Low-altitude airway flow field sensitive area dynamic identification optimization method and system based on set simulation

Abstract

The invention discloses a low-altitude airway flow field sensitive area dynamic identification optimization method based on set simulation, and the method comprises the steps: building a low-altitude flow field preprocessing data base with consistent time and space based on Beidou subdivision grids and multi-source heterogeneous data fusion; constructing a low-altitude airspace digital twinning environment based on the data; based on the low-altitude airspace digital twin environment and the cellular automaton-fluid coupling model, generating a diversified flow field evolution scene covering extreme weather and equipment faults; based on a set simulation result, extracting a high-conflict probability region through a spatio-temporal clustering algorithm and quantifying region risk features; generating an air route planning scheme meeting security constraints through a multi-objective evolutionary algorithm based on the quantitative regional risk features; on the basis of a low-altitude airspace digital twin environment and an air route planning scheme, verifying the feasibility of the air route planning scheme through historical data playback and virtual-real fusion test; and according to a verification feedback result, carrying out dynamic feedback optimization on the low-altitude air route flow field sensitive area identification and air route planning scheme.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of low-altitude economy, and particularly relates to a low-altitude air route flow field sensitive area dynamic identification optimization method and system based on ensemble simulation. BACKGROUND

[0002] The part provided in this part is merely background information related to the present disclosure, which is not necessarily prior art.

[0003] With the gradual opening of low-altitude airspace and the rapid growth of aircraft, the safe operation and efficient management of low-altitude airspace are facing severe challenges. Existing route planning relies on static maps and experience, and lacks dynamic perception and response capability to complex environment, emergencies and high-density flight behavior;

[0004] However, the existing low-altitude economic air route flow field sensitive area determination method still has certain defects. In the fusion of multi-source heterogeneous data, there is a lack of unified space-time reference, which leads to serious data island problem, and it is impossible to build a high-fidelity and dynamically updated low-altitude flow field data base, which affects the accurate quantification of flow field evolution law, relies on a single scene or limited disturbance, and cannot generate diversified flow field evolution scenes covering extreme weather and equipment failure, resulting in insufficient adaptability of the system to unexpected conditions, and difficulty in simulating the complex dynamic process of aircraft avoidance behavior and flow field congestion. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application proposes a low-altitude air route flow field sensitive area dynamic identification optimization method based on ensemble simulation, comprising the following steps:

[0006] Step 1: Based on the Beidou subdivision grid and the fusion of multi-source heterogeneous data, a space-time consistent low-altitude flow field preprocessing data base is constructed.

[0007] Step 2: Based on the data obtained in step 1, a low-altitude airspace digital twin environment is constructed.

[0008] Step 3: Based on the low-altitude airspace digital twin environment and the cellular automaton-fluid coupling model constructed in step 2, diversified flow field evolution scenes covering extreme weather and equipment failure are generated.

[0009] Step 4: Based on the ensemble simulation results of the diversified flow field evolution scenes generated in step 3, high-conflict probability areas are extracted by a space-time clustering algorithm and the area risk features are quantified.

[0010] Step 5: Based on the quantified area risk features obtained in step 4, a route planning scheme satisfying safety constraints is generated by a multi-objective evolutionary algorithm.

[0011] Step 6: Based on the low-altitude airspace digital twin environment constructed in step 2 and the flight path planning scheme generated in step 5, the feasibility of the flight path planning scheme is verified through historical data playback and virtual-real fusion testing.

[0012] Step 7: According to the verification feedback result of step 6, the low-altitude flight path flow field sensitive area identification and flight path planning scheme are dynamically feedback optimized.

[0013] In some embodiments, step 1 includes the following steps:

[0014] Step 1-1: According to the low-altitude airspace management requirements, the earth's surface is divided into multi-scale three-dimensional grids, each grid is assigned a unique code through a satellite system, and the grid boundary and attributes are updated regularly, including: spatial attributes, obstacle attributes, airspace type attributes, environmental attributes, safety attributes, time attributes;

[0015] Step 1-2: Real-time collection of flight data of aircraft through sensing system, integration of geographic information, weather information, airspace rules and traffic flow prediction model;

[0016] Step 1-3: Format unification, coordinate system conversion and timestamp alignment processing of collected multi-source heterogeneous data, denoising and interpolation completion of aircraft trajectory data through data smoothing and noise suppression algorithm;

[0017] Step 1-4: Mapping the processed aircraft position data to Beidou grid code through timestamp synchronization protocol, and calculating the dynamic distance between the aircraft and the ground obstacles combined with geographic information;

[0018] Step 1-5: Training multi-modal neural network through confidence weighted fusion of complementary data sources of radar detection and ADS-B broadcast, joint modeling of image, point cloud and text data to extract flow field features;

[0019] Step 1-6: Based on Beidou grid, the number of aircraft in each grid is counted to calculate the traffic density of low-altitude area, and the time series model is used to predict the future flow field change trend;

[0020] Step 1-7: According to the extracted flow field features and prediction results, construct conflict discrimination rules based on Beidou grid, dynamically adjust rule threshold according to real-time data, and form a spatiotemporally consistent low-altitude flow field preprocessing data base;

[0021] The spatio-temporal consistent representation ensures that the multi-source data in the spatial dimension and the temporal dimension can be accurately aligned, ensuring the spatio-temporal consistency of the data, for example, the aircraft position data, weather data, terrain data, etc. must be in the same spatial coordinate system, and the timestamps are aligned, so as to accurately reflect the real-time state of the low-altitude airspace. By using the Beidou subdivision grid as a unified spatio-temporal reference, seamless fusion of different source data can be achieved. If there is a lack of spatio-temporal consistency, it will lead to data fusion errors, affecting the accuracy of subsequent sensitive area identification and route planning.

[0022] In some embodiments, the smoothing and noise suppression algorithm in steps 1-3 is Kalman filtering or sliding window smoothing algorithm.

[0023] In some embodiments, step 2 includes the following steps:

[0024] Step 2-1: Based on the spatio-temporally consistent low-altitude flow field preprocessing data obtained in step 1, a dynamic and static element database of the low-altitude airspace is constructed;

[0025] Step 2-2: Through the GIS three-dimensional modeling tool, the terrain, building and obstacle static data are fused with the real-time dynamic data of the aircraft to generate a low-altitude airspace three-dimensional digital twin model;

[0026] Step 2-3: A physics simulation engine is introduced into the digital twin model to simulate the kinematic characteristics and aerodynamic behavior of the aircraft;

[0027] Step 2-4: Based on the computational fluid dynamics method, the spatio-temporal evolution law of the aircraft density, conflict probability and flow distribution in the low-altitude airspace is quantitatively analyzed;

[0028] Step 2-5: Based on the historical flight data, the digital twin model is simulated and played back to verify the model's ability to reproduce complex scenarios. By comparing the simulation results with the actual operation data, the flow field modeling algorithm parameters are adjusted.

