Pre-hospital emergency intelligent scheduling management method and system based on emergency resources

Abstract

The invention discloses a pre-hospital first-aid intelligent scheduling management method and system based on first-aid resources, and relates to the technical field of medical resource management, and the method comprises the steps: determining a first-aid task type according to the initial illness state of a patient, and determining a first-aid resource combination through the correlation characteristics of the historical illness state and the first-aid resources; determining candidate first-aid stations by using the real-time position of the patient, and screening a target first-aid station from the candidate first-aid stations based on periodic characteristics of resource occupation of the first-aid stations; generating an alternative first-aid route by using an actual road layout between the target first-aid station and the patient position, and determining a potential influence area of the current traffic event; evaluating the traffic capacity of the alternative first-aid routes after the alternative first-aid routes are affected by the potential influence areas through a road traffic capacity prediction model, and determining a target first-aid route from the alternative first-aid routes; according to the invention, intelligent and accurate scheduling of first-aid resources, stations and routes is realized, and timeliness and accuracy of pre-hospital first-aid response are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of medical resource management technology, specifically to a pre-hospital emergency intelligent dispatch management method and system based on emergency resources. Background Technology

[0002] Pre-hospital emergency care is a crucial component of the modern healthcare system, and its service efficiency and resource allocation quality directly impact patient treatment outcomes and survival rates. Currently, pre-hospital emergency dispatch systems typically rely on human dispatchers who, after initially assessing the patient's condition, determine the type of emergency task based on experience and manually assign appropriate emergency resource combinations. However, this traditional approach struggles to accurately assess the potential changes in a patient's condition when dealing with complex and dynamically changing emergency scenarios. It often fails to pre-determine a suitable combination of emergency resources to align with the patient's evolving condition, resulting in insufficient targeted resource allocation and consequently affecting treatment efficiency and effectiveness.

[0003] Furthermore, in actual emergency dispatching, traditional methods typically select emergency stations based primarily on proximity, failing to fully consider the cyclical patterns of resource usage at each station. This can lead to delays and waste of resources during periods of high resource scarcity, potentially preventing the provision of timely and effective emergency services. Simultaneously, current technologies often neglect the impact and propagation patterns of real-time traffic events when selecting emergency routes, frequently resulting in emergency vehicles getting stuck in traffic congestion and severely delaying crucial patient care, thus failing to meet the practical demands for rapid response.

[0004] Therefore, there is an urgent need in this field for a more intelligent and systematic method for the scheduling and management of pre-hospital emergency medical resources. This method should be able to comprehensively consider the patterns of changes in patients' conditions, the characteristics of emergency medical resource usage, the cyclical occupancy patterns of emergency medical station resources, and the spatial propagation characteristics of traffic incidents. It should be able to accurately determine the combination of emergency medical resources and rationally select target emergency medical stations and emergency routes to achieve the scientific scheduling and efficient utilization of emergency medical resources and improve the overall response efficiency and quality of pre-hospital emergency medical services. Summary of the Invention

[0005] The purpose of this invention is to provide a pre-hospital emergency medical service intelligent dispatch and management method and system based on emergency medical resources, so as to solve the problems in the background art mentioned above.

[0006] To achieve the above objectives, the present invention provides the following technical solution: In a first aspect, the present invention provides a pre-hospital emergency medical service intelligent dispatch and management method based on emergency medical resources, characterized in that it includes: The type of emergency task is determined based on the patient's initial condition, and the combination of emergency resources is determined by utilizing the correlation characteristics between the patient's historical condition and emergency resources to adapt to potential changes in the patient's condition. Candidate emergency medical stations are identified by utilizing the patient's real-time location, and target emergency medical stations that meet the combination of emergency medical resources are selected from the candidate emergency medical stations based on the periodic characteristics of emergency medical station resource occupancy. Alternative emergency routes are generated by utilizing the actual road layout between the target emergency station and the patient's location, and the potential impact area of ​​the current traffic event is determined based on the spatial propagation characteristics of historical traffic events. The road capacity prediction model is used to assess the traffic capacity of alternative emergency routes after they are affected by potential impact areas, and the target emergency route is determined from the alternative emergency routes based on the traffic capacity.

[0007] Secondly, the present invention provides a pre-hospital emergency medical service intelligent dispatch and management system based on emergency medical resources, implemented based on the aforementioned pre-hospital emergency medical service intelligent dispatch and management method based on emergency medical resources, including: The resource determination module is used to determine the type of emergency task based on the patient's initial condition and to determine the combination of emergency resources that are suitable for potential changes in the patient's condition by using the correlation characteristics between historical conditions and emergency resources. The site determination module is used to determine candidate emergency sites using the patient's real-time location, and to select target emergency sites that meet the combination of emergency resources from the candidate emergency sites based on the periodic characteristics of emergency site resource occupancy. The impact analysis module is used to generate alternative emergency routes by utilizing the actual road layout between the target emergency station and the patient's location, and to determine the potential impact area of ​​the current traffic event based on the spatial propagation characteristics of historical traffic events. The route determination module is used to assess the traffic capacity of candidate emergency routes after they are affected by potential impact areas using a preset road capacity prediction model, and to determine the target emergency route from the candidate emergency routes based on the traffic capacity.

