Digital twin deduction and dynamic capacity assessment method for low-altitude airspace operation scenario

Abstract

This invention relates to the field of low-altitude airspace management and digital twin simulation technology, and provides a method for digital twin simulation and dynamic capacity assessment of low-altitude airspace operation scenarios. This method constructs a digital twin model of the target low-altitude airspace and injects virtual aircraft agents with differentiated performance parameters into the twin environment in batches. It employs the Monte Carlo method for parallel simulation across multiple scenarios, automatically identifying capacity boundary inflection points by progressively increasing the number of virtual aircraft and monitoring changes in system conflict probability, thereby achieving a quantitative assessment of the ultimate capacity of low-altitude airspace. This method integrates the precise mapping capabilities of digital twins, the autonomous interaction characteristics of multiple agents, and the probabilistic simulation advantages of Monte Carlo simulation, providing a more scientific decision-making basis for low-altitude airspace planning, traffic management, and safety control.

Description

Technical Field

[0001] This invention relates to the field of low-altitude airspace management and digital twin simulation technology, and in particular to a method for digital twin simulation and dynamic capacity assessment of low-altitude airspace operation scenarios. Background Technology

[0002] With the rapid development of the low-altitude economy, the number of new low-altitude aircraft, such as drones and electric vertical take-off and landing (eVTOL) aircraft, has increased dramatically, significantly increasing the operational density and complexity of low-altitude airspace. Accurately assessing the carrying capacity of low-altitude airspace (the number of aircraft that the airspace can simultaneously accommodate while ensuring safety) has become a core foundation for airspace planning, traffic flow management, and safety control. The results of capacity assessments directly affect the rational allocation of airspace resources, flight plan approval decisions, and traffic control strategies in emergency situations.

[0003] Currently, low-altitude airspace capacity assessment methods are mainly based on analytical methods of mathematical modeling. By establishing route network models and capacity calculation mathematical models, theoretical capacity and operational capacity are solved. This method has the following shortcomings: (1) Capacity assessment is mostly static or deterministic deduction, focusing on verifying whether the system is in conflict under a given flight plan, rather than actively exploring the system's limit carrying capacity boundary, making it difficult to answer the core question of "how many aircraft can the airspace accommodate at most under the current environmental conditions"; (2) Assessment models are mostly based on simplified route networks or preset flight rules, failing to fully consider the high flexibility of low-altitude airspace aircraft tracks; (3) Meteorological factors and infrastructure status are usually simplified to static inputs or uniform external conditions, failing to fully reflect the actual impact of their spatial distribution unevenness and dynamic change characteristics on capacity; (4) Capacity assessment results are mostly single deterministic values, lacking quantitative characterization of the uncertainty of input data, making it difficult to judge the credibility of the assessment results.

[0004] Some solutions have introduced assessment methods based on digital twins and simulations, or capacity analysis methods based on spatiotemporal grids. However, most existing solutions rely on historical data and real-time conditions to predict the future state of factors such as aircraft flight trajectories, weather conditions, and signal interference, and then conduct low-altitude airspace capacity assessments. Due to the limited number of predicted scenarios, it is difficult to cover sufficient airspace conditions, resulting in insufficient accuracy in predicting airspace conflicts and thus limiting the reliability of the assessment results. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a digital twin simulation and dynamic capacity assessment method for low-altitude airspace operation scenarios to solve the problem of insufficient accuracy of capacity assessment in low-altitude airspace operation scenarios in the prior art.

[0006] A first aspect of this invention provides a method for digital twin simulation and dynamic capacity assessment of low-altitude airspace operation scenarios, comprising:

[0007] Construct a digital twin model of the target low-altitude airspace; the target low-altitude airspace digital twin model includes multiple airspace grid agents, multiple meteorological grid agents, multiple infrastructure grid agents and initial availability labels. Each airspace grid agent includes static attribute data of the airspace grid, each meteorological grid agent includes meteorological status data of the meteorological grid, and each infrastructure grid agent includes infrastructure service status data of the infrastructure grid.

[0008] Based on the current state of the target low-altitude airspace digital twin model under the current environmental conditions, and using the current environmental conditions as the extrapolation boundary, the virtual aircraft load level is progressively increased, and the system executes independently at each load level. The Monte Carlo stochastic simulation is repeated until the simulation termination condition is met, resulting in the load level-conflict probability response curve for low-altitude airspace. Under each virtual aircraft load level, the performance parameters of the virtual aircraft are generated based on differential sampling from a preset probability distribution model. The conflict probability under each load level is determined by... The conflict indicator variables obtained from the second simulation are determined by probability statistics and convergence judgment. It is a positive integer. greater than 0 and less than or equal to positive integers, To determine the number of aircraft load levels;

[0009] Multiscale curvature analysis is used to identify capacity boundary inflection points on the load level-conflict probability response curve. The validity of the capacity boundary inflection points is determined by statistical hypothesis testing, and the limit capacity assessment value at significant inflection points or the conservative limit capacity at no significant inflection points is obtained.

[0010] Calculate the confidence index of the limit capacity assessment value at the significant inflection point, and determine the significant inflection point and its confidence index as the capacity assessment result, or determine the conservative limit capacity as the capacity assessment result.

[0011] A second aspect of the present invention provides a digital twin simulation and dynamic capacity assessment device for low-altitude airspace operation scenarios, comprising:

[0012] The building module is configured to build a digital twin model of the target low-altitude airspace. The target low-altitude airspace digital twin model includes multiple airspace grid agents, multiple meteorological grid agents, multiple infrastructure grid agents, and initial availability labels. Each airspace grid agent includes static attribute data, each meteorological grid agent includes meteorological status data of the meteorological grid, and each infrastructure grid agent includes infrastructure service status data of the infrastructure grid.

[0013] The simulation module is configured to use the current environmental conditions as the basis for the state of the target low-altitude airspace digital twin model, with the current environmental conditions as the simulation boundary, progressively increasing the virtual aircraft load level, and executing independently at each load level. The Monte Carlo stochastic simulation is repeated until the simulation termination condition is met, resulting in the load level-conflict probability response curve for low-altitude airspace. Under each virtual aircraft load level, the performance parameters of the virtual aircraft are generated based on differential sampling from a preset probability distribution model. The conflict probability under each load level is determined by... The conflict indicator variables obtained from the second simulation are determined by probability statistics and convergence judgment. It is a positive integer. greater than 0 and less than or equal to positive integers, To determine the number of aircraft load levels;

[0014] The statistics module is configured to use multi-scale curvature analysis to identify capacity boundary inflection points on the load level-conflict probability response curve, and to determine the validity of capacity boundary inflection points through statistical hypothesis testing, thereby obtaining the limit capacity assessment value when there is a significant inflection point or the conservative limit capacity when there is no significant inflection point.

[0015] The evaluation module is configured to calculate the confidence index of the limit capacity assessment value at the significant inflection point, determine the significant inflection point and its confidence index as the capacity assessment result, or determine the conservative limit capacity as the capacity assessment result.

[0016] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows:

[0017] This invention constructs a digital twin model of the target low-altitude airspace and injects virtual aircraft agents with differentiated performance parameters into the twin environment in batches. Using the Monte Carlo method, it performs parallel simulations across multiple scenarios, automatically identifying capacity boundary inflection points by progressively increasing the number of virtual aircraft and monitoring changes in system conflict probability. This enables a quantitative assessment of the ultimate capacity of low-altitude airspace. This method integrates the precise mapping capabilities of digital twins, the autonomous interaction characteristics of multiple agents, and the probabilistic inference advantages of Monte Carlo simulation, providing a more scientific basis for decision-making in low-altitude airspace planning, traffic management, and safety control. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating a digital twin simulation and dynamic capacity assessment method for low-altitude airspace operation scenarios provided in an embodiment of the present invention.

[0020] Figure 2 This is a schematic diagram of the structure of a digital twin simulation and dynamic capacity assessment device for low-altitude airspace operation scenarios provided in an embodiment of the present invention.

[0021] Figure 3 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0022] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0023] The following will describe in detail, with reference to the accompanying drawings, a method and apparatus for digital twin simulation and dynamic capacity assessment of a low-altitude airspace operation scenario according to an embodiment of the present invention.

[0024] As mentioned above, existing analytical methods based on mathematical modeling for low-altitude airspace capacity assessment have many shortcomings. Meanwhile, some assessment schemes based on digital twins and simulations construct airspace digital twin models to simulate aircraft operational status and perform capacity calculations. Other capacity analysis schemes based on spatiotemporal grids generate capacity grid maps by discretizing the airspace and combine them with demand grid maps for conflict resolution. However, most of these schemes rely on historical data and real-time conditions for low-altitude airspace capacity assessment, resulting in insufficient accuracy in predicting airspace conflicts and limiting the reliability of the assessment results.

[0025] In view of this, embodiments of the present invention provide a digital twin simulation and dynamic capacity assessment method for low-altitude airspace operation scenarios. By constructing a digital twin model of the target low-altitude airspace, virtual aircraft agents are injected in batches into the twin environment. The Monte Carlo method is used to perform parallel simulations of multiple scenarios. By progressively increasing the number of virtual aircraft and monitoring changes in the system's conflict probability, the capacity boundary inflection point is automatically identified, thereby achieving a quantitative assessment of the ultimate capacity of low-altitude airspace under given environmental conditions. This provides a scientific basis for low-altitude airspace planning, traffic management, and safety control.

