Intelligent airport-oriented AED emergency equipment distribution optimization method

Abstract

The invention relates to the technical field of smart airport information service, in particular to a smart airport-oriented AED emergency equipment distribution optimization method. The method comprises the following steps: collecting airport mobile scene multi-source data; performing airport space behavior representation analysis processing through the airport mobile scene multi-source data to generate airport space behavior representation data; performing airport space first aid candidate point region feature analysis based on the airport space behavior representation data to generate airport space first aid candidate point region feature data; performing AED candidate region coverage arrival time domain analysis based on the airport space emergency candidate point region feature data to generate AED candidate region coverage arrival time domain data; and performing optimal distribution design of the AED first-aid equipment through the AED candidate area coverage arrival time domain data, and generating AED first-aid equipment distribution optimization data. According to the invention, the AED equipment distribution requirements of emergency first aid of the smart airport are met.

Description

Technical Field

[0001] This invention relates to the field of smart airport information service technology, and in particular to a method for optimizing the deployment of AED emergency medical devices in smart airports. Background Technology

[0002] As a critical intervention device for sudden emergency events such as sudden cardiac death, the rationality of AED (Automated External Defibrillator) deployment directly determines the response efficiency of airport emergency medical services. Smart airports are characterized by complex scene structures, drastic spatial and temporal fluctuations in passenger flow density, clear division of functional areas, and strong constraints on passenger flow paths due to flight scheduling cycles and security check processes. These characteristics place demands on the accuracy, dynamic adaptability, and practical operability of AED deployment. However, existing airport AED deployment schemes generally suffer from technical limitations. Most schemes employ an experience-based deployment method, relying on the past work experience of airport operators to determine the placement locations. A few optimized schemes use a uniform deployment method, distributing AEDs based solely on the physical spatial distance of the airport, completely ignoring the differences in the probability of emergency events occurring in different functional areas. Furthermore, they fail to consider the constraints of airport spatial accessibility, further exacerbating the irrationality of the deployment schemes. Moreover, existing AED deployment optimization technologies lack the ability to integrate multi-source dynamic data from smart airports, failing to accurately characterize the spatiotemporal correlation between personnel behavior and emergency events within the airport space. This results in a serious discrepancy between the theoretical emergency response efficiency and the actual application effect of existing deployment schemes, making it difficult to meet the high-quality emergency medical needs of smart airports. Summary of the Invention

[0003] Based on this, the present invention provides a method for optimizing the deployment of AED emergency medical devices in smart airports to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a method for optimizing the deployment of AED emergency medical devices in smart airports includes the following steps: Step S1: Obtain the multi-source data acquisition interface for airport mobility scenarios of the smart airport; perform multi-source data acquisition and preprocessing of airport mobility scenarios based on the multi-source data acquisition interface for airport mobility scenarios to generate multi-source data for airport mobility scenarios; Step S2: Perform airport spatial behavior representation analysis and processing using multi-source data of airport flow scenarios to generate airport spatial behavior representation data; Step S3: Based on the airport spatial behavior representation data, perform regional feature analysis of airport spatial emergency rescue candidate points to generate regional feature data of airport spatial emergency rescue candidate points; Step S4: Based on the regional feature data of airport spatial emergency rescue candidate points, perform AED candidate area coverage arrival time domain analysis to generate AED candidate area coverage arrival time domain data; Step S5: Optimize the deployment of AED emergency rescue devices by using the arrival time domain data of AED candidate areas to generate optimized deployment data for AED emergency rescue devices.

[0005] Furthermore, step S1 includes the following steps: Step S11: Obtain the multi-source data acquisition interface for airport flow scenarios in smart airports; Step S12: Collect preliminary multi-source data of airport flow scenarios according to the multi-source acquisition interface of airport flow scenarios to obtain preliminary multi-source data of airport flow scenarios. The preliminary multi-source data of airport flow scenarios includes airport passenger flow monitoring data, security check area queue monitoring data, flight operation scheduling data, staff allocation data, historical emergency rescue time record data, and airport scene spatial data. Step S13: Perform airport scene spatial structure modeling based on airport scene spatial data to generate an airport scene spatial structure model; Step S14: Map the initial multi-source data of the airport flow scene to the airport scene spatial structure model to perform time-series alignment and spatial mapping processing of heterogeneous data, and generate multi-source data of the airport flow scene.

[0006] Furthermore, step S2 includes the following steps: Step S21: Extract and process airport spatial behavior and emergency events data based on multi-source data of airport flow scenarios to generate airport spatial behavior data and emergency event data respectively; Step S22: Divide the airport scene spatial structure model into functional areas to generate airport spatial functional area data; Step S23: Based on the airport spatial functional area data and airport spatial behavior data, perform temporal feature analysis of regional spatial behavior attributes to generate temporal feature data of regional spatial behavior attributes; Step S24: Based on airport spatial functional area data and emergency event data, process the frequency of regional spatial emergency event patterns to generate regional spatial emergency event pattern frequency data; Step S25: Based on the temporal characteristic data of regional spatial behavior attributes and the frequency data of regional spatial emergency rescue events, airport spatial behavior representation is integrated and processed to generate airport spatial behavior representation data.

[0007] Furthermore, step S23 includes the following steps: Airport spatial behavior data is divided into airport spatial behavior attributes to generate airport spatial behavior attribute data; based on airport spatial functional area data, the temporal characteristics of spatial behavior attributes of each functional area are analyzed to generate regional spatial behavior attribute temporal characteristic data.

[0008] Furthermore, step S24 includes the following steps: Emergency event pattern analysis is performed on emergency event data to generate emergency event pattern data; based on airport spatial functional area data, frequency analysis of emergency event patterns in each functional area is performed on the emergency event pattern data to generate regional spatial emergency event pattern frequency data.

[0009] Furthermore, step S3 includes the following steps: Step S31: Map the airport spatial behavior representation data to the preset airport emergency medical event impact factors to perform airport spatial emergency medical event impact mapping feature analysis and generate airport spatial emergency medical event impact mapping feature data; Step S32: Perform probability analysis of airport space emergency rescue events based on the impact mapping feature data of airport space emergency rescue events, and generate airport space emergency rescue event probability data; Step S33: Analyze airport space emergency response candidate points using airport space emergency response event probability data to generate airport space emergency response candidate point data; Step S34: Perform regional feature analysis on the candidate point data for airport spatial emergency rescue to generate regional feature data for airport spatial emergency rescue candidate points.

[0010] Furthermore, the airport spatial emergency rescue candidate point area feature data mentioned in step S34 includes emergency rescue event probability intensity feature data of the candidate point area, deployability feature data of the candidate point area, and emergency rescue feature data related to the functional attributes of the candidate point area.

[0011] Furthermore, step S4 includes the following steps: Step S41: Perform airport spatial reachability path analysis based on the airport scene spatial structure model to generate airport spatial reachability path data; Step S42: Perform reachability path structure characteristic analysis on the airport spatial reachability path data to generate reachability path structure characteristic data; Step S43: Based on the reachable path structure characteristic data and the airport spatial emergency rescue candidate point area feature data, perform path travel constraint rate analysis of the AED candidate area to generate AED candidate area path travel constraint rate data. Step S44: Perform AED candidate area coverage arrival time domain analysis using AED candidate area path travel constraint rate data to generate AED candidate area coverage arrival time domain data.

[0012] Furthermore, step S43 includes the following steps: Step S431: Analyze the impact characteristics of passenger flow fluctuations on the reachable path structure based on the reachable path structure characteristic data, and generate passenger flow fluctuation impact characteristic data on the reachable path structure. Step S432: Based on the passenger flow fluctuation impact characteristic data of the reachable path structure, perform AED travel constraint rate analysis on the reachable path structure to generate AED travel constraint rate data for the reachable path structure; Step S433: Based on the reachable path structure AED travel constraint rate data, perform path travel constraint rate analysis on the airport spatial emergency rescue candidate point area feature data to generate AED candidate area path travel constraint rate data.

[0013] Furthermore, step S5 includes the following steps: Step S51: Design a multi-objective optimization function for AED deployment based on the preset AED deployment requirements; Step S52: Use the AED deployment multi-objective optimization function to perform global deployment evaluation of the AED candidate area coverage arrival time domain data, and generate global deployment evaluation data of the AED candidate area. Step S53: Based on the global deployment evaluation data of the AED candidate area, perform local deployment optimization and combination analysis of the candidate area to generate local deployment optimization and combination data of the AED candidate area; Step S54: Based on the local deployment optimization combination data of AED candidate areas, optimize the deployment design of AED emergency rescue equipment and generate AED emergency rescue equipment deployment optimization data.

