Intelligent AED trigger system

Abstract

The invention relates to the technical field of medical first aid, in particular to an intelligent AED trigger system which comprises a data acquisition and sensing unit, an intelligent matching and decision unit, a communication transmission unit, an actuator unit and a feedback and learning unit. Standardized data is generated through a work order creating module to guarantee integrity and traceability, high-precision position sensing is achieved by means of multi-source positioning, risk grading is completed through information preprocessing, an optimal AED operator is accurately screened through a distributed database and a multi-dimensional self-adaptive matching algorithm, a scheduling sequence is dynamically generated, and the accuracy and reliability of the system are improved. The AED scheduling response speed, the matching accuracy and the rescue operation specialty are remarkably improved, and the rescue timeliness and the success rate in emergency scenes such as sudden cardiac arrest are effectively guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical first aid, in particular to an intelligent AED triggering system. BACKGROUND

[0002] Intelligent AED (Automated External Defibrillator) is a new generation of first aid equipment and supporting system based on the core function of traditional AED "automatic analysis of heart rhythm and defibrillation", combined with artificial intelligence, Internet of Things, big data, positioning and navigation technologies, and is an indispensable part of the first aid system. However, the traditional AED still faces many technical bottlenecks in practical application, and it is difficult to form an efficient rescue closed loop, which seriously restricts the play of its first aid value. In the prior art:

[0003] 1. Insufficient positioning accuracy: the position information of AED equipment and first aid personnel often deviates;

[0004] 2. Serious response delay: traditional dispatch relies on manual coordination, and the average response time of first aid request is 47 seconds, far exceeding the time limit of "golden 4 minutes". The system needs to transfer the dispatcher, ambulance and hospital several times, and the redundant links cause the delay of key decisions;

[0005] 3. Resource state update lag: AED device power, fault state and personnel availability rely on manual reporting.

[0006] Therefore, the present application provides an intelligent AED triggering system to solve the above technical problems. SUMMARY

[0007] The present application aims to provide an intelligent AED triggering system. The work order creation module of the present application generates standardized data to ensure integrity and traceability, realizes high-precision position sensing with multi-source positioning, completes risk grading after information preprocessing, and accurately selects the optimal AED operator and dynamically generates a dispatch sequence relying on a distributed database and a multi-dimensional adaptive matching algorithm. Through multi-modal communication and redundant transmission mechanism, the present application ensures efficient and reliable transmission of instructions, provides support for on-site rescue by using high-priority reminders, AR navigation and remote consultation, and realizes continuous optimization of the system by using data warehouse storage, performance board evaluation and algorithm engine iteration, which significantly improves the AED dispatch response speed, matching accuracy and rescue operation professionalism, and effectively guarantees the rescue timeliness and success rate in first aid scenes such as cardiac arrest.

[0008] To achieve the above purpose, the present application provides the following technical solutions:

[0009] The application provides an intelligent AED triggering system, comprising a data acquisition and perception unit, an intelligent matching and decision unit, a communication transmission unit, an executor unit, and a feedback and learning unit, wherein:

[0010] The data acquisition and perception unit is used for acquiring and uploading the occurrence location and key information of a cardiac arrest event in real time.

[0011] The intelligent matching and decision unit is used for dynamically adjusting the scene perception weight, pre-judging the space-time trajectory, and making a multi-dimensional adaptive matching algorithm for multi-person collaborative decision-making according to the event location, patient information, and real-time resource state, so as to quickly calculate and decide one or more optimal AED operators for scheduling.

[0012] The communication transmission unit is used for transmitting instructions and state information between the intelligent matching and decision unit and the executor unit by using a multi-modal communication protocol and a redundant communication mechanism.

[0013] The executor unit is used for receiving a high-priority reminder alarm and providing navigation and guidance support through the intelligent terminal of a registered operator, and the operator carries out on-site rescue action according to the instructions.

[0014] The feedback and learning unit is used for collecting rescue process data, evaluating system performance, and optimizing the matching decision algorithm.