[0026] Figure 1 This is a flowchart illustrating a digital twin simulation and dynamic capacity assessment method for low-altitude airspace operation scenarios provided in an embodiment of the present invention. Figure 1 As shown, the method includes the following steps:

[0027] In step S101, a digital twin model of the target low-altitude airspace is constructed.

[0028] The target low-altitude airspace digital twin model includes multiple airspace grid agents, multiple meteorological grid agents, multiple infrastructure grid agents, and initial availability labels. Each airspace grid agent includes static attribute data of the airspace grid, each meteorological grid agent includes meteorological status data of the meteorological grid, and each infrastructure grid agent includes infrastructure service status data of the infrastructure grid.

[0029] In step S102, based on the state of the target low-altitude airspace digital twin model under the current environmental conditions, and using the current environmental conditions as the deduction boundary, the virtual aircraft load level is progressively increased, and execution is carried out independently at each load level. The Monte Carlo stochastic simulation is repeated until the simulation termination condition is met, resulting in the load level-conflict probability response curve for low-altitude airspace.

[0030] Among them, the performance parameters of the virtual aircraft are generated based on differential sampling of a preset probability distribution model under each virtual aircraft load level; the collision probability under each load level is obtained by... The conflict indicator variables obtained from the second simulation are determined by probability statistics and convergence judgment. It is a positive integer. greater than 0 and less than or equal to positive integers, The number of load levels for the determined aircraft.

[0031] In step S103, the multi-scale curvature analysis method is used to identify the capacity boundary inflection point on the load level-conflict probability response curve, and the validity of the capacity boundary inflection point is determined by statistical hypothesis testing to obtain the limit capacity assessment value when there is a significant inflection point or the conservative limit capacity when there is no significant inflection point.

[0032] In step S104, the confidence index of the limit capacity assessment value at the significant inflection point is calculated, and the significant inflection point and its confidence index are determined as the capacity assessment result, or the conservative limit capacity is determined as the capacity assessment result.

[0033] In some embodiments of the present invention, the method may be executed by a server or by a terminal device with certain processing capabilities.

[0034] In some embodiments of the present invention, a target low-altitude airspace digital twin model can be constructed first. The target low-altitude airspace digital twin model includes multiple airspace grid agents, multiple meteorological grid agents, multiple infrastructure grid agents, and initial availability labels. Each airspace grid agent includes static attribute data of the airspace grid, each meteorological grid agent includes meteorological status data of the meteorological grid, and each infrastructure grid agent includes infrastructure service status data of the infrastructure grid.

[0035] In some embodiments of the present invention, based on the state of the target low-altitude airspace digital twin model under the current environmental conditions, and using the current environmental conditions as the deduction boundary, the virtual aircraft load level is progressively increased, and execution is carried out independently at each load level. The Monte Carlo stochastic simulation is repeated until the simulation termination condition is met, resulting in the load level-conflict probability response curve for low-altitude airspace.

[0036] Among them, the performance parameters of the virtual aircraft are generated based on differential sampling of a preset probability distribution model under each virtual aircraft load level; the collision probability under each load level is obtained by... The conflict indicator variables obtained from the second simulation are determined by probability statistics and convergence judgment.

[0037] In some embodiments of the present invention, a multi-scale curvature analysis method is used to identify capacity boundary inflection points on the load level-conflict probability response curve, and the validity of the capacity boundary inflection points is determined by statistical hypothesis testing to obtain the conservative limit capacity when there is a significant inflection point or no significant inflection point.

[0038] If a significant inflection point can be obtained, the confidence index of the limit capacity assessment value at the significant inflection point can be further calculated, and the significant inflection point and its confidence index can be determined as the capacity assessment result. If a significant inflection point cannot be obtained, the conservative limit capacity can be determined as the capacity assessment result.

[0039] According to the technical solution provided in this embodiment of the invention, a digital twin model of the target low-altitude airspace is constructed, and virtual aircraft agents with differentiated performance parameters are injected in batches into the twin environment. The Monte Carlo method is used for parallel simulation across multiple scenarios. By progressively increasing the number of virtual aircraft and monitoring changes in system conflict probability, the capacity boundary inflection point is automatically identified, achieving a quantitative assessment of the ultimate capacity of low-altitude airspace. This method integrates the precise mapping capability of digital twins, the autonomous interaction characteristics of multiple agents, and the probabilistic inference advantages of Monte Carlo simulation, providing a more scientific decision-making basis for low-altitude airspace planning, traffic management, and safety control.

[0040] In some embodiments of the present invention, constructing a digital twin model of the target low-altitude airspace may include:

[0041] A discretized three-dimensional airspace grid is constructed for the target low-altitude airspace. An airspace grid agent is configured for each airspace grid, and static attribute data is assigned to each airspace grid agent. The static attribute data includes the spatial location, geometric dimensions, physical attributes, and regulatory attributes of the airspace grid.

[0042] Construct a meteorological grid intelligent agent, use the meteorological grid intelligent agent to obtain meteorological monitoring data of the target low-altitude airspace, and map the meteorological monitoring data to each airspace grid intelligent agent through spatial interpolation, and assign meteorological status data to each airspace grid intelligent agent;

[0043] Construct an infrastructure grid intelligent agent, use the infrastructure grid intelligent agent to obtain infrastructure status data of the target low-altitude airspace, map the infrastructure status data values ​​to each airspace grid intelligent agent based on the spatial coverage area of ​​each infrastructure, and assign infrastructure service status data to each airspace grid intelligent agent.

[0044] The static attribute data, meteorological status data, and infrastructure service status data of the target airspace grid agent are integrated to obtain the status data of the target airspace grid agent. The initial availability label of the target airspace grid agent is determined based on the status data of the target airspace grid agent using a preset airspace availability judgment rule.

[0045] Consistency verification is performed on the state data and initial availability labels of the target airspace grid agents. The state data and initial availability labels that fail the verification are updated. When the verification passes, the set of airspace grid agents composed of the state data and initial availability labels of each airspace grid agent is determined as the target low-altitude airspace digital twin model.

[0046] The target spatial grid agent can be any spatial grid agent.

[0047] In some implementations, mapping meteorological monitoring data to various spatial grid agents via spatial interpolation may include:

[0048] Constructing a set of meteorological elements based on meteorological monitoring data The meteorological element set includes... Meteorological elements, It is a positive integer;

[0049] Determine the first A set of neighboring meteorological data points for a spatial grid agent Determining the first by spatial interpolation Meteorological status data of a spatial grid agent for: ;in, Indicates the first A collection of meteorological elements from a set of meteorological data points. , Indicates the first The first of the meteorological data points Meteorological elements; The interpolation weight coefficients satisfy the following conditions: ; It is a positive integer greater than or equal to 1 and less than or equal to the total number of smart agents in the spatial grid. greater than or equal to 1 and less than or equal to 1 Positive integers.

[0050] Mapping infrastructure status data values ​​to spatial grid agents based on the spatial coverage area of ​​each infrastructure can include:

[0051] Obtain the service scope of the target infrastructure; the target infrastructure can be any type of infrastructure.

[0052] Determine the capacity degradation factor of the target infrastructure based on current environmental conditions;

[0053] The product of the service range and the capacity degradation coefficient is determined as the spatial coverage area of ​​the target infrastructure;

[0054] In response to determining that the grid center point of the target spatial grid agent falls within the spatial coverage area of ​​the target infrastructure, the state data values ​​of the target infrastructure are mapped to the target spatial grid agent.

[0055] In other words, when constructing a digital twin model of the target low-altitude airspace, multi-source environmental data of the target low-altitude airspace can be obtained, including static airspace data, real-time meteorological monitoring data, and infrastructure status data; using an agent-based multi-domain data fusion method, a digital twin model of the target low-altitude airspace can be constructed as the initial state for extreme capacity simulation.

[0056] When acquiring static airspace data, the target low-altitude airspace can be discretized into a three-dimensional airspace grid according to a preset spatial resolution (such as meter-level accuracy). Let the range of the target airspace be... (Length × Width × Height), the size of each grid is... The total number of spatial grids is Each three-dimensional spatial grid is defined as a spatial grid intelligent agent. Each spatial grid agent has a unique identifier. and spatial coordinates (Represented by the grid center point), and carries the following static attribute data:

[0057] Spatial location and geometric dimensions of the grid: coordinate range, center point coordinates, length, width and height dimensions;

[0058] The physical properties of the grid include: terrain type (buildings, water bodies, green spaces, roads, etc.), altitude, and whether it is a permanent no-fly zone or a temporary restricted area.

[0059] The regulatory attributes of the grid include: airspace classification (controlled airspace, monitored airspace, reporting airspace, etc.), altitude layer attributes, and speed limits.

[0060] The above static attribute data comes from urban geographic information systems, low-altitude airspace delineation specifications, and airspace control rules.