[0029] In some embodiments, step 3 includes the following steps:

[0030] Step 3-1: Based on the low-altitude airspace digital twin model constructed in step 2, the three-dimensional model and real-time flow field data are used as the initial flow field state to provide basic input for scene generation; the parameters of the cellular automaton-fluid coupling model are loaded, and the boundary conditions for scene generation are defined;

[0031] Step 3-2: Parameter perturbations including wind speed mutation, heavy rain and low visibility are introduced into the model to simulate aircraft avoidance behavior and flow field congestion phenomena, set fault modes and trigger corresponding emergency path planning logic;

[0032] Step 3-3: Generate regular flow field evolution scenarios based on historical flight data, superimpose random perturbations in the base scenarios through Monte Carlo method or Latin hypercube sampling technique to generate diversified extreme scenarios;

[0033] Step 3-4: Divide the low-altitude airspace into a cell grid, such as 500m x 500m, define the movement rules of the aircraft between the grids, simulate the discrete state transition process, and the movement probability of the aircraft between the cells is based on the speed, obstacle distance and conflict risk, which is realized as:

[0034] ,

[0035] In the formula, represents the probability of the aircraft moving to the adjacent cell, represents the maximum speed of the aircraft, represents the current speed of the aircraft, represents the distance to the nearest obstacle, represents the distance to the nearest high-risk area;

[0036] Step 3-5: Calculate the influence of air flow on the trajectory of the aircraft based on the Navier-Stokes equation, which is realized as:

[0037] ,

[0038] In the formula, represents the lateral offset of the aircraft trajectory due to wind speed, k represents the air flow influence coefficient, k takes the value range [0, 1], represents the wind speed, represents the speed of the aircraft itself, represents the time step; and coupled with the discrete rules of cellular automata, which is realized as:

[0039] ,

[0040] In the formula, represents the adjusted movement probability considering the influence of fluid, represents the coupling coefficient, takes the value range [0, 1], represents the maximum safe moving distance of the aircraft; through iterative calculation, the discrete rules of the cellular automaton and the continuous solution of the fluid equation are interacted to drive the evolution of the flow field, thereby generating a diversified flow field evolution scenario covering extreme weather and equipment failure; when the wind speed suddenly changes, for example, the Navier-Stokes equation calculates the influence of air flow on the aircraft trajectory, and the cellular automaton adjusts the aircraft movement rules according to the influence, forming a dynamic coupling, through iterative calculation, the discrete cellular rules and the continuous fluid equation are interacted to drive the evolution of the flow field, thereby generating a diversified flow field evolution scenario covering extreme weather and equipment failure.

[0041] In some embodiments, step 4 includes the following steps:

[0042] Step 4-1: access the diversified flow field evolution scenario set generated in step 3, extract key spatio-temporal features as input data for clustering and risk quantification, the key spatio-temporal features include spatial position features, time features, flow field environment features, conflict features, and grid features; through the OPTICS density clustering algorithm, spatio-temporal clustering is performed based on the spatial position and timestamp of the aircraft trajectory and conflict points, and high-conflict probability areas are identified, and the clustering parameters are dynamically adjusted according to different scenario types;

[0043] Step 4-2: count the occurrence frequency of conflict events in different scenarios in each clustering area, and calculate the regional comprehensive risk score combining the aircraft density, conflict intensity and environmental impact factors;

[0044] Step 4-3: evaluate the sensitivity of each risk indicator through single factor variation analysis method, and identify the key factors that have the greatest impact on the overall risk;

[0045] Step 4-4: compare the high-conflict areas identified by clustering with historical accident data to verify the accuracy of the identification of high-conflict areas, if the accuracy of the identification of high-conflict areas is greater than a preset threshold, then execute step 4-5, otherwise return to step 4-1, dynamically adjust the parameters of the clustering algorithm, re-perform spatio-temporal clustering analysis, and repeat the verification process of step 4-4 until the accuracy requirement is met;

[0046] Step 4-5: based on the digital twin model constructed in step 2, the identified sensitive areas are visualized and a structured risk report is generated; according to the newly generated scenario data continuously generated in step 3, the distribution of sensitive areas and risk characteristics are regularly updated.

[0047] In some embodiments, step 5 includes the following steps:

[0048] Step 5-1: access the high-conflict area coordinates, risk level and dynamic update results output by step 4, and obtain the sudden scenario parameters generated in step 3, and construct a multi-objective route optimization model including navigation safety, flight efficiency and environmental adaptability;

[0049] Step 5-2: Based on the optimization model, a multi-objective evolutionary algorithm is used to perform population iterative calculations to generate a Pareto optimal solution set, and the risk index of sensitive areas is transformed into a constraint condition to ensure that the generated route prioritizes avoiding high-risk areas;

[0050] Step 5-3: Based on the sudden changes in scenario parameters, adjust the objective function weights in real time and rerun the optimization algorithm. The comprehensive optimization objective function is as follows:

[0051] ,

[0052] In the formula, F represents the value of the comprehensive optimization objective function. Indicates risk score, Indicates the length of the voyage. Indicates the level of ambient noise. This represents the weighting coefficient, which monitors the airspace status in real time through a digital twin model and triggers route replanning when a risk threshold is detected to be exceeded.

[0053] Step 5-4: Based on the pre-computed Pareto optimal solution set, the optimal route that fits the current scenario is selected through a multi-attribute decision algorithm, and the generated route is imported into the digital twin environment to verify its safety and efficiency.

[0054] Safety constraints include prohibiting entry into high-risk areas (conflict probability ≥ 0.7 times / hour); minimum safe distance between aircraft and obstacles ≥ 50m; minimum safe distance between aircraft ≥ 100m; and constraints are implemented by converting risk indicators of sensitive areas into hard and soft constraints.

[0055] In some embodiments, the multi-attribute decision algorithm described in steps 5-4 is TOPSIS or a fuzzy decision method.

[0056] In some embodiments, step 6 includes the following steps:

[0057] Step 6-1: Based on the digital twin model built in Step 2, acquire historical flight data and integrate it with the route planning scheme and high-risk area coordinates output in Step 4;

[0058] Step 6-2: Reconstruct the historical airspace status in the digital twin model, overlay the route planning scheme and sensitive area markers, and build a virtual-real integrated test environment;

[0059] Step 6-3: Using the spatiotemporal synchronization function of the digital twin model, replay historical flight events and inject route planning schemes to simulate their execution effect in the original scenario;

[0060] Step 6-4: Combine the extreme weather and equipment failure scenarios generated in Step 3 to conduct extended tests, compare and analyze the simulation playback results with actual historical operation data, and verify the feasibility of optimizing the route strategy.

[0061] A dynamic identification and optimization system for sensitive areas of low-altitude airway flow field based on ensemble simulation includes a multi-source data acquisition and preprocessing module, a digital twin modeling flow field simulation module, an ensemble simulation multi-scenario generation module, a sensitive area identification and risk quantification module, a flight path optimization dynamic adjustment module, a digital twin verification simulation playback module, and a dynamic feedback optimization module.

[0062] In some embodiments, step 7 includes the following steps:

[0063] Step 7-1: Based on the verification feedback results, automatically execute the closed-loop optimization process: If the verification finds that the accuracy of high-risk area identification is lower than the threshold, the system automatically lowers the risk threshold of sensitive areas, expands the coverage of high-conflict areas, and re-executes the cluster analysis in step 4; If the conflict rate of the optimized route exceeds the standard in extreme scenarios, the system dynamically adjusts the weight coefficients of the multi-objective evolutionary algorithm.

[0064] Step 7-2: Update the parameters of the scenarios that failed during verification to the cellular automata-fluid coupling model in Step 3;

[0065] Step 7-3: Automatically trigger the verification replay of step 6 after each optimization.