[0015] The data acquisition and perception unit comprises a work order creation module, a multi-source positioning module, and an information preprocessing module, wherein:

[0016] The work order creation module is used for automatically receiving a 120 emergency call and generating a standardized emergency work order containing the event location, patient information, and time stamp.

[0017] The multi-source positioning module is used for fusing multi-source signals of GPS / Beidou, base station positioning, and network positioning to obtain high-precision geographic location information of the incident site.

[0018] The information preprocessing module is used for performing preliminary analysis and risk level classification on the emergency event by using voice recognition and keyword extraction technology.

[0019] The intelligent matching and decision unit comprises a registered personnel database, a multi-dimensional matching algorithm module, and a dynamic scheduling decision module, wherein:

[0020] The registered personnel database is used for storing and managing the identity qualification, skill level, real-time location, and equipment state information of all registered AED operators.

[0021] The multi-dimensional matching algorithm module is used for executing the multi-dimensional adaptive matching algorithm of scene perception weight dynamic adjustment, space-time trajectory pre-judgment, and multi-person collaborative decision-making to comprehensively calculate the optimal responder.

[0022] The dynamic scheduling decision module is used to generate the final scheduling instruction and manage the multi-level response mechanism.

[0023] The multi-dimensional matching algorithm module executes a multi-dimensional adaptive matching algorithm that dynamically adjusts scene perception weights, predicts spatiotemporal trajectories, and performs multi-person collaborative decision-making to comprehensively calculate the optimal responder. The specific operations are as follows:

[0024] A1: Obtain real-time environmental data of the incident site, including population density, traffic congestion index and weather level, and assign dynamic weights to the three core matching factors of "geographical distance", "skill level rating" and "real-time availability" according to preset rules. When the traffic congestion index is ≥80, the weight of "estimated arrival time" replaces the weight of "geographical distance" and is increased to 50%.

[0025] A2: Retrieve the location data of the registered AED operator for the past 5 minutes, calculate the movement speed and direction vector, combine road network data to predict the time deviation value of their arrival at the scene within 3 minutes, and filter candidates with a deviation value ≤ 60 seconds;

[0026] A3: Ranking candidates based on a dynamic weighted composite score formula:

[0027] ① When the traffic congestion index is <80, the score = geographical location weight score × 40% + skill level score × 30% + predicted arrival time score × 30%;

[0028] ② When the traffic congestion index is ≥80, the score = predicted arrival time score × 50% + skill level score × 30% + real-time availability score × 20%;

[0029] Generate a scheduling sequence based on the event risk level:

[0030] ①For high-risk events, generate 3 primary responders + 3 backup responders;

[0031] ② For a medium-risk event, generate 2 primary responders + 2 backup responders;

[0032] ③ For low-risk events, generate 1 primary responder + 2 backup responders;

[0033] If the primary responder fails to provide confirmation within 15-25 seconds, the first backup responder will be automatically promoted to the primary responder.

[0034] The multi-level response mechanism in the dynamic scheduling decision module is as follows:

[0035] ① Level 1 Alert "APP Strong Reminder": Automatically escalates if there is no response within 10 seconds;

[0036] ② Secondary alarm "SMS + phone call": 15 seconds of non-response triggers a backup staff call;

[0037] ③ Tertiary alarm "Competitive response": notify 3-5 people at the same time, and the first acceptor gets the task.

[0038] The communication transmission unit includes a multi-modal protocol stack, a redundant transmission control module, and a state confirmation relay module, wherein:

[0039] The multi-modal protocol stack supports multiple communication protocols such as 5G, APP push, and SMS.

[0040] The redundant transmission control module automatically and seamlessly switches to a backup channel when the main communication channel fails.

[0041] The state confirmation relay module receives feedback states from the executor unit and returns them to the intelligent matching and decision unit in real time.

[0042] The executor unit includes a strong reminder receiving module, an AR navigation and guidance module, and a remote video consultation module, wherein:

[0043] The strong reminder receiving module receives and displays rescue instructions on the intelligent terminal in the highest priority mode.

[0044] The AR navigation and guidance module provides real-time navigation and AR visualization guidance for responders to go to the incident location and the nearest AED storage point.

