A method and system for emergency rescue route planning on mountain highways

Abstract

This application provides a method and system for emergency rescue route planning on mountainous highways. It involves collecting traffic flow data, meteorological data, and on-site condition data from disaster sites; generating a dynamic risk map based on the meteorological data, on-site condition data, and traffic flow data; determining the rescue constraints for each rescue entity during the emergency rescue process based on the dynamic location information and rescue entity type near the disaster site; hierarchically scheduling each rescue entity based on the dynamic risk map and all rescue constraints to obtain a spatiotemporal conflict matrix; and performing multi-objective real-time route planning for each rescue entity based on the spatiotemporal conflict matrix to generate a set of spatiotemporally conflict-free collaborative rescue routes. Using the scheme of this application, efficient collaboration among multiple rescue entities can be achieved in the context of multi-disaster chain coupling evolution on mountainous highways to generate globally optimal spatiotemporally conflict-free rescue routes.

Description

Technical Field

[0001] This application relates to the field of traffic control technology, and more specifically, to a method and system for planning emergency rescue routes on mountain highways. Background Technology

[0002] Due to their complex terrain and fragile geological conditions, mountain highways are prone to geological disasters such as landslides and mudslides under extreme weather conditions (such as heavy rain and earthquakes). These disasters can also trigger secondary disasters such as traffic congestion and chain-reaction accidents. For example, a slope collapse caused by heavy rain on the Ya'an section of the Beijing-Kunming Expressway resulted in a five-hour traffic disruption in both directions. The conflict between the routes of rescue vehicles and stranded vehicles further delayed the golden rescue period. In this scenario, it is necessary to integrate disaster site data, traffic flow status, and rescue resource information in real time to achieve dynamic risk assessment and efficient rescue coordination in order to reduce disaster losses.

[0003] Traditional emergency rescue models rely on manual patrols and single-point monitoring, making it difficult to grasp the real-time evolution of disasters. Furthermore, the lack of global coordination in rescue resource allocation often leads to problems such as path conflicts and delayed responses. Existing intelligent rescue systems rely heavily on static models for disaster risk assessment (such as single meteorological warnings or geological disaster threshold judgments), failing to quantify the dynamic risks of the "geology-meteorology-accident" chain evolution (e.g., the Guizhou rainstorm model only focuses on rainfall). Moreover, rescue scheduling often employs single-objective optimization (such as shortest path), neglecting the differentiated constraints of heterogeneous rescue vehicles such as fire trucks, ambulances, and engineering vehicles (e.g., patient bump tolerance, bridge weight limits), resulting in insufficient path feasibility. Simultaneously, path collaborative planning lacks mechanisms to resolve spatiotemporal conflicts, easily causing secondary delays when multiple vehicles converge. Therefore, how to achieve efficient collaboration among multiple rescue vehicles in the context of multi-disaster chain coupling evolution on mountainous highways to generate spatiotemporally conflict-free globally optimal rescue paths has become a challenging problem for the industry. Summary of the Invention

[0004] This application provides a method and system for emergency rescue route planning on mountain highways, which can achieve efficient collaboration among multiple rescue entities to generate a globally optimal rescue route without spatiotemporal conflicts in the context of multiple disaster chain coupling and evolution on mountain highways.

[0005] Firstly, this application provides a method for planning emergency rescue routes on mountain highways, including:

[0006] Locate disaster sites on mountain highways, and then collect traffic flow data, meteorological data, and on-site condition data at the disaster sites;

[0007] Based on the meteorological data, a meteorological raster map of the area near the disaster site is determined. Based on the on-site status data and the traffic flow data, the spatiotemporal evolution of geological disasters and secondary accidents at the disaster site is determined through a pre-constructed disaster chain coupling inference model. The meteorological raster map and the spatiotemporal evolution are overlaid using raster algebra to generate a dynamic risk map near the disaster site.

[0008] The dynamic location information of each rescue entity near the disaster site is obtained, and then the rescue constraints of each rescue entity in the emergency rescue process are determined based on all the dynamic location information and the rescue entity type of each rescue entity.

[0009] Based on the dynamic risk map and all rescue constraints, each rescue entity is scheduled in a hierarchical manner, thereby obtaining the spatiotemporal conflict matrix of all rescue entities in the emergency rescue process.

[0010] Based on the spatiotemporal conflict matrix, multi-objective real-time path planning is performed on each rescue entity to generate a set of spatiotemporally conflict-free collaborative rescue paths.

[0011] In some embodiments, determining a meteorological raster map of an area near a disaster site based on the meteorological data specifically includes:

[0012] The area near the disaster site is divided into grids to obtain a climate analysis region centered on the disaster site;

[0013] The meteorological data is converted into raster values ​​and filled into the climate analysis area to generate an initial meteorological raster map.

[0014] Based on the initial meteorological raster map, the raster values ​​are dynamically updated in sync with the current meteorological data to form a meteorological raster map of the area near the disaster site.

[0015] In some embodiments, determining the spatiotemporal evolution of geological disasters and secondary accidents at a disaster location based on the on-site condition data and the traffic flow data using a pre-built disaster chain coupling simulation model specifically includes:

[0016] Extract surface deformation features near the disaster site from the on-site condition data;

[0017] Based on the surface deformation features and traffic flow data, a topographic-vehicle flow change map of the disaster site is constructed;

[0018] The terrain-vehicle flow correlation map is spatiotemporally extrapolated using a pre-constructed disaster chain coupling inference model to generate a spatiotemporal evolution probability field of the disaster chain at the disaster location.

[0019] By combining historical disaster and accident records, the spatiotemporal evolution probability field is dynamically calibrated to obtain the spatiotemporal evolution trend of geological disasters and secondary accidents occurring at the disaster site.

[0020] In some embodiments, performing raster algebraic overlay on the meteorological raster map and the spatiotemporal evolution trend to generate a dynamic risk map near the disaster location specifically includes:

[0021] The meteorological raster map is spatiotemporally aligned with the spatiotemporal evolution situation using a spatiotemporal raster aligner to obtain a spatiotemporal coupled tensor field for multimodal data fusion.

[0022] Perform grid algebra operations on the spatiotemporal coupled tensor field to obtain all risk grid values ​​in the spatiotemporal coupled tensor field of the disaster location;

[0023] A dynamic risk map of the vicinity of the disaster site is generated based on all risk grid values.

[0024] In some embodiments, determining the rescue constraints for each rescuer during the emergency rescue process based on all dynamic location information and the rescuer type of each rescuer specifically includes:

[0025] For each rescue entity, the location information at the current time point is extracted from the dynamic location information of the rescue entity based on its status tag.

[0026] Determine the inherent constraints of the rescue entity based on its type;

[0027] Calculate the time window constraint for the rescue team to reach the disaster site based on the location information;

[0028] By using a dynamic constraint adaptation algorithm to fuse the time constraint and the inherent constraint, the rescue constraint conditions of the rescue entity in the emergency rescue process are obtained, and then the rescue constraint conditions of each rescue entity in the emergency rescue process are obtained.

[0029] In some embodiments, the hierarchical scheduling of each rescue entity is performed based on the dynamic risk map and all rescue constraints, thereby obtaining the spatiotemporal conflict matrix of all rescue entities in the emergency rescue process, specifically including:

[0030] A three-dimensional scheduling and decision-making space for all rescue entities is constructed using the dynamic risk map and all rescue constraints.