[0014] The beneficial effects of this application are as follows: This invention specifically acquires multi-source data collection interfaces for smart airports and collects multi-dimensional information covering passenger flow monitoring, security check queuing, flight scheduling, staff allocation, historical emergency records, and airport spatial data. This overcomes the limitations of relying solely on single static data and ensures that the data comprehensively reflects the real situation of airport personnel flow, event occurrence, and spatial layout. Based on airport scene spatial data, a spatial structure model is constructed, providing a concrete spatial digital twin model for scattered data and avoiding data disconnect from the actual scene. Through temporal alignment and spatial mapping processing, heterogeneous preliminary data is standardized and integrated into the spatial model, eliminating the differences between different data in the time and spatial dimensions, forming structured, multi-source airport flow scene data that can be directly used for subsequent analysis. By extracting airport spatial behavior data and emergency medical event data from multi-source data, the core analytical objects were accurately separated, laying the foundation for subsequent targeted analysis. The airport spatial structure model was functionally divided into subdivided areas such as check-in area, security checkpoint, and boarding gate, avoiding the inaccuracies caused by holistic analysis and allowing subsequent analysis to focus on the characteristics of each functional area. By classifying spatial behavior attributes and analyzing the temporal characteristics of each functional area, the behavioral patterns of personnel flow intensity and dwell time in different areas at different times (such as morning rush hour and flight delay periods) can be dynamically captured. By analyzing emergency medical event patterns and the frequency of each area, the types and intensity of emergency medical events in different areas (such as densely populated transit areas and less crowded office areas) can be clearly identified. By integrating the airport spatial behavior representation data generated from these two types of features, the behavioral dynamics and differences in emergency medical needs of each area are clearly presented. By mapping airport spatial behavior data to pre-defined emergency event influencing factors (such as passenger flow density and regional functional attributes), the selection of candidate sites is ensured to always revolve around the core factors affecting emergency needs, avoiding blind site selection that is detached from actual needs. Based on the mapping feature analysis of the probability of emergency events in airport space, high-risk emergency areas can be accurately identified, allowing candidate sites to prioritize key areas that truly require AED coverage and reducing invalid candidates in low-risk areas. By extracting emergency event probability intensity, deployability, and functional attribute-related emergency features of candidate site areas, the priority of emergency needs for each candidate site is clarified, and "invalid candidate sites" that cannot be practically deployed due to insufficient space or lack of power supply are eliminated in advance. At the same time, the adaptability of candidate sites is further optimized by combining regional functional attributes (such as the personnel dwell characteristics of the boarding gate area). The final generated candidate site regional feature data has both reasonable demand and practical operability.Based on the analysis of reachable paths using the spatial structure model of the airport scenario, it is possible to fully reconstruct the actual movement routes of people within the airport (such as avoiding closed passages and matching dedicated routes for passengers / employees), avoiding deviations in time calculations based on abstract paths. Secondly, the combination of the analysis of path structure characteristics and the impact characteristics of passenger flow fluctuations can accurately capture the movement constraints of different time periods and different paths, thereby calculating the actual AED movement constraint rate, rather than relying on a fixed theoretical rate. The generated AED candidate area coverage arrival time domain data can truly reflect the actual emergency response time from the candidate point to its coverage area, providing a core time dimension basis for subsequent deployment optimization that conforms to the actual operation scenario of the airport. Based on a pre-defined AED deployment requirement decision design, a multi-objective optimization function is designed (typically encompassing minimizing average arrival time, maximizing emergency event coverage probability, and minimizing deployment cost). This ensures that the deployment plan meets emergency response efficiency while also considering the economics of resource investment, avoiding cost waste due to over-deployment or coverage gaps due to under-deployment. A global deployment assessment considers all candidate areas holistically, preventing the problem of optimal local deployments leading to global imbalances (e.g., dense deployment in one area resulting in insufficient coverage in adjacent high-risk areas). Combining the global assessment results with local deployment optimization allows for refined adjustments to address the specific needs of different areas (e.g., additional deployments are needed during peak passenger flow in transit areas, while fewer deployments are needed in office areas with low passenger flow). The generated AED emergency equipment deployment optimization data not only meets the overall emergency response needs of the airport but also adapts to the operational characteristics of each local area, effectively improving the scientific effectiveness of AED deployment.

[0015] Therefore, the AED emergency equipment deployment optimization method for smart airports in this invention precisely addresses the core problems in existing technologies, such as the subjectivity of experience-based and uniform deployment, the lack of spatial accessibility constraints, insufficient integration of multi-source dynamic data, and the absence of spatiotemporal correlation pattern representation, achieving comprehensive technological breakthroughs and performance improvements. Addressing the subjective blindness of experience-based deployment and the neglect of regional emergency response probability differences by uniform deployment, this method integrates multi-source data such as passenger flow monitoring, flight scheduling, and historical emergency records. Through spatial behavior characterization analysis, it uncovers the correlation patterns between personnel behavior and emergency events in each functional area. Furthermore, through emergency event probability analysis, it accurately identifies high-risk areas, replacing traditional experience-based judgments and average allocation methods. This ensures that AED deployment is entirely based on objective data, focusing on core areas of emergency needs, fundamentally solving the problem of deployment being disconnected from actual needs. Regarding airport spatial accessibility constraints not considered in existing technologies, this method generates AED candidate area coverage arrival time domain data that closely matches the actual airport operation scenario through accessibility path analysis and travel constraint rate calculation under the influence of passenger flow fluctuations. This breaks the deviation between theoretical path time and actual response efficiency, ensuring the practical operability of the emergency response time of the deployment plan. By aligning and mapping multi-source data in a temporal sequence and spatially, and by representing the spatiotemporal correlation between spatial behavior and emergency events, the shortcomings in the ability to integrate multi-source dynamic data of smart airports are compensated. In turn, multi-objective optimization is used to achieve the global optimization and local adaptation of the deployment scheme, so that the deployment of AEDs can meet the high-quality emergency rescue requirements of smart airports. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the steps of an AED emergency equipment deployment optimization method for smart airports according to the present invention. Figure 2 for Figure 1 A detailed flowchart illustrating the implementation steps of step S2. The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0017] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0018] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. Functional entities may be implemented in software, or in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods. The term "and / or" as used herein includes any and all combinations of one or more of the associated items listed.

[0019] To achieve the above objectives, please refer to Figures 1 to 2 This invention provides a method for optimizing the deployment of AED emergency medical devices in smart airports. In the embodiments of this invention, please refer to... Figure 1 The diagram shows a step-by-step flowchart of a method for optimizing the deployment of AED emergency medical devices in smart airports according to the present invention. The method includes the following steps: Step S1: Obtain the multi-source data acquisition interface for airport mobility scenarios of the smart airport; perform multi-source data acquisition and preprocessing of airport mobility scenarios based on the multi-source data acquisition interface for airport mobility scenarios to generate multi-source data for airport mobility scenarios; In this embodiment of the invention, the multi-source acquisition interface of the smart airport is specifically connected to the TCP / IP protocol interface of the airport passenger flow monitoring system, the RS485 serial port interface of the security check area queuing monitoring system, the database direct connection interface of the flight operation scheduling system, the HTTP interface of the employee personnel allocation system, the FTP file transfer interface of the historical emergency rescue time record system, and the BIM model export interface of the airport scene spatial data. Corresponding data acquisition tools were used for different interfaces: An industrial data acquisition gateway was used to collect infrared sensor data from the passenger flow monitoring interface in real time, with a collection frequency of once every 5 minutes; a serial data acquisition terminal was used to connect to the security check queuing interface to capture the number of people queuing and waiting time data every minute; a database synchronization tool was used to incrementally collect data from the MySQL database of the flight scheduling interface, with a synchronization frequency of once every 10 minutes; an interface calling tool was used to retrieve the on-duty location and number data of personnel assigned to the employee interface at 30-minute intervals; a file parsing tool was used to extract structured table data of historical emergency records from the FTP interface; and a laser point cloud scanning device combined with BIM modeling software was used to collect spatial data such as spatial coordinates, facility layout, and passage width of areas such as airport terminals, parking lots, and transfer areas, constructing an airport spatial basic dataset with an accuracy of 0.5 meters. The data was then preprocessed: outlier handling was performed using the 3σ principle, the mean and standard deviation of passenger flow data were calculated, and abrupt changes exceeding the mean ± 3 times the standard deviation were replaced with the mean of the three nearest time periods; for missing time periods in security check queuing data, linear interpolation was used to fit a straight line based on the five valid data points before and after the missing point to calculate the queuing data for the missing time period; all data were standardized by converting data of different dimensions such as passenger flow, queuing time, and flight number to the [0,1] interval through minimum-maximum normalization; finally, based on the airport's geocoding system (such as terminal number-floor-area code), the time series data was mapped to spatial coordinates, for example, the passenger flow data of the waiting area on the 3rd floor of Terminal 1 every 5 minutes was bound to the spatial attributes of the area such as latitude and longitude and floor height to generate multi-source data of airport flow scenarios.

[0020] Step S2: Perform airport spatial behavior representation analysis and processing using multi-source data of airport flow scenarios to generate airport spatial behavior representation data; In this embodiment of the invention, a data feature extraction algorithm is used to separate airport spatial behavior data and emergency event data from multi-source data of airport flow scenarios: behavioral features such as personnel movement trajectories, area stay duration, and personnel density are extracted from passenger flow and employee data; core information such as the location, time, and event type of emergency events are extracted from historical emergency records. Next, the airport scene spatial structure model is functionally divided into regions. The K-means spatial clustering algorithm is used, with the geographical coordinates and facility functional attributes of the airport space as clustering features. The number of clusters K is set to 9 (corresponding to nine functional areas: check-in area, security checkpoint, waiting area, boarding gate, baggage claim area, parking lot, transfer area, catering area, and office area). The number of iterations is 50, and the clustering error threshold is 0.01. The region division is completed by calculating the Euclidean distance between spatial points and cluster centers, generating airport spatial functional region data. Then, the temporal characteristics of regional spatial behavior attributes were analyzed. For behavioral attributes such as hourly personnel density, dwell time, and movement rate in each functional area, the ARIMA time series analysis model was used, with model parameters configured (e.g., autoregressive term p=2, difference order d=1, moving average term q=2). After data stabilization, the model was fitted, and temporal characteristics such as trend, periodicity, and fluctuation were extracted to generate temporal characteristic data of regional spatial behavior attributes. Simultaneously, the frequency of regional spatial emergency medical event patterns was analyzed. Frequency statistics algorithms were used to classify emergency medical event data by event type, such as cardiac arrest, syncope, and trauma, and the monthly occurrence frequency of each type of event in each functional area was statistically analyzed to generate regional spatial emergency medical event pattern frequency data. A weighted fusion method was used for feature integration, setting the weight of behavioral temporal features to 0.6 and the weight of emergency medical event frequency features to 0.4. The two types of data were bound by spatial region to generate airport spatial behavior representation data containing spatial location, behavioral temporal characteristics, and emergency medical event frequency.