[0045] The remote video consultation module establishes a low-latency video call with 120 emergency center experts during the rescue process to obtain real-time guidance.

[0046] The highest priority mode includes continuous strong vibration, sharp alarm sound, and full-screen pop-up window.

[0047] The feedback and learning unit includes a rescue data warehouse, a performance evaluation board, and an algorithm optimization engine, wherein:

[0048] The rescue data warehouse collects and stores response time, arrival time, operation steps, and medical result data for each rescue task.

[0049] The performance evaluation board visualizes the core indicators of system average response time and matching success rate and generates evaluation reports.

[0050] The algorithm optimization engine continuously trains and optimizes the algorithm model of the intelligent matching and decision unit based on historical rescue data through machine learning.

[0051] The algorithm optimization engine continuously trains and optimizes the algorithm model of the intelligent matching and decision unit based on historical rescue data through machine learning, and the specific operation is as follows:

[0052] B1: Based on the response time, arrival time and rescue result data in the historical rescue data, the weight coefficients and the pre-judgment model in the multi-dimensional matching algorithm module are iteratively updated to minimize the average response time as the optimization goal;

[0053] B2: The verified and optimized model parameters and rules are deployed to the online scheduling environment of the intelligent matching and decision unit;

[0054] B3: Part of the scheduling request is allocated to the test group decided by the optimized model, and the response efficiency is compared with the control group decided by the original model to evaluate the optimization effect.

[0055] Compared with the prior art, the beneficial effects of the present application are:

[0056] The present application generates standardized data through the work order creation module to ensure integrity and traceability, realizes high-precision position sensing with multi-source positioning, completes risk grading after information preprocessing, and relies on a distributed database and a multi-dimensional adaptive matching algorithm to accurately screen the optimal AED operator and dynamically generate a scheduling sequence. Through multi-modal communication and redundant transmission mechanism, it ensures efficient and reliable transmission of instructions, uses high-priority reminders, AR navigation and remote consultation to provide support for on-site rescue, and finally realizes continuous optimization of the system through data warehouse storage, performance board evaluation and algorithm engine iteration. The present application significantly improves the AED scheduling response speed, matching accuracy and rescue operation professionalism, effectively guarantees the rescue timeliness and success rate in emergency scenes such as cardiac arrest. BRIEF DESCRIPTION OF DRAWINGS

[0057] Fig. 1 The system diagram of the intelligent AED triggering system of the present application.

[0058] Fig. 2 The multi-level response and communication guarantee process diagram in the intelligent AED triggering system of the present application.

[0059] BRIEF DESCRIPTION OF DRAWINGS

[0060] 100. Data Acquisition and Sensing Unit; 101. Work Order Creation Module; 102. Multi-Source Positioning Module; 103. Information Preprocessing Module; 200. Intelligent Matching and Decision-Making Unit; 201. Registered Personnel Database; 202. Multi-Dimensional Matching Algorithm Module; 203. Dynamic Scheduling Decision-Making Module; 300. Communication Transmission Unit; 301. Multimodal Protocol Stack; 302. Redundant Transmission Control Module; 303. Status Confirmation Relay Module; 400. Actuator Unit; 401. Strong Alert Receiving Module; 402. AR Navigation and Guidance Module; 403. Remote Video Consultation Module; 500. Feedback and Learning Unit; 501. Rescue Data Warehouse; 502. Performance Evaluation Dashboard; 503. Algorithm Optimization Engine. Detailed Implementation

[0061] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0062] Example:

[0063] like Figs. 1-2 As shown, this embodiment provides an intelligent AED triggering system, including a data acquisition and sensing unit 100, an intelligent matching and decision-making unit 200, a communication transmission unit 300, an actuator unit 400, and a feedback and learning unit 500, wherein:

[0064] Data acquisition and sensing unit 100: used to acquire and upload the location and key information of cardiac arrest events in real time;

[0065] Intelligent Matching and Decision Unit 200: Based on the event location, patient information, and real-time resource status, it uses a multi-dimensional adaptive matching algorithm that dynamically adjusts scene perception weights, predicts spatiotemporal trajectories, and enables multi-person collaborative decision-making to quickly calculate and decide on one or more optimal AED operators for dispatch.