[0031] Based on the aforementioned three-dimensional scheduling decision space, a comprehensive rescue path is determined for each rescue entity from its current location to the disaster site.

[0032] Based on all the comprehensive rescue routes, dispatch echelons are divided to dispatch various rescue teams;

[0033] Spatiotemporal detection is performed on the spatiotemporal convergence point of the comprehensive rescue path of all rescue entities after dispatch, and a spatiotemporal conflict matrix of all rescue entities in the emergency rescue process is generated.

[0034] In some embodiments, performing multi-objective real-time path planning for each rescue entity based on the spatiotemporal conflict matrix to generate a set of spatiotemporally conflict-free collaborative rescue paths specifically includes:

[0035] The set of rescue entities whose rescue paths are affected is identified based on the spatiotemporal conflict matrix.

[0036] Rescue paths are replanned for each rescue entity in the set of rescue entities to obtain an optimized spatiotemporal matrix that eliminates conflicts;

[0037] Based on the optimized spatiotemporal matrix, a set of spatiotemporally conflict-free collaborative rescue paths for all rescue entities is generated.

[0038] Secondly, this application provides an emergency rescue route planning system for mountain highways, the system comprising:

[0039] The data acquisition module is used to locate disaster sites on mountain highways and then collect traffic flow data, meteorological data, and on-site condition data at the disaster sites.

[0040] The processing module is used to determine a meteorological raster map of the area near the disaster site based on the meteorological data, determine the spatiotemporal evolution of geological disasters and secondary accidents at the disaster site based on the on-site status data and the traffic flow data through a pre-constructed disaster chain coupling inference model, and perform raster algebra overlay on the meteorological raster map and the spatiotemporal evolution to generate a dynamic risk map near the disaster site.

[0041] The processing module is used to acquire the dynamic location information of each rescuer near the disaster site, and then determine the rescue constraints of each rescuer in the emergency rescue process based on all the dynamic location information and the rescuer type of each rescuer.

[0042] The processing module is used to perform hierarchical scheduling of each rescue entity based on the dynamic risk map and all rescue constraints, thereby obtaining the spatiotemporal conflict matrix of all rescue entities in the emergency rescue process.

[0043] The execution module is used to perform multi-objective real-time path planning for each rescue entity based on the spatiotemporal conflict matrix, and generate a set of spatiotemporally conflict-free collaborative rescue paths.

[0044] Thirdly, this application provides a computer device, the computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the above-described emergency rescue route planning method for mountain highways.

[0045] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described emergency rescue route planning method for mountain highways.

[0046] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects:

[0047] The method and system for emergency rescue route planning on mountain highways provided in this application first locates the disaster site on the mountain highway, and then collects traffic flow data, meteorological data, and on-site status data of the disaster site; based on the meteorological data, a meteorological raster map of the area near the disaster site is determined; based on the on-site status data and the traffic flow data, the spatiotemporal evolution of geological disasters and secondary accidents at the disaster site is determined through a pre-constructed disaster chain coupling inference model; the meteorological raster map and the spatiotemporal evolution are overlaid using raster algebra to generate a dynamic risk map near the disaster site; the dynamic location information of each rescue entity near the disaster site is obtained, and then the rescue constraints of each rescue entity in the emergency rescue process are determined based on all the dynamic location information and the rescue entity type of each rescue entity; based on the dynamic risk map and all the rescue constraints, each rescue entity is hierarchically scheduled, thereby obtaining a spatiotemporal conflict matrix of all rescue entities in the emergency rescue process; based on the spatiotemporal conflict matrix, multi-objective real-time route planning is performed on each rescue entity to generate a set of spatiotemporally conflict-free collaborative rescue routes.

[0048] Therefore, this application performs multi-objective real-time path planning for each rescue entity based on the aforementioned spatiotemporal conflict matrix, generating a set of spatiotemporally conflict-free collaborative rescue paths. Firstly, determining the dynamic risk map yields a visualized map displaying different spatiotemporal risk distributions near the disaster site and updating it in real time. The dynamic risk map is determined by overlaying meteorological raster maps with the spatiotemporal evolution of the disaster chain using raster algebra, transforming real-time meteorological data (such as rainfall and wind speed), on-site deformation characteristics (such as slope cracks and water accumulation areas), and traffic flow conditions (such as vehicle density and speed) into a calculable spatiotemporally continuous risk field, thus solving the problem of multiple... The dynamic fusion of heterogeneous source data provides an intuitive risk reference for rescue route planning, helping to avoid high-risk areas and improve rescue safety. Then, determining the rescue constraints reveals all the constraints on the rescue capabilities and time required for the rescue vehicle to perform the mission. These constraints can be determined by integrating rescue vehicle type and dynamic location information, setting targeted time constraints for different rescue vehicles (e.g., binding an ambulance to a "30-minute golden rescue window"), and by clearly defining the capability boundaries of each rescue vehicle (e.g., a drone's maximum payload of 50kg and 30-minute flight time) and mission objectives (e.g., rescuing...). (Ambulances are responsible for transporting the wounded, and engineering vehicles are responsible for clearing roads), providing a quantitative basis for tiered dispatching and ensuring that planned routes align with the actual capabilities and time requirements of the rescue teams. Finally, determining the spatiotemporal conflict matrix yields a matrix displaying the conflict intensity of all rescue teams at the spatiotemporal convergence point along the rescue route after dispatching. This matrix identifies potential conflicts between rescue teams in the spatiotemporal dimension, thus providing conflict resolution targets for subsequent multi-objective route planning by accurately locating high-conflict areas and time periods. Through efficient detection, quantitative evaluation, and coordinated adaptation, it becomes a bridge connecting dynamic risk assessment and multi-objective route planning. The key hub not only addresses the technical pain points of "inefficient conflict detection and lack of basis for collaborative decision-making" in the background technology, but also constructs a closed-loop system of "real-time conflict perception - intelligent conflict resolution - global optimization and collaboration" through deep integration with risk maps and rescue constraints. This enables emergency rescue on mountain highways to shift from "experience-driven" to "data-driven," providing core technical support for achieving "spatiotemporally conflict-free collaborative rescue paths." In summary, based on the above solution, efficient collaboration among multiple rescue entities can be achieved in the context of multi-hazard chain coupling and evolution on mountain highways to generate globally optimal rescue paths that are spatiotemporally conflict-free. Attached Figure Description

[0049] Figure 1 This is an exemplary flowchart of an emergency rescue route planning method for mountain highways according to some embodiments of this application;

[0050] Figure 2 This is an operation flowchart illustrating the determination of spatiotemporal evolution states according to some embodiments of this application;

[0051] Figure 3This is an exemplary flowchart illustrating the determination of rescue constraints according to some embodiments of this application;

[0052] Figure 4 This is a schematic diagram of the structure of an emergency rescue route planning system for mountain highways, as shown in some embodiments of this application;

[0053] Figure 5 This is an internal structural diagram of a computer device for implementing an emergency rescue route planning method for mountain highways, according to some embodiments of this application. Detailed Implementation

[0054] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0055] refer to Figure 1 The figure is an exemplary flowchart of an emergency rescue route planning method for mountainous highways according to some embodiments of this application. The emergency rescue route planning method for mountainous highways mainly includes the following steps:

[0056] In step 101, the disaster location on the mountain highway is located, and traffic flow data, meteorological data and on-site condition data of the disaster location are collected.