[0066] The multi-source data acquisition and preprocessing module is based on the fusion of BeiDou grid segmentation and multi-source heterogeneous data to construct a spatiotemporally consistent low-altitude flow field data base.

[0067] The digital twin modeling flow field simulation module is used to combine the high-quality data from the multi-source data acquisition and preprocessing module with three-dimensional holographic interaction technology to construct a reproducible low-altitude airspace digital twin environment.

[0068] The multi-scenario generation module of the collection simulation is based on the digital twin platform of the digital twin modeling flow field simulation module and the cellular automata-fluid coupling model to generate diverse flow field evolution scenarios covering extreme weather and equipment failures;

[0069] The sensitive area identification risk quantification module extracts high conflict probability areas and quantifies their risk characteristics by combining the set simulation results of the set simulation multi-scenario generation module with the spatiotemporal clustering algorithm.

[0070] The route optimization dynamic adjustment module generates a route planning scheme that meets safety constraints based on the risk quantification index of the sensitive area identification and risk quantification module and a multi-objective evolutionary algorithm, and dynamically adapts to parameter changes in sudden scenarios.

[0071] The digital twin verification simulation playback module, through the digital twin platform of the digital twin modeling flow field simulation module and the optimized scheme of the sensitive area identification risk quantification module, verifies the feasibility of the route strategy by playing back historical data and conducting virtual-real fusion tests.

[0072] The dynamic feedback optimization module optimizes based on the verification feedback from the digital twin verification simulation playback module;

[0073] Preferably, the multi-source data acquisition and preprocessing module is based on the fusion of BeiDou-based grid segmentation and multi-source heterogeneous data to construct a spatiotemporally consistent low-altitude flow field data base. According to the requirements of low-altitude airspace management, the Earth's surface is divided into multi-scale three-dimensional grids. Combining the centimeter-level positioning capability of the BeiDou-3 satellite navigation system, a unique code is assigned to each grid. Through the real-time dynamic positioning function of BeiDou satellites, the grid boundaries and attributes are updated regularly. Through 5G-A sensing base stations, ADS-B, lidar, and UAV airborne sensors, the position, speed, altitude, and mission type data of aircraft are collected in real time. The module integrates the topography, 3D building models, meteorological station data, and airspace rules of the geographic information system, and accesses external information such as traffic flow prediction models and emergency databases. The module performs format unification, coordinate system transformation, and timestamp alignment on the heterogeneous data to eliminate data silos.

[0074] Preferably, the multi-source data acquisition and preprocessing module filters noise points in the aircraft trajectory using Kalman filtering or sliding window smoothing algorithms, interpolates and completes missing data, cross-checks BeiDou grid IDs with GIS data, maps the aircraft position to BeiDou grid IDs using a timestamp synchronization protocol, calculates the distance between the aircraft and ground obstacles using a GIS 3D model, trains a multimodal neural network by weighting the complementary data sources of radar detection and ADS-B broadcast according to confidence level, jointly models images, point clouds, and text, extracts flow field features, statistically analyzes the distribution of aircraft numbers based on the BeiDou grid, calculates the flow density in the low-altitude region, predicts future flow field changes using a time series model, constructs conflict rules based on the BeiDou grid, and dynamically adjusts thresholds.

[0075] Preferably, the digital twin modeling flow field simulation module is used to combine the high-quality data from the multi-source data acquisition and preprocessing module with three-dimensional holographic interaction technology to construct a reproducible low-altitude airspace digital twin environment; based on the spatiotemporally consistent data provided by the multi-source data acquisition and preprocessing module, a dynamic and static element database of the low-altitude airspace is constructed; static data such as terrain, buildings, and obstacles are fused with dynamic data of aircraft status through GIS three-dimensional modeling tools to generate a high-fidelity three-dimensional digital twin model; a physical simulation engine is introduced to simulate the kinematic characteristics and aerodynamic behavior of the aircraft; and computational fluid dynamics methods are used to quantify the spatiotemporal evolution of aircraft density, conflict probability, and flow distribution in the low-altitude airspace.

[0076] Preferably, the digital twin modeling flow field simulation module, combined with Unity's XR module, renders the 3D model and flow field data into a holographic view in real time, supports multi-view interactive operation, injects real-time data from multi-source data acquisition modules into the digital twin model through the MQTT protocol, drives the model to update dynamically, performs simulation playback based on historical flight data, verifies the model's ability to reproduce complex scenarios, and adjusts the flow field modeling algorithm by comparing simulation results with actual airspace operation data.

[0077] Preferably, the multi-scenario generation module of the set simulation is based on the digital twin platform of the digital twin modeling flow field simulation module and the cellular automata-fluid coupling model to generate diverse flow field evolution scenarios covering extreme weather and equipment failures; based on the three-dimensional model and real-time flow field data of the digital twin modeling flow field simulation module, the initial flow field state is constructed as the basic input for scenario generation, the parameters of the cellular automata-fluid coupling model are loaded, and the boundary conditions for scenario generation are defined. The disturbances of wind speed change, heavy rainfall, and low visibility parameters are introduced to simulate the aircraft avoidance behavior and flow field congestion. Fault modes such as aircraft sensor failure and navigation system abnormality are set to trigger emergency path planning logic.

[0078] Preferably, the ensemble simulation multi-scenario generation module generates conventional flow field evolution scenarios based on historical flight data. Through Monte Carlo methods or Latin hypercube sampling, random perturbations are superimposed on the basic scenarios to generate diverse extreme scenarios. The low-altitude airspace is divided into cellular grids, the movement rules of the aircraft are defined, and its discrete state transitions between cells are simulated. The influence of airflow on the aircraft trajectory is calculated based on the Navier-Stokes equations and coupled with the discrete behavior of cellular automata. Through iterative calculation, the discrete rules of cellular automata interact with the continuous solutions of the fluid equations to drive the flow field evolution.

[0079] Preferably, the sensitive area identification and risk quantification module extracts high-conflict-probability areas and quantifies their risk characteristics through the ensemble simulation results of the ensemble simulation multi-scenario generation module and a spatiotemporal clustering algorithm; it accesses the flow field evolution scenario output by the ensemble simulation multi-scenario generation module, extracts key spatiotemporal features as clustering and quantification inputs, and uses the OPTICS density clustering algorithm to cluster based on the spatial location and timestamps of the aircraft trajectory and conflict points to identify high-conflict-probability areas. It dynamically adjusts the clustering parameters according to the scenario type, counts the frequency of conflict events in each scenario within the clustered area, calculates risk scores by combining aircraft density, conflict intensity, and environmental factors, assesses the sensitivity of risk indicators through single-factor changes, identifies the key factors with the greatest impact on risk, compares the clustering results with historical accident data to verify the accuracy of high-conflict area identification, displays the sensitive areas in the form of heat maps or 3D markers through a digital twin platform, and outputs a risk report. The distribution of sensitive areas is updated regularly based on the new scenario data added by the ensemble simulation module.