[0021] Step S3: Based on the airport spatial behavior representation data, perform regional feature analysis of airport spatial emergency rescue candidate points to generate regional feature data of airport spatial emergency rescue candidate points; In this embodiment of the invention, airport spatial behavior representation data is mapped to preset airport emergency response event influencing factors (personnel density, frequency of emergency response events, regional personnel age structure, and facility access complexity). A Gaussian kernel function mapping method is used, with the kernel function parameter σ=0.5, to convert the high-dimensional representation data into low-dimensional features matching the influencing factors, generating airport spatial emergency response event influence mapping feature data. Next, a logistic regression probability model is constructed to analyze the probability of airport spatial emergency response events, using the influence mapping features as independent variables and the occurrence of an emergency response event as the dependent variable. A regularization coefficient of 0.01 is set, and the number of iterations is 100. The model weight parameters are optimized using gradient descent to calculate the probability of an emergency response event occurring in each spatial grid of the airport, generating airport spatial emergency response event probability data. Then, a threshold screening method is used to identify spatial grids with a probability value higher than 0.7 as airport spatial emergency response candidate points, generating airport spatial emergency response candidate point data. Finally, regional feature analysis was conducted on the candidate points: the kernel density estimation method was used to calculate the probability intensity of emergency events in the candidate point area, with a bandwidth of 0.3, to calculate the probability distribution density within a 50-meter radius of the candidate point; based on airport spatial data, indicators such as the available deployment area, distance from power interfaces, and passage occupancy of the candidate points were statistically analyzed to generate regional deployability feature data for the candidate points, where the available deployment area must be greater than 0.5 square meters, the distance from power interfaces must be less than 5 meters, and the passage width must be no less than 2 meters; through correlation analysis, the matching relationship between the emergency event type and functional attributes of the functional area where the candidate point is located was statistically analyzed to generate emergency feature data related to the functional attributes of the candidate point area, and the three types of features were integrated to form regional feature data of airport spatial emergency candidate points.

[0022] Step S4: Based on the regional feature data of airport spatial emergency rescue candidate points, perform AED candidate area coverage arrival time domain analysis to generate AED candidate area coverage arrival time domain data; In this embodiment of the invention, based on an airport scene spatial structure model, the Dijkstra path planning algorithm is used. Using the distance between each airport passageway as weight and ignoring reverse paths in one-way passageways, the shortest path from each airport emergency rescue candidate point to all areas of the airport is calculated, generating airport spatial reachable path data. The path search precision is set to 1 meter. Then, a graph theory topology analysis method is used to extract structural characteristics of each reachable path, such as the number of nodes, edges, number of turns, passageway width, and passenger capacity, generating reachable path structural characteristic data. Then, the path travel constraint rate of the AED candidate area is analyzed. First, based on the reachable path structure characteristics data, a passenger flow fluctuation impact characteristic model is constructed using multiple linear regression to calculate the correlation between passenger flow density and travel rate, determining that for every 1 person / square meter increase in passenger flow density, the travel rate decreases by 0.2 meters / second. Next, combined with path structure characteristics, constraint rules are set: when the passage width is less than 3 meters, the travel rate decreases by 0.3 meters / second; for every additional turn, the travel rate decreases by 0.1 meters / second. The travel rate of each path is calculated, generating AED candidate area path travel constraint rate data. Finally, based on the travel rate and path length, a gridded time calculation method is used, dividing the airport into 10m × 10m spatial grids. The time from the candidate point carrying the AED to each grid is calculated, and the time data is classified into intervals of 0-4 minutes, 4-6 minutes, and more than 6 minutes. The coverage area and number of regions for each candidate point in different time intervals are statistically analyzed, generating AED candidate area coverage arrival time domain data.

[0023] Step S5: Optimize the deployment of AED emergency rescue devices by using the arrival time domain data of AED candidate areas to generate optimized deployment data for AED emergency rescue devices.

[0024] In this embodiment of the invention, a multi-objective optimization function is designed, defining three core objectives: AED coverage of areas with a high probability of emergency rescue (≥0.6) ≥90% of the area, average arrival time in the covered area ≤4 minutes, and minimum total deployment cost (cost per unit based on a preset market research price, total quantity not exceeding budget). The deployment status of candidate areas (0 = no deployment, 1 = deployment) is used as the decision variable, and weights are set using the analytic hierarchy process (coverage percentage 0.4, average time 0.35, cost 0.25, pairwise comparison matrix CR = 0.08 < 0.1) to form the optimization function. Then, the NSGA-II algorithm (population 100, iterations 200, crossover probability 0.8, mutation probability 0.05, crowding threshold 0.1) is used to globally evaluate the coverage arrival time domain data, calculating the coverage percentage (area percentage of high probability areas within the 0-4 minute domain), average time (probability-weighted mean), and total cost (cost per unit × ...) for each candidate combination. The system generates a global Pareto optimal solution set by first selecting solutions with coverage ≥90% and time ≤4 minutes and sorting them by cost. It then uses a spatial structure model to detect the spacing between deployment points (if the distance is <30 meters, points with low overlap and short time are retained) and checks for functional area redundancy (e.g., two overlapping points in the waiting area with time ≤3 minutes are removed due to high cost). This yields a locally optimized combination containing coordinates and coverage range. Finally, it binds spatial data, assigns unique numbers to deployment points, labels functional areas, surrounding facilities, and maintenance departments, generates a visual map of the coverage time domain (0-4 minutes / 4-6 minutes) and emergency response probability, calculates the annual expected usage frequency of each point (combining historical data and passenger flow forecasts), and integrates these to form optimized deployment data containing deployment location, coverage range, equipment, and maintenance parameters.

[0025] Furthermore, step S1 includes the following steps: Step S11: Obtain the multi-source data acquisition interface for airport flow scenarios in smart airports; In this embodiment of the invention, obtaining the multi-source data acquisition interface for airport mobility scenarios in a smart airport requires connecting to the interfaces of the airport's existing business systems and completing adaptation verification. The system integrates with the airport passenger flow monitoring system. The TCP / IP protocol interface is obtained from the airport information center. This interface supports the transmission of raw data collected by infrared sensors and video cameras. An interface access application must be submitted according to airport security regulations. After review, a dedicated IP address, port number, and data transmission key are obtained. The key uses AES-256 encryption to ensure data transmission security. Next, the system integrates with the security checkpoint queue monitoring system. This system uses an RS485 serial port interface, requiring a serial converter to connect to the data acquisition terminal. Serial communication parameters, including a baud rate of 9600bps, 8 data bits, 1 stop bit, and no parity, are obtained from the airport security management module. Real-time data read permissions are requested for this interface to ensure access to real-time queue data for each security checkpoint. Following this, the system integrates with the flight operation scheduling system. This system uses a direct MySQL database interface. The database server address, database name, access account, and password are obtained from the airport operations command center. The account is configured with read-only permissions, allowing only access to data tables such as flight plans, flight status, and gate allocation. Data access frequency limits are also set to avoid affecting normal system operation. Finally, the system integrates with the employee allocation system, which provides HTTP... The interface obtains the interface URL, request header parameters (including a Token field), and data return format (JSON format) from the airport's human resources module. The Token validity period is set to 24 hours and needs to be automatically updated daily. Next, it connects to the historical emergency medical record system, which uses an FTP file transfer interface. The FTP server address, login account, password, and file storage path are obtained, ensuring that this interface automatically generates emergency medical record files. Finally, it connects to the airport's scene spatial data system, obtaining its BIM model export interface, supporting the export of .dwg format spatial model files. The interface's model precision parameter configuration permissions are also obtained to ensure that the exported model contains millimeter-level spatial coordinates and facility attribute information. After obtaining all interfaces, connectivity testing is required. Only after the test passes can the multi-source data acquisition interface be acquired.

[0026] Step S12: Collect preliminary multi-source data of airport flow scenarios according to the multi-source acquisition interface of airport flow scenarios to obtain preliminary multi-source data of airport flow scenarios. The preliminary multi-source data of airport flow scenarios includes airport passenger flow monitoring data, security check area queue monitoring data, flight operation scheduling data, staff allocation data, historical emergency rescue time record data, and airport scene spatial data. In this embodiment of the invention, for the airport passenger flow monitoring interface, an industrial data acquisition gateway is used for data acquisition. The gateway is deployed in the airport information room and connects to the passenger flow monitoring system through a preset IP and port. The acquisition frequency is set to once every 5 minutes, and each acquisition includes the real-time number of people at each monitoring point, the direction of personnel movement, and the duration of stay. For the security checkpoint queue monitoring interface, a serial data acquisition terminal is used to connect to an RS485 interface, and the acquisition frequency is set to once every 1 minute. The acquired data includes the security checkpoint number, the current number of people in the queue, the average waiting time, and the channel open status (0 = closed, 1 = open). For the flight operation scheduling interface, a database synchronization tool is used for incremental acquisition, collecting only newly added or updated flight data. For the employee personnel allocation interface, the employee's employee number, department, current on-duty location (specific area number), and on-duty time period are obtained. For the historical emergency rescue time record interface, an emergency rescue record file is downloaded. The file includes the emergency rescue event number, occurrence time, occurrence location (specific coordinates), event type (01 = cardiac arrest, 02 = fainting, 03 = trauma), and processing time. For the airport scene spatial data interface, BIM is used. The model export tool obtains spatial model files, which contain spatial attribute data such as the 3D coordinates (X / Y / Z axes, in meters) of various areas of the airport, facility types (01 = passageway, 02 = waiting seats, 03 = shops, 04 = elevators), passageway width, and floor height. After all data is collected, independent storage directories are created according to data type, and subdirectories are created under each directory according to the collection date to ensure clear data classification and form preliminary multi-source data of airport flow scenarios.

[0027] Step S13: Perform airport scene spatial structure modeling based on airport scene spatial data to generate an airport scene spatial structure model; In this embodiment of the invention, airport scene spatial structure modeling is performed based on airport scene spatial data. CAD data processing tools are used to extract 3D coordinate data and facility attribute data from the files. During the conversion process, coordinate offset correction parameters are set to ensure that the model coordinates are consistent with the actual geographical coordinates of the airport. Next, a laser point cloud processing tool is used to denoise the converted point cloud data, removing isolated points caused by equipment errors (a point cloud density threshold is set, deleting isolated points with fewer than 3 points within a 5-meter radius). Simultaneously, point cloud resampling is performed, uniformly adjusting the point cloud density to 10 points per square meter to ensure uniform model accuracy. Then, spatial region division is performed. Based on the facility attribute data, a spatial clustering algorithm (K-means algorithm) is used to cluster the processed point cloud data. The number of clusters is set to 9 (corresponding to airport check-in area, security check area, waiting area, boarding gate, baggage claim area, parking lot, transfer area, catering area, and office area), the number of iterations is 50, and the clustering error threshold is 0.01. Region division is completed by calculating the Euclidean distance between the point cloud and the cluster centers, and each region is assigned a unique region code (e.g., T1-01). (Representing the T1 terminal check-in area); then, add facility attribute associations, binding facility data (passage width, elevator location, power interface coordinates, emergency exit location) in each area with the area code. For example, mark the coordinates of each seat, the location and operating floor of each elevator, and the coordinates of each power interface (accurate to 0.1 meters) in the waiting area model; use BIM modeling tools to build a 3D visualization model, import the clustered area data and facility attribute data into the modeling tool, set the model texture material (e.g., gray floor tile texture for passageways, blue fabric texture for waiting seats), and add spatial topology relationships (e.g., connectivity between passageways, connection between elevators and floors). After completing the modeling, perform model verification, and generate the airport scene spatial structure model after successful verification.

