, the specific operation steps of the resource scheduling and coordination module are as follows: Track the location and status of AED devices in real time and display them intuitively on a map through geographic information system technology; when an emergency call is received, quickly search for available AED devices in the surrounding area based on the location of the call, combined with traffic information and the distribution of emergency personnel; After searching for available AED devices in the surrounding area, determine the optimal AED device by analyzing and evaluating parameters; At the same time, a communication connection is established with the emergency system of surrounding medical institutions to inform the hospital of the patient's condition and estimated arrival time in advance, achieving seamless connection from the emergency site to the hospital and improving overall emergency efficiency.

[0012] Furthermore, the specific operation steps of the resource scheduling and coordination module to determine the optimal AED device by analyzing the evaluation parameters are as follows: Evaluation parameters include: Distance parameter: The call location is taken as the origin. The actual distance between each available AED device in the surrounding area and the call location is obtained through GIS technology and recorded as the distance judgment value JC; Traffic condition parameters: Connect with the intelligent traffic system of the traffic management department to obtain traffic condition information along the way from each AED device to the emergency call location; classify traffic conditions into different levels, including smooth, light congestion, moderate congestion, and heavy congestion, and assign different quantitative values ​​to different levels of traffic conditions, recorded as road rating values ​​LP; Equipment status parameters: real-time monitoring of the status of the AED equipment, including battery power and electrode validity period, and quantification to a value between 0 and 1 to obtain the status evaluation value ZK; The distance judgment value JC, road evaluation value LP and condition evaluation value ZK are normalized and then input into the following formula: To obtain the optimal judgment value YPZ, and use the obtained optimal judgment value as the selection standard for AED equipment, where are the preset weight coefficients of the distance judgment value JC, the road evaluation value LP and the condition evaluation value ZK, respectively, and The sum is 1; The superiority values ​​YPZ obtained by analyzing different AED devices are sorted by size, and the AED device with the largest superiority value YPZ is selected as the final selected device.

[0013] Furthermore, the execution process of the first aid training and assessment module is as follows: Comprehensively record the detailed data of emergency personnel in the process of performing tasks in each module, including emergency response time, execution order and timestamp of operation steps, and equipment operation parameter information in the emergency decision support module and resource scheduling and coordination module, and classify and organize these data according to different emergency events and personnel; Analyze and quantify the obtained data to obtain performance parameters, use the performance parameters to determine the comprehensive evaluation value of the first aid personnel, compare the comprehensive evaluation value ziu obtained by analyzing all first aid personnel with the preset comprehensive evaluation threshold, and when the comprehensive evaluation value ziu of the first aid personnel is less than the preset comprehensive evaluation threshold, mark the first aid personnel as a person to be trained; The personnel to be trained are classified into first aid personnel and AED equipment maintenance personnel. The training content includes the operating principles of AED equipment, the steps of first aid skills, and the response strategies in different cardiac arrest scenarios. It is presented in various forms such as video demonstrations, animation simulations, and virtual reality training.