[0008] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention constructs a disease evolution tree based on the patient's initial condition and utilizes frequent sequence patterns to mine stable correlation features between historical disease evolution paths and emergency resources. This allows for the accurate early determination of emergency resource combinations that match the patient's potential disease changes, effectively avoiding the problem of mismatched emergency resources caused by misjudgment of the condition in traditional methods, and significantly improving the accuracy of emergency response and the timeliness of treatment.

[0009] This invention constructs a spatiotemporal data cube of emergency medical station resource occupancy and employs autocorrelation and mutual information analysis methods to accurately mine the periodic characteristics of emergency medical resource occupancy. This enables the determination of target emergency medical stations based on real-time resource availability rather than simply distance factors, overcoming the scheduling delays caused by the periodic changes in resource occupancy in traditional technologies and significantly improving the utilization efficiency and response speed of emergency medical resources.

[0010] This invention constructs a road capacity prediction model based on the spatial propagation characteristics of traffic incidents. This model can predict the potential impact area of ​​traffic incidents in advance and accurately assess the actual capacity of alternative emergency routes. This effectively avoids emergency vehicles getting stuck in traffic congestion, allows for the targeted selection of the optimal emergency route, significantly reduces emergency delays caused by traffic, and ensures the timeliness and effectiveness of pre-hospital emergency care. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0012] Figure 1 This is a flowchart of a pre-hospital emergency medical service intelligent dispatch and management method based on emergency medical resources according to the present invention; Figure 2 This is a framework diagram of a pre-hospital emergency medical service intelligent dispatch and management system based on emergency medical resources according to the present invention. Detailed Implementation

[0013] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided to make the description of this application more complete and comprehensive, and to fully convey the concept of the exemplary embodiments to those skilled in the art. The drawings are merely illustrative illustrations of this application 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.

[0014] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more exemplary embodiments. Numerous specific details are provided in the following description to give a full understanding of the exemplary embodiments disclosed in this application. However, those skilled in the art will recognize that the technical solutions disclosed in this application can be practiced with one or more specific details omitted, or other methods, components, steps, etc., can be employed. In other instances, well-known structures, methods, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of the disclosure of this application.

[0015] Example 1 like Figure 1 As shown, this embodiment discloses a pre-hospital emergency medical service intelligent dispatch and management method based on emergency medical resources, including: S101: Determine the type of emergency task based on the patient's initial condition, and use the correlation characteristics between historical conditions and emergency resources to determine the combination of emergency resources that are suitable for potential changes in the patient's condition. It should be noted that the core objective of this step is to determine in advance the combination of emergency resources that match the possible path of change in the patient's condition based on the preliminary assessment of the patient's current condition and the experience data on the changes in the patient's condition, so as to achieve intelligent and precise dispatch of emergency resources.

[0016] Specifically, the step of determining the type of emergency task based on the patient's initial condition and determining a combination of emergency resources suitable for potential changes in the patient's condition using the correlation characteristics between historical medical conditions and emergency resources includes: A corresponding disease evolution tree is built based on the patient's initial condition information, and the tree structure is used to represent the possible paths of disease development. It should be noted that, in this embodiment, the patient's initial condition information refers to the patient's current chief complaint symptoms and vital signs parameters obtained on-site. Typical information includes clear symptom descriptions and medical indicators such as chest pain, dyspnea, altered consciousness, and abnormal electrocardiogram.

[0017] In the specific implementation process, firstly, a large number of real emergency cases' records of disease progression are analyzed through historical case databases and medical knowledge bases to statistically analyze the probability of disease transfer between conditions, forming a reliable database of disease evolution paths. Then, using the patient's initial condition as the root node, a disease evolution tree is constructed using the disease transfer probabilities obtained from the database. Nodes in the tree represent specific disease states, and the connections between nodes and the probability values ​​along those connections reflect the possible paths and trends in disease development.

[0018] For example, if a patient's initial symptom is "chest pain," the tree structure can be specifically represented as follows: Root node: "Chest pain"; Child node: "Acute myocardial infarction", path transfer probability is 0.60; Child node: "Aortic dissection", path transition probability is 0.25; Child node: "Arrhythmia", path transition probability is 0.15.

[0019] The construction process of the aforementioned tree structure is supported by clear probabilistic data, which can accurately describe the potential trend of disease changes and provide a clear basis for subsequent resource combination and matching.

[0020] Using frequent sequence pattern mining methods, we can identify the association features between different disease pathways and combinations of emergency resources from historical disease evolution records; It should be noted that the frequent sequence pattern mining method used in this embodiment is the classic FP-Growth frequent pattern mining algorithm, which is used to efficiently mine the correlation between stable and recurring disease change paths and corresponding emergency resource combinations from historical emergency event data.