[0057] It should be noted that, in this application, traffic flow data refers to a dataset consisting of vehicle flow, vehicle density, and vehicle speed information near the disaster site (e.g., within a diameter of 10 kilometers). This traffic flow data can provide accurate traffic status input for disaster chain coupling simulation, thereby improving the spatiotemporal accuracy and reliability of geological disaster and secondary accident prediction. Meteorological data refers to a dataset consisting of temperature, precipitation, and wind speed information near the disaster site (e.g., within a diameter of 10 kilometers). This meteorological data can be used to construct a meteorological raster map near the disaster site and overlay it with the disaster evolution probability, thereby dynamically reflecting the impact of meteorological conditions on risk distribution. Field condition data is image data that displays surface deformation characteristics (e.g., slope fissures, landslide precursors, etc.) near the disaster site (e.g., within a diameter of 10 kilometers). This field condition data can reflect changes in the geological state of the disaster site and enhance the real perception capability of disaster evolution trend simulation.

[0058] In practical implementation, locating disaster sites on mountainous highways and collecting traffic flow data, meteorological data, and on-site status data can be achieved in the following way: First, the precise coordinates (latitude and longitude) of the disaster sites on the mountainous highways can be obtained through the BeiDou high-precision positioning system, and these precise coordinates (latitude and longitude) can be matched to a pre-constructed highway vector grid to achieve precise positioning at the kilometer marker level. Next, vehicle traffic data (including timestamps, license plates, vehicle speeds, lane information, etc.) near the disaster sites (e.g., within a 10-kilometer radius) can be obtained through pre-installed roadside induction coils and electronic toll collection system interfaces on the highway. After denoising and smoothing the vehicle traffic data using the Kalman filter algorithm, real-time vehicle flow, vehicle density, and vehicle speed near the disaster sites are calculated every preset time interval (e.g., 1 second). All vehicle flow and vehicle density data are then combined and analyzed. The data is compiled from vehicle density and speed information to form traffic flow data. Then, the observation interface of the automatic weather station of the National Meteorological Administration near the disaster site can be accessed at preset time intervals (e.g., 1 second) to acquire temperature, precipitation, and wind speed information at the disaster site, and the compilation of all temperature, precipitation, and wind speed information is used as meteorological data. Finally, a swarm of multi-rotor drones can be dispatched to perform oblique photography of the ground surface near the disaster site (e.g., within a diameter of 10 kilometers), and the captured high-definition optical images are used as on-site status data. The drones are equipped with cameras and multispectral sensors and can fly along preset routes (e.g., grid-like coverage) to collect high-definition optical images with a resolution of 0.1 meters. Simultaneously, a sliding time window technique can be used to perform quality control on the meteorological data, removing outliers (e.g., wind speed exceeding the sensor's range), and a linear interpolation algorithm is used to interpolate the meteorological data to a 1-second resolution, ensuring that the meteorological data is synchronized with the traffic flow data.

[0059] In step 102, a meteorological raster map of the area near the disaster site is determined based on the meteorological data. Based on the on-site status data and the traffic flow data, the spatiotemporal evolution of geological disasters and secondary accidents at the disaster site is determined through a pre-constructed disaster chain coupling inference model. The meteorological raster map and the spatiotemporal evolution are overlaid using raster algebra to generate a dynamic risk map near the disaster site.

[0060] In some embodiments, determining a meteorological raster map of an area near a disaster site based on the meteorological data can be achieved using the following steps:

[0061] The area near the disaster site is divided into grids to obtain a climate analysis region centered on the disaster site;

[0062] The meteorological data is converted into raster values ​​and filled into the climate analysis area to generate an initial meteorological raster map.

[0063] Based on the initial meteorological raster map, the raster values ​​are dynamically updated in sync with the current meteorological data to form a meteorological raster map of the area near the disaster site.

[0064] It should be noted that in this application, the meteorological raster map is a real-time meteorological map that dynamically updates the raster values ​​based on the latest meteorological data synchronized with the initial meteorological raster map. This meteorological raster map can reflect the changes in meteorological elements near the disaster site in real time, provide real-time meteorological parameters for generating dynamic risk maps, ensure that risk assessments are in line with the latest meteorological conditions, and improve the timeliness and safety of rescue routes.

[0065] In practice, the area near the disaster site is divided into grids to obtain a climate analysis region centered on the disaster site. This can be achieved in the following way: the coordinates of the center point of the disaster site can be obtained through BeiDou high-precision positioning, and a square analysis region with a side length of 10 kilometers can be constructed with the coordinates of the center point as the origin. This square analysis region is then used as the climate analysis region centered on the disaster site. The climate analysis region is the spatial range centered on the disaster site. This climate analysis region defines the analysis boundary of meteorological data and can provide a structured storage carrier for meteorological data, ensuring that meteorological analysis covers the core area affected by the disaster.

[0066] In specific implementation, the meteorological data is converted into raster values ​​and filled into the climate analysis area to generate an initial meteorological raster map. This can be achieved in the following way: First, an equidistant grid partitioning algorithm can be used to divide the square analysis area into multiple 10m × 10m square raster units, and each raster unit is assigned a unique identifier to generate a three-dimensional matrix containing multiple addressable raster units. Then, geostatistical analysis tools in a geographic information system (such as Arc GIS) can be used to calculate the meteorological risk index of each raster unit based on the earliest recorded temperature, precipitation, and wind speed information in the meteorological data, and all meteorological risk indices are used as raster values ​​to fill into the corresponding raster units to generate the initial meteorological raster map. The initial meteorological raster map is an initial meteorological map that shows the initial spatial distribution of meteorological data in the climate analysis area. This initial meteorological raster map can preliminarily quantify the spatial distribution of meteorological elements, provide a benchmark layer for the dynamic updating of meteorological data, and ensure data continuity by providing a clear spatial reference for subsequent updates. Geostatistical analysis tools can use the Kriging method or the inverse distance weighting algorithm to fuse multi-source meteorological data into a unified meteorological risk index.

[0067] In specific implementation, the meteorological raster map of the area near the disaster site is dynamically updated based on the meteorological data of the current moment according to the initial meteorological raster map. This can be achieved in the following way: a meteorological data subscription mechanism based on the message queue telemetry transmission protocol can be established. The meteorological station data update is received every preset time interval (e.g., 1 second). The raster calculator of the geographic information system uses the Kalman filter algorithm to predict the raster value change trend and perform incremental updates of the raster value. When the raster value change exceeds the preset change threshold (e.g., 0.5), the ripple diffusion update algorithm is triggered to perform 3×3 neighborhood smoothing processing with the change point as the center, thereby obtaining a real-time refreshed meteorological raster map of the area near the disaster site.

[0068] In some embodiments, reference Figure 2 The figure is a flowchart illustrating the operation of determining the spatiotemporal evolution of a disaster situation according to some embodiments of this application. The determination of the spatiotemporal evolution of geological disasters and secondary accidents at a disaster location based on the on-site state data and traffic flow data using a pre-constructed disaster chain coupling inference model can be achieved through the following steps:

[0069] Extract surface deformation features near the disaster site from the on-site condition data;

[0070] Based on the surface deformation features and traffic flow data, a topographic-vehicle flow change map of the disaster site is constructed;

[0071] The terrain-vehicle flow correlation map is spatiotemporally extrapolated using a pre-constructed disaster chain coupling inference model to generate a spatiotemporal evolution probability field of the disaster chain at the disaster location.