[0028] Step S14: Map the initial multi-source data of the airport flow scene to the airport scene spatial structure model to perform time-series alignment and spatial mapping processing of heterogeneous data, and generate multi-source data of the airport flow scene.

[0029] In this embodiment of the invention, preliminary multi-source data of airport flow scenarios are mapped to an airport scenario spatial structure model to complete the temporal alignment and spatial mapping of heterogeneous data. First, temporal alignment processing is performed, and various preliminary data are standardized in time: for passenger flow monitoring data, the original timestamps are directly retained; for security check queuing data, a mean aggregation method is used to calculate the mean of 5 data points within every 5 minutes, serving as the representative data for that time period; for flight operation scheduling data, an interpolation method is used to supplement the predicted data for the middle 5 minutes within a 10-minute interval (based on linear fitting of the preceding and following 10-minute data); for employee allocation data, a proximity matching method is used to match employee location data within 30 minutes to the corresponding 5-minute time period, ensuring that the timestamps of all data are unified to one node every 5 minutes, while deleting data with duplicate or invalid timestamps (such as data with timestamps exceeding airport operating hours). Next, spatial mapping processing is performed. Based on the regional coding system of the airport scene spatial structure model, spatial labels are added to various preliminary data: For passenger flow monitoring data, the coordinates of the monitoring points in the model and their respective regional codes are queried according to the monitoring point number, thus binding the passenger flow data with the regional codes; for security check queuing data, the regional code of the security check area is determined according to the security check lane number; for flight operation scheduling data, the coordinates of the aircraft stands in the model and their respective regional codes are queried according to the aircraft stand number; for employee allocation data, the regional code of the on-duty position is directly matched with the regional code of the model; for historical emergency rescue time record data, the corresponding regional code is queried in the model according to the coordinates of the location where the emergency occurred. Finally, the data is integrated, and the time-aligned and spatially mapped data is stored in the spatiotemporal database using "regional code - timestamp" as the primary key, forming structured multi-source data of the airport flow scene.

[0030] Furthermore, as an embodiment of the present invention, reference is made to... Figure 2 As shown, Figure 1 A detailed flowchart illustrating the implementation steps of step S2 is provided in this embodiment. Step S2 includes: Step S21: Extract and process airport spatial behavior and emergency events data based on multi-source data of airport flow scenarios to generate airport spatial behavior data and emergency event data respectively; In this embodiment of the invention, airport spatial behavior data and emergency event data are extracted from multi-source data of airport movement scenarios. The core is to achieve the separation and structuring of the two types of data through targeted data filtering and feature extraction. A spatiotemporal data extraction tool is used to connect to the database storing multi-source data. For airport spatial behavior data, the extracted fields include the population density every 5 minutes in each area, population movement trajectory, area stay duration, and population flow intensity. Among them, the population density is calculated using a kernel density estimation algorithm, with a search radius of 5 meters, and the number of people at the monitoring points is converted into regional grid density values. The population movement trajectory is extracted using a trajectory tracking algorithm. Based on the time difference and spatial distance between adjacent monitoring points, the movement path and time data of people from area A to area B are generated. The trajectory data must meet the requirement that the continuous timestamp interval does not exceed 10 minutes (to ensure trajectory integrity). For emergency medical incident data, the extracted fields include emergency medical incident number, occurrence area code, occurrence timestamp, incident type (01 = cardiac arrest, 02 = fainting, 03 = trauma), and incident response time (time from incident occurrence to arrival of emergency personnel). Data cleaning rules were employed during extraction: invalid data with timestamps exceeding airport operating hours were deleted, and missing area codes due to coordinate ambiguity were supplemented (inferred through correlation with surrounding monitoring points, with errors controlled within one adjacent area). After extraction, the two types of data were structured separately to generate independent airport spatial behavior data and emergency medical incident data.

[0031] Step S22: Divide the airport scene spatial structure model into functional areas to generate airport spatial functional area data; In this embodiment of the invention, a three-dimensional spatial analysis tool is used to load the spatial structure model of the airport scene, and the basic spatial attribute data in the model is extracted, including the facility type (01 = passageway, 02 = waiting seats, 03 = shops, 04 = elevator), passageway width, floor height, and capacity of each area. Simultaneously, airport spatial behavior data is correlated to obtain the average passenger flow intensity, peak hours, and other behavioral characteristics of each area, forming a multi-dimensional feature matrix for area division. Next, an improved K-means spatial clustering algorithm is used for area division, setting the number of clusters K=9 (corresponding to 9 core functional areas: airport check-in area, security check area, waiting area, boarding gate, baggage claim area, parking lot, transfer area, catering area, and office area). The initial cluster centers are determined by randomly selecting the feature vectors of 9 typical areas. The number of iterations is set to 50, and the clustering error threshold (sum of squared errors) is 0.01. After each iteration, the Euclidean distance between the feature vector of each area and the cluster center is calculated, and the category of the area is adjusted until the error is less than the threshold or the maximum number of iterations is reached. After clustering, regional boundary verification is performed: using the on-site spatial marking method, 10 boundary marker points are selected in each candidate functional area of ​​the airport (such as the boundary line between the check-in area and the passageway, and the isolation zone between the security check area and the waiting area). The actual coordinates of the marker points are obtained and compared with the regional boundary coordinates generated by clustering in the model. The error is controlled within 0.3 meters. If the error exceeds the limit, the clustering parameters are readjusted (such as increasing the number of iterations to 60) and recalculated. A unique code is assigned to each clustered functional area (coding rule: terminal number - functional area number, such as T1-01 representing the check-in area of ​​terminal T1). The spatial range (minimum / maximum X / Y / Z coordinates), core facility list, average personnel capacity, and other attributes of the functional area are marked to form structured airport spatial functional area data.

[0032] Step S23: Based on the airport spatial functional area data and airport spatial behavior data, perform temporal feature analysis of regional spatial behavior attributes to generate temporal feature data of regional spatial behavior attributes; In this embodiment of the invention, the time granularity and period of the time series analysis are determined: the time granularity is divided into 3 levels (hourly, daily, and weekly), and the analysis period covers 3 months. For each functional area, three core behavioral attribute data categories—personnel gathering density, dwell time, and flow intensity—are extracted at the corresponding time granularity. For example, the hourly data for the waiting area of ​​Terminal 1 (code T1-03) includes gathering density values ​​for 18 time points from 6:00 to 24:00 every day. Next, time-series feature decomposition was performed on each type of behavioral attribute data: an ARIMA time series model (autoregressive term p=2, difference order d=1, moving average term q=2) was used to decompose the time-series data into trend terms (long-term population flow trends, such as growth / decline trends before and after holidays), periodic terms (periodic fluctuation patterns, such as peak periods at 8 am, 12 pm, and 6 pm each day), and fluctuation terms (random fluctuation components, such as short-term fluctuations caused by sudden flight delays). During the model fitting process, a residual threshold of 0.05 was set. When the residual exceeded the threshold, the model parameters were adjusted (such as adjusting the p value to 3) and the model was refitted. Then, the time-series characteristic indicators are calculated: for the trend item, the slope value (unit: person / square meter / day) is calculated to represent the trend strength, with positive values ​​indicating an increasing trend and negative values ​​indicating a decreasing trend; for the periodic item, peak periods (e.g., 8:00-10:00 AM in the waiting area is the morning peak, and 12:00-2:00 PM is the afternoon peak) and trough periods (e.g., 0:00-6:00 AM is the trough) are identified, and the periodic fluctuation amplitude (the difference between the peak and trough) is calculated; for the fluctuation item, the fluctuation coefficient (standard deviation / mean) is calculated, with a coefficient greater than 0.3 indicating a high fluctuation area and less than 0.1 indicating a low fluctuation area. The analyzed data are then integrated to generate time-series characteristic data of regional spatial behavior attributes.

[0033] Step S24: Based on airport spatial functional area data and emergency event data, process the frequency of regional spatial emergency event patterns to generate regional spatial emergency event pattern frequency data; In this embodiment of the invention, based on airport spatial functional area data and emergency medical event data, the pattern classification and frequency statistics of regional spatial emergency medical events are completed. An event feature classification tool is used to classify the emergency medical event data into patterns, determining pattern categories based on three dimensions: event type, occurrence scenario, and personnel characteristics. According to event type, three basic patterns are defined (01 = cardiac arrest, 02 = fainting, 03 = trauma). Sub-patterns are further subdivided based on functional area scenarios (e.g., fainting in the waiting area, trauma in the security checkpoint), resulting in a total of 27 sub-patterns (9 functional areas × 3 basic patterns). This classification is used to classify emergency medical event patterns and to perform frequency statistics for regional emergency medical events. A frequency statistics algorithm is used to count the frequency according to "functional area code - emergency medical mode - time period," with the time period divided into monthly and quarterly periods. Statistical dimensions include total frequency, daily average frequency, and frequency density. For each functional area, the proportion of each emergency response mode is calculated (frequency of a certain mode / total emergency response frequency of that functional area), the dominant mode is identified, and the frequency change trend is calculated. A growth rate > 10% is identified as a frequency increasing mode, and < -10% is identified as a frequency decreasing mode. The data is integrated to form regional spatial emergency response event mode frequency data.

[0034] Step S25: Based on the temporal characteristic data of regional spatial behavior attributes and the frequency data of regional spatial emergency rescue events, airport spatial behavior representation is integrated and processed to generate airport spatial behavior representation data.