[0021] The method of using frequent sequence pattern mining to determine the association features between different disease pathways and emergency resource combinations includes: Extract patient condition change events and resource response events from historical emergency medical records, and construct patient condition change and resource response event pairs; In practice, the first step is to select complete records from the pre-hospital emergency database that clearly contain the evolution of the patient's condition and the corresponding emergency resource dispatch. Each historical record must include the nodes of the change in the patient's condition (e.g., chest pain → acute myocardial infarction) and the corresponding resource allocation information (e.g., specific dispatch records of ambulances, defibrillators, electrocardiogram monitors, and cardiologists).

[0022] For example, a history record could be: "The patient progressed from chest pain to acute myocardial infarction, and the system allocated resources including 1 ambulance, 1 defibrillator, 1 electrocardiogram monitor, and 1 cardiologist."

[0023] The event pairs are mapped to the associated transactions of the path of disease change and the combination of emergency resources, and the disease-resource association transaction library is constructed using the prefix tree method; In the specific implementation process, the above-mentioned clearly defined event pairs are represented as standard related transactions, such as: [chest pain → acute myocardial infarction: ambulance, defibrillator, electrocardiogram monitor, cardiologist]. Then, the path of disease change is used as the transaction prefix, and the transactions are stored in the disease-resource related transaction database in the form of a prefix tree (Trie tree).

[0024] It should be noted that the prefix tree storage structure is as follows: Chest pain → Acute myocardial infarction →{Ambulance, defibrillator, cardiac monitor, cardiologist} →Aortic dissection →{Ambulance, Mobile CT Unit, Surgeon} → Cardiac arrhythmia →{Ambulance, ECG monitor, cardiologist} The specific storage format of the prefix tree facilitates efficient querying and frequent pattern statistics, clearly supporting the subsequent mining process.

[0025] In the associated transaction database, a frequent pattern growth algorithm with suffix expansion is used to mine stable association patterns between disease paths and resource combinations. Understandably, this step uses the FP-Growth algorithm to specifically mine frequent patterns. First, an appropriate minimum support threshold is determined, for example, set to 0.7, which means that a certain association pattern is considered a stable association pattern only when it accounts for at least 70% of all historical transactions.

[0026] In practice, the algorithm flow includes the following steps: First, scan the related transaction database to count the frequency of occurrence of disease paths and their corresponding resource combinations in historical transactions; Then, the resource combination items under each disease path prefix are continuously expanded and identified through suffix expansion until all patterns with support greater than the set threshold are discovered.

[0027] For example, if the support of the resource combination "ambulance, defibrillator, ECG monitor, cardiologist" corresponding to the "chest pain → acute myocardial infarction" path is 0.75, then the pattern is identified as a stable association pattern.

[0028] Stable correlation patterns are used as the correlation features between different disease pathways and combinations of emergency resources; It should be noted that the stable correlation patterns obtained from the above mining represent a clear and reliable relationship between changes in the patient's condition and resource response, and are an important basis for emergency resource dispatch decisions.

[0029] For example, the specific forms of stable association patterns are shown in Table 1 below: Table 1: Example Table of Stable Association Patterns

[0030] The aforementioned stable model provides a clear and effective reference for resource allocation, ensuring that resources are accurately matched to patient needs during actual emergency care.

[0031] Based on the disease progression tree and associated features, the path of potential changes in the patient's condition is matched to determine the combination of emergency resources that are appropriate for the potential changes in the patient's condition. It should be noted that in the specific implementation of this step, the system first retrieves the disease evolution tree based on the patient's initial condition and selects the path with the highest probability as the potential path of disease change; then, combined with stable association features, it determines the combination of emergency resources corresponding to the path for the next step of specific dispatch.

[0032] For example, if a patient's initial condition is "chest pain" and the disease progression tree shows that the probability of the condition evolving into "acute myocardial infarction" is the highest (0.60), then by querying the stable association pattern table, the corresponding emergency resource combination is determined to be "ambulance, defibrillator, electrocardiogram monitor, and cardiologist", and resource scheduling preparations are made in advance to achieve rapid response.

[0033] S102: Utilize the patient's real-time location to determine candidate emergency medical stations, and based on the periodic characteristics of emergency medical station resource occupancy, select target emergency medical stations from the candidate emergency medical stations that meet the requirements of emergency medical resource combination. It should be noted that the purpose of this step is to accurately identify the target emergency medical station that can meet the patient's needs for emergency medical resources based on the patient's current real-time location and the periodic pattern of emergency medical station resource occupancy, so as to effectively shorten resource response and dispatch time.

[0034] Specifically, the step of selecting target emergency medical stations that meet the requirements of emergency medical resource combinations from candidate emergency medical stations based on the periodic characteristics of emergency medical station resource occupancy includes: Obtain resource usage data of candidate emergency medical stations over multiple periods and construct a spatiotemporal data cube of resource usage for candidate emergency medical stations; It should be noted that, in this embodiment, based on the patient's real-time location, a preset spatial search method is used, such as a spatial search algorithm based on a geographic information system (GIS), to determine multiple candidate emergency sites within a set range (e.g., within a radius of 5 kilometers around the patient).

[0035] Furthermore, for each candidate emergency medical service station, this embodiment collects its historical resource usage data over multiple consecutive periods (e.g., 7 consecutive days, 24 hours a day, in 1-hour increments). The resource usage data specifically includes the following: Ambulances: Number of available vehicles and number of vehicles currently in use; Medical staff: Number of medical staff on duty and their areas of expertise (e.g., cardiology, surgery). Emergency medical equipment: including but not limited to the available quantity and occupancy status of defibrillators, electrocardiogram monitors, mobile CT units, etc.