[0072] By combining historical disaster and accident records, the spatiotemporal evolution probability field is dynamically calibrated to obtain the spatiotemporal evolution trend of geological disasters and secondary accidents occurring at the disaster site.

[0073] It should be noted that, in this application, the surface deformation features are quantitative characteristics that reflect the on-site ground disaster conditions (such as slope inclination, roadbed settlement, and water depth) at the disaster site. These surface deformation features accurately depict the physical state of the disaster site, providing basic parameters for constructing a terrain-vehicle flow correlation map and ensuring that subsequent simulations are consistent with the actual situation on site.

[0074] In specific implementation, the extraction of surface deformation features near disaster sites from the on-site state data can be achieved in the following way: a pre-trained deep learning-based target detection algorithm (such as YOLOv8) can be used to automatically identify surface deformation features (such as slope cracks, waterlogged areas, etc.) near disaster sites based on the on-site state data. The target detection algorithm can be trained on historical disaster datasets. For example, after setting the batch size to 16, the learning rate to 0.001, and the optimizer to train for 100 rounds, the input on-site state data is cut into 640×640 pixel sub-blocks and normalized. Then, the trained model is called to calculate the crack outline pixel length, convert it to the actual length, and calculate the crack expansion rate by comparing the results at different time points. Then, morphological dilation is performed on the waterlogged area to calculate the inundation depth, and finally, surface deformation features with geographic coordinates are generated.

[0075] In specific implementation, the topography-vehicle flow change map of the disaster site based on the surface deformation characteristics and traffic flow data can be constructed in the following way: a three-layer node map can be constructed based on the surface deformation characteristics and traffic flow data. For example, the slope layer nodes can be set as 8 key monitoring points (such as the top of the slope, the middle of the slope, etc.), the roadbed layer nodes can be set as 5 road segment units (such as the road segment 200 meters before and after the disaster point), and the vehicle flow layer nodes can be set as 3 vehicle clusters (such as congested sections, unobstructed sections, etc.). The association within the layer is represented by an adjacency matrix (such as the displacement transmission coefficient between slope nodes), and the association between layers is represented by a coupling matrix (such as the influence of slope displacement on the roadbed). The stability factor is 0.6, and the impact of roadbed damage on traffic flow speed is 0.8. The traffic flow data is then converted into vehicle flow layer node attributes (such as congestion index) to generate a terrain-vehicle flow correlation map of the disaster site. The terrain-vehicle flow correlation map is a structured map that correlates the dynamic influence relationships between various elements at the disaster site. This terrain-vehicle flow correlation map uses "slope-roadbed-vehicle flow" as the hierarchy and includes node attributes and inter-layer coupling weights. By quantifying inter-layer correlations (such as the impact weight of slope on roadbed), it provides a logical framework for spatiotemporal extrapolation, enabling multi-source data to form an organic whole and supporting disaster chain transmission analysis.

[0076] In specific implementation, the spatiotemporal evolution probability field of the disaster chain at the disaster location can be generated by performing spatiotemporal extrapolation on the terrain-vehicle flow correlation map using a pre-constructed disaster chain coupling extrapolation model. This can be achieved in the following way: an existing spatiotemporal extrapolation model (such as a Python-based cellular automata-multi-agent system coupling model) can be loaded as the disaster chain coupling extrapolation model to perform spatiotemporal extrapolation on the terrain-vehicle flow correlation map, generating the spatiotemporal evolution probability field of the disaster chain at the disaster location. Initial weights (e.g., 0.6) of the terrain-vehicle flow correlation map can be input at the initial moment. The disaster chain coupling extrapolation model extrapolates at a time step of 5 minutes, using inter-layer coupling rules (e.g., for every 1° increase in slope angle, the probability of roadbed damage increases by 5%). (For every 1-level increase in roadbed damage, the probability of rear-end collisions increases by 8%). The state parameters of each node in the terrain-traffic change map are calculated, and 1000 possible evolution paths are generated using Monte Carlo simulation. This allows for the statistical calculation of the probability of geological disasters (such as landslides) and secondary accidents (such as chain-reaction collisions) occurring at each grid cell near the disaster site within a future time period (e.g., 1 hour), forming a spatiotemporal evolution probability field. This spatiotemporal evolution probability field is a probability time series showing the likelihood of geological disasters and secondary accidents occurring near the disaster site. It reflects the spatiotemporal development possibility of the disaster chain, providing a quantitative basis for subsequent calibration. The probability distribution clearly identifies high-risk areas and time periods, improving the precision of predictions.

[0077] It should be noted that in this application, the spatiotemporal evolution trend is a dynamic trend predicting the time, location, and type of geological disasters and secondary accidents near the disaster site. This spatiotemporal evolution trend presents the complete evolution process of the disaster chain, including the disaster type and probability of occurrence at different times and locations within a future time period (e.g., 1 hour), and is presented in a three-dimensional dynamic map. This can provide a basis for disaster evolution for dynamic risk maps, ensuring that risk assessments are updated in real time as disasters develop, and enhancing the foresight of rescue and dispatch. In specific implementation, the spatiotemporal evolution probability field is dynamically calibrated in conjunction with historical disaster and accident records. The spatiotemporal evolution trend of geological disasters and secondary accidents occurring at the disaster site can be obtained in the following way: A calibration sample library can be constructed by pre-collecting historical disaster and accident records of mountain highways over the past 10 years (including slope parameters, traffic flow status, actual disaster types and occurrence times, etc.). The least squares method is used to adjust the coupling weights of the disaster chain coupling inference model (e.g., historical data shows that the actual weight of the roadbed's influence on traffic flow is 0.7, so the model parameters are corrected) to correct the deviation of the spatiotemporal evolution probability field (e.g., if the predicted probability of a certain grid cell is 20%, and the actual probability of similar historical scenarios is 15%, then it is calibrated to 17%). At the same time, the calibration coefficients are updated with the latest traffic flow data, meteorological data and on-site status data every preset update time interval (e.g., 5 minutes), and finally the calibrated spatiotemporal evolution situation is output.

[0078] In some embodiments, generating a dynamic risk map near a disaster location by performing raster algebraic overlay on the meteorological raster map and the spatiotemporal evolution trend can be achieved by the following steps:

[0079] The meteorological raster map is spatiotemporally aligned with the spatiotemporal evolution situation using a spatiotemporal raster aligner to obtain a spatiotemporal coupled tensor field for multimodal data fusion.

[0080] Perform grid algebra operations on the spatiotemporal coupled tensor field to obtain all risk grid values ​​in the spatiotemporal coupled tensor field of the disaster location;

[0081] A dynamic risk map of the vicinity of the disaster site is generated based on all risk grid values.