[0014] Furthermore, the process of determining the comprehensive evaluation value of the first aid personnel using the performance parameters in the first aid training and evaluation module is as follows: Performance parameters include: Timely emergency response: The time interval from receiving an emergency call to the time when emergency personnel start to take effective emergency measures is used as a measurement indicator; standard response time thresholds are set in different scenarios, and the difference is calculated, recorded as the negative evaluation value, and the negative evaluation value is compared with several preset negative evaluation value intervals. Several negative evaluation value intervals are respectively set with different response scores to determine the response score of the emergency personnel, and this response score is used to measure the timeliness of the emergency response of the emergency personnel; Operation accuracy: including the accuracy of defibrillator energy setting, electrode pasting accuracy and operation process integrity. The accuracy of defibrillator energy setting is calculated by calculating the deviation ratio between the defibrillator energy actually set by the first aid personnel and the energy value recommended by the system, which is recorded as the setting ratio. The electrode pasting accuracy uses image recognition technology or sensor feedback to determine whether the electrode pasting position meets the standard. The deviation area or distance ratio of the actual pasting position is calculated based on the standard pasting area, which is recorded as the pasting ratio. The operation process integrity counts the completeness of the first aid personnel in completing the AED standard operation process, which is measured by analyzing the ratio of the number of missed steps of the first aid personnel to the total number of steps, and is recorded as the operation ratio. After normalizing the obtained set ratio, pasting ratio and operation ratio, the set ratio, pasting ratio and operation ratio are used as the three sides of the triangle to establish a triangle, and a triangular pyramid model is established with the preset correction factor as the height, and the volume of the triangular pyramid model is calculated and recorded as the balance value, which is used as a standard to measure the accuracy of the operation; Accuracy of cardiopulmonary resuscitation operation: The accuracy of compression frequency is calculated by calculating the deviation ratio between the actual compression frequency and the compression frequency recommended by the system according to the patient's condition, and recorded as the frequency deviation value; the accuracy of compression depth is obtained by obtaining the average value of the actual compression depth through the pressure sensor, calculating the deviation ratio from the standard depth range, and recording it as the depth deviation value; the accuracy of compression position uses the human body model or sensor positioning technology to determine whether the compression position is accurately at the junction of the middle and lower 1 / 3 of the sternum, and measures it according to the degree of deviation, which is recorded as the deviation value; the obtained frequency deviation value, depth deviation value and deviation value are calibrated as pp, sm and pw respectively, and after normalization, they are entered into the following formula: To obtain the heart judgment value xzp, and use the heart judgment value xzp as a standard for measuring the accuracy of cardiopulmonary resuscitation operation; Then normalize the obtained response score xpi, balance value chy, and judgment value xzp and enter them into the following formula: To obtain the comprehensive evaluation value ziu, where They are the preset weight coefficients of response score xpi, balance value chy, and judgment value xzp respectively, and this comprehensive evaluation value is used as the standard to measure the performance of emergency personnel in emergency operations.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention integrates multiple sensors into AED equipment and on-site terminals to collect and transmit patient, equipment and environmental data to the cloud in real time, providing comprehensive and accurate data support for emergency decision-making; the emergency decision support module uses big data analysis and artificial intelligence algorithms to quickly determine the type and severity of cardiac arrest, and generates personalized emergency plans and pushes them to emergency personnel's equipment; enabling emergency personnel to operate accurately and in a standardized manner during the tense rescue process, thereby significantly improving the emergency response speed and success rate, and gaining precious treatment time for patients; In the present invention, the resource scheduling and coordination module tracks AED equipment and personnel information in real time, plans the deployment path and attendance route in combination with road conditions, and connects with the hospital system to achieve efficient coordination of first aid. By analyzing and evaluating parameters through intelligent algorithms, the optimal AED equipment and first aid personnel deployment plan is determined, which avoids waste and repeated deployment of resources and improves the efficiency of the use of first aid resources. At the same time, the seamless connection with the hospital system enables the hospital to make first aid preparations in advance, ensures that patients can receive further medical treatment in a timely manner, and improves the coordination and coherence of the overall first aid system; In the present invention, the first aid training and evaluation module comprehensively records and analyzes the performance data of first aid personnel, screens the personnel to be trained after quantitative evaluation, and classifies them for targeted training. Through detailed data records and scientific evaluation methods, it is possible to accurately identify the deficiencies of first aid personnel in actual operations, providing an accurate basis for training. The use of various forms of training methods, such as video demonstrations, animation simulations, virtual reality training, etc., improves the fun and effectiveness of training, helps to improve the professional skills and first aid quality of first aid personnel, and thus provides patients with higher quality first aid services. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings; Figure 1 This is the overall system block diagram of the present invention. DETAILED DESCRIPTION

[0017] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] It should be understood that the terms “include” and “comprising” used in the specification and claims of the present disclosure indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0019] It should also be understood that the terms used in this disclosure are only for the purpose of describing specific embodiments and are not intended to limit the disclosure. As used in this disclosure and claims, the singular forms of "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should also be further understood that the term "and / or" used in this disclosure and claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations.