[0036] Subsequently, a spatiotemporal data cube is constructed according to three dimensions: time (period), space (site location), and resource category. Each data unit in the data cube is determined by a specific three-dimensional index, and the specific data unit can be represented as follows: Where t represents a specific time period (e.g., Monday 09:00-10:00); s represents the spatial location of the emergency station; and r represents the specific resource category (e.g., ambulance, medical personnel, emergency equipment). This indicates the available quantity of the resource within the corresponding period; This indicates the amount of the resource used within the corresponding period.

[0037] By constructing the aforementioned data cube, the resource occupancy status of each candidate emergency medical station is clearly expressed, ensuring that data analysis and prediction have sufficient data support.

[0038] Based on autocorrelation and mutual information analysis methods, the periodic characteristics of resource occupancy are extracted from the data cube to determine the spatiotemporal evolution law of resource occupancy; In practice, this embodiment uses a specific data analysis method to uncover the periodic patterns in resource usage at each candidate site. The specific implementation process is as follows: First, from the spatiotemporal data cube constructed above, historical occupancy data sequences of various resources are extracted for each candidate emergency medical station. For example, the hourly usage sequence of ambulance resources over the past 7 days.

[0039] Secondly, using autocorrelation analysis, the periodic characteristics of resource occupancy in the historical occupancy data sequence are analyzed. The specific calculation formula is as follows: ,in, This represents the resource usage data value at time t. It represents the average value of the data sequence, and k represents the number of lag periods, such as in hours, to analyze the periodic patterns of 24 hours, 48 ​​hours, etc.

[0040] Furthermore, mutual information analysis is used to uncover the dependencies and correlations between different periods of the data sequence. The specific formula for calculating mutual information is as follows:

[0041] In the formula: X and Y represent different periodic sequences of resource usage data, For joint probability distribution, This represents a marginal probability distribution.

[0042] By using the aforementioned autocorrelation and mutual information analysis methods, the periodic patterns and spatiotemporal evolution characteristics of resource occupancy are revealed. For example, it can be clearly identified whether the ambulance usage rate increases or decreases significantly during a specific time period, providing reliable data for subsequent target site selection.

[0043] Based on the current types of emergency medical missions and the spatiotemporal evolution of resource consumption, target emergency medical stations that meet the combination of emergency medical resources are selected. It should be noted that, based on clearly defining the current emergency rescue task type, this embodiment utilizes the characteristics and patterns of resource occupancy cycles to accurately predict the current resource availability of each candidate site.

[0044] In practice, the resource combination required for the emergency rescue mission is first clarified. For example, according to the mission type "chest pain → acute myocardial infarction" determined in step S101, the required resource combination is: 1 ambulance, 1 defibrillator, 1 electrocardiogram monitor, and 1 cardiologist.

[0045] Secondly, based on the above resource usage patterns, predict the available resources for each candidate emergency medical station in the current time period (e.g., Monday 09:00-10:00).

[0046] Finally, the predicted number of available candidate site resources is matched one by one with the clearly defined resource requirements of the task. The specific matching rules are as follows: If the available resources of a candidate site fully meet or exceed the explicit requirements of the emergency response mission, then the site is explicitly marked as the target emergency response site. If the number of resources available at a site cannot fully meet the demand, then that site will be explicitly excluded.

[0047] Finally, a list of all target emergency medical stations that meet the resource combination requirements is output for use in the next step of the dispatch plan implementation.

[0048] S103: Generate alternative emergency routes using the actual road layout between the target emergency station and the patient's location, and determine the potential impact area of ​​the current traffic event based on the spatial propagation characteristics of historical traffic events; It should be noted that this step aims to generate multiple alternative emergency routes using road network data, and combine this with the clear propagation patterns of historical traffic incidents to predict the potential impact area of ​​the current traffic incident on the alternative routes, so as to improve the accuracy of route selection and avoid emergency delays caused by traffic congestion.

[0049] Specifically, the step of generating alternative emergency routes using the actual road layout between the target emergency medical station and the patient's location includes: Acquire road network data between the target emergency medical station and the patient's location, and construct a network connectivity topology; In practice, the first step is to use a Geographic Information System (GIS) database to obtain specific road network data, which includes at least the detailed attributes of road nodes (such as intersections) and road segments (such as road segment length, number of lanes, and speed limit level). Furthermore, the network topology is clearly constructed using the aforementioned road network data, where nodes represent road network intersections or key locations, and edges represent actual road segments connecting the nodes. For example, the road network topology is specifically shown in the figure. , where V is the set of nodes, representing each intersection or junction; E is the set of edges, representing the specific road segments connecting the nodes, and the attributes of the edges include at least the road segment length, road segment level, and speed limit.

[0050] A probability transition model for route selection is established based on the actual operating trajectory of emergency vehicles to obtain the priority selection probability for different road segments; It should be noted that in this embodiment, a probability transition model is constructed using historical emergency vehicle actual operation trajectory data to determine the degree of vehicle preference for each road segment during actual driving.