[0035] In this embodiment of the invention, regional spatial behavioral attribute temporal feature data and regional spatial emergency event pattern frequency data are integrated. The behavioral temporal feature data is aggregated by timestamp, and the emergency event pattern frequency data is mapped by timestamp. A feature fusion tool is used to load the two types of data to establish an association mapping table. Next, feature weighted fusion is performed: fusion weights are set according to the importance of the two types of data. The weight of regional spatial behavioral attribute temporal feature is 0.6 (reflecting basic characteristics of personnel activities), and the weight of regional spatial emergency event pattern frequency feature is 0.4 (reflecting core characteristics of emergency needs). A linear weighted algorithm is used to calculate the fused feature value, for example, "Regional Emergency Risk Index" = (behavioral fluctuation coefficient × 0.6) + (emergency frequency density × 0.4), where the behavioral fluctuation coefficient is standardized to the [0,1] interval (original value / maximum value), and the emergency frequency density is standardized to the [0,1] interval (original value / maximum value). Simultaneously, supplementary derivative features are added: based on the fused data, "behavior-emergency response correlation" (the time overlap rate between peak behavior and emergency response events within a certain time period) and "regional vulnerability index" (a comprehensive score for high behavioral volatility and high emergency response frequency) are calculated. The derivative feature calculation adopts a correlation analysis algorithm, and the time overlap rate is determined by calculating the proportion of the intersection time between the peak behavior period and the emergency response event period. The vulnerability index is calculated using a weighted summation (volatility coefficient × 0.5 + frequency density × 0.5). These are then integrated to generate airport spatial behavior representation data.

[0036] Furthermore, step S23 includes the following steps: Airport spatial behavior data is divided into airport spatial behavior attributes to generate airport spatial behavior attribute data; based on airport spatial functional area data, the temporal characteristics of spatial behavior attributes of each functional area are analyzed to generate regional spatial behavior attribute temporal characteristic data.

[0037] In this embodiment of the invention, airport spatial behavior data is classified by attributes, with the core being the structuring of behavioral attributes through multi-dimensional feature decomposition. Airport spatial behavior data is divided into three main categories based on "personnel activity status," each containing specific feature indicators. The first category is "personnel aggregation attributes," including aggregation density, aggregation uniformity (calculated using the standard deviation of grid density; smaller values ​​indicate greater uniformity), and peak aggregation duration. The second category is "personnel movement attributes," including flow intensity, movement speed, and concentration of movement direction. The third category is "personnel stay attributes," including average stay duration, percentage of people staying, and frequency of repeated stays. The classification process employs rule-based extraction: aggregation density is converted using a kernel density estimation algorithm, with a search radius of 5 meters and a grid size of 1 meter × 1 meter. The original number of people at each monitoring point is assigned to the corresponding grid, and the density is calculated. Movement speed is derived from the spatial distance between adjacent monitoring points (calculated based on spatial model coordinates) and the time difference (time stamp interval of monitoring data). Stay duration is calculated using the time stamp difference between personnel entering and leaving the area; if only entry is recorded and departure is not, the average stay duration for the area is used to complete the calculation. Finally, airport spatial behavior attribute data is generated. Combined with airport spatial functional area data, time-series analysis is performed on the airport spatial behavior attribute data. Nine functional areas are defined based on the airport spatial functional area data. For each functional area, corresponding airport spatial behavior attribute data is matched. An ARIMA model is used to fit each attribute indicator for each functional area, decomposing the time-series data into trend terms (long-term variation patterns, such as the increasing trend of cluster density during holidays), periodic terms, and fluctuation terms (random fluctuations, such as sudden changes in waiting area flow intensity due to flight delays). Model fitting must satisfy a residual threshold ≤ 0.05. If the residual exceeds the threshold, the p-value or q-value is adjusted to 3 and recalculated. The slope value is calculated for the trend term, the peak / trough periods are identified for the periodic term and the fluctuation amplitude is calculated, and the fluctuation coefficient is calculated for the fluctuation term. After analysis, the data is integrated to generate spatially coded regional spatial behavior attribute time-series feature data.

[0038] Furthermore, step S24 includes the following steps: Emergency event pattern analysis is performed on emergency event data to generate emergency event pattern data; based on airport spatial functional area data, frequency analysis of emergency event patterns in each functional area is performed on the emergency event pattern data to generate regional spatial emergency event pattern frequency data.

[0039] In this embodiment of the invention, emergency medical event data is classified into patterns. Event patterns are structured through multi-dimensional feature association. An event pattern classification tool is used to interface with the emergency medical event data table, determining three core dimensions for pattern analysis: "Event Type Dimension" (refined based on the original event type, 01 = cardiac arrest, 02 = syncope, 03 = trauma), "Trigger Feature Dimension" (classified by event-related factors, 01 = fatigue, 02 = illness, 03 = collision, 04 = environmental factors), and "Personnel Status Dimension" (classified by patient status, 01 = elderly passenger, 02 = child passenger, 03 = staff, 04 = ordinary adult). Triggers are associated through the event description field. Classification is based on age and identity information in the event records; when no clear information is available, it is inferred from the event's location. The three dimensions are combined into a unique pattern code, with the encoding rule being "Event Type - Trigger - Personnel Status." This code is compared with a historical emergency medical case database to calculate the classification accuracy of each pattern (number of correctly classified cases / total number of cases). Patterns are assigned to each emergency medical event, generating emergency medical event pattern data. By combining airport spatial functional area data, frequency statistics are performed on emergency medical event pattern data, quantifying the distribution patterns of patterns according to functional area dimensions. Based on the nine functional areas of the airport spatial functional area data, each functional area is matched with corresponding emergency medical event pattern data. Events of each pattern are assigned to the corresponding functional area by matching the "event occurrence area code" (emergency medical event data field) with the "functional area code" (functional area data field). The total frequency, daily average frequency or weekly average frequency, and frequency density of each pattern in each functional area are statistically analyzed for each period, and the pattern proportion of each functional area is calculated. The analysis results are then integrated to generate regional spatial emergency medical event pattern frequency data.

[0040] Furthermore, step S3 includes the following steps: Step S31: Map the airport spatial behavior representation data to the preset airport emergency medical event impact factors to perform airport spatial emergency medical event impact mapping feature analysis and generate airport spatial emergency medical event impact mapping feature data; In this embodiment of the invention, a pre-defined airport emergency medical incident impact factor system is defined, comprising four core factors: personnel density factor (reflecting the density of personnel in the area), emergency medical incident frequency factor (reflecting the distribution of historical emergency medical incidents), regional flow intensity factor (reflecting the activity of personnel movement), and facility access complexity factor (reflecting the degree to which facilities in the area hinder emergency medical access). Each factor is set to a quantification range of [0,1], where 0 represents the lowest impact and 1 represents the highest impact. Next, a spatiotemporal feature mapping tool is used to load airport spatial behavior representation data, dividing each functional area into a 10m × 10m spatial grid (ensuring spatial accuracy). Four basic features are extracted from each grid: "standardized value of personnel density" and "standardized value of flow intensity" are extracted from the behavior representation data; "standardized value of emergency medical incident frequency" is extracted from the regional spatial emergency medical incident pattern frequency data; and "facility access complexity" is extracted from the airport spatial functional area data. Subsequently, the Gaussian kernel function mapping method was used to complete the data association. The kernel function parameter σ=0.5 was set, and the four basic features were mapped to the corresponding influencing factors (x is the basic feature value, and μ is the historical mean of the factor) by the formula "mapping feature value = exp (-(x-μ)² / (2σ²))". After mapping, all factor values ​​were normalized to the interval [0,1] to avoid the influence of dimensional differences, and finally the airport space emergency rescue event influence mapping feature data was generated.

[0041] Step S32: Perform probability analysis of airport space emergency rescue events based on the impact mapping feature data of airport space emergency rescue events, and generate airport space emergency rescue event probability data; In this embodiment of the invention, the probability data of emergency medical events is calculated based on the impact mapping feature data of airport spatial emergency medical events. First, the modeling system for probability analysis is determined: four impact mapping features (personnel density mapping value, emergency medical event frequency mapping value, regional flow intensity mapping value, and facility access complexity mapping value) are used as independent variables (X1-X4), and "whether an emergency medical event occurred in this spatial grid" is used as the dependent variable (Y, 1 = occurred, 0 = did not occur). The dependent variable data is derived from emergency medical event data (historical occurrence records matched according to grid coordinates). Next, a probabilistic model is constructed using a logistic regression modeling tool, with the following model parameters set: regularization coefficient of 0.01 (to avoid overfitting), number of iterations of 100, learning rate of 0.005, and convergence threshold of 1e-5. The impact mapping feature data and corresponding dependent variable data from the past three months were divided into a training set (80%) and a test set (20%) in an 8:2 ratio. The model weight parameters (W1-W4) and bias term (b) were optimized using gradient descent, with the objective function being "minimize cross-entropy loss". The accuracy of the test set was calculated every 10 iterations until the accuracy stabilized (fluctuation ≤1% over 3 consecutive iterations). After the model training was completed, the impact mapping feature data of airport space emergency rescue events was loaded, and the probability of emergency rescue events occurring in each 10m × 10m spatial grid was calculated using the model to generate airport space emergency rescue event probability data.

[0042] Step S33: Analyze airport space emergency response candidate points using airport space emergency response event probability data to generate airport space emergency response candidate point data; In this embodiment of the invention, candidate points are selected based on emergency event probability data. The core of this method is to determine reasonable candidate points through threshold screening and spatial optimization. A dual standard is set for candidate point screening: the first is a probability threshold standard, selecting spatial grids with an emergency event probability value ≥ 0.7 as "high-risk grids" (based on historical data, the probability of emergency events occurring in the area corresponding to this threshold is more than 2.5 times the average); the second is a spatial distribution standard, requiring a straight-line distance of ≥ 30 meters between candidate points (combining the 25-meter AED coverage radius with pedestrian traffic efficiency to avoid redundant coverage and resource waste). The coordinates of the center points of all high-risk grids in the airport spatial emergency response probability data are extracted to form an initial candidate point set. The DBSCAN spatial clustering algorithm is used to cluster this initial set, with algorithm parameters ε=30 meters (cluster radius, i.e., minimum spacing between candidate points) and MinPts=2 (minimum number of cluster points). Points with a distance <30 meters are grouped into the same cluster, and only the point with the highest probability value in each cluster is retained as an "optimized candidate point." The coverage of the optimized candidate point set is compared with that of the high-risk grids across the entire airport area. If a high-risk area is not covered by any candidate points, a point with the second highest probability (≥0.65) is added to that area as a "complete candidate point." During the screening process, airport spatial functional area data is considered, prioritizing grids at functional area boundaries (such as the junction of waiting areas and passageways) and areas with long passenger dwell times (such as next to seating in dining areas) as candidate points to generate airport spatial emergency response candidate point data.