[0061] When acquiring real-time meteorological monitoring data, meteorological grid agents can be used to obtain meteorological monitoring data for the target low-altitude airspace, including but not limited to meteorological elements such as wind speed, wind direction, temperature, air pressure, humidity, visibility, and precipitation intensity. Meteorological data sources include weather radar, ground weather stations, radiosondes, and numerical weather prediction models. The acquired meteorological monitoring data can also be spatially interpolated and mapped to each airspace grid agent, assigning meteorological state data to each airspace grid agent. This meteorological state data can serve as a description of the current meteorological conditions within the airspace grid agent.

[0062] When acquiring infrastructure status data, infrastructure grid agents can be used to obtain infrastructure status data for the target low-altitude airspace, including the coverage and signal strength of communication base stations, the service range and signal availability of navigation beacons, the detection range and accuracy of surveillance radars, and the equipment status of low-altitude flight service stations. Communication base stations include 5G base stations and Long Term Evolution (LTE) base stations; navigation beacons include VHF omnidirectional radio ranges (VORs) and distance measuring equipment (DMEs); and surveillance radars include primary radars and secondary radars.

[0063] When mapping infrastructure status data values ​​to a spatial grid agent, it can be assumed that there are shared features within the low-altitude airspace. Class of infrastructure, the first Common infrastructure If there are 100 facility entities, then the total number of infrastructure entities is 100. . , and All are positive integers. greater than or equal to 1 and less than or equal to 1 Positive integers.

[0064] For the The first in the class of infrastructure A facility entity, whose service range is defined as centered on the facility location and extending over a nominal coverage distance. spatial region with radius The service capacity of infrastructure may be degraded due to environmental factors (such as building obstruction and weather conditions), defined as the actual coverage distance. ,in Capability degradation factor ( This indicates that the facility is operating normally. This indicates a downgrade in ability. (This indicates that the facility is completely inoperable). greater than or equal to 1 and less than or equal to 1 Positive integers.

[0065] For any spatial grid agent Its infrastructure service status is determined by the grid center point. Whether it falls within the actual service scope of each infrastructure entity To determine. If Then determine the first The first in the class of infrastructure Each facility entity is capable of providing services to the grid; otherwise, it is determined that no service can be provided. The service availability (available / unavailable) and service quality (such as signal strength, coverage redundancy, etc.) of various infrastructure components for the grid are assigned to the spatial grid agent. The dynamic attribute layer.

[0066] Furthermore, the static attribute data, meteorological status data, and infrastructure service status data of each airspace grid agent can be uniformly integrated to generate a complete status description for each airspace grid agent. For the first... A spatial grid intelligent agent Its complete state is defined as: ;in This is static attribute data. This is meteorological status data, including values ​​of various meteorological elements and data timestamps. This provides infrastructure service status data, including service availability and service quality parameters for various types of infrastructure.

[0067] In some embodiments of the present invention, it can be based on each spatial grid agent. According to the preset airspace availability determination rules (such as unavailability if wind speed exceeds a threshold, unavailability if infrastructure services are missing, or unavailability if located in a no-fly zone), an initial availability label is generated for each airspace grid. The state descriptions and availability labels of all airspace grid agents are organized according to their spatial location to form the initial state of the target low-altitude airspace digital twin model, which serves as the environmental boundary condition for subsequent limit capacity simulation.

[0068] In some embodiments of the present invention, consistency verification includes at least spatial consistency verification, logical consistency verification, and data integrity verification.

[0069] Spatial consistency verification includes checking for unreasonable abrupt changes in meteorological and infrastructure service status between adjacent spatial grids (such as wind speed differences between adjacent grids exceeding a preset threshold). If such a change is found, re-interpolation or manual review is triggered. Logical consistency verification includes checking for logical contradictions between the availability label of each grid and its basic data (such as wind speeds below a threshold being marked as unavailable). If such contradictions are found, abnormal grids are recorded and corrections are triggered. Data integrity verification includes checking whether each grid fully contains all the status fields required for the simulation, and marking and completing any missing fields.

[0070] After the consistency verification is passed, the initial state of the complete target low-altitude airspace digital twin model can be stored in the database for subsequent simulation and retrieval.

[0071] In some embodiments of the present invention, the virtual aircraft is represented by a virtual aircraft intelligent agent; the first 'Attributes of a virtual aircraft intelligent agent' for: ;in, It is a unique identifier; It is a set of flight performance parameters, including at least cruise speed, maximum rate of climb, maximum yaw angle, and maximum flight altitude; For communication capability parameters, representing the first... The degree of dependence of a virtual aircraft on infrastructure services; A set of meteorological tolerance thresholds, including at least the maximum tolerable wind speed and the minimum visibility requirement; attributes , and The corresponding performance parameters of the virtual aircraft;

[0072] Under the various virtual aircraft load levels, the cruise speed is generated based on differential sampling of the normal distribution model; the attributes in the meteorological tolerance threshold set are generated based on differential sampling of the uniform distribution or Weibull distribution model.

[0073] Execute independently at each load level Sub-Monte Carlo stochastic simulations may include:

[0074] by Randomly generated for the current load level Each virtual aircraft agent instance is analyzed, and its attributes are generated independently through sampling. A virtual aircraft intelligent agent; the current load level is any load level before the termination condition is met;

[0075] Will A virtual aircraft agent is injected into the target's low-altitude airspace digital twin model, causing it to be distributed at random initial positions in the airspace;

[0076] The Monte Carlo method is used for A virtual aircraft agent generates multiple flight mission samples;

[0077] In each Monte Carlo stochastic simulation, flight mission parameters are randomly generated for each virtual aircraft agent; the flight mission parameters include at least the takeoff point coordinates and destination coordinates, the planned route and the mission execution time window.

[0078] The flight mission samples are executed based on the flight mission parameters, and the conflict indicator variables for each simulation are recorded. The conflict indicator variable takes a value of 1 when at least one unsafe event occurs in the simulation, and takes a value of 0 when no unsafe event occurs in the simulation. Unsafe events include at least conflict events, environmental mismatch events, and communication loss events.

[0079] To achieve the current load level After the Monte Carlo stochastic simulation, the estimated conflict probability at this load level is calculated; the estimated conflict probability is... The proportion of simulations in which unsafe events occurred out of the total number of simulations;

[0080] The convergence of the inference results at this level is determined when the standard deviation of the determined conflict probability estimate is less than the preset convergence threshold.

[0081] In response to the standard deviation of the determined conflict probability estimate being greater than or equal to a preset convergence threshold, the number of simulations is increased. Re-execute independently at this load level The Monte Carlo stochastic simulation is repeated until the results of this level of simulation converge.

[0082] In response to the convergence of the current level's simulation results, the current level's load level is saved. Number of simulations The collision probability estimate and convergence standard deviation at convergence; the convergence standard deviation is the standard deviation of the collision probability estimate at convergence.

[0083] In some implementations, progressively increasing the virtual aircraft load level may include:

[0084] If the estimated conflict probability at the time of convergence of the current level is less than a preset safety threshold, the preset step size is used. Increase the load level to obtain the next load level. ;

[0085] The following load levels To obtain the new current load level, perform the Monte Carlo stochastic simulation described above. The estimated probability of conflict when the result converges;

[0086] If the estimated conflict probability at convergence under the new current load level is less than a preset safety threshold, then the iteration will proceed with a preset step size. Increase the load level, take the increased load level as the new current load level, perform the Monte Carlo stochastic inference described above, and obtain the estimated value of the conflict probability when the result converges under the new current load level, until the estimated value of the conflict probability when the result converges under the new current load level is greater than or equal to the preset safety threshold, and determine that the inference termination condition is met.

[0087] The load level-conflict probability response curve is a curve formed by the estimated conflict probability obtained from each load level and the corresponding load level.

[0088] In other words, based on the initial state of the constructed target low-altitude airspace digital twin model, and using the current environmental conditions as the extrapolation boundary, the load level of the virtual aircraft can be progressively increased and Monte Carlo stochastic extrapolation can be performed at each load level to obtain the load level-conflict probability response curve, providing a data foundation for the subsequent identification of capacity boundary inflection points.

[0089] A generic attribute template for the virtual aircraft agent can be defined, and the first... The attributes of a virtual aircraft agent are represented as follows: In this system, the performance parameters corresponding to each attribute of each virtual aircraft agent are not uniformly assigned values, but are generated through differentiated sampling based on a preset probability distribution model. For example, cruising speed can follow a normal distribution, weather tolerance threshold can follow a uniform distribution or a Weibull distribution, and so on. By sampling differentiated parameters, the virtual aircraft group covers the performance distribution range of various real aircraft in low-altitude airspace, avoiding deviations from reality in capacity assessment results due to homogeneous assumptions.

[0090] You can set the initial load level. (i.e., the number of virtual aircraft) is randomly generated in the digital twin model of the target low-altitude airspace. Each virtual aircraft agent instance is independently sampled according to the parameterized generation method defined in the above steps to obtain its performance parameter combination, thereby determining the value of each attribute. The generated virtual aircraft swarm can be injected into the target low-altitude airspace digital twin model, distributing it at random initial positions in the airspace.

[0091] Let the current simulated load level be... The injected set of virtual aircraft agents is then... .