[0080] Preferably, the route optimization dynamic adjustment module generates a route planning scheme that meets safety constraints based on the risk quantification indicators and multi-objective evolutionary algorithm of the sensitive area identification and risk quantification module, and dynamically adapts to changes in parameters of sudden scenarios; it accesses the coordinates, risk level, and dynamic update results of high-conflict areas output by the sensitive area identification and risk quantification module as safety constraints for route planning, obtains the sudden scenario parameters from the set simulation multi-scenario generation module as external variables for dynamic adjustment, constructs an optimization model that includes navigation safety, flight efficiency, and environmental adaptability objectives, and generates a Pareto optimal solution set through population iteration to balance safety and efficiency. The objectives of safety, efficiency, and environmental adaptability are achieved by transforming risk indicators in sensitive areas into hard and soft constraints. This ensures that generated routes prioritize avoiding high-risk areas. Based on changes in parameters during sudden scenarios, the objective function weights are updated in real time, and the optimization algorithm is rerun. The airspace status is monitored in real time through a digital twin platform. When a risk threshold is exceeded or scenario parameters change, route replanning is triggered. Based on a pre-calculated Pareto optimal solution set, the optimal route suitable for the current scenario is quickly selected using TOPSIS or fuzzy decision-making methods. The generated route is then imported into the digital twin modeling flow field simulation module to verify its safety and efficiency in the simulated scenario.

[0081] Preferably, the digital twin verification simulation playback module uses the digital twin platform of the digital twin modeling flow field simulation module and the optimization scheme of the sensitive area identification risk quantification module to verify the feasibility of the route strategy through historical data playback and virtual-real fusion testing. It obtains historical flight data from the digital twin modeling flow field simulation module and connects the route planning scheme and high-risk area coordinates of the sensitive area identification risk quantification module. Based on the three-dimensional model of the digital twin platform, it reconstructs the historical airspace state, overlays the optimized route scheme and sensitive area markers to form a virtual-real fusion test scenario. Through the spatiotemporal synchronization function of the digital twin platform, it replays historical flight events and injects the route planning scheme to simulate its execution effect in the original scenario. Combined with the extreme weather and equipment failure scenarios generated by the multi-scenario simulation module, the simulation playback results are compared with the actual historical operation data to verify the feasibility of the optimized route strategy.

[0082] By combining a digital twin modeling flow field simulation module with 3D holographic interaction and a physical simulation engine, high-fidelity modeling of dynamic and static elements in low-altitude airspace is achieved, the evolution law of flow field is accurately quantified, and the ability to reproduce and predict complex scenarios is improved. Based on the cellular automata-fluid coupling model and Monte Carlo sampling, diverse flow field scenarios covering extreme weather and equipment failures are generated, breaking through the limitation of a single scenario and enhancing the system's adaptability to emergencies.

[0083] By employing spatiotemporal clustering algorithms and dynamic parameter adjustments, high-conflict-probability areas are accurately extracted and risk characteristics are quantified. Combined with historical data verification, dynamic updates and risk warnings for sensitive areas are achieved, enhancing airspace management security. Multi-objective evolutionary algorithms and Pareto optimal solution sets are used to balance safety, efficiency, and environmental adaptability. Flight routes are dynamically adjusted in conjunction with real-time scene changes to ensure the robustness and real-time response capability of flight paths under complex conditions. Attached Figure Description

[0084] Fig. 1 This is a schematic diagram of a dynamic identification and optimization method for sensitive areas of low-altitude airway flow field based on ensemble simulation.

[0085] Fig. 2 This is a flowchart illustrating the operation of a dynamic identification and optimization method for sensitive areas of low-altitude airway flow field based on ensemble simulation. Detailed Implementation

[0086] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments, and the advantages of the present invention in the above and / or other aspects will become clearer.

[0087] like Figs. 1-2 As shown, a dynamic identification and optimization method for sensitive areas of low-altitude airway flow field based on ensemble simulation includes the following steps:

[0088] Step 1: Based on the fusion of BeiDou-based grid partitioning and multi-source heterogeneous data, construct a spatiotemporally consistent low-altitude flow field preprocessing data base.

[0089] Step 2: Construct a low-altitude airspace digital twin environment based on the data obtained in Step 1.

[0090] Step 3: Based on the low-altitude airspace digital twin environment and cellular automata-fluid coupling model constructed in Step 2, generate diverse flow field evolution scenarios covering extreme weather and equipment failures.

[0091] Step 4: Based on the ensemble simulation results of the diverse flow field evolution scenarios generated in Step 3, high conflict probability regions are extracted and regional risk characteristics are quantified using a spatiotemporal clustering algorithm.

[0092] Step 5: Based on the quantified regional risk characteristics obtained in Step 4, a route planning scheme that meets safety constraints is generated using a multi-objective evolutionary algorithm.

[0093] Step 6: Based on the low-altitude airspace digital twin environment constructed in Step 2 and the route planning scheme generated in Step 5, the feasibility of the route planning scheme is verified through historical data playback and virtual-real fusion testing.

[0094] Step 7: Based on the verification feedback results of Step 6, perform dynamic feedback optimization on the identification of sensitive areas of low-altitude airfield flow and the airway planning scheme.

[0095] Step 1 includes the following steps:

[0096] Step 1-1: Based on the requirements of low-altitude airspace management, the Earth's surface is divided into a multi-scale three-dimensional grid. Each grid is assigned a unique code through a satellite system, and the grid boundaries and attributes are updated regularly. The attributes include: spatial attributes, obstacle attributes, airspace type attributes, environmental attributes, safety attributes, and time attributes.

[0097] Steps 1-2: Collect flight data of the aircraft in real time through the sensor system, and integrate geographic information, meteorological information, airspace rules and traffic flow prediction models.

[0098] Steps 1-3: Standardize the format, transform the coordinate system, and align the timestamps of the collected multi-source heterogeneous data. Use data smoothing and noise suppression algorithms to denoise and interpolate the aircraft trajectory data.

[0099] Steps 1-4: Map the processed aircraft position data to the BeiDou grid code through a timestamp synchronization protocol, and calculate the dynamic distance between the aircraft and ground obstacles by combining geographic information.

[0100] Steps 1-5: Train a multimodal neural network by using confidence-weighted fusion of complementary data sources, including radar detection and ADS-B broadcast, to jointly model image, point cloud, and text data in order to extract flow field features.

[0101] Steps 1-6: Based on the BeiDou grid statistics, the distribution of the number of aircraft in each grid is statistically analyzed, the flow density in the low-altitude region is calculated, and the future flow field change trend is predicted through time series model.

[0102] Steps 1-7: Based on the extracted flow field features and prediction results, construct conflict discrimination rules based on the BeiDou grid, dynamically adjust the rule thresholds according to real-time data, and form a spatiotemporally consistent low-altitude flow field preprocessing data base.

[0103] The smoothing and noise suppression algorithms described in steps 1-3 are Kalman filtering or sliding window smoothing algorithms.

[0104] Step 2 includes the following steps:

[0105] Step 2-1: Based on the spatiotemporally consistent low-altitude flow field preprocessing data obtained in Step 1, construct a dynamic and static element database for the low-altitude airspace.

[0106] Step 2-2: Using GIS 3D modeling tools, static data of terrain, buildings and obstacles are fused with real-time dynamic data of the aircraft to generate a 3D digital twin model of the low-altitude airspace.

[0107] Steps 2-3: Introduce a physical simulation engine into the digital twin model to simulate the kinematic characteristics and aerodynamic behavior of the aircraft.

[0108] Steps 2-4: Based on computational fluid dynamics, quantitative analysis is performed on the spatiotemporal evolution of aircraft density, collision probability, and flow distribution in low-altitude airspace.