[0043] Step S34: Perform regional feature analysis on the candidate point data for airport spatial emergency rescue to generate regional feature data for airport spatial emergency rescue candidate points.

[0044] In this embodiment of the invention, the candidate point regional feature analysis is performed on the airport spatial emergency rescue candidate point data to determine the probability intensity characteristics of emergency events in the candidate point region. A kernel density estimation algorithm is used, with a bandwidth of 0.3 (influence range coefficient). Taking the candidate point as the center, the weighted density value of the probability of emergency events in all spatial grids within a radius of 50 meters is calculated. The higher the value, the more concentrated the emergency risk around the candidate point. Three key indicators are extracted: available deployment area, distance from power interface, and channel reserved width. If the available area of ​​a candidate point is <0.5 square meters or the distance from the power source is >5 meters, it is marked as "deployability level = low", otherwise it is marked as "high". The frequency data of spatial emergency event modes in the associated region are used to count the dominant emergency mode of the functional area to which the candidate point belongs (e.g., the dominant mode in the waiting area is the fainting mode, accounting for 60%), and the monthly occurrence frequency of this mode within 50 meters of the candidate point. For example, the dominant mode of CP003 (located in the waiting area) is the "fainting mode", with a monthly frequency of 8 times. At the same time, the emergency equipment demand corresponding to this mode is recorded to generate structured airport spatial emergency rescue candidate point regional feature data.

[0045] Furthermore, the airport spatial emergency rescue candidate point area feature data mentioned in step S34 includes emergency rescue event probability intensity feature data of the candidate point area, deployability feature data of the candidate point area, and emergency rescue feature data related to the functional attributes of the candidate point area.

[0046] Furthermore, step S4 includes the following steps: Step S41: Perform airport spatial reachability path analysis based on the airport scene spatial structure model to generate airport spatial reachability path data; In this embodiment of the invention, reachability path analysis is carried out based on the airport scene spatial structure model. The core is to generate spatial reachability path data for the entire airport through a path planning algorithm. A three-dimensional path analysis tool is used to load the airport scene spatial structure model, and the model is decomposed into 5m × 5m three-dimensional grid units according to "floor-functional area-spatial grid" (X / Y axis is planar coordinate, Z axis is floor height, and the unit is meters). Each grid unit is labeled with facility type (01 = barrier-free passage, 02 = elevator, 03 = stairs, 04 = shop, 05 = wall), where wall (05) is set as an impassable obstacle, and elevator (02) needs to be associated with the floor information. Then, starting from the airport spatial emergency rescue candidate point and ending with all spatial grid units of the airport, a path analysis matrix of "candidate point-grid" is constructed. Subsequently, an improved Dijkstra algorithm was used for path planning. The algorithm parameters were set as follows: path weights were based on "grid accessibility," the search radius covered the entire airport (maximum distance not exceeding 200 meters to ensure coverage of all functional areas), and path accuracy error was controlled within 0.5 meters. Obstacle avoidance was required during path calculation: grid cells labeled "walls" or "shop interiors" were designated as impassable nodes. If a unique path existed from a candidate point to a grid cell but required traversing an obstacle, a detour path was replanned. Simultaneously, the floor connections of elevators and staircases were preserved. After analysis, airport spatial reachability path data was generated.

[0047] Step S42: Perform reachability path structure characteristic analysis on the airport spatial reachability path data to generate reachability path structure characteristic data; In this embodiment of the invention, structural characteristic analysis is performed on airport spatial reachability path data. The core of this analysis is to quantify the path traversal difficulty by extracting topological features. A graph theory topology analysis tool is used to load the airport spatial reachability path data. The structural information of each path is decomposed one by one according to the "path ID" to determine the core characteristic indicators: number of path nodes (total number of spatial grid nodes contained in the path), number of path edges (number of connecting segments between adjacent nodes, edge count = number of nodes - 1), number of turns (number of times the direction of a node in the path changes by ≥90 degrees, such as turning from east to south counts as 1 turn), minimum passage width (minimum width of all passage grids in the path, unit: meters), and number of critical facilities (total number of facilities such as elevators and stairs in the path that require additional time). Next, each indicator is quantified: the number of path nodes is directly obtained by counting the number of entries in the path node sequence; the number of turns is determined by calculating the angle between the direction vectors of adjacent nodes (the direction vector is composed of the X / Y coordinate difference between the preceding and following nodes, and an angle ≥ 90 degrees is counted as one turn); the minimum channel width is obtained by extracting the width values ​​of all channel grids in the path and taking the minimum value (e.g., if the path contains channels with widths of 4 meters, 3.5 meters, and 3 meters, the minimum channel width = 3 meters); the number of critical facilities is obtained by counting the total number of nodes in the path node sequence whose facility type is "elevator" or "stairs". Subsequently, derived characteristic indicators are added: path tortuosity (the ratio of the actual path length to the straight path length from the start to the end point; the closer the ratio is to 1, the smoother the path, and it should be ≤ 1.5) and average node spacing. After the calculation is completed, the rationality of the characteristic data needs to be verified: if the tortuosity of a certain path is >1.8 or the minimum passage width is <2 meters, it is marked as a "high difficulty path", and the path analysis results of S41 are traced back to confirm whether there is a better path; if the number of critical facilities is >3, it is necessary to check whether there is an alternative path without critical facilities (such as prioritizing the use of barrier-free passages instead of elevator + staircase combination) to generate reachable path structure characteristic data.

[0048] Step S43: Based on the reachable path structure characteristic data and the airport spatial emergency rescue candidate point area feature data, perform path travel constraint rate analysis of the AED candidate area to generate AED candidate area path travel constraint rate data. In this embodiment of the invention, the travel constraint rate of an AED is analyzed by combining reachability path structure characteristic data and airport spatial emergency rescue candidate point area feature data (including candidate point probability intensity, deployability, and other attributes). First, the benchmark and constraint system for rate analysis are determined: the benchmark travel rate is set at 1.5 m / s (based on the normal travel speed of emergency personnel carrying an AED, the standard rate without any constraints). Constraint factors are divided into two categories: path structure constraints and passenger flow fluctuation constraints. Next, a rate constraint model is constructed using a multiple linear regression tool. The model inputs are "path structure constraint factors" (number of turns, minimum passage width, number of critical facilities) and "passenger flow fluctuation constraint factors" (standardized values ​​of flow intensity), and the output is the "actual travel rate." Subsequently, the rate is calculated according to the "candidate point-path" dimension: for all paths corresponding to each candidate point, the structural constraint factors of the path are extracted (e.g., number of turns = 2, minimum passage width = 3 meters, number of key facilities = 1), and the "normalized flow intensity value" of the functional area where the path is located is extracted from the passenger flow time series characteristic data (e.g., the normalized flow intensity value of the waiting area during the morning peak period = 0.8), and substituted into the rate constraint model to calculate the actual travel rate. A lower limit for the rate needs to be set during the calculation process: the actual travel rate must not be lower than 0.3 m / s (to avoid rate distortion due to excessive constraints; if the calculation result is <0.3 m / s, use 0.3 m / s); at the same time, the rate is adjusted in conjunction with the "deployability level" (high / low). Finally, the rate data is time-series calibrated: the actual AED device travel rate for different time periods is calculated according to the passenger flow time series characteristic data, generating the number of AED candidate area path travel constraint rates.

[0049] Step S44: Perform AED candidate area coverage arrival time domain analysis using AED candidate area path travel constraint rate data to generate AED candidate area coverage arrival time domain data.

[0050] In this embodiment of the invention, based on the travel constraint rate data of AED candidate area paths, the arrival time domain of AED coverage is analyzed to determine the time domain division criteria and calculation logic. The time domain is divided into three intervals: 0-4 minutes (optimal coverage time level), 4-6 minutes (second-best coverage time level), and more than 6 minutes (exceeding effective emergency rescue time, ineffective coverage time level). The arrival time calculation logic is "Arrival Time = Total Path Length / Actual Travel Rate" (unit: seconds, converted to minutes and rounded to one decimal place). Next, using spatiotemporal coverage analysis tools, airport spatial reachable path data (including total path length) and travel constraint rate data (including actual travel rate), the arrival time of each path endpoint grid is calculated. During the calculation process, spatial matching is required, associating the coordinates (X / Y / Z) of each path endpoint grid with the airport spatial functional area data, and labeling the functional area code to which the grid belongs, ensuring that the time domain data corresponds to the functional area. Subsequently, the coverage range is statistically analyzed according to candidate points: for each candidate point, within each time period, the number and area of ​​grids covered in different time domains are statistically analyzed. Simultaneously, coverage overlap analysis is performed: if multiple candidate points cover the same grid in the time domain, the candidate point with the shortest arrival time is retained to avoid duplicate statistics, so as to generate AED candidate area coverage arrival time domain data.