[0092] For a given load level A large number of flight mission samples can be generated using the Monte Carlo method. Specifically, the number of independent repeated simulations is set to... (For example or In each independent simulation, for Each virtual aircraft agent in the system randomly generates flight mission parameters, including:

[0093] Takeoff point coordinates and destination coordinates—Based on the actual takeoff and landing field distribution and historical flight demand data in the low-altitude airspace, a spatial distribution probability density function for takeoff and landing points and a spatial distribution probability density function for destinations are constructed. Non-uniform random sampling is performed according to the above probability density functions to make the sampling distribution approximate the takeoff and destination point distribution characteristics of flight missions in actual operation scenarios. When historical data is lacking, uniform random sampling within the airspace is adopted as the default strategy.

[0094] Route planning: Taking into account terrain obstacles and no-fly zone constraints, a three-dimensional trajectory is generated according to a preset route planning algorithm.

[0095] Mission execution time window: determines the departure time of each aircraft, reflecting the time dynamics of the airspace.

[0096] A group of virtual aircraft with random mission parameters is placed in a digital twin model of the target low-altitude airspace and run synchronously, performing flight mission simulations according to a set simulation time length (e.g., 1 hour). During the simulation, the spatial interval between each virtual aircraft is monitored in real time. When the spatial distance between any two aircraft is less than the preset safety interval, a conflict event is recorded; when any aircraft enters an area where the weather conditions exceed its tolerance threshold, an environmental mismatch event is recorded; when the infrastructure service status of any aircraft's grid is "unavailable," a communication loss event is recorded.

[0097] These three types of events are collectively referred to as unsafe events. For the first... Sub-independent simulation, defining conflict indicator variables for: .

[0098] Complete the current load level All of the following After each independent simulation, the conflict probability estimate is calculated. : This refers to the proportion of simulations in which unsafe events occur out of the total number of simulations.

[0099] To ensure the reliability of the statistical results, it is necessary to... The convergence of the parameters is determined. The standard deviation of the conflict probability estimate is calculated. .

[0100] like Less than the preset convergence threshold (like If the calculated load level has converged, then the result is considered to have converged and accepted. This is used as the conflict probability statistic for this load level, and this load level is... Conflict probability statistics Number of simulations convergence standard deviation All metadata is stored in the simulation results database; otherwise, the number of simulations is increased. Continue performing the Monte Carlo sampling and extrapolation process described above until the convergence condition is met. The structured extrapolation results can be reused for sensitivity analysis in subsequent steps—when the combination of input parameters changes, only the affected load levels need to be recalculated using Monte Carlo methods, while the extrapolation results for unaffected load levels can be directly retrieved from the database without requiring a complete recalculation.

[0101] Complete the current load level After the deduction, the following method can be used to determine whether the termination condition is met:

[0102] If the current conflict probability Less than the preset safety threshold (like If the preset step size is used, then... Increase the load level, i.e. Re-execute the initial load level setting and virtual aircraft swarm injection steps, and repeat the virtual aircraft swarm injection and Monte Carlo simulation under the new load level;

[0103] If the current conflict probability The preset safety threshold has been reached or exceeded. If the simulation fails, stop the simulation and record the current load level and the corresponding conflict probability statistics for each load level.

[0104] Through the above step-by-step incremental process, a set of data points is obtained. ,in ,and , , The values ​​are positive integers. This set of data points forms the load level-conflict probability response curve, which serves as input data for subsequent capacity boundary inflection point identification.

[0105] In some embodiments of the present invention, using a multi-scale curvature analysis method to identify capacity boundary inflection points on the load level-conflict probability response curve may include:

[0106] Smoothing filters are applied to each conflict probability estimate in the load level-conflict probability response curve to eliminate statistical fluctuations introduced by Inmont Carlo random sampling;

[0107] The filtered load level-conflict probability response curve is smoothed using a locally weighted regression method; whereby the smoothed value of each data point is determined by its neighboring values. The weighted average of the data points is obtained; To preset a smooth window, It is a positive integer;

[0108] By load level With the independent variable and the conflict probability estimate as the dependent variable, the smoothed load level-conflict probability response curve is fitted with a continuous function using the cubic spline interpolation method to obtain a continuously differentiable response curve.

[0109] set up Different smoothing scales ,in , It is a positive integer greater than 2;

[0110] At each smoothing scale Below, a Gaussian kernel is used. The continuous differentiable response curve is smoothed by convolution to obtain the smoothed continuous differentiable response curve; for to Any smoothing scale in;

[0111] Calculate the curvature of the smoothed, continuously differentiable response curve obtained at each smoothing scale, and obtain... Local maxima of curvature;

[0112] right The capacity boundary inflection point is obtained by weighting the local maxima of curvature; where the smoothing scale is inversely proportional to the weight.

[0113] In other words, the data points in the load level-conflict probability response curve can first be preprocessed and fitted with a function. Preprocessing includes: processing the conflict probability estimates... Smoothing filtering is performed to eliminate statistical fluctuations introduced by Monte Carlo random sampling; a local weighted regression method is used to smooth the data points, with the smoothing window set to [value missing]. (like That is, the smoothed value of each data point is determined by its neighboring values. The weighted average of the data points is obtained.

[0114] After smoothing, based on load level Independent variable, conflict probability As the dependent variable, cubic spline interpolation is used to fit a continuous function to the data points, resulting in a continuously differentiable response curve. ,in Cubic spline interpolation guarantees It has second-order continuity, that is and and It is continuous throughout the entire domain, ensuring the validity of subsequent curvature calculations.

[0115] The continuous response curve obtained by fitting Calculate its curvature Curvature is defined as: Curvature This reflects the curvature of the curve at a certain point. When the curve transitions from a flat section to a steep rise, the curvature near that point reaches a local maximum, and this local maximum point is a candidate location for the capacity boundary inflection point.

[0116] To achieve robust identification of inflection points, this invention employs a multi-scale curvature analysis method. (Setting...) Different smoothing scales ( ), at each scale Below, a Gaussian kernel is used. For response curve Perform convolution smoothing to obtain the smoothed curve. Calculate its curvature It also identifies local maxima of curvature at various scales.

[0117] The local maxima of curvature identified at each scale are weighted and combined, with the weights inversely proportional to the scale (smaller scales are given higher weights to preserve fine features; larger scales are given lower weights to avoid over-smoothing and inflection point shift), to determine the candidate positions of the final inflection points on the response curve. .

[0118] In some embodiments of the present invention, determining the validity of the capacity boundary inflection point through statistical hypothesis testing may include:

[0119] The conflict probability estimates derived from the load levels are divided into a first group of data and a second group of data. The first group of data includes conflict probability estimates that are less than or equal to the capacity boundary inflection point, and the second group of data includes conflict probability estimates that are greater than the capacity boundary inflection point.

[0120] Calculate the U-statistic for the first and second sets of data: ;in, , , Let the sum of the rank of the first set of data be . Let the sum of the ranks of the second set of data be . This is the sample size for the first set of data. This is the sample size for the second set of data;

[0121] Determine the p-values ​​for the first and second sets of data based on the U statistic:

[0122] In response to the determination that the p-value is less than the preset significance level, the capacity boundary inflection point is determined to be valid and significant.

[0123] In response to determining that the p-value is greater than or equal to the preset significance level, reset Different smoothing scales , use as Figure 1 The method provided in the illustrated embodiment redetermines the capacity boundary inflection point;

[0124] Based on the redefined capacity boundary inflection point, the estimated conflict probability values ​​derived under each load level are reclassified into a first group of data and a second group of data, and the p-values ​​of the first group of data and the second group of data after reclassification are calculated.

[0125] In response to the determination that the p-value is less than the preset significance level, the redefined capacity boundary inflection point is determined to be valid and the redefined capacity boundary inflection point is a significant inflection point. The capacity value corresponding to the significant inflection point is determined to be the limit capacity assessment value when the significant inflection point is reached.

[0126] In response to determining that the p-value is greater than or equal to the preset significance level, it is determined that there is no significant inflection point at present, and the load level corresponding to the preset safety threshold on the continuously differentiable response curve is determined to be the conservative limit capacity.

[0127] Furthermore, the confidence index for calculating the limit capacity assessment value at the significant inflection point can include:

[0128] Calculate the slope of the continuously differentiable response curve at significant inflection points. ;

[0129] Obtain the first conflict probability estimate after the significant inflection point. ;

[0130] Confidence index for determining the limit capacity assessment value at a significant inflection point for ;in, This represents the estimated probability of conflict at a significant inflection point. This is a significant inflection point.

[0131] In other words, to determine the inflection point identified above To determine whether the difference in response curve behavior before and after the inflection point is statistically significant and to avoid misjudgment due to random fluctuations, statistical hypothesis testing methods are used to test the significance of the difference.

[0132] All data points in the load level-conflict probability response curve can be divided into two groups according to load level: the group before the inflection point ( ) and the group after the inflection point ( ), For any data point, construct the null hypothesis. The conflict probability distributions of the two sets of data show no significant difference, i.e., the inflection point. The changes in the response curve before and after can be explained by random fluctuations; alternative hypothesis The conflict probability distributions of the two sets of data are significantly different.