[0082] In specific implementation, the spatiotemporal alignment of the meteorological raster map with the spatiotemporal evolution trend, obtained by using a spatiotemporal raster aligner, can be achieved in the following way: First, the spatiotemporal raster aligner of a geographic information system can be used to unify the coordinate system of the meteorological raster map and the spatiotemporal evolution trend based on the latitude and longitude of the disaster location. Then, the time step of the spatiotemporal evolution trend is adjusted to be synchronized with that of the meteorological raster map through linear interpolation (such as bilinear interpolation), ensuring that the time granularity of the two is consistent and that the meteorological raster map is aligned with the spatiotemporal evolution trend. The edge grids of the meteorological raster map and the spatiotemporal evolution status are resampled to eliminate resolution bias and generate a spatiotemporal coupled tensor field for multimodal data fusion. The spatiotemporal coupled tensor field is a multimodal data set containing the latitude and longitude of the disaster location, meteorological values ​​at different time points, and the probability of disaster evolution. This spatiotemporal coupled tensor field is formed by spatiotemporally aligning the meteorological raster map and the spatiotemporal evolution status, which can unify the spatiotemporal reference of multi-source data, provide a consistent data carrier for subsequent raster algebra operations, ensure the spatiotemporal matching of meteorological and disaster evolution data, and improve fusion accuracy.

[0083] In specific implementation, performing raster algebra operations on the spatiotemporal coupled tensor field to obtain all risk raster values ​​in the spatiotemporal coupled tensor field of the disaster location can be achieved in the following way: Call the raster calculator tool in the geographic information system to perform raster algebra operations on the spatiotemporal coupled tensor field, obtaining the risk raster value of each raster unit in the spatiotemporal coupled tensor field of the disaster location at the current time step. For example, the proportion of meteorological factors (such as rainfall and wind speed) can be set to 0.3, and the proportion of spatiotemporal evolution (such as the probability of geological disasters and the probability of secondary accidents) can be set to 0.7. Then the calculation formula is: Risk... The raster value is equal to the meteorological raster value multiplied by 0.3 plus the evolution probability value multiplied by 0.7. Then, all calculation results are normalized using existing normalization methods (such as max-min normalization) to map them to the 0-10 range, eliminating dimensional differences, and obtaining the risk raster value, which is all the risk raster values ​​in the spatiotemporal coupled tensor field of the disaster location. The risk raster value is a rasterized value that quantifies the comprehensive risk intensity of a single raster unit in the spatiotemporal coupled tensor field. This risk raster value can provide a quantitative basis for risk classification, enabling risk assessment to shift from qualitative to quantitative and enhancing accuracy.

[0084] In practice, generating a dynamic risk map near a disaster location based on all risk raster values ​​can be achieved in the following way: classify all risk raster values ​​based on the spatiotemporal coupled tensor field. For example, the risk raster values ​​can be divided into 5 risk levels: 0-2 for low risk, 3-4 for relatively low risk, 5-6 for medium risk, 7-8 for relatively high risk, and 9-10 for high risk. Then, use the symbol system tools in the geographic information system to visualize and render the classification results, and add a timeline control to achieve dynamic playback (such as updating one frame every 5 minutes) to obtain a dynamic risk map near the disaster location.

[0085] It should be noted that in this application, the dynamic risk map is a visual map that displays the risk distribution in different times and spaces near the disaster site and updates it in real time. This dynamic risk map supports zooming in and querying the risk level and corresponding time of any grid cell, which can provide an intuitive risk reference for rescue route planning, help avoid high-risk areas, and improve rescue safety.

[0086] In step 103, the dynamic location information of each rescue body near the disaster site is obtained, and then the rescue constraints of each rescue body in the emergency rescue process are determined based on all the dynamic location information and the rescue body type of each rescue body.

[0087] In practice, obtaining the dynamic location information of various rescue entities near the disaster site can be achieved in the following way: A Beidou dual-mode positioning terminal, uniformly equipped for each rescue entity (such as ambulances, drones, and engineering vehicles), receives satellite signals and generates in real-time data such as latitude and longitude coordinates, movement speed, and remaining battery life of rescue entities near the disaster site (e.g., within a straight-line distance of 10 kilometers). This generates timestamp-based location information for each rescue entity (including latitude and longitude coordinates, movement speed, and remaining battery life). Then, using a message queue telemetry transmission protocol, all location information is encapsulated in the format of "rescue entity number + location information + status tag" and transmitted in real-time to a distributed database to obtain the dynamic location information of each rescue entity near the disaster site. In this distributed database, all rescue entities are categorized by rescue entity type (e.g., ambulances, drones, engineering vehicles, etc.). The system categorizes and stores data for rescue vehicles (such as drones and engineering vehicles), and synchronously tags the status of these vehicles (e.g., 1 for idle, ∞ for busy). The distributed database automatically updates the stored data every 30 seconds and pushes the latest location information to the rescue dispatch platform in real time via an interface, ensuring that the acquired dynamic location information is consistent with the current status of the rescue vehicles. The dynamic location information reflects the real-time location, movement speed, and status (e.g., idle / busy) of rescue vehicles (such as ambulances, drones, and engineering vehicles) over time. This dynamic location information can track the spatial movement of rescue vehicles in real time, providing a location benchmark (e.g., time window for distance to disaster point) for determining rescue constraints, and providing real-time spatial data for hierarchical dispatch and route planning. This ensures that the dispatch of rescue vehicles is aligned with their dynamic locations, reduces spatiotemporal conflicts, and improves the efficiency of collaborative rescue.

[0088] In some embodiments, reference Figure 3 The figure is an exemplary flowchart illustrating the determination of rescue constraints according to some embodiments of this application. The determination of rescue constraints for each rescuer during emergency rescue based on all dynamic location information and the rescuer type can be achieved through the following steps:

[0089] In step 1031, for each rescuer, the location information at the current time point is extracted from the dynamic location information of the rescuer based on the rescuer's status tag;

[0090] In step 1032, the inherent constraints of the rescue body are determined according to the type of rescue body.

[0091] In step 1033, the time window constraint for the rescue vehicle to reach the disaster site is calculated based on the location information;

[0092] In step 1034, the time constraint and the inherent constraint are fused by the dynamic constraint adaptation algorithm to obtain the rescue constraint conditions of the rescue body in the emergency rescue process, and then obtain the rescue constraint conditions of each rescue body in the emergency rescue process.

[0093] In specific implementation, the location information at the current time point can be extracted from the dynamic location information of the rescue body based on its status tag. This can be achieved in the following way: when the status tag of the rescue body is "idle", the location information of the rescue body at the current time point is extracted from the dynamic location information of the rescue body. The location information is real-time data reflecting the current spatial location of the rescue body (such as latitude and longitude, movement speed and remaining range). The spatial relationship between the rescue body and the disaster point can be determined through the location information, providing basic data for calculating straight-line distance and generating time window constraints, and ensuring the accuracy of the spatial reference for time window calculation.

[0094] In specific implementation, determining the inherent constraints of a rescue vehicle based on its type can be achieved in the following way: Physical constraints (e.g., a drone's maximum range of 100 km and payload of 50 kg; a fire truck's maximum water capacity of 8 tons and endurance of 4 hours) and functional constraints (e.g., an ambulance can only perform medical transport, and an engineering vehicle requires an operator) can be extracted from an existing rescue vehicle capability feature database (containing 12 categories such as ambulances, fire trucks, and drones) based on the rescue vehicle type. The set of these physical and functional constraints is then used as the inherent constraints of the rescue vehicle. These inherent constraints are fixed limitations determined by the rescue vehicle type (e.g., drone payload, ambulance passenger capacity). These inherent constraints define the capability boundaries of the rescue vehicle, preventing the dispatch of tasks exceeding its capabilities (e.g., using drones to transport supplies exceeding their payload), thus ensuring rescue feasibility.