[0051] Furthermore, step S43 includes the following steps: Step S431: Analyze the impact characteristics of passenger flow fluctuations on the reachable path structure based on the reachable path structure characteristic data, and generate passenger flow fluctuation impact characteristic data on the reachable path structure. In this embodiment of the invention, based on reachable path structure characteristic data (including fields such as path ID, number of turns, minimum channel width, etc.), the impact characteristics of passenger flow fluctuations on the path structure are analyzed. The impact of passenger flow on path passage is quantified through spatiotemporal correlation, generating characteristic data of passenger flow fluctuations affecting reachable path structures. The analytical dimensions of passenger flow fluctuation impact are clearly defined, including passenger flow density fluctuation characteristics (reflecting changes in population density along the path), peak passenger flow period distribution characteristics (reflecting periods of concentrated traffic pressure on the path), and passenger flow direction characteristics (reflecting the degree of interference of personnel movement within the path on emergency response). All three types of characteristics must be matched with the path structure characteristic data in both spatial and temporal dimensions. Extract the coordinates (X / Y / Z) of all 10m × 10m spatial grids along each path and the corresponding path structure attributes of the grid (such as the width of the passage and the type of facilities in the grid segment); extract the "standardized value of personnel gathering density" (range [0,1]), "peak passenger flow period marker" (1 = peak period, 0 = off-peak period, peak period is defined as the period with gathering density ≥0.6, such as 8:00-10:00, 12:00-14:00, 18:00-20:00), and "concentration of personnel flow direction" (range [0,1], the higher the value, the more uniform the personnel flow direction, and the less interference to the emergency response). Subsequently, a "path-grid-passenger flow" correlation model was established: For each path, passenger flow data of the corresponding grid was matched one by one according to the grid sequence it passed through, and passenger flow fluctuation characteristic indicators at the path level were calculated: First, the average passenger flow density fluctuation coefficient of the path was calculated by the formula "fluctuation coefficient = standard deviation of passenger flow density of grids passed through the path / mean of passenger flow density of grids passed through the path". A coefficient ≥ 0.3 was marked as "high fluctuation path", and < 0.3 was marked as "low fluctuation path". The standard deviation calculation needs to cover the data of the entire 24 hours of the day; Second, the proportion of peak hours of the path was calculated by counting the number of hours marked as "peak hours" in the grids passed through the path out of the entire 24 hours of the day. A proportion ≥ 30% (i.e. ≥ 7.2 hours) was marked as "peak pressure path"; Third, the interference value of the path flow direction was calculated by the formula "interference value = 1 - mean of flow direction concentration of grids passed through the path". An interference value ≥ 0.5 was marked as "high interference path". Spatial verification is required during the calculation process: If a route passes through multiple functional areas (such as from the waiting area through the passage to the security check area), the passenger flow characteristics need to be calculated segment by segment according to the functional area and then the weighted average is taken (the weight is the proportion of the length of the segment in that functional area to the total length of the route) to avoid feature distortion caused by differences in passenger flow across functional areas, and finally generate characteristic data on the impact of passenger flow fluctuations on the reachable path structure.

[0052] Step S432: Based on the passenger flow fluctuation impact characteristic data of the reachable path structure, perform AED travel constraint rate analysis on the reachable path structure to generate AED travel constraint rate data for the reachable path structure; In this embodiment of the invention, based on the characteristic data of passenger flow fluctuations affecting the reachable path structure, the AED travel constraint rate of the path structure is analyzed. The core is to construct a rate calculation model by combining path structure constraints and passenger flow fluctuation constraints. First, the benchmark system and constraint factors for rate calculation are determined: the benchmark travel rate is set at 1.5 m / s (based on the measured travel speed of emergency personnel carrying AEDs in an unobstructed environment without passenger flow interference). The constraint factors are divided into two categories: path structure constraint factors (derived from S42 data, including the number of turns, minimum passage width, and number of key facilities) and passenger flow fluctuation constraint factors (derived from S431 data, including the average passenger flow density fluctuation coefficient of the path, the proportion of peak hours, and the interference value of the flow direction). Each type of factor needs to be quantified into a rate attenuation coefficient. Next, a rate constraint model was constructed using a multiple linear regression modeling tool. The model input consisted of six constraint factors, and the output was the "actual travel rate of the path." The model formula was defined as: Actual travel rate = Baseline rate × [1 - 0.1 × Number of turns - 0.08 × (4 - Minimum lane width) - 0.05 × Number of critical facilities - 0.2 × Passenger flow density fluctuation coefficient - 0.15 × Peak hour percentage - 0.1 × Flow direction interference value]. Each coefficient in the formula was calibrated using historical emergency medical service travel data. For each path, the actual travel rate was calculated for both peak and off-peak hours. Based on the output results, AED travel constraint rate data for reachable path structures was generated.

[0053] Step S433: Based on the reachable path structure AED travel constraint rate data, perform path travel constraint rate analysis on the airport spatial emergency rescue candidate point area feature data to generate AED candidate area path travel constraint rate data.

[0054] In this embodiment of the invention, based on the reachable path structure AED travel constraint rate data and combined with the regional feature data of airport spatial emergency rescue candidate points (including candidate point number, deployability level, functional area, etc.), the path travel constraint rate of AED candidate areas is analyzed. First, the correlation logic of the candidate area rate analysis is clarified: each airport spatial emergency rescue candidate point corresponds to multiple reachable paths. The travel rate of each path needs to be combined with the regional features of the corresponding candidate point to supplement the fine-tuning effect of candidate point attributes on the rate, ensuring that the rate data more closely matches the actual deployment scenario. Candidate point data is bound to path rate data to ensure that all path rates corresponding to each candidate point are associated with that candidate point's attributes. The actual travel rate of the path is multiplied by the rate adjustment coefficient of the corresponding candidate point. Based on the original peak / off-peak periods, four time periods are further divided: morning peak (8:00-10:00), afternoon peak (12:00-14:00), evening peak (18:00-20:00), and off-peak (other times). The rates for each time period are recalculated using passenger flow data to ensure a more accurate match between rates and actual passenger flow changes. Finally, data integrity is verified: if the path rate corresponding to a candidate point is missing (e.g., a newly created path has no historical passenger flow data), the average rate of paths in the same functional area and with similar structural characteristics (e.g., similar number of turns, similar passage width) is used to complete the data. This ultimately generates the AED candidate area path travel constraint rate data.

[0055] Furthermore, step S5 includes the following steps: Step S51: Design a multi-objective optimization function for AED deployment based on the preset AED deployment requirements; In this embodiment of the invention, a multi-objective optimization function is designed based on a preset AED deployment demand decision. The core is to construct a mathematical model by quantifying deployment objectives and constraints to generate a computable multi-objective optimization function for AED deployment. First, the preset deployment demand decision system is defined, which includes three core objectives: First, maximizing coverage efficiency (ensuring effective coverage of high-risk areas), quantified by the "percentage of 0-4 minute coverage area in high-probability emergency areas," with a target value ≥ 95%; second, minimizing time cost (shortening emergency arrival time), quantified by the "average arrival time of AEDs in all covered areas," with a target value ≤ 4 minutes; and third, minimizing economic cost (controlling equipment investment), quantified by the "total cost of AED deployment." Next, the analytic hierarchy process (AHP) is used to determine the objective weights: a pairwise comparison matrix of "coverage efficiency - time cost - economic cost" is constructed, and the weights are calculated to be 0.45, 0.35, and 0.2 respectively, through matrix consistency testing (consistency ratio CR = 0.07 < 0.1, meeting the consistency requirement). Subsequently, the decision variables and constraints of the optimization function are defined: the decision variable is the "deployment status of candidate points" (x_i, x_i=1 indicates deployment of AED, x_i=0 indicates no deployment, i is the candidate point number, there are a total of n candidate points, n is determined by S33); the constraints include "single candidate point deployability constraint" (only candidate points with a deployability level of "high" are allowed x_i=1), "total number constraint" (Σx_i≤50), and "overlap constraint" (the straight-line distance between any two deployment points is ≥30 meters). Construct a multi-objective optimization function: The main function is "Max F (x)=0.45×C(x)+0.35×(1-T (x) / 4)+0.2×(1-Cost (x) / (50×1.2))", where C (x) is the coverage area ratio (taken from the ratio of the 0-4 minute coverage area to the total area of ​​the high-risk area in S44), T (x) is the average arrival time (taken from the weighted average of the arrival times of all coverage grids in S44, with the weight being the grid emergency rescue probability), and Cost (x) is the total cost (Cost (x)=1.2×Σx_i); at the same time, the constraint function "st x_i∈{0,1},Σx_i≤50,d_ij≥30 meters (i is not equal to j, d_ij is the straight-line distance between candidate points i and j)" is added. All indicators in the function are standardized to the [0,1] interval to ensure collaborative optimization of objectives and generate a clearly quantified multi-objective optimization function for AED deployment.

[0056] Step S52: Use the AED deployment multi-objective optimization function to perform global deployment evaluation of the AED candidate area coverage arrival time domain data, and generate global deployment evaluation data of the AED candidate area. In this embodiment of the invention, a multi-objective optimization function is used to perform a global deployment evaluation of AED candidate area coverage arrival time domain data (including candidate point number, time period, coverage area, arrival time, etc.). First, the evaluation algorithm and parameters are determined: NSGA-II (Non-dominated sorting genetic algorithm II) is adopted, with the following parameter configurations: population size 100 (generating 100 deployment combinations in each iteration), number of iterations 200 (ensuring sufficient search), crossover probability 0.8 (controlling gene exchange frequency), mutation probability 0.05 (avoiding local optima), and crowding distance threshold 0.1 (maintaining population diversity). Next, the core data required for the evaluation are loaded: first, AED candidate area coverage arrival time domain data, extracting the "number of coverage grids within 0-4 minutes" and "average arrival time" for each candidate point in the four time periods of off-peak, morning peak, noon peak, and evening peak; second, airport spatial emergency candidate point area feature data, extracting the "deployability level" and "emergency event probability intensity" (as coverage weights) for each candidate point; and third, candidate point coordinate data, used to calculate deployment point spacing constraints. Subsequently, a global evaluation process is executed: The first step is population initialization, randomly generating 100 placement combinations that satisfy the "deployability constraint" and "quantity constraint" (each combination is a 0-1 sequence of x_i); the second step is non-dominated sorting, substituting the optimization function into each combination to calculate the F(x) value, and dividing the Pareto levels according to the rule of "not being dominated by other combinations" (i.e., the C(x), (1-T(x) / 4), and (1-Cost(x) / 60) of a certain combination are all not lower than another combination), retaining the first 3 levels as valid solutions; the third step is crowding calculation, evaluating the dispersion of solutions using the formula "crowding degree = Σ|F_k+1 - F_k-1|" (F_k is the objective function value of the k-th combination), and eliminating duplicate solutions with excessively high crowding (>0.1); the fourth step is iterative evolution, repeating the sorting and filtering process until it terminates after 200 iterations. After the evaluation is completed, global placement evaluation data is generated.