[0133] The Mann-Whitney U test (also known as the Wilcoxon rank-sum test) is used to test the above hypotheses. This test is a non-parametric method that does not require the data to follow a normal distribution, and is suitable for situations where the data in Monte Carlo simulations may not be normally distributed. The Mann-Whitney U statistic and its corresponding p-value are calculated.

[0134] If the p-value is less than the preset significance level (like If the null hypothesis is rejected, then the null hypothesis is rejected. Determine the inflection point If the p-value is statistically significant, the inflection point is accepted as the effective capacity boundary inflection point; if the p-value is greater than or equal to If the statistical significance of the inflection point is insufficient, the smoothing scale needs to be readjusted or the current response curve is determined to have no obvious inflection point, that is, under the current environmental conditions, the probability of airspace conflict increases smoothly with the increase of load, without obvious boundary characteristics.

[0135] If, after readjusting the smoothing scale, a valid inflection point still cannot be identified, i.e., the response curve... In the domain If the upward trend is smooth and there is no obvious curvature maxima, then the no-inflection-point handling rule is triggered: a preset safety threshold is used. The corresponding load level is used as the conservative limit capacity, and interpolation is performed using an inverse function. The calculation was obtained, and a confidence index for determining the limit capacity was set. This result is identified as a conservative estimate under conditions without an inflection point.

[0136] If the inflection point is confirmed through the above steps If the load level at the inflection point is statistically significant, it is determined as the minimum low-altitude airspace capacity under the current environmental conditions, denoted as . If the above steps trigger a processing rule without a clear inflection point, then the conservative limit capacity will be applied. As the limit capacity assessment value, it is denoted as The result was then labeled as a conservative estimate.

[0137] To further improve the reliability of the judgment results, a comprehensive judgment is made by combining the local trends before and after the inflection point. The slope of the tangent line to the response curve at the inflection point is calculated. The probability of conflict with the first data point after the inflection point Define the confidence index of the limit capacity assessment value at the significant inflection point. This confidence index comprehensively reflects both the steepness of the rise (tangent slope) and the magnitude of the rise (relative change in conflict probability) at the inflection point. A higher value indicates a more reliable inflection point determination. As a supplementary indicator for limit capacity assessment, it is provided for reference by downstream airspace control systems.

[0138] Furthermore, the determined capacity boundary inflection point and its confidence index can be used to output the ultimate capacity assessment results of low-altitude airspace under the current environmental conditions. Simultaneously, through quantitative analysis of the uncertainties in the simulation input data, a comprehensive confidence characterization of the ultimate capacity assessment results is provided, enabling the assessment results to directly serve airspace control decisions.

[0139] On the one hand, the ultimate capacity assessment results can be used as the basic output, specifically including: the above ultimate capacity assessment values (If a valid inflection point exists, then it is) If there is no obvious inflection point, then it is a conservative estimate. Output the corresponding collision probability value. and the confidence index for determining the limit capacity. (If a valid inflection point exists, calculate using the original formula; if no inflection point exists, calculate using the original formula.) The output is appended with a "No Inflection Point - Conservative Estimation" label to indicate to downstream systems that the result is based on inference from a safety threshold rather than inflection point identification. This ensures that the probability of conflict remains below the safety threshold at this capacity level, but it has not undergone statistical significance verification of boundary inflection points. The above information is encapsulated in a structured data format (such as JSON or XML).

[0140] Among them, the maximum capacity This represents the maximum number of simultaneously operating aircraft that low-altitude airspace can support within an acceptable safety level, under current meteorological conditions, infrastructure status, and airspace structure constraints; critical conflict probability. This represents an estimate of the probability of an unsafe event occurring at this load level; its physical meaning is the overall safety risk level faced by the system when operating at its limit capacity boundary; the confidence index is used to determine the level of security. The overall reliability of this limit capacity determination is characterized by a value range of [value range missing]. The closer the result is to 1, the more reliable the judgment is.

[0141] On the other hand, the method can further perform a limit capacity confidence interval estimation based on input uncertainty. That is, after obtaining the conservative limit capacity at or without a significant inflection point, the method can also include:

[0142] Determine the set of uncertain parameters; the uncertain parameters should include at least meteorological state parameters, infrastructure state parameters, and virtual aircraft performance distribution parameters.

[0143] The Monte Carlo method is used to randomly sample each uncertainty parameter in the set of uncertainty parameters. Group input parameters;

[0144] At least based on The group input parameters are executed respectively. Next Figure 1 The Monte Carlo stochastic extrapolation and capacity boundary inflection point identification and verification provided in the illustrated embodiment yield the following results: Conservative limit capacity when there is a significant inflection point or no significant inflection point; wherein, the target low-altitude airspace digital twin model is reused in each simulation, and only the uncertainty parameters are updated incrementally;

[0145] Determine the conservative limit capacity when there is a significant inflection point or no significant inflection point obtained from the initial calculation. The standard deviation of the conservative limit capacity when there is a significant inflection point or no significant inflection point is the limit capacity assessment value.

[0146] Calculate the confidence interval of the ultimate capacity assessment value based on a pre-set confidence level. ;in, This is the lower limit of the confidence interval. This represents the upper limit of the confidence interval;

[0147] Confidence index for calculating the limit capacity assessment value at the significant inflection point ;

[0148] Calculate the input sensitivity index: ;in For input sensitivity indicators, The limit capacity assessment value is used; the sensitivity ranking of each uncertainty parameter is calculated based on the input sensitivity index.

[0149] Overall reliability of the calculated limit capacity assessment: ;in, To assess the overall reliability of the ultimate capacity assessment value, These are the weighting coefficients;

[0150] Packaging Critical conflict probability corresponding to the limit capacity assessment value , , The capacity assessment results are obtained by ranking the sensitivity of each uncertainty parameter.

[0151] In other words, the boundary conditions for the inference inputs in this embodiment of the invention, including meteorological monitoring data, infrastructure status data, and the distribution of virtual aircraft performance parameters, all exhibit varying degrees of uncertainty. To quantify the impact of these input uncertainties on the ultimate capacity assessment results, Monte Carlo sensitivity analysis can be used to estimate the confidence interval of the ultimate capacity. This analysis is an offline post-assessment process that does not change the ultimate capacity baseline value determined in the above steps, but only quantifies its uncertainty range around the baseline value.

[0152] The specific steps are as follows: Set the set of input uncertainty parameters. This includes meteorological data errors (such as wind speed measurement errors). Infrastructure status estimation error (such as coverage estimation error) The estimation errors include those related to the performance distribution parameters of the virtual aircraft and the estimation errors of the spatial distribution probability density function of the origin and destination points (such as historical data sampling bias, changes in demand patterns, etc.). For each uncertainty parameter... Set a reasonable fluctuation range .

[0153] Under the reference input conditions (i.e., determining the above-mentioned limit capacity assessment value) Based on the input data used at the time, the baseline limit capacity assessment value has been obtained. The Monte Carlo method is now used to independently and randomly sample from the fluctuation range of each uncertainty parameter. Group input parameter combination For each set of sampled input parameter combinations Perform a complete limit capacity assessment process, recalculate the limit capacity value under the input conditions using the same set of input parameters, and obtain the corresponding limit capacity assessment value. Since the combinations of input parameters are independent of each other, the inferences under different parameter combinations can be executed in parallel on the computing cluster (such as using a distributed computing framework or multi-threaded parallel computing). The total time for the group simulation is approximately equal to the time for a single group simulation multiplied by 1. By dividing by the number of parallel computing units, the computation time can be reduced to an acceptable range by properly allocating parallel resources.

[0154] Based on this, a data reuse strategy is further adopted: under various combinations of input parameters, digital twin models and virtual aircraft agent instances can be pre-generated in batches and cached, and incremental updates are only performed on key parameters that are affected (such as meteorological data errors, infrastructure coverage estimation errors, etc.), avoiding the need to completely rebuild the model from scratch for each simulation.

[0155] It should be noted that both the construction of the digital twin model and the Monte Carlo simulation of the virtual aircraft agent involved a large amount of random sampling. The logical relationship between the two is as follows: The evaluation is a complete execution that determines the baseline limit capacity under nominal input conditions; each Monte Carlo simulation generates a large number of random flight samples at the same load level to estimate the collision probability; and based on Sensitivity analysis of the simulation results involves repeating the entire evaluation process under different input conditions to observe the impact of input uncertainty on the final limit capacity output value. The former characterizes the statistical fluctuations of the simulation, while the latter characterizes the propagation of input uncertainty. The two address different issues at different stages, with clear logical levels and no overlap.

[0156] statistics Distribution characteristics of the group's limit capacity assessment values. Calculate the standard deviation of the limit capacity assessment values. and with a preset information level (like (Confidence level) to determine the confidence interval for the limiting capacity .