[0109] Steps 2-5: Simulate and replay the digital twin model based on historical flight data to verify the model's ability to reproduce complex scenarios. Adjust the flow field modeling algorithm parameters by comparing the simulation results with actual operating data.

[0110] Step 3 includes the following steps:

[0111] Step 3-1: Based on the low-altitude airspace digital twin model constructed in Step 2, use its 3D model and real-time flow field data as the initial flow field state to provide basic input for scene generation; load the parameters of the cellular automata-fluid coupling model and define the boundary conditions for scene generation.

[0112] Step 3-2: Introduce parameter disturbances, including sudden wind speed changes, heavy rainfall, and low visibility, into the model to simulate aircraft avoidance behavior and flow field congestion, set fault modes, and trigger corresponding emergency path planning logic.

[0113] Step 3-3: Generate conventional flow field evolution scenarios based on historical flight data, and superimpose random perturbations on the basic scenarios using the Monte Carlo method or Latin hypercube sampling technique to generate diverse extreme scenarios.

[0114] Steps 3-4: Divide the low-altitude airspace into cellular grids, define the movement rules of the aircraft between grids, and simulate its discrete state transition process.

[0115] Steps 3-5: Calculate the impact of airflow on the aircraft trajectory based on the Navier-Stokes equations and couple it with the discrete rules of cellular automata; through iterative calculation, make the discrete rules of cellular automata interact with the continuous solutions of the fluid equations to drive the flow field evolution, thereby generating diverse flow field evolution scenarios covering extreme weather and equipment failures.

[0116] Step 4 includes the following steps:

[0117] Step 4-1: Access the diverse flow field evolution scenario set generated in Step 3, extract key spatiotemporal features as input data for clustering and risk quantification; use the OPTICS density clustering algorithm to perform spatiotemporal clustering based on the spatial location and timestamp of the aircraft trajectory and conflict point, identify high conflict probability areas, and dynamically adjust clustering parameters according to different scenario types.

[0118] Step 4-2: Calculate the frequency of conflict events in different scenarios within each cluster area, and combine aircraft density, conflict intensity and environmental impact factors to calculate the comprehensive risk score of the area.

[0119] Step 4-3: Assess the sensitivity of each risk indicator using single-factor variation analysis to identify the key factors that have the greatest impact on overall risk.

[0120] Step 4-4: Compare the high-conflict regions identified by clustering with historical accident data to verify the accuracy of high-conflict region identification. If the accuracy of high-conflict region identification is greater than the preset threshold, proceed to step 4-5; otherwise, return to step 4-1, dynamically adjust the parameters of the clustering algorithm, re-perform spatiotemporal clustering analysis, and repeat the verification process of step 4-4 until the accuracy requirements are met.

[0121] Steps 4-5: Based on the digital twin model built in step 2, visualize the identified sensitive areas and generate a structured risk report; based on the newly generated scenario data in step 3, regularly update the distribution and risk characteristics of sensitive areas.

[0122] Step 5 includes the following steps:

[0123] Step 5-1: Integrate the coordinates, risk level, and dynamic update results of the high-conflict area output in Step 4, and simultaneously obtain the emergency scenario parameters generated in Step 3 to construct a multi-objective route optimization model that includes navigation safety, flight efficiency, and environmental adaptability.

[0124] Step 5-2: Based on the optimization model, perform population iterative calculations using a multi-objective evolutionary algorithm to generate a Pareto optimal solution set, and transform the risk index of sensitive areas into constraints to ensure that the generated routes prioritize avoiding high-risk areas.

[0125] Step 5-3: Adjust the objective function weights in real time and rerun the optimization algorithm based on changes in parameters of sudden scenarios. Monitor the airspace status in real time through the digital twin model and trigger route replanning when a risk threshold is detected to be exceeded.

[0126] Step 5-4: Based on the pre-computed Pareto optimal solution set, the optimal route suitable for the current scenario is selected through a multi-attribute decision algorithm, and the generated route is imported into the digital twin environment to verify its safety and efficiency.

[0127] The multi-attribute decision algorithm described in step 5-4 is TOPSIS or a fuzzy decision method.

[0128] Step 6 includes the following steps:

[0129] Step 6-1: Based on the digital twin model built in Step 2, acquire historical flight data and integrate it with the route planning scheme and high-risk area coordinates output in Step 4.

[0130] Step 6-2: Reconstruct the historical airspace status in the digital twin model, overlay the route planning scheme and sensitive area markers, and build a virtual-real integrated test environment.

[0131] Step 6-3: Using the spatiotemporal synchronization function of the digital twin model, replay historical flight events and inject route planning schemes to simulate their execution effect in the original scenario.

[0132] Step 6-4: Combine the extreme weather and equipment failure scenarios generated in Step 3 to conduct extended tests, compare and analyze the simulation playback results with actual historical operation data, and verify the feasibility of optimizing the route strategy.

[0133] Step 7 includes the following steps:

[0134] Step 7-1: Based on the verification feedback results, automatically execute the closed-loop optimization process: If the verification finds that the accuracy of high-risk area identification is lower than the threshold, the system automatically lowers the risk threshold of sensitive areas, expands the coverage of high-conflict areas, and re-executes the cluster analysis in step 4; If the conflict rate of the optimized route exceeds the standard in extreme scenarios, the system dynamically adjusts the weight coefficients of the multi-objective evolutionary algorithm.

[0135] Step 7-2: Update the parameters of the scenarios that failed during verification to the cellular automata-fluid coupling model in Step 3;

[0136] Step 7-3: Automatically trigger the verification replay of step 6 after each optimization.

[0137] A dynamic identification and optimization system for sensitive areas of low-altitude airway flow field based on ensemble simulation includes a multi-source data acquisition and preprocessing module, a digital twin modeling flow field simulation module, an ensemble simulation multi-scenario generation module, a sensitive area identification and risk quantification module, a flight path optimization dynamic adjustment module, a digital twin verification simulation playback module, and a dynamic feedback optimization module.

[0138] The multi-source data acquisition and preprocessing module is based on the fusion of BeiDou grid segmentation and multi-source heterogeneous data to construct a spatiotemporally consistent low-altitude flow field data base.

[0139] The digital twin modeling flow field simulation module is used to combine the high-quality data from the multi-source data acquisition and preprocessing module with three-dimensional holographic interaction technology to construct a reproducible low-altitude airspace digital twin environment.

[0140] The multi-scenario generation module of the collection simulation is based on the digital twin platform of the digital twin modeling flow field simulation module and the cellular automata-fluid coupling model to generate diverse flow field evolution scenarios covering extreme weather and equipment failures;

[0141] The sensitive area identification risk quantification module extracts high conflict probability areas and quantifies their risk characteristics by combining the set simulation results of the set simulation multi-scenario generation module with the spatiotemporal clustering algorithm.

[0142] The route optimization dynamic adjustment module generates a route planning scheme that meets safety constraints based on the risk quantification index of the sensitive area identification and risk quantification module and a multi-objective evolutionary algorithm, and dynamically adapts to parameter changes in sudden scenarios.

[0143] The digital twin verification simulation playback module, through the digital twin platform of the digital twin modeling flow field simulation module and the optimized scheme of the sensitive area identification risk quantification module, verifies the feasibility of the route strategy by playing back historical data and conducting virtual-real fusion tests.