[0066] Communication transmission unit 300: used to transmit instructions and status information between intelligent matching and decision-making unit 200 and actuator unit 400 using multimodal communication protocols and redundant communication mechanisms;

[0067] Actuator Unit 400: Used to receive high-priority alerts and provide navigation and guidance support through the registered operator's smart terminal, so that the operator can carry out on-site rescue operations according to the instructions;

[0068] Feedback and Learning Unit 500: Used to collect data on the rescue process, evaluate system performance, and optimize matching decision algorithms.

[0069] The data acquisition and perception unit 100 comprises a work order creation module 101, a multi-source positioning module 102, and an information preprocessing module 103, wherein:

[0070] The work order creation module 101 is used for automatically receiving 120 emergency calls and generating a standardized emergency work order containing the event location, patient information, and time stamp;

[0071] The multi-source positioning module 102 is used for fusing multi-source signals of GPS / Beidou, base station positioning, and network positioning to obtain high-precision geographic location information of the incident site;

[0072] The information preprocessing module 103 is used for performing preliminary analysis and risk level classification of emergency events through voice recognition and keyword extraction technology.

[0073] The intelligent matching and decision-making unit 200 comprises a registered personnel database 201, a multi-dimensional matching algorithm module 202, and a dynamic scheduling decision-making module 203, wherein:

[0074] The registered personnel database 201 is used for storing and managing the identity qualifications, skill levels, real-time locations, and device status information of all registered AED operators;

[0075] The multi-dimensional matching algorithm module 202 is used for executing a multi-dimensional adaptive matching algorithm for scene perception weight dynamic adjustment, spatiotemporal trajectory prediction, and multi-person collaborative decision-making to comprehensively calculate the optimal responder;

[0076] The dynamic scheduling decision-making module 203 is used for generating final scheduling instructions and managing multi-level response mechanisms.

[0077] The multi-dimensional adaptive matching algorithm in the multi-dimensional matching algorithm module 202 for scene perception weight dynamic adjustment, spatiotemporal trajectory prediction, and multi-person collaborative decision-making comprehensively calculates the optimal responder, with the specific operations as follows:

[0078] A1: Obtain real-time environmental data of the incident site, including population density, traffic congestion index, and weather level, and assign dynamic weights to the three core matching factors of "geographic location distance", "skill level rating", and "real-time availability" according to preset rules, wherein when the traffic congestion index ≥ 80, the "estimated arrival time" weight replaces the "geographic location distance" weight and is increased to 50%;

[0079] A2: Retrieve the positioning data of registered AED operators within the last 5 minutes, calculate the moving speed and direction vector, and predict the time deviation value of the incident site within 3 minutes combined with road network data to screen candidates with a deviation value ≤ 60 seconds;

[0080] A3: Score and sort the candidates according to the dynamic weight comprehensive score formula:

[0081] ① When traffic congestion index < 80, score = geographical location weight score x 40% + skill level score x 30% + pre-judgment arrival time efficiency score x 30%;

[0082] ② When traffic congestion index ≥ 80, score = pre-judgment arrival time efficiency score x 50% + skill level score x 30% + real-time availability score x 20%;

[0083] Generate dispatch sequence according to event risk level:

[0084] ① High-risk event generates 3 main responders + 3 standby responders;

[0085] ② Medium-risk event generates 2 main responders + 2 standby responders;

[0086] ③ Low-risk event generates 1 main responder + 2 standby responders;

[0087] And when the main responder does not respond within 15-25 seconds, the first standby responder is automatically upgraded to the main responder.

[0088] The multi-level response mechanism in the dynamic dispatch decision module 203 is:

[0089] ① First-level alarm "APP strong reminder": automatically upgrade if not responded within 10 seconds;

[0090] ② Second-level alarm "SMS + phone": trigger standby personnel call if not responded within 15 seconds;

[0091] ③ Third-level alarm "competitive response": notify 3-5 people at the same time, and the first acceptor gets the task.