[0157] In some embodiments of the present invention, the assessment value based on the ultimate capacity can be comprehensively considered. Determined confidence index and input uncertainty confidence interval This forms a multi-dimensional confidence characterization system for the ultimate capacity assessment results, including:

[0158] Confidence index of the limit capacity assessment value at the significant inflection point The confidence index for the limit capacity assessment value at this significant inflection point is the statistical confidence level, which is used to characterize the statistical significance of identifying the capacity boundary inflection point and reflects the quality of the inferred data itself. The higher the value, the more significant the jump in conflict probability before and after the inflection point, and the less affected the inflection point identification is by random fluctuations;

[0159] Input sensitivity index : Characterizes the sensitivity of the limit capacity assessment results to the uncertainty of the input data, which is defined as That is, the ratio of the half-width of the limit capacity confidence interval to the benchmark value. The smaller the value, the less sensitive the evaluation results are to input uncertainty, and the more robust the results are. The larger the value, the greater the impact of input uncertainty on the evaluation results, requiring greater caution in practical applications;

[0160] Overall reliability Through weighted synthesis and get, ,in These are weighting coefficients, which can be set to emphasize different aspects of statistical significance and robustness depending on the specific application scenario. Overall reliability. It provides a unified confidence reference index for downstream airspace control systems, when If the result is below the preset threshold, the system may prompt the user or system that the uncertainty of the evaluation result is high and suggest supplementing data collection or adopting a conservative strategy.

[0161] Finally, the limit capacity baseline value Critical conflict probability Confidence interval Overall reliability The system also sorts the sensitivity of each input uncertainty parameter, encapsulates it according to a preset data format, and forms a standardized output result. The output format can adopt JavaScript Object Notation (JSON) data structure, which facilitates data exchange and integration with downstream airspace control systems (such as flight plan approval systems, traffic management modules, path planning engines, etc.).

[0162] All of the above-mentioned optional technical solutions can be combined in any way to form optional embodiments of the present invention, and will not be described in detail here.

[0163] The following are embodiments of the apparatus of the present invention, which can be used to execute embodiments of the method of the present invention. For details not disclosed in the embodiments of the apparatus of the present invention, please refer to the embodiments of the method of the present invention.

[0164] Figure 2 This is a schematic diagram of the structure of a digital twin simulation and dynamic capacity assessment device for low-altitude airspace operation scenarios provided in an embodiment of the present invention. Figure 2 As shown, the device includes:

[0165] Module 201 is configured to construct a digital twin model of the target low-altitude airspace. The digital twin model of the target low-altitude airspace includes multiple airspace grid agents, multiple meteorological grid agents, multiple infrastructure grid agents, and initial availability labels. Each airspace grid agent includes static attribute data, each meteorological grid agent includes meteorological status data of the meteorological grid, and each infrastructure grid agent includes infrastructure service status data of the infrastructure grid.

[0166] The simulation module 202 is configured to use the current environmental conditions as the basis for the state of the target low-altitude airspace digital twin model, with the current environmental conditions as the simulation boundary, to progressively increase the virtual aircraft load level, and to execute independently at each load level. The Monte Carlo stochastic simulation is repeated until the simulation termination condition is met, resulting in the load level-conflict probability response curve for low-altitude airspace. Under each virtual aircraft load level, the performance parameters of the virtual aircraft are generated based on differential sampling from a preset probability distribution model. The conflict probability under each load level is determined by... The conflict indicator variables obtained from the second simulation are determined by probability statistics and convergence judgment. It is a positive integer. greater than 0 and less than or equal to positive integers, The number of load levels for the determined aircraft.

[0167] The statistics module 203 is configured to use a multi-scale curvature analysis method to identify capacity boundary inflection points on the load level-conflict probability response curve, and to determine the validity of the capacity boundary inflection points through statistical hypothesis testing, so as to obtain the limit capacity assessment value when there is a significant inflection point or the conservative limit capacity when there is no significant inflection point.

[0168] The evaluation module 204 is configured to calculate the confidence index of the limit capacity assessment value at the significant inflection point, determine the significant inflection point and its confidence index as the capacity assessment result, or determine the conservative limit capacity as the capacity assessment result.

[0169] According to the technical solution provided in this embodiment of the invention, a digital twin model of the target low-altitude airspace is constructed, and virtual aircraft agents with differentiated performance parameters are injected in batches into the twin environment. The Monte Carlo method is used for parallel simulation across multiple scenarios. By progressively increasing the number of virtual aircraft and monitoring changes in system conflict probability, the capacity boundary inflection point is automatically identified, achieving a quantitative assessment of the ultimate capacity of low-altitude airspace. This method integrates the precise mapping capability of digital twins, the autonomous interaction characteristics of multiple agents, and the probabilistic inference advantages of Monte Carlo simulation, providing a more scientific decision-making basis for low-altitude airspace planning, traffic management, and safety control.

[0170] In some implementations, constructing a digital twin model of the target low-altitude airspace includes: constructing a discretized three-dimensional airspace grid of the target low-altitude airspace, configuring an airspace grid agent for each airspace grid, and assigning static attribute data to each airspace grid agent; the static attribute data includes the spatial location, geometric dimensions, physical attributes, and regulatory attributes of the airspace grid; constructing a meteorological grid agent, using the meteorological grid agent to acquire meteorological monitoring data of the target low-altitude airspace, and mapping the meteorological monitoring data to each airspace grid agent through spatial interpolation, assigning meteorological state data to each airspace grid agent; and constructing an infrastructure grid agent, using the infrastructure grid agent to acquire infrastructure state data of the target low-altitude airspace, and mapping the infrastructure state data values ​​to each infrastructure based on the spatial coverage area of ​​each infrastructure. An airspace grid agent is assigned infrastructure service status data to each airspace grid agent. The target airspace grid agent's status data is obtained by fusing static attribute data, meteorological status data, and infrastructure service status data. An initial availability label for the target airspace grid agent is determined based on its status data using preset airspace availability judgment rules. Consistency checks are performed on the target airspace grid agent's status data and initial availability labels. Status data and initial availability labels that fail the check are updated. Upon successful check, the set of airspace grid agents composed of the status data and initial availability labels of each airspace grid agent is determined as the target low-altitude airspace digital twin model. The target airspace grid agent can be any airspace grid agent.

[0171] In some implementations, meteorological monitoring data is mapped to various spatial grid agents through spatial interpolation, including:

[0172] Constructing a set of meteorological elements based on meteorological monitoring data Among them, the meteorological element set includes Meteorological elements, It is a positive integer; determine the first... A set of neighboring meteorological data points for a spatial grid agent Determining the first by spatial interpolation Meteorological status data of a spatial grid agent for: ;in, Indicates the first A collection of meteorological elements from a set of meteorological data points. , Indicates the first The first of the meteorological data points Meteorological elements; The interpolation weight coefficients satisfy the following conditions: ; It is a positive integer greater than or equal to 1 and less than or equal to the total number of smart agents in the spatial grid. greater than or equal to 1 and less than or equal to 1 Positive integers; mapping infrastructure status data values ​​to each spatial grid agent based on the spatial coverage area of ​​each infrastructure, including: obtaining the service range of the target infrastructure; the target infrastructure is any infrastructure; determining the capability degradation coefficient of the target infrastructure based on the current environmental conditions; determining the product of the service range and the capability degradation coefficient as the spatial coverage area of ​​the target infrastructure; in response to determining that the grid center point of the target spatial grid agent falls within the spatial coverage area of ​​the target infrastructure, mapping the status data values ​​of the target infrastructure to the target spatial grid agent; consistency verification includes at least spatial consistency verification, logical consistency verification and data integrity verification.

[0173] In some implementations, the virtual aircraft is represented by a virtual aircraft agent; Attributes of a virtual aircraft intelligent agent for: ;in, It is a unique identifier; It is a set of flight performance parameters, including at least cruise speed, maximum rate of climb, maximum yaw angle, and maximum flight altitude; For communication capability parameters, representing the first... The degree of dependence of a virtual aircraft on infrastructure services; A set of meteorological tolerance thresholds, including at least the maximum tolerable wind speed and the minimum visibility requirement; attributes , and The performance parameters of the corresponding virtual aircraft; the cruise speed under each level of virtual aircraft load is generated by differential sampling based on a normal distribution model; the attributes in the meteorological tolerance threshold set are generated by differential sampling based on a uniform distribution or a Weibull distribution model.

[0174] In some implementations, execution is performed independently at each load level. Sub-Monte Carlo randomization, including: Randomly generated for the current load level Each virtual aircraft agent instance is analyzed, and its attributes are generated independently through sampling. A virtual aircraft intelligent agent; the current load level is any load level before reaching the termination condition; will A virtual aircraft agent is injected into the target low-altitude airspace digital twin model, distributing it at random initial positions within the airspace; the Monte Carlo method is used for... Each virtual aircraft agent generates multiple flight mission samples; in each Monte Carlo stochastic simulation, flight mission parameters are randomly generated for each virtual aircraft agent; flight mission parameters include at least takeoff point coordinates and destination coordinates, planned route, and mission execution time window; flight mission samples are executed based on flight mission parameters, and conflict indicator variables are recorded for each simulation; the conflict indicator variable takes a value of 1 when at least one unsafe event occurs in the simulation, and a value of 0 when no unsafe event occurs in the simulation; unsafe events include at least conflict events, environment mismatch events, and communication loss events; [The last sentence appears to be incomplete and possibly refers to a different scenario.] After the Monte Carlo stochastic simulation, the estimated conflict probability at this load level is calculated; the estimated conflict probability is... The proportion of simulations in which unsafe events occur out of the total number of simulations; the simulation result is considered convergent if the standard deviation of the determined conflict probability estimate is less than a preset convergence threshold; the number of simulations is increased if the standard deviation of the determined conflict probability estimate is greater than or equal to the preset convergence threshold. Re-execute independently at this load level The Monte Carlo stochastic simulation continues until the current level's simulation results converge; in response to confirming convergence of the current level's simulation results, the current level's load level is saved. Number of simulations The collision probability estimate and convergence standard deviation at convergence; the convergence standard deviation is the standard deviation of the collision probability estimate at convergence.