[0020] like Figure 1 As shown, the AED intelligent emergency management system based on the Internet of Things includes a data acquisition and transmission module, an emergency decision support module, a resource scheduling and coordination module, and an emergency training and evaluation module; The data acquisition and transmission module is used to integrate multiple sensors into AED equipment and field terminals, collect and transmit patient, equipment and environmental data to the cloud, and provide a basis for subsequent decision-making; A variety of sensors are integrated into the AED device, including high-precision ECG sensors for detecting the patient's heart rhythm, environmental sensors for measuring ambient temperature and humidity (to ensure that the device is operating in a suitable environment), and device status sensors for monitoring the device's own status (such as battery power, electrode connection status, etc.). These sensors transmit data to the cloud server in real time via low-power Bluetooth or NB-IoT modules; at the same time, a portable data acquisition terminal is set up at the emergency site, carried by emergency personnel or installed on an emergency cart; the terminal is equipped with a camera and a microphone to collect on-site images, videos, and voice information, including the patient's appearance symptoms and surrounding environment, and upload them to the server instantly via the 4G / 5G network, providing comprehensive on-site data for remote medical guidance and subsequent analysis.

[0021] The emergency decision support module is used to process the collected data using big data and AI algorithms, determine the type and severity of cardiac arrest, and generate personalized emergency plans and push them to the emergency personnel's equipment; After receiving data from AED equipment and on-site collection terminals, it is processed using big data analysis and artificial intelligence algorithms. Based on different historical emergency cases and medical research data, an intelligent diagnosis model is established, which can quickly determine the type of cardiac arrest (such as ventricular fibrillation, pulseless ventricular tachycardia, etc.) and severity of the patient according to the patient's real-time heart rhythm data, symptoms and environmental factors; the specific process is as follows: In the data receiving stage, the patient's real-time heart rhythm data obtained from the AED device is presented in the form of a continuous electrocardiogram (ECG) waveform with a sampling frequency of , in the time interval The discrete data points collected in , ,in ; If the value of N is a decimal, it is rounded up according to the rules; The symptom data provided by the on-site collection terminal is coded according to the pre-set symptom classification To quantify; Represents loss of consciousness. Represents respiratory arrest, etc.; each symptom has a specific numerical weight ,Environmental factors including temperature T and humidity H are recorded with actual measured values; Then perform data preprocessing. For the heart rate data, use wavelet transform to perform denoising and select the corresponding wavelet basis function , using Daubechies wavelet, Decompose and reconstruct to remove high-frequency noise interference and obtain pure heart rhythm signals At the same time, the symptom data and environmental factor data were normalized so that their values ​​were within the range of Between, for subsequent calculations; In terms of intelligent diagnosis model construction, a deep convolutional neural network (CNN) and a long short-term memory network (LSTM) were introduced; the CNN layer is used to automatically extract the features of heart rhythm data, and its convolution kernel size is , the step length is , through the combination of multiple convolutional layers and pooling layers, the raw heart rate data is converted into a high-level abstract feature vector ; The LSTM layer is used to process the symptom data and environmental factor data in the time series, which have different effects on the diagnosis results at different time points; Suppose the number of LSTM units is , after being processed by the LSTM layer, the feature vector is obtained ; Then, and Perform splicing and fusion to obtain a comprehensive feature vector ; Classification is performed through a fully connected layer (FC) and a softmax classifier to calculate the probability that the patient belongs to different types of cardiac arrest ,in is the category number of cardiac arrest type; for ventricular fibrillation type, the probability calculation formula is: ,in is the score of the corresponding ventricular fibrillation category output by the fully connected layer, is a natural constant; is the score corresponding to different types of cardiac arrest output by the fully connected layer (FC); Represents different category indexes, from 1 to ( is the number of categories of cardiac arrest types); the weights are adjusted by continuously training the model so that when the input data is given, the probability distribution of various types of cardiac arrest can be accurately output, thereby determining the type of cardiac arrest of the patient.