[0051] In practice, a large amount of historical driving trajectory data of emergency vehicles is extracted from the emergency management information system database. Each trajectory clearly includes the road nodes passed through, the speed of the corresponding road segment, and the number of times it passes through. More specifically, the number of times each road segment is explicitly passed through in the historical trajectory of emergency vehicles is statistically analyzed, and the probability of each road segment being selected as the priority is calculated. The specific calculation formula is as follows: ,in, Indicates road segment The probability of prioritization; This indicates the route taken by emergency vehicles in historical data. The number of times; n is the total number of possible route segments between the target site and the patient's location.

[0052] The above probability transition model can clearly indicate the degree of preference of emergency vehicles for different road segments during actual emergency rescue, forming a clear basis for road segment selection preferences.

[0053] By combining network topology and probabilistic transition model, a random walk algorithm is used to generate a set of alternative emergency rescue routes; In its implementation, this embodiment employs an acyclic weighted random walk algorithm, starting from the road node corresponding to the target emergency medical station and ending at the road node corresponding to the patient's location, to perform a random path search based on the explicit priority selection probability of the aforementioned road segments; the specific process includes: Initialization: Add the starting point to the current path and mark the starting point as visited.

[0054] Next node selection: Among the current node's neighboring nodes, select the set of "unvisited" nodes; if the set is not empty, select the next node based on the probability calculated by the formula. Perform weighted random sampling on the nodes in the set and select the next hop node. Repeat and Termination: Repeat the above selection steps until the destination node is reached, forming a complete alternative route; if a dead end is reached (no unvisited neighbors and the destination has not been reached), the walk is backtracked or invalidated.

[0055] Set generation: Repeat the above random walk process multiple times (e.g., 100 times) to generate several specific sets of alternative emergency routes to ensure diverse coverage of routes.

[0056] Perform path redundancy removal on the set of alternative emergency routes and output the alternative emergency routes; It should be noted that this embodiment performs path similarity analysis on the generated set of candidate routes and removes redundant routes; In practice, the formula for calculating similarity is as follows: ,in, Indicates two routes Similarity; Indicates the length of the overlapping section between two routes; These are the total lengths of the two routes; Set a specific similarity threshold (e.g., 0.8). When the route similarity exceeds the threshold, only one route is retained. Using the above method, output the alternative emergency rescue routes after redundancy removal.

[0057] Specifically, determining the potential impact area of ​​the current traffic event based on the spatial propagation characteristics of historical traffic events includes: Based on historical traffic incident location and congestion spread trajectory data, a spatial impact topology network for traffic incidents is established. It should be noted that this step aims to clearly construct the spatial topology network of the area affected by traffic incidents using historical traffic incident data, in order to accurately analyze the spatial propagation path and patterns of traffic incidents.

[0058] In practice, the specific locations of historical traffic events (e.g., road intersections or clear road segment markers) and trajectory data of the spread of congestion are first extracted from the traffic event database. The trajectory data clearly records the specific spatial path of the congestion from the initial event location to the surrounding road nodes. Furthermore, the above data is used to construct a clear spatial impact topology network of traffic events, with specific road nodes as network nodes and clear congestion propagation paths as directed edges of the network.

[0059] For example, if a historical traffic event occurs at road node A, and historical data shows that its congestion propagation trajectory is A→B→D or A→C→E, then the following spatial influence topology network can be specifically constructed: Node: A (Initial event location) Directed edge: A→B→D Directed edge: A→C→E The clearly defined spatial topology network structure described above can accurately express the propagation path and impact range of traffic incidents.

[0060] The graph embedding method is used to mine typical diffusion patterns of the spatial impact network of traffic incidents and determine spatial propagation characteristics; It should be noted that this embodiment specifically uses graph embedding algorithms (such as the Node2Vec algorithm) to conduct in-depth analysis of the constructed spatial impact topology network, to uncover typical traffic congestion diffusion patterns, and thus to identify the spatial propagation characteristics of traffic events.

[0061] The method of using graph embedding to mine typical diffusion patterns of spatial impact networks of traffic events includes: Extract the spatial propagation paths of historical traffic incidents and construct corresponding event propagation path maps; In practice, based on the clear propagation paths recorded in historical traffic incident records, such as the clear path being node A → node B → node D, these paths are extracted to construct independent event propagation path maps. We use network embedding methods to perform low-dimensional feature mapping on the event propagation path graph, generating a path propagation pattern embedding space. It should be noted that this embodiment uses the Node2Vec algorithm for graph embedding, mapping the explicit topological relationships between nodes to a low-dimensional continuous vector space; the specific mapping process of the Node2Vec algorithm includes: Execute a random walk sampling algorithm to explicitly generate a node sequence from the nodes; Based on the node sequence, construct the optimization objective function, as shown in the following formula: ;in, This represents the specific embedding vector of node u. Represents the set of neighbors of node u; By training the above model, a low-dimensional embedding feature vector space of the event propagation path is obtained; Perform spectral clustering in the path propagation pattern embedding space to obtain typical diffusion patterns and determine spatial propagation characteristics; In practice, within a defined embedding feature space, spectral clustering is used to perform cluster analysis on the node feature vectors, resulting in several cluster sets, each representing a typical traffic incident diffusion pattern. Furthermore, the above cluster sets are regarded as explicit spatial propagation characteristics, and each typical pattern is explicitly represented as a group of nodes, representing the frequently occurring traffic congestion diffusion path within a specific area.