[0175] In some implementations, the virtual aircraft load level is increased incrementally, including: in response to the estimated conflict probability being less than a preset safety threshold when the current level's results converge, increasing the load level in preset steps. Increase the load level to obtain the next load level. The following load levels To obtain the new current load level, perform the Monte Carlo stochastic simulation described above. The estimated conflict probability at convergence of the results; if the estimated conflict probability at convergence of the results under the new current load level is less than the preset safety threshold, then the iteration proceeds with a preset step size. Increase the load level, use the increased load level as the new current load level, and perform the Monte Carlo stochastic inference described above to obtain the estimated conflict probability value when the result converges under the new current load level. Continue until the estimated conflict probability value when the result converges under the new current load level is greater than or equal to a preset safety threshold, and determine that the inference termination condition is met. The load level-conflict probability response curve is the curve formed by the estimated conflict probability value obtained from the inference at each load level and the corresponding load level.

[0176] In some implementations, a multi-scale curvature analysis method is used to identify capacity boundary inflection points on the load level-conflict probability response curve. This includes: smoothing and filtering each conflict probability estimate in the load level-conflict probability response curve to eliminate statistical fluctuations introduced by Inmont Carlo random sampling; and smoothing the filtered load level-conflict probability response curve using a local weighted regression method; wherein the smoothed value of each data point is derived from its neighboring values. The weighted average of the data points is obtained; To preset a smooth window, Positive integer; based on load level Using the independent variable and the conflict probability estimate as the dependent variable, a cubic spline interpolation method is used to continuously fit the smoothed load level-conflict probability response curve to obtain a continuously differentiable response curve; setting... Different smoothing scales ,in , A positive integer greater than 2; at each smoothing scale Below, a Gaussian kernel is used. The continuous differentiable response curve is smoothed by convolution to obtain the smoothed continuous differentiable response curve; for to Given any smoothing scale; calculate the curvature of the smoothed, continuously differentiable response curve obtained at each smoothing scale, and obtain... Local maxima of curvature; for The capacity boundary inflection point is obtained by weighting the local maxima of curvature; where the smoothing scale is inversely proportional to the weight.

[0177] In some implementations, the validity of the capacity boundary inflection point is determined through statistical hypothesis testing, including: dividing the conflict probability estimates derived from various load levels into a first group of data and a second group of data; wherein the first group of data includes conflict probability estimates less than or equal to the capacity boundary inflection point, and the second group of data includes conflict probability estimates greater than the capacity boundary inflection point; calculating the U-statistic for the first group of data and the second group of data: ;in, , , Let the sum of the rank of the first set of data be . Let the sum of the ranks of the second set of data be . This is the sample size for the first set of data. The sample size for the second set of data is given; the p-values ​​for the first and second sets of data are determined based on the U statistic: in response to determining that the p-value is less than the preset significance level, the capacity boundary inflection point is determined to be valid and significant; in response to determining that the p-value is greater than or equal to the preset significance level, the p-value is reset. Different smoothing scales , use as Figure 1 The method provided in the illustrated embodiment redetermines the capacity boundary inflection point; based on the redetermined capacity boundary inflection point, the conflict probability estimates derived from each load level are reclassified into a first group of data and a second group of data, and the p-values ​​of the reclassified first group of data and second group of data are calculated; in response to determining that the p-value is less than a preset significance level, the redetermined capacity boundary inflection point is determined to be valid and a significant inflection point; in response to determining that the p-value is greater than or equal to the preset significance level, it is determined that there is currently no significant inflection point, and the load level corresponding to the preset safety threshold on the continuously differentiable response curve is determined to be the conservative limit capacity; the confidence index of the significant inflection point is calculated, including: calculating the slope of the continuously differentiable response curve at the significant inflection point. Obtain the first conflict probability estimate after the significant inflection point. Confidence index for determining significant inflection points for ;in, This represents the estimated probability of conflict at a significant inflection point. This is a significant inflection point.

[0178] In some implementations, after obtaining the conservative limit capacity at or without a significant inflection point, the method further includes: determining a set of uncertainty parameters; the uncertainty parameters include at least meteorological state parameters, infrastructure state parameters, and virtual aircraft performance distribution parameters; and randomly sampling each uncertainty parameter in the set using the Monte Carlo method. Group input parameters; at least based on The group input parameters are executed respectively. Next Figure 1 The Monte Carlo stochastic extrapolation and capacity boundary inflection point identification and verification provided in the illustrated embodiment yield the following results: The conservative limit capacity is determined for either a significant inflection point or no significant inflection point. In each simulation, the target low-altitude airspace digital twin model is reused, with only the uncertainty parameters updated incrementally. The estimated limit capacity value at the significant inflection point or the conservative limit capacity at the absence of a significant inflection point, obtained from the initial calculation, is determined. The limit capacity assessment value is the standard deviation of the limit capacity at the significant inflection point or the conservative limit capacity without a significant inflection point; the confidence interval of the limit capacity assessment value is calculated based on a pre-set confidence level. ;in, This is the lower limit of the confidence interval. The upper limit of the confidence interval; a confidence index for calculating the limit capacity assessment value at a significant inflection point. ; Calculate the input sensitivity index: ;in For input sensitivity indicators, The limit capacity assessment value is used; the sensitivity ranking of each uncertainty parameter is calculated based on the input sensitivity index; the overall reliability of the limit capacity assessment value is calculated. ;in, To assess the overall reliability of the ultimate capacity assessment value, Weighting coefficients; encapsulation Critical conflict probability corresponding to the limit capacity assessment value , , The capacity assessment results are obtained by ranking the sensitivity of each uncertainty parameter.

[0179] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0180] Figure 3 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. For example... Figure 3 As shown, the electronic device 3 of this embodiment includes a processor 301, a memory 302, and a computer program 303 stored in the memory 302 and executable on the processor 301. When the processor 301 executes the computer program 303, it implements the steps in the various method embodiments described above. Alternatively, when the processor 301 executes the computer program 303, it implements the functions of each module / unit in the various device embodiments described above.

[0181] Electronic device 3 can be a desktop computer, laptop, handheld computer, cloud server, or other electronic device. Electronic device 3 may include, but is not limited to, processor 301 and memory 302. Those skilled in the art will understand that... Figure 3 This is merely an example of electronic device 3 and does not constitute a limitation on electronic device 3. It may include more or fewer components than shown, or different components.

[0182] The processor 301 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0183] The memory 302 can be an internal storage unit of the electronic device 3, such as a hard disk or RAM of the electronic device 3. The memory 302 can also be an external storage device of the electronic device 3, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, Flash Card, etc., equipped on the electronic device 3. The memory 302 can also include both internal and external storage units of the electronic device 3. The memory 302 is used to store the computer program 303 and other programs and data required by the electronic device.

[0184] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0185] If integrated modules / units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program may include computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. Computer-readable media may include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0186] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A digital twin simulation and dynamic capacity assessment method for low-altitude airspace operation scenarios, characterized in that, include: Construct a digital twin model of the target low-altitude airspace; The target low-altitude airspace digital twin model includes multiple airspace grid agents, multiple meteorological grid agents, multiple infrastructure grid agents, and an initial availability label. Each airspace grid agent includes static attribute data of the airspace grid, each meteorological grid agent includes meteorological status data of the meteorological grid, and each infrastructure grid agent includes infrastructure service status data of the infrastructure grid. Based on the current state of the target low-altitude airspace digital twin model under the current environmental conditions, and using the current environmental conditions as the extrapolation boundary, the virtual aircraft load level is progressively increased, and the system executes independently at each load level. The Monte Carlo stochastic simulation is repeated until the simulation termination condition is met, resulting in the load level-conflict probability response curve for low-altitude airspace. Under each virtual aircraft load level, the performance parameters of the virtual aircraft are generated based on differential sampling from a preset probability distribution model. The conflict probability under each load level is determined by... The conflict indicator variables obtained from the second simulation are determined by probability statistics and convergence judgment. It is a positive integer. greater than 0 and less than or equal to positive integers, To determine the number of aircraft load levels; Multiscale curvature analysis is used to identify capacity boundary inflection points on the load level-conflict probability response curve. The validity of the capacity boundary inflection points is determined by statistical hypothesis testing, and the limit capacity assessment value at significant inflection points or the conservative limit capacity at no significant inflection points is obtained. Calculate the confidence index of the limit capacity assessment value at the significant inflection point, and determine the significant inflection point and its confidence index as the capacity assessment result, or determine the conservative limit capacity as the capacity assessment result.