[0144] The dynamic feedback optimization module optimizes based on the verification feedback from the digital twin verification simulation playback module.

[0145] The multi-source data acquisition and preprocessing module is based on the fusion of BeiDou-based grid subdivision and multi-source heterogeneous data to construct a spatiotemporally consistent low-altitude flow field data base. According to the requirements of low-altitude airspace management, the Earth's surface is divided into multi-scale three-dimensional grids. Combining the centimeter-level positioning capability of the BeiDou-3 satellite navigation system, a unique code is assigned to each grid. Through the real-time dynamic positioning function of BeiDou satellites, grid boundaries and attributes are updated periodically. Real-time data on aircraft position, speed, altitude, and mission type are collected via 5G-A sensing base stations, ADS-B, lidar, and UAV onboard sensors. The module integrates topographic data, 3D building models, weather station data, and airspace rules from the geographic information system, and accesses external information from traffic flow prediction models and emergency databases. Heterogeneous data is formatted, coordinate-system converted, and timestamp-aligned to eliminate data silos.

[0146] The multi-source data acquisition and preprocessing module filters noise points in the aircraft trajectory using Kalman filtering or sliding window smoothing algorithms, interpolates and completes missing data, cross-checks BeiDou grid IDs with GIS data, maps the aircraft position to BeiDou grid IDs using a timestamp synchronization protocol, and calculates the distance between the aircraft and ground obstacles using a GIS 3D model. It also trains a multimodal neural network by weighting radar detection and ADS-B broadcast complementary data sources according to confidence levels, jointly models images, point clouds, and text, extracts flow field features, statistically analyzes the distribution of aircraft numbers based on the BeiDou grid, calculates the flow density in low-altitude areas, predicts future flow field changes using a time series model, constructs conflict rules based on the BeiDou grid, and dynamically adjusts thresholds.

[0147] The digital twin modeling flow field simulation module is used to combine the high-quality data from the multi-source data acquisition and preprocessing module with 3D holographic interaction technology to construct a reproducible low-altitude airspace digital twin environment. Based on the spatiotemporally consistent data provided by the multi-source data acquisition and preprocessing module, a dynamic and static element database of the low-altitude airspace is constructed. Through GIS 3D modeling tools, static data such as terrain, buildings, and obstacles are fused with dynamic data of aircraft status to generate a high-fidelity 3D digital twin model. A physical simulation engine is introduced to simulate the kinematic characteristics and aerodynamic behavior of the aircraft. Computational fluid dynamics methods are used to quantify the spatiotemporal evolution of aircraft density, collision probability, and flow distribution in the low-altitude airspace.

[0148] The digital twin modeling flow field simulation module, combined with Unity's XR module, renders the 3D model and flow field data into a holographic view in real time, supports multi-view interactive operation, and injects real-time data from the multi-source data acquisition module into the digital twin model through the MQTT protocol to drive the model to update dynamically. Simulation playback is performed based on historical flight data to verify the model's ability to reproduce complex scenarios. By comparing the simulation results with actual airspace operation data, the flow field modeling algorithm is adjusted.

[0149] The multi-scenario generation module of the collection simulation is based on the digital twin platform of the digital twin modeling flow field simulation module and the cellular automata-fluid coupling model to generate diverse flow field evolution scenarios covering extreme weather and equipment failures. Based on the 3D model and real-time flow field data of the digital twin modeling flow field simulation module, the initial flow field state is constructed as the basic input for scenario generation. The parameters of the cellular automata-fluid coupling model are loaded, and the boundary conditions for scenario generation are defined. The disturbances of parameters such as sudden wind speed changes, heavy rainfall, and low visibility are introduced to simulate the avoidance behavior of aircraft and flow field congestion. Fault modes such as aircraft sensor failure and navigation system abnormality are set to trigger emergency path planning logic.

[0150] The ensemble simulation multi-scenario generation module generates conventional flow field evolution scenarios based on historical flight data. Through Monte Carlo methods or Latin hypercube sampling, random perturbations are superimposed on the basic scenarios to generate diverse extreme scenarios. The low-altitude airspace is divided into cellular grids, the movement rules of the aircraft are defined, and its discrete state transitions between cells are simulated. The influence of airflow on the aircraft trajectory is calculated based on the Navier-Stokes equations and coupled with the discrete behavior of cellular automata. Through iterative calculation, the discrete rules of cellular automata interact with the continuous solutions of the fluid equations to drive the flow field evolution.

[0151] The sensitive area identification and risk quantification module extracts high-conflict-probability areas and quantifies their risk characteristics by combining the ensemble simulation results from the ensemble simulation multi-scenario generation module with a spatiotemporal clustering algorithm. It also accesses the flow field evolution scenario output by the ensemble simulation multi-scenario generation module, extracting key spatiotemporal features as clustering and quantification inputs. The OPTICS density clustering algorithm clusters areas based on the spatial location and timestamps of the aircraft trajectory and conflict points, identifying high-conflict-probability areas. Clustering parameters are dynamically adjusted according to the scenario type, and the frequency of conflict events in each scenario within the clustered area is statistically analyzed. Risk scores are calculated by combining aircraft density, conflict intensity, and environmental factors. The sensitivity of risk indicators is assessed through single-factor changes, identifying the key factors with the greatest impact on risk. The clustering results are compared with historical accident data to verify the accuracy of high-conflict area identification. Sensitive areas are displayed as heatmaps or 3D markers through a digital twin platform, and a risk report is output. The distribution of sensitive areas is updated periodically based on new scenario data added by the ensemble simulation module.

[0152] The route optimization dynamic adjustment module, based on the risk quantification indicators and multi-objective evolutionary algorithm of the sensitive area identification and risk quantification module, generates route planning schemes that meet safety constraints and dynamically adapts to parameter changes in sudden scenarios. It uses the coordinates, risk levels, and dynamic update results of high-conflict areas output by the sensitive area identification and risk quantification module as safety constraints for route planning, and obtains sudden scenario parameters from the multi-scenario generation module as external variables for dynamic adjustment. It constructs an optimization model encompassing navigation safety, flight efficiency, and environmental adaptability objectives, and generates a Pareto optimal solution set through population iteration to balance safety, efficiency, and environmental factors. The efficiency and environmental adaptability objectives transform the risk indicators of sensitive areas into hard and soft constraints, ensuring that the generated routes prioritize avoiding high-risk areas. Based on changes in parameters during sudden scenarios, the objective function weights are updated in real time and the optimization algorithm is rerun. The airspace status is monitored in real time through a digital twin platform. When a risk threshold is exceeded or scenario parameters change, route replanning is triggered. Based on the pre-calculated Pareto optimal solution set, the optimal route that adapts to the current scenario is quickly selected through TOPSIS or fuzzy decision-making methods. The generated route is then imported into the digital twin modeling flow field simulation module to verify its safety and efficiency in the simulated scenario.

[0153] The digital twin verification simulation playback module, through the digital twin platform of the digital twin modeling flow field simulation module and the optimization scheme of the sensitive area identification risk quantification module, verifies the feasibility of the route strategy through historical data playback and virtual-real fusion testing. It acquires historical flight data from the digital twin modeling flow field simulation module and integrates the route planning scheme and high-risk area coordinates from the sensitive area identification risk quantification module. Based on the 3D model of the digital twin platform, it reconstructs the historical airspace state, overlays the optimized route scheme and sensitive area markers, forming a virtual-real fusion test scenario. Through the spatiotemporal synchronization function of the digital twin platform, it replays historical flight events and injects the route planning scheme, simulating its execution effect in the original scenario. Combined with extreme weather and equipment failure scenarios generated by the multi-scenario simulation module, the simulation playback results are compared with actual historical operating data to verify the feasibility of the optimized route strategy.