[0062] Based on the current location and spatial propagation characteristics of the traffic incident, predict the potential impact area of ​​the current traffic incident; In practice, when a traffic event occurs, the initial location node of the event is first determined, and its corresponding embedded feature vector is obtained. Furthermore, the similarity between the embedded feature vector of this node and the obtained typical diffusion pattern feature vector is calculated, specifically using the cosine similarity calculation formula: ;in, is the feature vector of the node at the current event location, and v is the cluster center vector of the typical diffusion pattern; The above calculations determine the typical diffusion pattern that clearly matches the current traffic incident, thereby predicting the specific potential impact area of ​​the current traffic incident and identifying a series of road nodes that may be congested. Identify the potential impact areas from the forecasts, which will then serve as important input information for the next step of assessing the traffic capacity of specific routes in S104.

[0063] S104: Evaluate the traffic capacity of alternative emergency routes after they are affected by potential impact areas using a preset road traffic capacity prediction model, and determine the target emergency route from the alternative emergency routes based on the traffic capacity. It should be noted that this step aims to use historical traffic flow data, road segment attribute information, and predicted potential impact areas of traffic incidents to establish an accurate road capacity assessment model, accurately quantify the capacity of alternative emergency routes, and achieve automatic selection of the optimal route.

[0064] Specifically, the step of evaluating the traffic capacity of alternative emergency routes after they have been affected by the potential impact area using a preset road capacity prediction model includes: Historical traffic flow data, road segment attribute information, and identified potential impact area information of alternative emergency routes are obtained to construct a road segment capacity feature space that includes event impact factors; In practice, historical traffic flow data for each specific segment of each alternative emergency route is first collected from the historical traffic database. Key parameters such as vehicle speed and vehicle density are clearly defined, and the attribute information of each segment is clearly recorded, including but not limited to parameters such as the number of lanes, speed limit level, road type, and segment length. At the same time, data on the potential impact area of ​​the current traffic incident is obtained from step S103, and road segments within the impact area of ​​the alternative emergency rescue routes are clearly marked. Furthermore, based on the above information, a clear road segment capacity feature space is constructed, specifically defined as follows: ;in, This represents the average vehicle speed of the road segment based on historical statistics. Indicates the historical average traffic density of the road segment; Indicates the number of lanes in a road segment; Indicates the speed limit level for a road section; Indicates the length of the road segment; This represents the event's impact factor, marked with a binary value: 1 indicates that the event is located within the potential impact area, and 0 indicates that the event is not located within the potential impact area.

[0065] The road segment capacity feature space is input into a preset road capacity prediction model. The model is based on a convolutional neural network with a dual-branch fusion structure, wherein the first branch is used to process the structural features of road segment attributes, and the second branch is used to process the temporal features of traffic flow and the event interference features of potential impact areas. In the specific implementation process, this embodiment constructs a dual-branch convolutional neural network model, the specific structure of which includes: The first branch (structural feature branch): The input includes fixed feature data such as road segment attribute features, such as the number of lanes, speed limit level, and road type. These fixed structural features are extracted through multiple convolutional layers. The second branch (temporal features and event interference branch): The input includes historical traffic flow data sequences and event impact markers. Temporal dynamic features are extracted through a one-dimensional convolutional layer, and the weight distribution of event interference features is explicitly captured using an attention mechanism.

[0066] In the model, structural features and temporal features are fused through a cross-branch feature interaction mechanism to obtain a fused feature representation; The design logic of the cross-branch feature interaction mechanism lies in using the static structural features of road segments (such as road width and grade) as attention guides to focus on the truly critical fluctuations in the temporal features; specifically, it includes: The features output from the first and second branches are remapped along the channel dimension to generate dimension-aligned intermediate features. In specific implementation, a fully connected layer is used to map the channel dimensions of the output features of the two branches respectively, and intermediate feature representations with the same dimensions are generated to facilitate subsequent cross-branch feature interaction; An attention mechanism is used to calculate the correlation between features in two branches, generating a cross-branch feature interaction matrix. In the specific implementation process, this embodiment explicitly adopts the dot product attention method in the attention mechanism to calculate the correlation between the features of two branches. The specific formula is expressed as follows: In this context, the first branch features are used as the query matrix Q, and the second branch features are used as the key matrix K. As the feature dimension, the cross-branch feature interaction matrix A is obtained through the above calculation; The two branch features are weighted and fused based on the feature interaction matrix to obtain a fused feature representation for accessibility prediction. In specific implementation, this embodiment performs an explicit weighted fusion operation on the feature matrix V output by the second branch based on the obtained cross-branch feature interaction matrix A. The specific calculation method is as follows: ;in, This represents the fused feature representation matrix, used for subsequent traffic capacity prediction; the feature matrix V specifically originates from the output of the second branch convolutional network, and is fused with the interaction matrix after remapping. It should be noted that this fusion method fully considers the correlation between road segment structural features and temporal dynamic features, and the fused feature representation more accurately reflects the actual traffic capacity of the road segment. Based on the fusion feature representation, the traffic capacity assessment results of the alternative emergency rescue routes under the influence of the potential impact area are output; In specific implementation, this embodiment further uses the fusion feature matrix obtained above. The input is fed into the fully connected network prediction layer, where prediction calculations are performed, and the output is the predicted traffic capacity value for each segment of each alternative emergency route, such as the predicted average vehicle speed or the time required to pass through a segment. The specific calculation formula is as follows: Where Y represents the specific predicted traffic capacity index value output by the model, such as the predicted total travel time or predicted vehicle speed of a road segment. , These are the weight matrix and bias vector of the output prediction layer, respectively.