[0022] For severity judgment, a multi-factor regression model was constructed; the patient's cardiac arrest type C (represented by a unique coding form, where different cardiac arrest types, including ventricular fibrillation, pulseless ventricular tachycardia, pulseless electrical activity and ventricular asystole, have C values ​​of 1, 2, 3 and 4 respectively after coding), the sum of symptom weights (in The number of symptoms that occur) and a comprehensive index of environmental factors (obtained by weighted summation of environmental factors such as temperature and humidity) as input, that is, , the model outputs a severity score , the regression formula used is: ;in is the regression coefficient, obtained through historical data training, is the error term; according to The numerical range of the disease is used to divide the severity into different levels, including mild, moderate and severe. In the process of model training, the historical emergency case database is used , Represents the total number of cases in the historical emergency case database, where each case Contains complete heart rhythm data, symptoms, environmental factors, and corresponding diagnosis results and severity labels. Use the stochastic gradient descent (SGD) algorithm or its variants, including Adagrad and Adadelta, to continuously adjust the model parameters to minimize the loss function between the predicted results and the true labels; specifically: For classification tasks, the cross entropy loss function is used: in It is a case The true category label in 1 if the class is valid, 0 otherwise). is the probability predicted by the model; for regression tasks, the mean square error (MSE) loss function is used: in It is a case The true severity score in It is the score predicted by the model. Through continuous iterative training, the model can quickly and accurately determine the type and severity of the patient's cardiac arrest when faced with new input data, providing a reliable basis for the subsequent generation of personalized first aid plans.

[0023] Based on the diagnosis results, a personalized first aid plan is automatically generated, including AED defibrillation energy setting recommendations, cardiopulmonary resuscitation (CPR) compression frequency and depth guidance, and pushed to the first aid personnel's smart devices through voice prompts and mobile applications to ensure that first aid personnel can operate accurately and in a standardized manner during the tense rescue process; the defibrillation energy setting recommendations include: After the type of cardiac arrest (such as ventricular fibrillation, pulseless ventricular tachycardia, etc.) and severity of the patient are determined through the intelligent diagnostic model, the AED defibrillation energy setting recommendation is determined according to different situations; For ventricular fibrillation (VF): Assume the severity of the patient It is divided into three levels: mild, moderate and severe, corresponding to different defibrillation energy ranges; if In the mild range ( ), initial defibrillation energy Set to a lower value, the formula is: ,in It is the basic defibrillation energy of the AED device (usually the lowest energy preset by the device). is a The random factors between are used to consider individual differences among patients and environmental uncertainties. Their distribution can be obtained through training models and can be initially set to 0.5 for fine-tuning. ), defibrillation energy for: , where is another random factor, which can also be optimized by training the model and is initially set to 0.6; for severe cases ( ), defibrillation energy Set to: ,in Initially set to 0.7; For pulseless ventricular tachycardia (VT): If the VT is sustained and the patient is ,in Indicates the severity of the patient in the case of pulseless ventricular tachycardia (VT); defibrillation energy use: For more severe pulseless ventricular tachycardia , defibrillation energy Set to: .

[0024] CPR compression rate and depth guidance: According to the severity and weight of the patient (Unit: kg), providing first aid personnel with personalized CPR compression frequency and compression depth Instructions: Compression frequency : Establish a rate adjustment factor taking into account the severity of the patient and the type of cardiac arrest For mild patients, Set to 1.0; moderate patients, Set to 1.1; severe patients, Set to 1.2; then the pressing frequency calculation formula is: ,in is the basic compression frequency, which is 100-120 times / minute; compression depth :For adult patients, the compression depth is related to body weight. The compression depth formula is: ; In the formula, for mild patients, is 0.5 and is 2.5 (unit: cm); for moderate patients, is 0.6 and is 3.0; for severe patients, is 0.7 and is 3.5; For children and infants, the compression depth is adjusted based on age and weight. Between 1 and 8 years old), the formula is modified to: ,in and Different settings are made according to different age groups. For children aged 1-3, is 1.0 and is 1.5; for children aged 4-8, is 0.8 and is 2.0.