[0154] This solution integrates multi-source heterogeneous data based on BeiDou-based grid partitioning. Through spatiotemporal alignment, noise filtering, and feature extraction, it constructs a unified low-altitude flow field data foundation, eliminating data silos. The pre-processed data is combined with 3D holographic technology to build a high-fidelity digital twin airspace model containing dynamic aircraft status and static geographic information. This model simulates aircraft kinematics and aerodynamics, quantifies flow field evolution, and uses a cellular automata-fluid coupling model to overlay disturbances such as extreme weather and equipment failures onto the digital twin platform. This generates diverse flow field evolution scenarios, simulating complex dynamic behaviors such as aircraft avoidance and congestion. Spatiotemporal features are extracted from the simulation, high-conflict-probability areas are identified using density clustering algorithms, and risk scores are calculated by combining aircraft density, environmental factors, etc. The accuracy of sensitive areas is verified by historical accident data, and the risk distribution is dynamically updated. Based on the risk quantification results, a multi-objective evolutionary algorithm is used to generate a route plan that takes into account safety, efficiency, and environmental adaptability. Dynamic replanning is triggered by real-time monitoring of emergency scenario parameters to quickly select the optimal path. Historical data is replayed through a digital twin platform and injected into the route planning scheme. The feasibility is verified by combining virtual and real fusion testing, and the effect is evaluated by comparing with actual operation data. The verification results are fed back to the optimization module.

[0155] This invention provides a method and system for dynamic identification and optimization of sensitive areas in low-altitude flight paths based on ensemble simulation. Many methods and approaches exist for implementing this technical solution; the above description is merely a preferred embodiment. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this invention, and these improvements and modifications should also be considered within the scope of protection of this invention. All components not explicitly stated in this embodiment can be implemented using existing technologies.

Claims

1. A dynamic identification and optimization method for sensitive areas of low-altitude flight path flow field based on ensemble simulation, characterized in that, Includes the following steps: Step 1: Based on the fusion of BeiDou-based grid partitioning and multi-source heterogeneous data, construct a spatiotemporally consistent low-altitude flow field preprocessing data base; Step 2: Construct a low-altitude airspace digital twin environment based on the data obtained in Step 1; Step 3: Based on the low-altitude airspace digital twin environment and cellular automata-fluid coupling model constructed in Step 2, generate diverse flow field evolution scenarios covering extreme weather and equipment failures, and obtain parameters for sudden scenarios; Step 4: Based on the ensemble simulation results of the diverse flow field evolution scenarios generated in Step 3, high conflict probability regions are extracted and regional risk characteristics are quantified using a spatiotemporal clustering algorithm; Step 5: Based on the quantified regional risk characteristics obtained in Step 4 and the emergency scenario parameters generated in Step 3, a route planning scheme that meets safety constraints is generated through a multi-objective evolutionary algorithm; Step 6: Based on the low-altitude airspace digital twin environment constructed in Step 2 and the route planning scheme generated in Step 5, the feasibility of the route planning scheme is verified through historical data playback and virtual-real fusion testing. Step 7: Based on the verification feedback results of Step 6, perform dynamic feedback optimization on the identification of sensitive areas of low-altitude airfield flow and the airway planning scheme.

2. The method for dynamic identification and optimization of sensitive areas of low-altitude flight path flow field based on ensemble simulation according to claim 1, characterized in that, Step 1 includes the following steps: Step 1-1: Based on the requirements of low-altitude airspace management, the Earth's surface is divided into a multi-scale three-dimensional grid. Each grid is assigned a unique code through a satellite system, and the grid boundaries and attributes are updated regularly. Steps 1-2: Collect flight data of the aircraft in real time through the sensor system, and integrate geographic information, meteorological information, airspace rules and traffic flow prediction models; Steps 1-3: Perform format unification, coordinate system transformation and timestamp alignment on the collected multi-source heterogeneous data, and perform noise reduction and interpolation on the aircraft trajectory data through data smoothing and noise suppression algorithms; Steps 1-4: Map the processed aircraft position data to the BeiDou grid code through a timestamp synchronization protocol, and calculate the dynamic distance between the aircraft and ground obstacles by combining geographic information; Steps 1-5: Train a multimodal neural network by using the complementary data sources of radar detection and ADS-B broadcast with confidence weighting, and jointly model image, point cloud and text data to extract flow field features; Steps 1-6: Based on the BeiDou grid, statistically analyze the distribution of aircraft numbers within each grid, calculate the flow density in the low-altitude region, and predict future flow field trends using time series models; Steps 1-7: Based on the extracted flow field features and prediction results, construct conflict discrimination rules based on the BeiDou grid, dynamically adjust the rule thresholds according to real-time data, and form a spatiotemporally consistent low-altitude flow field preprocessing data base.

3. The method for dynamic identification and optimization of sensitive areas of low-altitude flight path flow field based on ensemble simulation according to claim 2, characterized in that, The smoothing and noise suppression algorithms described in steps 1-3 are Kalman filtering or sliding window smoothing algorithms.

4. The method for dynamic identification and optimization of sensitive areas of low-altitude flight path flow field based on ensemble simulation according to claim 1, characterized in that, Step 2 includes the following steps: Step 2-1: Based on the spatiotemporally consistent low-altitude flow field preprocessing data obtained in Step 1, construct a dynamic and static element database for the low-altitude airspace; Step 2-2: Using GIS 3D modeling tools, static data of terrain, buildings and obstacles are fused with real-time dynamic data of the aircraft to generate a 3D digital twin model of the low-altitude airspace. Steps 2-3: Introduce a physics simulation engine into the digital twin model to simulate the kinematic characteristics and aerodynamic behavior of the aircraft; Steps 2-4: Based on computational fluid dynamics, quantitative analysis is performed on the spatiotemporal evolution of aircraft density, collision probability, and flow distribution in low-altitude airspace; Steps 2-5: Simulate and replay the digital twin model based on historical flight data to verify the model's ability to reproduce complex scenarios. Adjust the flow field modeling algorithm parameters by comparing the simulation results with actual operating data.