[0067] Through the above specific implementation, we can obtain clear road segment traffic capacity assessment data for each alternative emergency rescue route under the current traffic event, so as to conduct a comparative analysis of the overall route performance in the future.

[0068] Finally, this embodiment ranks the candidate emergency routes based on the output traffic capacity prediction index (e.g., the total predicted passage time of all segments of each route). Specifically, calculate the overall capacity evaluation index for each candidate route, such as the predicted total travel time for the entire route: ;in, Indicates the overall predicted travel time for the route; This represents the length of the i-th segment in the route; This represents the predicted average speed of the i-th road segment; m represents the total number of road segments included in the route. Finally, from all alternative emergency routes, the route with the best traffic capacity index (e.g., the shortest total predicted travel time) is selected as the final target emergency route for this emergency mission, and detailed route information, including the road segments traversed, total travel time, and travel speed of each segment, is output to guide the actual navigation and dispatch of pre-hospital emergency vehicles.

[0069] Example 2 like Figure 2 As shown in the example, the parts not detailed in this embodiment are as shown in Example 1. This embodiment discloses a pre-hospital emergency medical resource-based intelligent dispatch and management system, including: The resource determination module 201 is used to determine the type of emergency task based on the patient's initial condition and to determine the combination of emergency resources that are suitable for the patient's potential changes in condition by using the correlation characteristics between historical conditions and emergency resources. The site determination module 202 is used to determine candidate emergency sites using the patient's real-time location, and to select target emergency sites that meet the combination of emergency resources from the candidate emergency sites based on the periodic characteristics of emergency site resource occupancy. The impact analysis module 203 is used to generate alternative emergency routes using the actual road layout between the target emergency station and the patient's location, and to determine the potential impact area of ​​the current traffic event based on the spatial propagation characteristics of historical traffic events. The route determination module 204 is used to evaluate the traffic capacity of the candidate emergency routes after they are affected by the potential impact area through a preset road traffic capacity prediction model, and to determine the target emergency route from the candidate emergency routes based on the traffic capacity.

[0070] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters, weights, and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0071] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A pre-hospital emergency medical service intelligent dispatch and management method based on emergency medical resources, characterized in that, include: The type of emergency task is determined based on the patient's initial condition, and the combination of emergency resources is determined by utilizing the correlation characteristics between the patient's historical condition and emergency resources to adapt to potential changes in the patient's condition. Candidate emergency medical stations are identified by utilizing the patient's real-time location, and target emergency medical stations that meet the combination of emergency medical resources are selected from the candidate emergency medical stations based on the periodic characteristics of emergency medical station resource occupancy. Alternative emergency routes are generated by utilizing the actual road layout between the target emergency station and the patient's location, and the potential impact area of ​​the current traffic event is determined based on the spatial propagation characteristics of historical traffic events. The road capacity prediction model is used to assess the traffic capacity of alternative emergency routes after they are affected by potential impact areas, and the target emergency route is determined from the alternative emergency routes based on the traffic capacity.

2. The intelligent dispatch and management method for pre-hospital emergency care based on emergency resources according to claim 1, characterized in that, The process of determining the type of emergency task based on the patient's initial condition and determining a combination of emergency resources suitable for potential changes in the patient's condition using the correlation characteristics between historical medical conditions and emergency resources includes: A corresponding disease evolution tree is built based on the patient's initial condition information, and the tree structure is used to represent the possible paths of disease development. Using frequent sequence pattern mining methods, we can identify the association features between different disease pathways and combinations of emergency resources from historical disease evolution records; By matching the path of potential changes in a patient's condition with the disease progression tree and associated features, an appropriate combination of emergency resources can be determined to match the potential changes in the patient's condition.

3. The intelligent dispatch and management method for pre-hospital emergency care based on emergency resources according to claim 2, characterized in that, The method of using frequent sequence pattern mining to determine the association features between different disease pathways and combinations of emergency resources includes: Extract patient condition change events and resource response events from historical emergency medical records, and construct patient condition change and resource response event pairs; The event pairs are mapped to the associated transactions of the path of disease change and the combination of emergency resources, and the disease-resource association transaction library is constructed using the prefix tree method; In the associated transaction database, a frequent pattern growth algorithm with suffix expansion is used to mine stable association patterns between disease paths and resource combinations. Stable association patterns are used as the association features between different disease pathways and combinations of emergency resources.