[0095] In specific implementation, the time window constraint for the rescuer to reach the disaster site based on the location information can be implemented in the following way: First, the precise coordinates (i.e., latitude and longitude) of the disaster site can be obtained. Then, the straight-line distance between the rescuer and the disaster site can be calculated using the spherical distance formula based on the precise coordinates of the disaster site and the latitude and longitude in the rescuer's location information. Next, the estimated arrival time of the rescuer to the disaster site can be calculated based on the movement speed in the rescuer's location information, and a time window constraint can be generated by combining it with the golden time for emergency rescue (e.g., life rescue requires arrival within 30 minutes). For example, the earliest arrival time can be calculated by dividing the distance by the average speed in the movement speed, and the latest arrival time can be calculated by dividing the distance by the minimum speed in the movement speed. If the latest arrival time exceeds the golden time for emergency rescue, the latest arrival time is forcibly set as the golden time for emergency rescue. The time window constraint is the earliest and latest time range for the rescuer to reach the disaster site. This time window constraint limits the time boundary for the rescuer to carry out rescue operations, ensuring that the rescuer arrives within the effective time (neither delayed nor waiting too early), thus balancing rescue efficiency and resource utilization.

[0096] It should be noted that, in this application, the rescue constraints are all the constraints on the rescue capability and rescue time of the rescue entity when performing a rescue mission. These rescue constraints can provide clear boundaries for path planning (such as avoiding high-risk areas beyond the rescue entity's capabilities), ensuring that the planned path conforms to the actual capabilities and time requirements of the rescue entity. In specific implementation, the time constraints and the inherent constraints are fused through a dynamic constraint adaptation algorithm. The rescue constraints of the rescue entity in the emergency rescue process can be implemented in the following way: existing dynamic constraint adaptation algorithms (such as the improved Lagrange relaxation algorithm) can be used to fuse the time constraints and the inherent constraints. Constraint fusion yields the rescue constraints for the rescue entity during emergency rescue. The objective function of the dynamic constraint adaptation algorithm can be set as: min(time window violation × time constraint weight + inherent constraint violation × inherent constraint weight). The time constraint weight and inherent constraint weight can be dynamically adjusted according to the disaster level. Conflicts between time window constraints and inherent constraints are resolved through iterative optimization: if the time window requirement is too tight, resulting in insufficient drone endurance, the time window is extended; if the load constraint conflicts with the time window, life rescue tasks are prioritized. Finally, a rescue constraint vector containing time constraint factors and inherent constraint factors is generated as the rescue constraint condition.

[0097] In step 104, each rescue entity is hierarchically scheduled based on the dynamic risk map and all rescue constraints, thereby obtaining the spatiotemporal conflict matrix of all rescue entities in the emergency rescue process.

[0098] In some embodiments, the hierarchical scheduling of each rescue entity based on the dynamic risk map and all rescue constraints, thereby obtaining the spatiotemporal conflict matrix of all rescue entities during the emergency rescue process, can be achieved through the following steps:

[0099] A three-dimensional scheduling decision space for all rescue entities is constructed using the dynamic risk map and all rescue constraints.

[0100] Based on the aforementioned three-dimensional scheduling decision space, a comprehensive rescue path is determined for each rescue entity from its current location to the disaster site.

[0101] Based on all the comprehensive rescue routes, dispatch echelons are divided to dispatch various rescue teams;

[0102] Spatiotemporal detection is performed on the spatiotemporal convergence point of the comprehensive rescue path of all rescue entities after dispatch, and a spatiotemporal conflict matrix of all rescue entities in the emergency rescue process is generated.

[0103] In specific implementation, constructing a three-dimensional scheduling decision space for all rescue entities based on the dynamic risk map and all rescue constraints can be achieved in the following way: A three-dimensional coordinate system (e.g., X representing longitude and Y representing latitude) is established based on the dynamic risk map with the disaster location as the origin, and all risk grid values ​​in the dynamic risk map are mapped to the Z-axis of the three-dimensional coordinate system. Then, a comprehensive rescue weight vector for each grid cell is generated through the weight calculation logic of the improved Floyd algorithm, thereby obtaining the three-dimensional scheduling decision space for all rescue entities. The comprehensive rescue weight vector is composed of the comprehensive rescue weights of all rescue entities, calculated according to the weight calculation logic. The calculation formula for the comprehensive rescue weight of a rescue entity is: (Euclidean distance from the grid cell to the disaster site + risk grid cell value × time window constraint + time window constraint × inherent constraint factor) × type weight of the rescue entity (e.g., 1.5 for life rescue). This formula can be used to calculate the comprehensive rescue weight of each rescue entity. The three-dimensional scheduling decision space is a three-dimensional decision model that provides multi-dimensional quantitative benchmarks for rescue entity path planning. This three-dimensional scheduling decision space transforms risk, time, and capability constraints into calculable spatial weight parameters, enabling path planning to simultaneously meet safety, timeliness, and feasibility, and avoiding scheduling deviations caused by a single factor.

[0104] In specific implementation, determining the comprehensive rescue path from the location of each rescuer to the disaster site based on the three-dimensional scheduling decision space can be achieved in the following way: For each rescuer, an existing optimal path search algorithm (such as the improved A* algorithm) can be used to search for the optimal path to the disaster site in the three-dimensional scheduling decision space based on the current location coordinates of the rescuer, thus obtaining the comprehensive rescue path from the location of the rescuer to the disaster site. Through the above steps, the comprehensive rescue path from the location of each rescuer to the disaster site can be obtained; wherein, the heuristic function of the optimal path search algorithm can be set as f(n) = the actual cost from the current location to the current grid cell (i.e., the cumulative comprehensive rescue weight) + the estimated cost from the current grid cell to the destination (i.e., the risk-weighted straight-line distance). Different movement rules are set for different types of rescuers (e.g., drones can traverse areas with a risk grid value ≤ 7, while ground rescuers can only pass through areas with a risk grid value ≤ 5). The optimal path search algorithm dynamically avoids high-risk grids during the path search process based on obstacle avoidance strategies, obtaining a comprehensive rescue path from the rescuer's current location to the disaster site. The comprehensive rescue path is the optimal route for the rescuer to reach the disaster site from its current location. This comprehensive rescue path includes the coordinates of path points, the estimated arrival time, and the risk grid value, taking into account risk avoidance, time window constraints, and rescuer capability limitations. Differentiated movement rules are set for different types of rescuers to ensure that each path meets both inherent constraints (such as load capacity and range) and dynamic risk avoidance requirements, thereby improving the reliability of rescue mission execution.

[0105] In practice, the dispatching of various rescue entities based on the division of dispatch echelons according to all comprehensive rescue paths can be achieved in the following way: three dispatch echelons can be divided according to rescue priority (life rescue > material transportation > road repair). The first dispatch echelon can be drones and ambulances, the second dispatch echelon can be fire trucks and engineering vehicles, and the third dispatch echelon can be transport vehicles. Based on a hierarchical strategy of dynamic programming (e.g., the first echelon needs to start dispatching within 5 minutes after the disaster occurs, the second echelon within 10 minutes, and the third echelon within 20 minutes), rescue dispatching is carried out according to the comprehensive rescue paths of the rescue entities in each dispatch echelon.