[0092] The communication transmission unit 300 includes a multi-modal protocol stack 301, a redundant transmission control module 302, and a state confirmation relay module 303, wherein:

[0093] Multi-modal protocol stack 301: used to support multiple communication protocols such as 5G, APP push, and SMS;

[0094] Redundant transmission control module 302: used to automatically switch to a standby channel seamlessly when the main communication channel fails;

[0095] State confirmation relay module 303: used to receive feedback states from the executor unit 400 and return them to the intelligent matching and decision unit 200 in real time.

[0096] The executor unit 400 includes a strong reminder receiving module 401, an AR navigation and guidance module 402, and a remote video consultation module 403, wherein:

[0097] Strong reminder receiving module 401: for receiving and displaying rescue instructions on the intelligent terminal in the highest priority mode;

[0098] AR navigation and guidance module 402: for providing real-time navigation and AR visualization guidance for responders to go to the incident site and the nearest AED storage point;

[0099] Remote video consultation module 403: for establishing a low-latency video call with 120 emergency center experts during the rescue process to obtain real-time guidance.

[0100] The highest priority mode includes continuous strong vibration, sharp alarm sound, and pop-up window covering the full screen.

[0101] The feedback and learning unit 500 includes a rescue data warehouse 501, a performance evaluation board 502, and an algorithm optimization engine 503, wherein:

[0102] Rescue data warehouse 501: for collecting and storing the response time, arrival time, operation steps, and medical result data of each rescue task comprehensively;

[0103] Performance evaluation board 502: for visualizing the core indicators of system average response time and matching success rate, and generating evaluation reports;

[0104] Algorithm optimization engine 503: based on historical rescue data, continuously train and optimize the algorithm model of the intelligent matching and decision unit through machine learning.

[0105] The algorithm optimization engine 503 continuously trains and optimizes the algorithm model of the intelligent matching and decision unit based on historical rescue data through machine learning, and the specific operation is as follows:

[0106] B1: Based on the response time, arrival time, and rescue result data in the historical rescue data, the weight coefficients and prediction model in the multi-dimensional matching algorithm module 202 are iteratively updated to minimize the average response time;

[0107] B2: Deploy the verified and optimized model parameters and rules to the online scheduling environment of the intelligent matching and decision unit 200;

[0108] B3: Distribute part of the dispatching requests to the test group decided by the optimized model, and compare their response efficiency with the control group decided by the original model to evaluate the optimization effect.

[0109] In this embodiment, an intelligent AED triggering system is provided, and the specific method is as follows: first, a work order creation module 101 is used to automatically receive 120 emergency calls and generate a standardized emergency work order containing the event location, patient information and timestamp; the standardized emergency work order contains 12 core fields: event unique ID, incident location latitude and longitude (accurate to 6 decimal places), patient gender / age / underlying disease history (such as coronary heart disease), caller contact information, call timestamp (accurate to milliseconds), preliminary symptom description, dispatch priority, AED requirement identification, affiliated emergency partition, data collection device number, field update time, interface check code, to ensure data integrity and traceability. A multi-source positioning module 102 is used to fuse multi-source signals of GPS / Beidou (accuracy 3-15 meters), base station positioning and network positioning to obtain high-precision geographic location information of the incident site; an information preprocessing module 103 is used to preliminarily analyze and risk level classify the emergency event through voice recognition and keyword extraction technology. The risk level classification is divided into three levels: high risk (such as cardiac arrest), medium risk (such as weak breathing), and low risk (such as confusion). Secondly, a registered personnel database 201 is used to store and manage the identity qualifications, skill levels, real-time positions and device status information of all registered AED operators; a distributed database is used to store operator information, including identity qualifications (training certificate number and validity period), skill levels (A-E level, based on training performance and rescue history evaluation), real-time position (updated every 10 seconds), device status (whether carrying a portable AED or knowing the location of a nearby AED), supporting million-level data second-level query. A multi-dimensional matching algorithm module 202 is used to execute a multi-dimensional adaptive matching algorithm for scene awareness weight dynamic adjustment, space-time trajectory prediction and multi-person collaborative decision-making, and comprehensively calculate the optimal responder; the specific operation is as follows: A1: obtain real-time environmental data of the incident site, including population density, traffic congestion index and weather level, and assign dynamic weights to the three core matching factors of "geographic location distance", "skill level rating" and "real-time availability" according to the preset rules, wherein when the traffic congestion index is greater than or equal to 80, the "estimated arrival time" weight replaces the "geographic location distance" weight and is increased to 50%; A2: call the positioning data of the registered AED operators within the last 5 minutes, calculate the moving speed and direction vector, and combine with the road network data to predict the time deviation value of the incident site within 3 minutes, and select the candidates with a deviation value of less than or equal to 60 seconds; the calculation method of the time deviation value T error is as follows:

[0110] T error = |T predicted -T actual |

[0111] Wherein, T predicted is estimated by the following path planning model:

[0112]

[0113] is the length of the i-th segment of the path, is the average speed of the i-th segment based on real-time traffic data, delay jA3: Score and rank the candidates according to the dynamic weight integrated score formula: ① When the traffic congestion index < 80, the score = geographical location weight score × 40% + skill level score × 30% + pre-judgment arrival timeliness score × 30%; ② When the traffic congestion index ≥ 80, the score = pre-judgment arrival timeliness score × 50% + skill level score × 30% + real-time availability score × 20%; generate a dispatch sequence according to the event risk level: ① High-risk events generate 3 main responders + 3 standby responders; ② Medium-risk events generate 2 main responders + 2 standby responders; ③ Low-risk events generate 1 main responder + 2 standby responders; and if the main responder does not respond within 15-25 seconds, the first standby is automatically upgraded to a main responder. The dynamic dispatch decision module 203 is used to generate the final dispatch instruction and manage the multi-level response mechanism. The multi-level response mechanism is: ① Level 1 alarm "APP strong reminder": automatically upgraded within 10 seconds if not responded; ② Level 2 alarm "SMS + phone": trigger standby personnel call if not responded within 15 seconds; ③ Level 3 alarm "competitive response": notify 3-5 people at the same time, and the first acceptor gets the task. The multi-modal protocol stack 301 is used to support various communication protocols such as 5G, APP push, and SMS; the redundant transmission control module 302 is used to automatically switch to the standby channel (SMS + satellite communication, suitable for remote mountainous areas) within 50ms when the main communication channel (5G / 4G) fails; the state confirmation relay module 303 is used to receive the feedback state of the executor unit 400 and return it to the intelligent matching and decision unit 200 in real time. The strong reminder receiving module 401 is used to receive and display the rescue instruction on the intelligent terminal in the highest priority mode; the highest priority mode includes continuous strong vibration, sharp alarm sound, and full-screen pop-up window. After triggering, the terminal automatically enters "emergency mode", starts continuous strong vibration (frequency 200Hz), sharp alarm sound (volume ≥ 85dB), and full-screen pop-up window (displays the incident location map, AED location, and patient information), and the pop-up window can be closed only by clicking "accept" or "reject" to avoid ignoring the instruction. The AR navigation and guidance module 402 is used to provide real-time navigation and AR visual guidance for responders to go to the incident site and the nearest AED storage point; integrates ARCore / ARKit technology to generate a combined navigation path "get AED → go to the scene", superimposes AED storage point markers (distance, direction) and walking / biking navigation lines on the real scene, and automatically displays AR guidance for CPR pressing position and AED electrode patch sticking position after arriving at the scene, reducing the operation error rate. The remote video consultation module 403 is used to establish a low-latency video call with 120 emergency center experts during the rescue process to obtain real-time guidance. WebRTC technology is used to establish a low-latency video call (delay ≤ 2 seconds) to support 120 experts to view the scene in real time and guide AED operation and cardiopulmonary resuscitation through voice + picture.The rescue data warehouse 501 is used for comprehensively collecting and storing response time, arrival time, operation steps and medical result data of each rescue task; a time series database is used to store full-process data, including event information (position, risk level), dispatch data (responder information, matching score), execution data (response time, arrival time, operation steps) and medical results (defibrillation success rate, patient prognosis), and the data storage period is greater than or equal to 3 years. The performance evaluation board 502 is used to visually display the core indicators of the average response time and the matching success rate of the system, and generate an evaluation report; the algorithm optimization engine 503 continuously trains and optimizes the algorithm model of the intelligent matching and decision unit based on historical rescue data through machine learning. The specific operation is as follows: B1: based on the response time, arrival time and rescue result data in the historical rescue data, the weight coefficients and the pre-judgment model in the multi-dimensional matching algorithm module 202 are iteratively updated with the optimization goal of minimizing the average response time; B2: the verified and optimized model parameters and rules are deployed in the online scheduling environment of the intelligent matching and decision unit 200; B3: part of the dispatch requests are allocated to a test group decided by the optimized model, and the response efficiency is compared with a control group decided by the original model to evaluate the optimization effect.