[0057] Step S53: Based on the global deployment evaluation data of the AED candidate area, perform local deployment optimization and combination analysis of the candidate area to generate local deployment optimization and combination data of the AED candidate area; In this embodiment of the invention, based on the global deployment evaluation data of AED candidate areas, a local deployment optimization combination analysis is performed on the candidate areas. The core is to screen out conflict-free and highly adaptable deployment combinations through local conflict detection and coverage completion. The core constraints of local optimization are clearly defined, including spatial conflict constraints (deployment point spacing ≥ 30 meters to avoid overlapping and wasted coverage), functional area adaptation constraints (the number of deployments in the same functional area ≤ the area of ​​the functional area / 1000 square meters, such as ≤ 2 units in a 2000 square meter waiting area), and coverage blind spot constraints (0-4 minute coverage rate ≥ 98% in high-risk areas, filling in uncovered blind spots). Next, a greedy-local search hybrid algorithm is used for optimization: The first step is initial solution screening, selecting the top 5 placement combinations with F(x) from the global Pareto optimal solution set as the initial solution, prioritizing combinations with the highest C(x) and lowest T(x); the second step is spatial conflict detection, loading the coordinates of the deployment candidate points of the initial solution, calculating the distance between any two points using the Euclidean distance formula, and retaining candidate points with higher probability of emergency events if there are conflict points with d_ij < 30 meters; the third step is functional area adaptation adjustment, checking the deployment quantity of each functional area in the initial solution based on the functional area area data; the fourth step is coverage blind spot completion, comparing the 0-4 minute coverage grid of the adjusted combination with the high-risk area grid to identify uncovered blind spots, and selecting candidate points that are closest to the blind spots and have a "high" deployability level to supplement the combination. After optimization, the above process is repeated for each initial solution to generate local optimization combinations. Each combination contains "optimization combination ID, final deployment candidate point list, conflict elimination record, blind spot completion record, and adjusted C (x) / T (x) / Cost (x)", ensuring that all combinations meet the requirements of no spatial conflict, functional area adaptation, and no coverage blind spots, and are stored in a database table associated with the global evaluation data.

[0058] Step S54: Based on the local deployment optimization combination data of AED candidate areas, optimize the deployment design of AED emergency rescue equipment and generate AED emergency rescue equipment deployment optimization data.

[0059] In this embodiment of the invention, based on the optimized combination data of local AED candidate area deployment, the optimized deployment design of AED emergency rescue devices is completed. The core is through spatial positioning, attribute labeling, and data integration. The core output dimensions of the deployment design are determined as follows: including spatial positioning information (precise coordinates and functional area association), coverage area labeling (0-4 minute / 4-6 minute coverage area), device attribute information (number and maintenance parameters), and visualization map (spatial distribution and time domain mapping). Each dimension must be associated with the preceding data to ensure accuracy. Next, the deployment design process is executed: The first step is spatial positioning and numbering, with local optimization and combination. The 3D coordinates of each deployment point are extracted from the candidate point data, and a unique device number is assigned to each deployment point. The second step is coverage and time domain labeling. Combining the AED candidate area coverage arrival time domain data, the 0-4 minute coverage grid range and 4-6 minute coverage grid range of each deployment point are extracted and labeled into the spatial coordinate system. Simultaneously, the proportion of high-risk areas covered by each deployment point is calculated. The third step is the integration of device attributes and maintenance information. Based on the candidate point area feature data, the "deployable location type," "power interface distance," and "maintenance responsible department" of each deployment point are labeled. The fourth step is visualization and structured output. A 3D spatial modeling tool is used to map the deployment point coordinates and coverage range to the airport scene spatial structure model, generating a deployment visualization map (0-4 minute coverage area is marked in green, 4-6 minute in yellow), while simultaneously constructing structured deployment data. Finally, optimized AED emergency equipment deployment data is generated, including spatial coordinates and attribute information directly usable for construction, as well as operation and maintenance layout parameters.

[0060] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0061] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for optimizing the deployment of AED emergency medical devices in smart airports, characterized in that, Includes the following steps: Step S1: Obtain the multi-source data acquisition interface for airport mobility scenarios of the smart airport; perform multi-source data acquisition and preprocessing of airport mobility scenarios based on the multi-source data acquisition interface for airport mobility scenarios to generate multi-source data for airport mobility scenarios; Step S2: Perform airport spatial behavior representation analysis and processing using multi-source data of airport flow scenarios to generate airport spatial behavior representation data; Step S3: Based on the airport spatial behavior representation data, perform regional feature analysis of airport spatial emergency rescue candidate points to generate regional feature data of airport spatial emergency rescue candidate points; Step S4: Based on the regional feature data of airport spatial emergency rescue candidate points, perform AED candidate area coverage arrival time domain analysis to generate AED candidate area coverage arrival time domain data; Step S5: Optimize the deployment of AED emergency rescue devices by using the arrival time domain data of AED candidate areas to generate optimized deployment data for AED emergency rescue devices.

2. The method for optimizing the deployment of AED emergency medical devices in smart airports according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Obtain the multi-source data acquisition interface for airport flow scenarios in smart airports; Step S12: Collect preliminary multi-source data of airport flow scenarios according to the multi-source acquisition interface of airport flow scenarios to obtain preliminary multi-source data of airport flow scenarios. The preliminary multi-source data of airport flow scenarios includes airport passenger flow monitoring data, security check area queue monitoring data, flight operation scheduling data, staff allocation data, historical emergency rescue time record data, and airport scene spatial data. Step S13: Perform airport scene spatial structure modeling based on airport scene spatial data to generate an airport scene spatial structure model; Step S14: Map the initial multi-source data of the airport flow scene to the airport scene spatial structure model to perform time-series alignment and spatial mapping processing of heterogeneous data, and generate multi-source data of the airport flow scene.

3. The method for optimizing the deployment of AED emergency medical devices in smart airports according to claim 2, characterized in that, Step S2 includes the following steps: Step S21: Extract and process airport spatial behavior and emergency events data based on multi-source data of airport flow scenarios to generate airport spatial behavior data and emergency event data respectively; Step S22: Divide the airport scene spatial structure model into functional areas to generate airport spatial functional area data; Step S23: Based on the airport spatial functional area data and airport spatial behavior data, perform temporal feature analysis of regional spatial behavior attributes to generate temporal feature data of regional spatial behavior attributes; Step S24: Based on airport spatial functional area data and emergency event data, process the frequency of regional spatial emergency event patterns to generate regional spatial emergency event pattern frequency data; Step S25: Based on the temporal characteristic data of regional spatial behavior attributes and the frequency data of regional spatial emergency rescue events, airport spatial behavior representation is integrated and processed to generate airport spatial behavior representation data.

4. The method for optimizing the deployment of AED emergency medical devices in smart airports according to claim 3, characterized in that, Step S23 includes the following steps: Airport spatial behavior data is divided into airport spatial behavior attributes to generate airport spatial behavior attribute data; based on airport spatial functional area data, the temporal characteristics of spatial behavior attributes of each functional area are analyzed to generate regional spatial behavior attribute temporal characteristic data.

5. The method for optimizing the deployment of AED emergency medical devices in smart airports according to claim 3, characterized in that, Step S24 includes the following steps: Emergency event pattern analysis is performed on emergency event data to generate emergency event pattern data; based on airport spatial functional area data, frequency analysis of emergency event patterns in each functional area is performed on the emergency event pattern data to generate regional spatial emergency event pattern frequency data.

6. The method for optimizing the deployment of AED emergency medical devices in smart airports according to claim 1, characterized in that, Step S3 includes the following steps: Step S31: Map the airport spatial behavior representation data to the preset airport emergency medical event impact factors to perform airport spatial emergency medical event impact mapping feature analysis and generate airport spatial emergency medical event impact mapping feature data; Step S32: Perform probability analysis of airport space emergency rescue events based on the impact mapping feature data of airport space emergency rescue events, and generate airport space emergency rescue event probability data; Step S33: Analyze airport space emergency response candidate points using airport space emergency response event probability data to generate airport space emergency response candidate point data; Step S34: Perform regional feature analysis on the candidate point data for airport spatial emergency rescue to generate regional feature data for airport spatial emergency rescue candidate points.

7. The method for optimizing the deployment of AED emergency medical devices in smart airports according to claim 6, characterized in that, The airport spatial emergency rescue candidate point area feature data mentioned in step S34 includes emergency event probability intensity feature data of the candidate point area, deployability feature data of the candidate point area, and emergency rescue feature data related to the functional attributes of the candidate point area.

8. The method for optimizing the deployment of AED emergency medical devices in smart airports according to claim 2, characterized in that, Step S4 includes the following steps: Step S41: Perform airport spatial reachability path analysis based on the airport scene spatial structure model to generate airport spatial reachability path data; Step S42: Perform reachability path structure characteristic analysis on the airport spatial reachability path data to generate reachability path structure characteristic data; Step S43: Based on the reachable path structure characteristic data and the airport spatial emergency rescue candidate point area feature data, perform path travel constraint rate analysis of the AED candidate area to generate AED candidate area path travel constraint rate data. Step S44: Perform AED candidate area coverage arrival time domain analysis using AED candidate area path travel constraint rate data to generate AED candidate area coverage arrival time domain data.

9. The method for optimizing the deployment of AED emergency medical devices in smart airports according to claim 8, characterized in that, Step S43 includes the following steps: Step S431: Analyze the impact characteristics of passenger flow fluctuations on the reachable path structure based on the reachable path structure characteristic data, and generate passenger flow fluctuation impact characteristic data on the reachable path structure. Step S432: Based on the passenger flow fluctuation impact characteristic data of the reachable path structure, perform AED travel constraint rate analysis on the reachable path structure to generate AED travel constraint rate data for the reachable path structure; Step S433: Based on the reachable path structure AED travel constraint rate data, perform path travel constraint rate analysis on the airport spatial emergency rescue candidate point area feature data to generate AED candidate area path travel constraint rate data.

10. The method for optimizing the deployment of AED emergency medical devices in smart airports according to claim 1, characterized in that, Step S5 includes the following steps: Step S51: Design a multi-objective optimization function for AED deployment based on the preset AED deployment requirements; Step S52: Use the AED deployment multi-objective optimization function to perform global deployment evaluation of the AED candidate area coverage arrival time domain data, and generate global deployment evaluation data of the AED candidate area. Step S53: Based on the global deployment evaluation data of the AED candidate area, perform local deployment optimization and combination analysis of the candidate area to generate local deployment optimization and combination data of the AED candidate area; Step S54: Based on the local deployment optimization combination data of AED candidate areas, optimize the deployment design of AED emergency rescue equipment and generate AED emergency rescue equipment deployment optimization data.

Images