[0106] It should be noted that in this application, the spatiotemporal conflict matrix is ​​a matrix that displays the conflict intensity of all rescue entities at the spatiotemporal convergence point of the rescue path after rescue dispatch. This spatiotemporal conflict matrix can identify potential conflicts between rescue entities in the spatiotemporal dimension, thereby providing conflict resolution targets for subsequent multi-objective path planning by accurately locating high-conflict areas and time periods, avoiding efficiency losses or safety risks caused by the aggregation of rescue entities in the same spatiotemporal region. In specific implementation, spatiotemporal detection is performed on the spatiotemporal convergence point of the comprehensive rescue path of all rescue entities after dispatch, and a spatiotemporal conflict matrix of all rescue entities in the emergency rescue process is generated. This can be achieved by using existing spatiotemporal conflict detectors (such as spatiotemporal conflict detection algorithms based on quadtree indexes) to detect the time difference (e.g., less than 5 minutes) and spatial difference (e.g., less than 100 meters) at the intersection points of all integrated rescue paths. The detection process is accelerated by using quadtree spatial indexes, which divides the integrated rescue path into multiple sub-regions according to spatial range, and only compares the paths in adjacent regions in detail to generate a spatiotemporal conflict matrix of all rescue entities during the emergency rescue process. Here, the matrix element Cij represents the conflict intensity between rescue entities i and j at the spatiotemporal convergence point.

[0107] In step 105, multi-objective real-time path planning is performed on each rescue entity based on the spatiotemporal conflict matrix to generate a set of spatiotemporally conflict-free collaborative rescue paths.

[0108] In some embodiments, generating a set of spatiotemporally conflict-free collaborative rescue paths by performing multi-objective real-time path planning for each rescue entity based on the spatiotemporal conflict matrix can be achieved through the following steps:

[0109] The set of rescue entities whose rescue paths are affected is identified based on the spatiotemporal conflict matrix.

[0110] Rescue paths are replanned for each rescue entity in the set of rescue entities to obtain an optimized spatiotemporal matrix that eliminates conflicts;

[0111] Based on the optimized spatiotemporal matrix, a set of spatiotemporally conflict-free collaborative rescue paths for all rescue entities is generated.

[0112] In specific implementation, identifying the set of rescue entities affected by the rescue path based on the spatiotemporal conflict matrix can be achieved in the following way: the spatiotemporal conflict matrix can be threshold filtered to extract all rescue entities with conflict intensity greater than a preset conflict intensity threshold (e.g., 30), and these rescue entities can be merged into conflict connected components using a disjoint-set data structure algorithm. Each conflict connected component represents a group of mutually conflicting rescue entities, and the set of all conflict connected components is then used as the set of rescue entities affected by the rescue path. Here, the set of rescue entities is the group of rescue entities affected by the rescue path conflict during the rescue process. This set of rescue entities clearly identifies the objects that need priority path adjustment. By focusing on conflict-related rescue entities, indiscriminate adjustments can be avoided, the efficiency of path optimization can be improved, and targeted targets can be provided for precise obstacle avoidance.

[0113] In specific implementation, replanning the rescue paths for each rescuer in the rescuer set to obtain the optimized spatiotemporal matrix for conflict elimination can be achieved in the following way: existing path planning optimization (such as the improved fast exploratory random tree algorithm) can be used to replan the paths of rescuers under different conflicting connected components in the rescuer set to generate the optimized spatiotemporal matrix for conflict elimination; wherein, the fast exploratory random tree algorithm can construct a search tree in three-dimensional space, assign a dedicated height layer to each rescuer (such as 300 meters for drone 1 and 350 meters for drone 2) and set a vertical safety interval (such as 50 meters), and then add conflict time constraints to the path cost function: If rescuer i and rescuer j conflict at time t, then the new path of i must satisfy the condition that the time to reach the conflict point is not equal to t ± Δt (e.g., 5 minutes) and a conflict penalty term. Iterative optimization continues until all paths satisfy the spatiotemporal separation constraints (e.g., spatial interval greater than or equal to 150 meters, time interval greater than or equal to 8 minutes), thereby generating an optimized spatiotemporal matrix that eliminates the conflict. The optimized spatiotemporal matrix is ​​a matrix that records the path parameters of each rescuer after replanning the conflicting rescuer paths. This optimized spatiotemporal matrix stores the basic data of conflict-free paths. The spatiotemporal separation constraints verified by quantification provide a reliable benchmark for generating the final collaborative path, ensuring that there are no secondary conflicts in subsequent paths.

[0114] In specific implementation, the set of spatiotemporally conflict-free collaborative rescue paths for all rescue entities based on the optimized spatiotemporal matrix can be generated in the following way: the path parameters (such as longitude, latitude, altitude, and time) in the optimized spatiotemporal matrix can be input into the 3D path rendering engine to generate a set of collaborative rescue paths with a time dimension; wherein, a spatiotemporal consistency verification algorithm can be used to verify the feasibility of the paths: check whether each rescue path meets its own rescue constraints (such as load capacity and endurance), and confirm whether all rescue paths meet the spatiotemporal separation constraints.

[0115] It should be noted that, in this application, the set of collaborative rescue paths is the sum of all spatiotemporally conflict-free paths taken by all rescue entities to reach the disaster site during the rescue process. This set of collaborative rescue paths can guide rescue entities to coordinate their actions efficiently. By integrating the rescue paths of all rescue entities and ensuring spatiotemporal consistency, it avoids congestion at spatiotemporal convergence points, thereby improving the overall rescue response speed and execution safety.

[0116] In another aspect, in some embodiments, this application provides an emergency rescue route planning system for mountainous highways, with reference to... Figure 4 The figure is a schematic diagram of the structure of an emergency rescue route planning system for mountainous highways according to some embodiments of this application. The emergency rescue route planning system 400 for mountainous highways includes: a data acquisition module 401, a processing module 402, and an execution module 403, which are described below:

[0117] The data acquisition module 401 in this application is mainly used to locate disaster sites on mountain highways, and then collect traffic flow data, meteorological data and on-site status data of the disaster sites;

[0118] Processing module 402 in this application is mainly used to determine the meteorological raster map of the area near the disaster site based on the meteorological data, determine the spatiotemporal evolution of geological disasters and secondary accidents at the disaster site through a pre-constructed disaster chain coupling inference model based on the on-site status data and the traffic flow data, and perform raster algebra overlay on the meteorological raster map and the spatiotemporal evolution to generate a dynamic risk map near the disaster site.

[0119] It should be noted that the processing module 402 in this application is also used to obtain the dynamic location information of each rescue body near the disaster site, and then determine the rescue constraints of each rescue body in the emergency rescue process based on all the dynamic location information and the rescue body type of each rescue body.

[0120] In addition, it should be noted that the processing module 402 in this application is also used to perform hierarchical scheduling of each rescue entity based on the dynamic risk map and all rescue constraints, thereby obtaining the spatiotemporal conflict matrix of all rescue entities in the emergency rescue process.