[0025] The resource scheduling and coordination module is used to track AED equipment and personnel information, plan dispatch routes and attendance routes based on road conditions, and connect with the hospital system to achieve efficient coordination of emergency rescue; Track the location and status of AED devices in real time and display them intuitively on a map through geographic information system (GIS) technology; when an emergency call is received, quickly search for available AED devices in the surrounding area based on the location of the call, and combine traffic information (obtained by connecting to the intelligent transportation system of the traffic management department) and the distribution of emergency personnel; after searching for available AED devices in the surrounding area, determine the optimal AED device by analyzing and evaluating parameters, including: Distance parameters: Taking the call location as the origin, the actual distance between each available AED device and the call location is obtained through GIS technology, and recorded as the distance judgment value JC; Traffic condition parameters: Connect with the intelligent transportation system of the traffic management department to obtain traffic condition information along the way from each AED device to the call location; and divide traffic conditions into different levels, including smooth, light congestion, moderate congestion, and heavy congestion, and assign different quantitative values ​​to different levels of road conditions, recorded as road evaluation values ​​LP; Equipment status parameters: Real-time monitoring of the status of the AED device, including battery power and electrode validity period, and quantify it to a value between 0 and 1 to obtain the status evaluation value ZK; The distance judgment value JC, road evaluation value LP and condition evaluation value ZK are normalized and then input into the following formula: To obtain the optimal judgment value YPZ, and use the obtained optimal judgment value as the selection standard for AED equipment, where are respectively the preset weight coefficients of the distance judgment value JC, the road evaluation value LP and the condition evaluation value ZK, and The sum is 1; the superiority values ​​YPZ obtained by analyzing different AED devices are sorted according to size, and the AED device with the largest superiority value YPZ is selected as the final selected device.

[0026] At the same time, a communication connection is established with the emergency system of surrounding medical institutions to inform the hospital of the patient's condition and estimated arrival time in advance, so that the hospital can make corresponding emergency preparations, achieve seamless connection from the emergency site to the hospital, and improve overall emergency efficiency.

[0027] The first aid training and assessment module is used to record and analyze the performance data of first aid personnel, screen the personnel to be trained after quantitative assessment, and carry out targeted training according to categories; Comprehensively record the detailed data of emergency personnel in the process of performing tasks in various modules, including emergency decision support module and resource scheduling and coordination module, including emergency response time, execution order and timestamp of operation steps, equipment operation parameters (such as AED defibrillation energy setting value, actual value of cardiopulmonary resuscitation compression frequency and depth), and classify and organize these data according to different emergency events and personnel; analyze and quantify the obtained data to obtain performance parameters, including: Timeliness of emergency response: The time interval from receiving an emergency call to the time when emergency personnel begin to take effective emergency measures (such as starting AED or starting cardiopulmonary resuscitation) is used as a measurement indicator, in seconds. Set standard response time thresholds for different scenarios, and calculate the difference, recorded as the negative rating value, and compare the negative rating value with several preset negative rating value intervals. Several negative rating value intervals are respectively set with different response scores to determine the response score of the emergency personnel, and use this response score as a measure of the timeliness of emergency response by emergency personnel; Operation accuracy: including defibrillation energy setting accuracy, electrode patch pasting accuracy and operation process integrity. The defibrillation energy setting accuracy is calculated by calculating the deviation ratio between the defibrillation energy actually set by the emergency personnel and the energy value recommended by the system, recorded as the setting ratio; the electrode patch pasting accuracy uses image recognition technology or sensor feedback to determine whether the electrode patch pasting position meets the standard; based on the standard pasting area, calculate the deviation area or distance ratio of the actual pasting position, recorded as the pasting ratio; the operation process integrity statistics the completeness of the emergency personnel to complete the AED standard operation process (such as power on, self-test, electrode patch pasting, heart rhythm analysis, discharge, etc.) The degree of operation accuracy is measured by analyzing the ratio of the number of missed steps of the first aid personnel to the total number of steps, and recorded as the operation ratio; the obtained set ratio, paste ratio and operation ratio are normalized, and the set ratio, paste ratio and operation ratio are used as the three sides of the triangle to establish a triangle, and a triangular pyramid model is established with the preset correction factor as the height, and the volume of the triangular pyramid model is calculated and recorded as the balance value, which is used as the standard for measuring operation accuracy; Cardiopulmonary resuscitation operation accuracy: The compression frequency accuracy is calculated by calculating the deviation ratio between the actual compression frequency and the compression frequency recommended by the system according to the patient's condition, and recorded as the frequency deviation value; The compression depth accuracy obtains the actual compression depth average value through the pressure sensor, calculates the deviation ratio from the standard depth range, and records it as the depth deviation value; The compression position accuracy uses the human body model or sensor positioning technology to determine whether the compression position is accurate at the junction of the middle and lower 1 / 3 of the sternum, and is measured according to the degree of deviation, and recorded as the deviation value; The obtained frequency deviation value, depth deviation value and position deviation value are calibrated as pp, sm and pw respectively, and then normalized and input into the following formula: To obtain the heart judgment value xzp, and use this heart judgment value xzp as a standard for measuring the accuracy of cardiopulmonary resuscitation operation; then normalize the obtained response score xpi, balance value chy, and heart judgment value xzp and enter them into the following formula: To obtain the comprehensive evaluation value ziu, where are the preset weight coefficients of the response score xpi, the balance value chy, and the judgment value xzp, respectively, and this comprehensive evaluation value is used to measure the performance of the first aid personnel during the first aid operation; the comprehensive evaluation value ziu obtained by analyzing all first aid personnel is compared with the preset comprehensive evaluation threshold. When the comprehensive evaluation value ziu of the first aid personnel is less than the preset comprehensive evaluation threshold, the first aid personnel is marked as a person to be trained.