4. The intelligent dispatch and management method for pre-hospital emergency care based on emergency resources according to claim 3, characterized in that, The process of selecting target emergency medical stations that meet the requirements of emergency medical resource combinations from candidate emergency medical stations based on the periodic characteristics of emergency medical station resource occupancy includes: Obtain resource usage data of candidate emergency medical stations over multiple periods and construct a spatiotemporal data cube of resource usage for candidate emergency medical stations; Based on autocorrelation and mutual information analysis methods, the periodic characteristics of resource occupancy are extracted from the data cube to determine the spatiotemporal evolution law of resource occupancy; Based on the current emergency rescue mission types and the spatiotemporal evolution of resource consumption, target emergency rescue sites that meet the requirements of emergency rescue resource combinations are selected.

5. The intelligent dispatch and management method for pre-hospital emergency care based on emergency resources according to claim 4, characterized in that, The process of generating alternative emergency routes using the actual road layout between the target emergency station and the patient's location includes: Acquire road network data between the target emergency medical station and the patient's location, and construct a network connectivity topology; A probability transition model for route selection is established based on the actual operating trajectory of emergency vehicles to obtain the priority selection probability for different road segments; By combining network topology and probabilistic transition model, a random walk algorithm is used to generate a set of alternative emergency rescue routes; Perform path redundancy removal on the set of alternative emergency routes and output the alternative emergency routes; It should be noted that this embodiment performs path similarity analysis on the generated set of alternative routes and eliminates redundant routes.

6. The intelligent dispatch and management method for pre-hospital emergency care based on emergency resources according to claim 5, characterized in that, The method of determining the potential impact area of ​​a current traffic event based on the spatial propagation characteristics of historical traffic events includes: Based on historical traffic incident location and congestion spread trajectory data, a spatial impact topology network for traffic incidents is established. The graph embedding method is used to mine typical diffusion patterns of the spatial impact network of traffic incidents and determine spatial propagation characteristics; Based on the current location and spatial propagation characteristics of the traffic incident, the potential impact area of ​​the current traffic incident is predicted.

7. The intelligent dispatch and management method for pre-hospital emergency care based on emergency resources according to claim 6, characterized in that, The method of using graph embedding to mine typical diffusion patterns of spatial impact networks of traffic events includes: Extract the spatial propagation paths of historical traffic incidents and construct corresponding event propagation path maps; We use network embedding methods to perform low-dimensional feature mapping on the event propagation path graph, generating a path propagation pattern embedding space. Perform spectral clustering in the path propagation pattern embedding space to obtain typical diffusion patterns and determine spatial propagation characteristics.

8. The intelligent dispatch and management method for pre-hospital emergency care based on emergency resources according to claim 7, characterized in that, The assessment of the traffic capacity of alternative emergency routes after being affected by potential impact areas using a preset road capacity prediction model includes: Historical traffic flow data, road segment attribute information, and identified potential impact area information of alternative emergency routes are obtained to construct a road segment traffic capacity feature space that includes event impact factors; The road segment capacity feature space is input into a preset road capacity prediction model. The model is based on a convolutional neural network with a dual-branch fusion structure, wherein the first branch is used to process the structural features of road segment attributes, and the second branch is used to process the temporal features of traffic flow and the event interference features of potential impact areas. In the model, structural features and temporal features are fused through a cross-branch feature interaction mechanism to obtain a fused feature representation; Based on the fusion feature representation, the output shows the traffic capacity assessment results of the alternative emergency rescue routes under the influence of the potential impact area.

9. The intelligent dispatch and management method for pre-hospital emergency care based on emergency resources according to claim 8, characterized in that, The method of fusing structural features and temporal features through cross-branch feature interaction includes: The features output from the first and second branches are remapped along the channel dimension to generate dimension-aligned intermediate features. An attention mechanism is used to calculate the correlation between features in two branches, generating a cross-branch feature interaction matrix. The two branch features are weighted and fused based on the feature interaction matrix to obtain a fused feature representation for traffic capacity prediction.

10. A pre-hospital emergency medical resource-based intelligent dispatch and management system, implemented based on any one of claims 1-9, characterized in that, include: The resource determination module is used to determine the type of emergency task based on the patient's initial condition and to determine the combination of emergency resources that are suitable for potential changes in the patient's condition by using the correlation characteristics between historical conditions and emergency resources. The site determination module is used to determine candidate emergency sites using the patient's real-time location, and to select target emergency sites that meet the combination of emergency resources from the candidate emergency sites based on the periodic characteristics of emergency site resource occupancy. The impact analysis module is used to generate alternative emergency routes by utilizing the actual road layout between the target emergency station and the patient's location, and to determine the potential impact area of ​​the current traffic event based on the spatial propagation characteristics of historical traffic events. The route determination module is used to assess the traffic capacity of candidate emergency routes after they are affected by potential impact areas using a preset road capacity prediction model, and to determine the target emergency route from the candidate emergency routes based on the traffic capacity.

Images