[0114] In the description of the present specification, the description referring to the terms "one embodiment", "an example", "a specific example" and the like means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0115] The preferred embodiments of the application disclosed above are only used to help explain the application. The preferred embodiments do not describe all the details and limit the application to the specific embodiments described. Obviously, many modifications and variations can be made according to the content of the present specification. The present specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the application, so that those skilled in the art can well understand and utilize the application. The application is limited only by the claims and their full scope and equivalents.

Claims

1. An intelligent AED triggering system, characterized in that, It comprises a data acquisition and perception unit (100), an intelligent matching and decision-making unit (200), a communication transmission unit (300), an executor unit (400), and a feedback and learning unit (500), wherein: The data acquisition and perception unit (100) is used for acquiring and uploading the occurrence location and key information of the cardiac arrest event in real time. The intelligent matching and decision-making unit (200) is used for quickly calculating and deciding one or more optimal AED operators for scheduling through a multi-dimensional adaptive matching algorithm of scene perception weight dynamic adjustment, space-time trajectory prediction, and multi-person collaborative decision-making according to the event location, patient information, and real-time resource state. The communication transmission unit (300) is used for transmitting instructions and state information between the intelligent matching and decision-making unit (200) and the executor unit (400) by using a multi-modal communication protocol and a redundant communication mechanism. The executor unit (400) is used for receiving a high-priority reminder alarm and providing navigation and guidance support through the intelligent terminal of the registered operator, and the operator carries out on-site rescue action according to the instructions. The feedback and learning unit (500) is used for collecting rescue process data, evaluating system performance, and optimizing the matching decision-making algorithm.

2. The intelligent AED triggering system of claim 1, wherein, The data acquisition and perception unit (100) comprises a work order creation module (101), a multi-source positioning module (102), and an information preprocessing module (103), wherein: The work order creation module (101) is used for automatically receiving a 120 emergency call and generating a standardized emergency work order containing the event location, patient information, and timestamp. The multi-source positioning module (102) is used for fusing multi-source signals of GPS / Beidou, base station positioning, and network positioning to obtain high-precision geographic location information of the incident site. The information preprocessing module (103) is used for performing preliminary analysis and risk level classification on the emergency event through voice recognition and keyword extraction technology.

3. The intelligent AED triggering system of claim 1, wherein, The intelligent matching and decision-making unit (200) comprises a registered personnel database (201), a multi-dimensional matching algorithm module (202), and a dynamic scheduling decision-making module (203), wherein: The registered personnel database (201) is used for storing and managing the identity qualification, skill level, real-time location, and equipment state information of all registered AED operators. The multi-dimensional matching algorithm module (202) is used for executing a multi-dimensional adaptive matching algorithm of scene perception weight dynamic adjustment, space-time trajectory prediction, and multi-person collaborative decision-making to comprehensively calculate the optimal responder. The dynamic scheduling decision-making module (203) is used for generating a final scheduling instruction and managing a multi-level response mechanism.