[0028] The personnel to be trained are classified into categories, including the general public, professional first aid personnel, and AED equipment maintenance personnel. The training content includes the basic operating principles of AED equipment, detailed steps of first aid skills (CPR, Heimlich maneuver, etc.), coping strategies in different cardiac arrest scenarios, etc. It is presented in various forms such as video demonstrations, animation simulations, and virtual reality (VR) training to make the training more interesting and effective.

[0029] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. AED intelligent emergency management system based on the Internet of Things, characterized by: include: Data collection and transmission module, which is used to integrate multiple sensors into AED equipment and field terminals, collect and transmit patient, equipment and environmental data to the cloud to provide a basis for subsequent decision-making; The emergency decision support module is used to process the collected data using big data and AI algorithms, determine the type and severity of cardiac arrest, and generate personalized emergency plans and push them to the emergency personnel's equipment; Resource scheduling and coordination module, used to track AED equipment and personnel information, plan dispatch routes and attendance routes based on road conditions, and connect with the hospital system to achieve efficient coordination of emergency rescue; The first aid training and evaluation module is used to record and analyze the performance data of first aid personnel, screen the personnel to be trained after quantitative evaluation, and carry out targeted training according to categories. The specific process is as follows: After receiving data from AED devices and on-site collection terminals, it is processed using big data analysis and artificial intelligence algorithms; Based on different historical emergency cases and medical research data, an intelligent diagnosis model is established to determine the type and severity of a patient's cardiac arrest based on the patient's real-time heart rhythm data, symptoms, and environmental factors; Based on the diagnosis results, a personalized first aid plan is automatically generated, including AED defibrillation energy setting recommendations and CPR compression frequency and depth guidance, and pushed to the first aid personnel's smart devices through voice prompts and mobile applications.

2. The AED intelligent emergency management system based on the Internet of Things according to claim 1 is characterized in that: The specific operation steps of the emergency decision support module for determining the type of cardiac arrest of the patient are as follows: First, the patient's real-time heart rhythm data presented in the form of a continuous electrocardiogram waveform is obtained from the AED device, and discrete data points are collected within a preset time interval. At the same time, the symptom data provided by the on-site collection terminal is quantified through a preset code. Each symptom has a corresponding numerical weight, and environmental factors record temperature and humidity. Then, the heart rhythm data was denoised using Daubechies wavelet to obtain a pure heart rhythm signal. The symptom data and environmental factor data were also normalized to the range of 0-1. Then, an architecture combining a deep convolutional neural network and a long short-term memory network is constructed. The deep convolutional neural network layer uses convolution kernels and step sizes to extract features of heart rhythm data and convert them into high-level abstract feature vectors. The long short-term memory network layer processes symptom manifestations and environmental factor data in time series to obtain feature vectors. The two are concatenated and fused, and then classified by a fully connected layer and a softmax classifier to calculate the probability that the patient belongs to different types of cardiac arrest. The type is accurately determined by continuously training the model and adjusting the weights.