[0121] The execution module 403 in this application is mainly used to perform multi-target real-time path planning for each rescue entity based on the spatiotemporal conflict matrix, and generate a set of spatiotemporally conflict-free collaborative rescue paths.

[0122] The modules in the aforementioned emergency rescue route planning system for mountainous highways can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0123] In another embodiment, this application provides a computer device, which may be a server, and its internal structure diagram may be as follows. Figure 5 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data on emergency rescue route planning methods for mountainous highways. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements an emergency rescue route planning method for mountainous highways.

[0124] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0125] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described embodiment of the emergency rescue route planning method for mountain highways.

[0126] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps described in the above embodiment of the emergency rescue route planning method for mountain highways.

[0127] In one embodiment, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps described in the embodiment of the emergency rescue route planning method for mountain highways.

[0128] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0129] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0130] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for planning emergency rescue routes on mountain highways, characterized in that, Includes the following steps: Locate disaster sites on mountain highways, and then collect traffic flow data, meteorological data, and on-site condition data at the disaster sites; Based on the meteorological data, a meteorological raster map of the area near the disaster site is determined. Based on the on-site status data and the traffic flow data, the spatiotemporal evolution of geological disasters and secondary accidents at the disaster site is determined through a pre-constructed disaster chain coupling inference model. The meteorological raster map and the spatiotemporal evolution are overlaid using raster algebra to generate a dynamic risk map near the disaster site. The dynamic location information of each rescue entity near the disaster site is obtained, and then the rescue constraints of each rescue entity in the emergency rescue process are determined based on all the dynamic location information and the rescue entity type of each rescue entity. Based on the dynamic risk map and all rescue constraints, each rescue entity is scheduled in a hierarchical manner, thereby obtaining the spatiotemporal conflict matrix of all rescue entities in the emergency rescue process. Based on the spatiotemporal conflict matrix, multi-objective real-time path planning is performed on each rescue entity to generate a set of spatiotemporally conflict-free collaborative rescue paths; Specifically, generating a dynamic risk map near a disaster location by performing raster algebraic overlay on the meteorological raster map and the spatiotemporal evolution trend includes: The meteorological raster map is spatiotemporally aligned with the spatiotemporal evolution situation using a spatiotemporal raster aligner to obtain a spatiotemporal coupled tensor field for multimodal data fusion. Perform grid algebra operations on the spatiotemporal coupled tensor field to obtain all risk grid values ​​in the spatiotemporal coupled tensor field of the disaster location; A dynamic risk map of the vicinity of the disaster site is generated based on all risk grid values.

2. The method as described in claim 1, characterized in that, Determining the meteorological raster map of the area near the disaster site based on the aforementioned meteorological data specifically includes: The area near the disaster site is divided into grids to obtain a climate analysis region centered on the disaster site; The meteorological data is converted into raster values ​​and filled into the climate analysis area to generate an initial meteorological raster map. Based on the initial meteorological raster map, the raster values ​​are dynamically updated in sync with the current meteorological data to form a meteorological raster map of the area near the disaster site.

3. The method as described in claim 1, characterized in that, Based on the on-site condition data and traffic flow data, a pre-constructed disaster chain coupling simulation model is used to determine the spatiotemporal evolution of geological disasters and secondary accidents at the disaster location. Specifically, this includes: Extract surface deformation features near the disaster site from the on-site condition data; Based on the surface deformation features and traffic flow data, a topographic-vehicle flow correlation map of the disaster site is constructed; The terrain-vehicle flow correlation map is spatiotemporally extrapolated using a pre-constructed disaster chain coupling inference model to generate a spatiotemporal evolution probability field of the disaster chain at the disaster location. By combining historical disaster and accident records, the spatiotemporal evolution probability field is dynamically calibrated to obtain the spatiotemporal evolution trend of geological disasters and secondary accidents occurring at the disaster site.

4. The method as described in claim 1, characterized in that, Based on all dynamic location information and the rescuer type of each rescuer, the specific rescue constraints for each rescuer during the emergency rescue process include: For each rescue entity, the location information at the current time point is extracted from the dynamic location information of the rescue entity based on its status tag. Determine the inherent constraints of the rescue entity based on its type; Calculate the time window constraint for the rescue team to reach the disaster site based on the location information; By using a dynamic constraint adaptation algorithm to fuse the time window constraint and the inherent constraint, the rescue constraint conditions of the rescue entity in the emergency rescue process are obtained, and then the rescue constraint conditions of each rescue entity in the emergency rescue process are obtained.

5. The method as described in claim 1, characterized in that, Based on the dynamic risk map and all rescue constraints, each rescue entity is scheduled in a hierarchical manner, thereby obtaining the spatiotemporal conflict matrix of all rescue entities in the emergency rescue process, specifically including: A three-dimensional scheduling and decision-making space for all rescue entities is constructed using the dynamic risk map and all rescue constraints. Based on the aforementioned three-dimensional scheduling decision space, a comprehensive rescue path is determined for each rescue entity from its current location to the disaster site. Based on all the comprehensive rescue routes, dispatch echelons are divided to dispatch various rescue teams; Spatiotemporal detection is performed on the spatiotemporal convergence point of the comprehensive rescue path of all rescue entities after dispatch, and a spatiotemporal conflict matrix of all rescue entities in the emergency rescue process is generated.

6. The method as described in claim 1, characterized in that, Based on the aforementioned spatiotemporal conflict matrix, multi-objective real-time path planning is performed on each rescue entity to generate a set of spatiotemporally conflict-free collaborative rescue paths, specifically including: The set of rescue entities whose rescue paths are affected is identified based on the spatiotemporal conflict matrix. Rescue paths are replanned for each rescue entity in the set of rescue entities to obtain an optimized spatiotemporal matrix that eliminates conflicts; Based on the optimized spatiotemporal matrix, a set of spatiotemporally conflict-free collaborative rescue paths for all rescue entities is generated.

7. A mountain highway emergency rescue route planning system, which uses the method described in any one of claims 1 to 6 for emergency rescue route planning, characterized in that, The system includes: The data acquisition module is used to locate disaster sites on mountain highways and then collect traffic flow data, meteorological data, and on-site condition data at the disaster sites. The processing module is used to determine a meteorological raster map of the area near the disaster site based on the meteorological data, determine the spatiotemporal evolution of geological disasters and secondary accidents at the disaster site based on the on-site status data and the traffic flow data through a pre-constructed disaster chain coupling inference model, and perform raster algebra overlay on the meteorological raster map and the spatiotemporal evolution to generate a dynamic risk map near the disaster site. The processing module is used to acquire the dynamic location information of each rescuer near the disaster site, and then determine the rescue constraints of each rescuer in the emergency rescue process based on all the dynamic location information and the rescuer type of each rescuer. The processing module is used to perform hierarchical scheduling of each rescue entity based on the dynamic risk map and all rescue constraints, thereby obtaining the spatiotemporal conflict matrix of all rescue entities in the emergency rescue process. The execution module is used to perform multi-objective real-time path planning for each rescue entity based on the spatiotemporal conflict matrix, and generate a set of spatiotemporally conflict-free collaborative rescue paths.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the mountain highway emergency rescue route planning method as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the emergency rescue route planning method for mountain highways as described in any one of claims 1 to 6.

Images