4. The intelligent AED triggering system of claim 3, wherein, The multi-dimensional adaptive matching algorithm of scene perception weight dynamic adjustment, space-time trajectory prediction, and multi-person collaborative decision-making in the multi-dimensional matching algorithm module (202) comprehensively calculates the optimal responder, and the specific operation is as follows: A1: Obtain real-time environmental data at the incident location, including population density, traffic congestion index, and weather level, and assign dynamic weights to the three core matching factors of "geographical location distance", "skill level rating", and "real-time availability" according to the preset rules. When the traffic congestion index is ≥80, the "expected arrival time" weight replaces the "geographical location distance" weight and is increased to 50%; A2: Retrieve the positioning data of registered AED operators within the last 5 minutes, calculate the moving speed and direction vector, and predict the time deviation value within 3 minutes of arriving at the incident location by combining road network data to filter candidates with a deviation value ≤60 seconds; A3: Score and rank the candidates according to the dynamic weight comprehensive score formula: ① When the traffic congestion index is <80, the score = geographical location weight score × 40% + skill level score × 30% + predicted arrival time efficiency score × 30%; ② When the traffic congestion index is ≥80, the score = predicted arrival time efficiency score × 50% + skill level score × 30% + real-time availability score × 20%; Generate a dispatch sequence based on the event risk level: ① High-risk events generate 3 main responders + 3 backup responders; ② Medium-risk events generate 2 main responders + 2 backup responders; ③ Low-risk events generate 1 main responder + 2 backup responders; And if the main responder does not respond within 15-25 seconds, automatically upgrade the first backup to the main responder.

5. The intelligent AED triggering system of claim 3, wherein, The multi-level response mechanism in the dynamic dispatch decision module (203) is: ① First-level alarm "APP strong reminder": automatically upgrade if not responded within 10 seconds; ② Second-level alarm "SMS + phone": trigger backup personnel call if not responded within 15 seconds; ③ Third-level alarm "competitive response": notify 3-5 people simultaneously, and the first acceptor gets the task.

6. The intelligent AED triggering system of claim 1, wherein, The communication transmission unit (300) includes a multi-modal protocol stack (301), a redundant transmission control module (302), and a state confirmation relay module (303), wherein: The multi-modal protocol stack (301) supports multiple communication protocols such as 5G, APP push, and SMS; The redundant transmission control module (302) automatically switches to the backup channel seamlessly when the main communication channel fails; The state confirmation relay module (303) receives feedback states from the executor unit (400) and transmits them back to the intelligent matching and decision unit (200) in real time.

7. The intelligent AED triggering system of claim 1, wherein, The executor unit (400) includes a strong reminder receiving module (401), an AR navigation and guidance module (402), and a remote video consultation module (403), wherein: The strong reminder receiving module (401) receives and displays rescue instructions on the intelligent terminal in the highest priority mode; The AR navigation and guidance module (402) provides real-time navigation and AR visualization guidance for responders to go to the incident location and the nearest AED storage point; The remote video consultation module (403) establishes a low-latency video call with experts at the 120 emergency center during the rescue process to obtain real-time guidance.

8. The intelligent AED triggering system of claim 7, wherein, The highest priority mode includes continuous strong vibration, sharp alarm sound, and full-screen pop-up window.

9. The intelligent AED triggering system of claim 1, wherein, The feedback and learning unit (500) comprises a rescue data warehouse (501), a performance evaluation board (502), and an algorithm optimization engine (503), wherein: The rescue data warehouse (501) is used for collecting and storing the response time, arrival time, operation steps, and medical result data of each rescue task comprehensively; The performance evaluation board (502) is used for visualizing the core indicators of the system average response time and matching success rate, and generating an evaluation report; The algorithm optimization engine (503) continuously trains and optimizes the algorithm model of the intelligent matching and decision unit based on historical rescue data through machine learning.

10. The intelligent AED triggering system of claim 9, wherein, The algorithm optimization engine (503) continuously trains and optimizes the algorithm model of the intelligent matching and decision unit based on historical rescue data through machine learning, and the specific operation is as follows: B1: Based on the response time, arrival time, and rescue result data in the historical rescue data, the weight coefficients and the pre-judgment model in the multi-dimensional matching algorithm module (202) are iteratively updated with the optimization goal of minimizing the average response time; B2: The optimized model parameters and rules verified are deployed to the online scheduling environment of the intelligent matching and decision unit (200); B3: Part of the scheduling requests is allocated to a test group decided by the optimized model, and the response efficiency is compared with a control group decided by the original model to evaluate the optimization effect.

Images