3. The AED intelligent emergency management system based on the Internet of Things according to claim 1 is characterized in that: The specific operation steps of the emergency decision support module for judging the severity are as follows: For severity judgment, the emergency decision support module constructs a multi-factor-based regression model, taking the patient's cardiac arrest type, the sum of symptom weights, and a comprehensive index of environmental factors as input; The sum of symptom weights is calculated based on the number of symptoms and their corresponding weights, and the comprehensive index of environmental factors is obtained by weighted summation of temperature, humidity, etc.; The model outputs a severity score, which is trained through a historical emergency case database and continuously adjusts model parameters using the stochastic gradient descent algorithm or its variants. The cross entropy loss function is used for classification tasks, and the mean square error loss function is used for regression tasks to minimize the loss between the prediction and the true label. The severity levels are divided into mild, moderate, and severe according to the score range.

4. The AED intelligent emergency management system based on the Internet of Things according to claim 1 is characterized in that: The first aid decision support module includes the following steps: If the patient is in ventricular fibrillation, the severity of the condition will be classified into different levels; In mild cases, the initial defibrillation energy is increased by a preset ratio based on the basic defibrillation energy of the AED device, and the proportionality coefficient contains a random factor distributed through the training model; In moderate and severe cases, only the random factors of the proportional coefficients are different and have preset initial values; For pulseless ventricular tachycardia, the defibrillation energy is calculated using different formulas for sustained low-severity and high-severity ventricular tachycardia; In terms of cardiopulmonary resuscitation, the compression frequency is determined by the adjustment factor based on the severity of the patient and the type of cardiac arrest, combined with the basic compression frequency; The compression depth for adults is related to their weight, and different coefficients correspond to different degrees of severity. The compression depth for children and infants is adjusted according to age and weight, and each age group is analyzed using different calculation formulas.

5. The AED intelligent emergency management system based on the Internet of Things according to claim 1 is characterized in that: The specific operation steps of the resource scheduling and coordination module are as follows: Track the location and status of AED devices in real time and display them intuitively on a map through geographic information system technology; when an emergency call is received, quickly search for available AED devices in the surrounding area based on the location of the call, combined with traffic information and the distribution of emergency personnel; After searching for available AED devices in the surrounding area, determine the optimal AED device by analyzing and evaluating parameters; At the same time, a communication connection is established with the emergency system of surrounding medical institutions to inform the hospital of the patient's condition and estimated arrival time in advance, achieving seamless connection from the emergency site to the hospital and improving overall emergency efficiency.

6. The AED intelligent emergency management system based on the Internet of Things according to claim 5 is characterized in that: The specific operation steps of the resource scheduling and coordination module to determine the optimal AED device by analyzing the evaluation parameters are as follows: Evaluation parameters include: Distance parameter: The call location is taken as the origin. The actual distance between each available AED device in the surrounding area and the call location is obtained through GIS technology and recorded as the distance judgment value JC; Traffic condition parameters: Connect with the intelligent traffic system of the traffic management department to obtain traffic condition information along the way from each AED device to the emergency call location; classify traffic conditions into different levels, including smooth, light congestion, moderate congestion, and heavy congestion, and assign different quantitative values ​​to different levels of traffic conditions, recorded as road rating values ​​LP; Equipment status parameters: real-time monitoring of the status of the AED equipment, including battery power and electrode validity period, and quantification to the sum of values ​​between 0 and 1 to obtain the status evaluation value ZK; The distance judgment value JC, road evaluation value LP and condition evaluation value ZK are normalized and then input into the following formula: To obtain the optimal judgment value YPZ, and use the obtained optimal judgment value as the selection standard for AED equipment, where are respectively the preset weight coefficients of the distance judgment value JC, the road evaluation value LP and the condition evaluation value ZK, and The sum is 1; The superiority values ​​YPZ obtained by analyzing different AED devices are sorted by size, and the AED device with the largest superiority value YPZ is selected as the final selected device.