5. The method for dynamic identification and optimization of sensitive areas of low-altitude flight path flow field based on ensemble simulation according to claim 1, characterized in that, Step 3 includes the following steps: Step 3-1: Based on the low-altitude airspace digital twin model constructed in Step 2, use its 3D model and real-time flow field data as the initial flow field state to provide basic input for scene generation; load the parameters of the cellular automata-fluid coupling model and define the boundary conditions for scene generation; Step 3-2: Introduce parameter disturbances, including sudden wind speed changes, heavy rainfall, and low visibility, into the model to simulate aircraft avoidance behavior and flow field congestion, set fault modes, and trigger corresponding emergency path planning logic; Step 3-3: Generate conventional flow field evolution scenarios based on historical flight data, and superimpose random perturbations on the basic scenarios using the Monte Carlo method or Latin hypercube sampling technique to generate diverse extreme scenarios; Steps 3-4: Divide the low-altitude airspace into a cellular grid, define the movement rules of the aircraft between the grids, and simulate its discrete state transition process; obtain the probability of the aircraft moving to an adjacent cell. : in, Indicates the maximum speed of the aircraft. Indicates the current speed of the aircraft. Indicates the distance to the nearest obstacle. Indicates the distance to the nearest high-risk area; Steps 3-5: Calculate the impact of airflow on the aircraft trajectory based on the Navier-Stokes equations, and obtain the lateral deviation of the aircraft trajectory due to wind speed. : ， Where k represents the airflow influence coefficient, and the value of k ranges from [0,1]. Indicates wind speed. Indicates the speed of the aircraft itself. Indicates the time step; By coupling the influence of airflow on the aircraft trajectory with the discrete rules of cellular automata, the adjusted movement probability considering the fluid effects is obtained. : in, Represents the coupling coefficient. The value range is [0,1]. Indicates the maximum safe distance the aircraft can travel; Through iterative calculations, the discrete rules of cellular automata interact with the continuous solutions of fluid equations, driving the evolution of the flow field and generating diverse flow field evolution scenarios covering extreme weather and equipment failures, thereby obtaining parameters for sudden scenarios.

6. The method for dynamic identification and optimization of sensitive areas of low-altitude flight path flow field based on ensemble simulation according to claim 1, characterized in that, Step 4 includes the following steps: Step 4-1: Access the diverse flow field evolution scenario set generated in Step 3, extract key spatiotemporal features as input data for clustering and risk quantification; use the OPTICS density clustering algorithm to perform spatiotemporal clustering based on the spatial location and timestamp of the aircraft trajectory and conflict point, identify high conflict probability areas, and dynamically adjust clustering parameters according to different scenario types. Step 4-2: Calculate the frequency of conflict events in different scenarios within each cluster area, and combine aircraft density, conflict intensity and environmental impact factors to calculate the comprehensive risk score of the area; Step 4-3: Assess the sensitivity of each risk indicator using single-factor variation analysis to identify the key factors that have the greatest impact on overall risk; Step 4-4: Compare the high-conflict regions identified by clustering with historical accident data to verify the accuracy of high-conflict region identification; if the high-conflict region identification accuracy is greater than the preset threshold, proceed to step 4-5; otherwise, return to step 4-1, dynamically adjust the parameters of the clustering algorithm, re-perform spatiotemporal clustering analysis, and repeat the verification process of step 4-4 until the accuracy requirements are met. Steps 4-5: Based on the digital twin model built in step 2, visualize the identified sensitive areas and generate a structured risk report; based on the newly generated scenario data in step 3, regularly update the distribution and risk characteristics of sensitive areas.

7. The method for dynamic identification and optimization of sensitive areas of low-altitude flight path flow field based on ensemble simulation according to claim 1, characterized in that, Step 5 includes the following steps: Step 5-1: Integrate the coordinates, risk level and dynamic update results of the high conflict area output in step 4, and at the same time obtain the emergency scenario parameters generated in step 3 to construct a multi-objective route optimization model that includes navigation safety, flight efficiency and environmental adaptability. Step 5-2: Based on the optimization model, a multi-objective evolutionary algorithm is used to perform population iterative calculations to generate a Pareto optimal solution set, and the risk index of sensitive areas is transformed into a constraint condition to ensure that the generated route prioritizes avoiding high-risk areas; Step 5-3: Adjust the objective function weights in real time and rerun the optimization algorithm based on changes in parameters of the sudden scenario. Monitor the airspace status in real time through the digital twin model and trigger route replanning when the risk threshold is detected to be exceeded. Step 5-4: Based on the pre-computed Pareto optimal solution set, the optimal route suitable for the current scenario is selected through a multi-attribute decision algorithm, and the generated route is imported into the digital twin environment to verify its safety and efficiency.

8. The method for dynamic identification and optimization of sensitive areas of low-altitude flight path flow field based on ensemble simulation according to claim 1, characterized in that, Step 6 includes the following steps: Step 6-1: Based on the digital twin model built in Step 2, acquire historical flight data and integrate it with the route planning scheme and high-risk area coordinates output in Step 4; Step 6-2: Reconstruct the historical airspace status in the digital twin model, overlay the route planning scheme and sensitive area markers, and build a virtual-real integrated test environment; Step 6-3: Using the spatiotemporal synchronization function of the digital twin model, replay historical flight events and inject route planning schemes to simulate their execution effect in the original scenario; Step 6-4: Combine the extreme weather and equipment failure scenarios generated in Step 3 to conduct extended tests, compare and analyze the simulation playback results with actual historical operation data, and verify the feasibility of optimizing the route strategy.

9. The method for dynamic identification and optimization of sensitive areas of low-altitude flight path flow field based on ensemble simulation according to claim 1, characterized in that, Step 7 includes the following steps: Step 7-1: Based on the verification feedback results, automatically execute the closed-loop optimization process: If the verification finds that the accuracy of identifying high-risk areas is lower than the preset threshold, the system automatically lowers the risk threshold of sensitive areas, expands the coverage of high-conflict areas, and re-executes the cluster analysis in step 4; If the conflict rate of the optimized route exceeds the standard in extreme scenarios, the system dynamically adjusts the weight coefficients of the multi-objective evolutionary algorithm. Step 7-2: Update the parameters of the scenarios that failed during verification to the cellular automata-fluid coupling model in Step 3; Step 7-3: Automatically trigger the verification replay of step 6 after each optimization.

10. A dynamic identification and optimization system for sensitive areas of low-altitude flight path flow field based on ensemble simulation, characterized in that, The system uses the method described in any one of claims 1-9, including a multi-source data acquisition and preprocessing module, a digital twin modeling flow field simulation module, a multi-scenario generation module for ensemble simulation, a sensitive area identification and risk quantification module, a route optimization and dynamic adjustment module, a digital twin verification simulation playback module, and a dynamic feedback optimization module; The multi-source data acquisition and preprocessing module is based on the fusion of BeiDou grid segmentation and multi-source heterogeneous data to construct a spatiotemporally consistent low-altitude flow field data base. The digital twin modeling flow field simulation module is used to combine the high-quality data from the multi-source data acquisition and preprocessing module with three-dimensional holographic interaction technology to construct a reproducible low-altitude airspace digital twin environment. The multi-scenario generation module of the collection simulation is based on the digital twin platform of the digital twin modeling flow field simulation module and the cellular automata-fluid coupling model to generate diverse flow field evolution scenarios covering extreme weather and equipment failures; The sensitive area identification risk quantification module extracts high conflict probability areas and quantifies their risk characteristics by combining the set simulation results of the set simulation multi-scenario generation module with the spatiotemporal clustering algorithm. The route optimization dynamic adjustment module generates a route planning scheme that meets safety constraints based on the risk quantification index of the sensitive area identification and risk quantification module and a multi-objective evolutionary algorithm, and dynamically adapts to parameter changes in sudden scenarios. The digital twin verification simulation playback module, through the digital twin platform of the digital twin modeling flow field simulation module and the optimized scheme of the sensitive area identification risk quantification module, verifies the feasibility of the route strategy by playing back historical data and conducting virtual-real fusion tests. The dynamic feedback optimization module optimizes based on the verification feedback from the digital twin verification simulation playback module.

Images