7. The AED intelligent emergency management system based on the Internet of Things according to claim 1 is characterized in that: The execution process of the first aid training and assessment module is as follows: Comprehensively record the detailed data of emergency personnel in the process of performing tasks in each module, including emergency response time, execution order and timestamp of operation steps, and equipment operation parameter information in the emergency decision support module and resource scheduling and coordination module, and classify and organize these data according to different emergency events and personnel; Analyze and quantify the obtained data to obtain performance parameters, use the performance parameters to determine the comprehensive evaluation value of the first aid personnel, compare the comprehensive evaluation value ziu obtained by analyzing all first aid personnel with the preset comprehensive evaluation threshold, and when the comprehensive evaluation value ziu of the first aid personnel is less than the preset comprehensive evaluation threshold, mark the first aid personnel as a person to be trained; The personnel to be trained are classified into first aid personnel and AED equipment maintenance personnel. The training content includes the operating principles of AED equipment, the steps of first aid skills, and the response strategies in different cardiac arrest scenarios. It is presented in various forms such as video demonstrations, animation simulations, and virtual reality training.

8. The AED intelligent emergency management system based on the Internet of Things according to claim 7 is characterized in that: The process of determining the comprehensive evaluation value of the first aid personnel using the performance parameters in the first aid training and evaluation module is as follows: Performance parameters include: Timely emergency response: The time interval from receiving an emergency call to the time when emergency personnel start to take effective emergency measures is used as a measurement indicator; standard response time thresholds are set in different scenarios, and the difference is calculated, recorded as the negative evaluation value, and the negative evaluation value is compared with several preset negative evaluation value intervals. Several negative evaluation value intervals are respectively set with different response scores to determine the response score of the emergency personnel, and this response score is used to measure the timeliness of the emergency response of the emergency personnel; Operation accuracy: including the accuracy of defibrillator energy setting, electrode pasting accuracy and operation process integrity. The accuracy of defibrillator energy setting is calculated by calculating the deviation ratio between the defibrillator energy actually set by the first aid personnel and the energy value recommended by the system, which is recorded as the setting ratio. The electrode pasting accuracy uses image recognition technology or sensor feedback to determine whether the electrode pasting position meets the standard. The deviation area or distance ratio of the actual pasting position is calculated based on the standard pasting area, which is recorded as the pasting ratio. The operation process integrity counts the completeness of the first aid personnel in completing the AED standard operation process, which is measured by analyzing the ratio of the number of missed steps of the first aid personnel to the total number of steps, and is recorded as the operation ratio. After normalizing the obtained set ratio, pasting ratio and operation ratio, the set ratio, pasting ratio and operation ratio are used as the three sides of the triangle to establish a triangle, and a triangular pyramid model is established with the preset correction factor as the height, and the volume of the triangular pyramid model is calculated and recorded as the balance value, which is used as a standard to measure the accuracy of the operation; Accuracy of cardiopulmonary resuscitation operation: The accuracy of compression frequency is calculated by calculating the deviation ratio between the actual compression frequency and the compression frequency recommended by the system according to the patient's condition, and recorded as the frequency deviation value; the accuracy of compression depth is obtained by obtaining the average value of the actual compression depth through the pressure sensor, calculating the deviation ratio from the standard depth range, and recording it as the depth deviation value; the accuracy of compression position uses the human body model or sensor positioning technology to determine whether the compression position is accurately at the junction of the middle and lower 1 / 3 of the sternum, and measures it according to the degree of deviation, which is recorded as the deviation value; the obtained frequency deviation value, depth deviation value and deviation value are calibrated as pp, sm and pw respectively, and after normalization, they are entered into the following formula: To obtain the heart judgment value xzp, and use the heart judgment value xzp as a standard for measuring the accuracy of cardiopulmonary resuscitation operation; Then normalize the obtained response score xpi, balance value chy, and judgment value xzp and enter them into the following formula: To obtain the comprehensive evaluation value ziu, where They are the preset weight coefficients of response score xpi, balance value chy, and judgment value xzp respectively, and this comprehensive evaluation value is used as the standard to measure the performance of emergency personnel in emergency operations.

Images