2. The digital twin simulation and dynamic capacity assessment method for low-altitude airspace operation scenarios according to claim 1, characterized in that, Constructing a digital twin model of the target low-altitude airspace, including: A discretized three-dimensional airspace grid of the target low-altitude airspace is constructed, an airspace grid agent is configured for each airspace grid, and static attribute data is assigned to each airspace grid agent; the static attribute data includes the spatial location, geometric dimensions, physical attributes and regulatory attributes of the airspace grid. A meteorological grid intelligent agent is constructed, which is used to acquire meteorological monitoring data of the target low-altitude airspace. The meteorological monitoring data is then mapped to each airspace grid intelligent agent through spatial interpolation, and meteorological status data is assigned to each airspace grid intelligent agent. Construct an infrastructure grid intelligent agent, use the infrastructure grid intelligent agent to obtain infrastructure status data of the target low-altitude airspace, map the infrastructure status data value to each airspace grid intelligent agent based on the spatial coverage area of ​​each infrastructure, and assign infrastructure service status data to each airspace grid intelligent agent. The static attribute data, meteorological status data, and infrastructure service status data of the target airspace grid agent are integrated to obtain the status data of the target airspace grid agent. The initial availability label of the target airspace grid agent is determined based on the status data of the target airspace grid agent using a preset airspace availability judgment rule. The consistency of the state data and initial availability labels of the target airspace grid agents is verified. The state data and initial availability labels that fail the verification are updated. When the verification passes, the set of airspace grid agents composed of the state data and initial availability labels of each airspace grid agent is determined as the target low-altitude airspace digital twin model. The target spatial grid agent can be any spatial grid agent.

3. The digital twin simulation and dynamic capacity assessment method for low-altitude airspace operation scenarios according to claim 2, characterized in that, The meteorological monitoring data is mapped to each spatial grid agent through spatial interpolation, including: Constructing a set of meteorological elements based on meteorological monitoring data The meteorological element set includes... Meteorological elements, It is a positive integer; Determine the first A set of neighboring meteorological data points for a spatial grid agent Determining the first by spatial interpolation Meteorological status data of a spatial grid agent for: ;in, Indicates the first A collection of meteorological elements from a set of meteorological data points. , Indicates the first The first of the meteorological data points Meteorological elements; The interpolation weight coefficients satisfy the following conditions: ; It is a positive integer greater than or equal to 1 and less than or equal to the total number of smart agents in the spatial grid. greater than or equal to 1 and less than or equal to 1 Positive integers; Mapping the infrastructure status data values ​​to each spatial grid agent based on the spatial coverage area of ​​each infrastructure includes: Obtain the service scope of the target infrastructure; the target infrastructure can be any type of infrastructure. Determine the capacity degradation factor of the target infrastructure based on current environmental conditions; The product of the service range and the capacity degradation coefficient is determined as the spatial coverage area of ​​the target infrastructure; In response to determining that the grid center point of the target spatial grid agent falls within the spatial coverage area of ​​the target infrastructure, the state data values ​​of the target infrastructure are mapped to the target spatial grid agent; Consistency checks include at least spatial consistency checks, logical consistency checks, and data integrity checks.

4. The digital twin simulation and dynamic capacity assessment method for low-altitude airspace operation scenarios according to claim 1, characterized in that, Virtual aircraft are represented by virtual aircraft intelligent agents; the first Attributes of a virtual aircraft intelligent agent for: ;in, It is a unique identifier; It is a set of flight performance parameters, including at least cruise speed, maximum rate of climb, maximum yaw angle, and maximum flight altitude; For communication capability parameters, representing the first... The degree of dependence of a virtual aircraft on infrastructure services; A set of meteorological tolerance thresholds, including at least the maximum tolerable wind speed and the minimum visibility requirement; attributes , and The corresponding performance parameters of the virtual aircraft; Under the various virtual aircraft load levels, the cruise speed is generated based on differential sampling of the normal distribution model; the attributes in the meteorological tolerance threshold set are generated based on differential sampling of the uniform distribution or Weibull distribution model.

5. The digital twin simulation and dynamic capacity assessment method for low-altitude airspace operation scenarios according to claim 4, characterized in that, Execute independently at each load level Sub-Monte Carlo stochastic simulations include: by Randomly generated for the current load level Each virtual aircraft agent instance is analyzed, and its attributes are generated independently through sampling. A virtual aircraft intelligent agent; the current load level is any load level before the termination condition is met; The A virtual aircraft agent is injected into the target's low-altitude airspace digital twin model, causing it to be distributed at random initial positions in the airspace; The Monte Carlo method is used for the above A virtual aircraft agent generates multiple flight mission samples; In each Monte Carlo stochastic simulation, flight mission parameters are randomly generated for each virtual aircraft agent; the flight mission parameters include at least the takeoff point coordinates and destination coordinates, the planned route, and the mission execution time window. The flight mission sample is executed based on the flight mission parameters, and the conflict indicator variable for each simulation is recorded; wherein, the conflict indicator variable takes a value of 1 when at least one unsafe event occurs in the simulation, and takes a value of 0 when no unsafe event occurs in the simulation; the unsafe event includes at least a conflict event, an environmental mismatch event, and a communication loss event; To achieve the current load level After the Monte Carlo stochastic simulation, the estimated conflict probability value under the current load level is calculated; the estimated conflict probability value is... The proportion of simulations in which unsafe events occurred out of the total number of simulations; In response to the fact that the standard deviation of the estimated conflict probability is less than a preset convergence threshold, the inference result of this level is determined to be converged. In response to the standard deviation of the determined conflict probability estimate being greater than or equal to a preset convergence threshold, the number of iterations is increased. Re-execute independently at this load level The Monte Carlo stochastic simulation is repeated until the results of this level of simulation converge. In response to the convergence of the current level's simulation results, the current level's load level is saved. Number of simulations The results include the estimated conflict probability and the convergence standard deviation at the time of convergence; the convergence standard deviation is the standard deviation of the estimated conflict probability at the time of convergence.

6. The digital twin simulation and dynamic capacity assessment method for low-altitude airspace operation scenarios according to claim 5, characterized in that, The virtual aircraft load level is increased progressively, including: If the estimated conflict probability at the time of convergence of the current level is less than a preset safety threshold, the preset step size is used. Increase the load level to obtain the next load level. ; The following load levels To obtain the new current load level, perform the Monte Carlo stochastic deduction as described in claim 5. The estimated probability of conflict when the result converges; If the estimated conflict probability at convergence under the new current load level is less than a preset safety threshold, then the iteration will proceed with a preset step size. Increase the load level, take the increased load level as the new current load level, perform the Monte Carlo stochastic inference as described in claim 5, and obtain the collision probability estimate when the result converges under the new current load level, until the collision probability estimate when the result converges under the new current load level is greater than or equal to a preset safety threshold, and determine that the inference termination condition is met. The load level-conflict probability response curve is a curve formed by the estimated conflict probability obtained from each load level and the corresponding load level.

7. The digital twin simulation and dynamic capacity assessment method for low-altitude airspace operation scenarios according to claim 6, characterized in that, Multi-scale curvature analysis was used to identify capacity boundary inflection points on the load level-conflict probability response curve, including: Smoothing filters are applied to each conflict probability estimate in the load level-conflict probability response curve to eliminate statistical fluctuations introduced by Inmont Carlo random sampling; The filtered load level-conflict probability response curve is smoothed using a locally weighted regression method; whereby the smoothed value of each data point is determined by its neighboring values. The weighted average of the data points is obtained; To preset a smooth window, It is a positive integer; By load level With the independent variable and the conflict probability estimate as the dependent variable, the smoothed load level-conflict probability response curve is fitted with a continuous function using the cubic spline interpolation method to obtain a continuously differentiable response curve. set up Different smoothing scales ,in , It is a positive integer greater than 2; At each smoothing scale Below, a Gaussian kernel is used. The continuously differentiable response curve is convolved and smoothed to obtain a smoothed continuously differentiable response curve; for to Any smoothing scale in; Calculate the curvature of the smoothed, continuously differentiable response curve obtained at each smoothing scale, and obtain... Local maxima of curvature; right The capacity boundary inflection point is obtained by weighting the local maxima of curvature; wherein the smoothing scale is inversely proportional to the weight.

8. The digital twin simulation and dynamic capacity assessment method for low-altitude airspace operation scenarios according to claim 7, characterized in that, The validity of capacity boundary inflection points is determined through statistical hypothesis testing, including: The conflict probability estimates derived from the load levels are divided into a first group of data and a second group of data. The first group of data includes conflict probability estimates that are less than or equal to the capacity boundary inflection point, and the second group of data includes conflict probability estimates that are greater than the capacity boundary inflection point. Calculate the U-statistic for the first and second sets of data: ;in, , , Let the sum of the rank of the first set of data be . Let the sum of the ranks of the second set of data be . This is the sample size for the first set of data. This is the sample size for the second set of data; Determine the p-values ​​for the first and second sets of data based on the U statistic: In response to determining that the p-value is less than a preset significance level, the capacity boundary inflection point is determined to be valid and that the capacity boundary inflection point is a significant inflection point; In response to determining that the p-value is greater than or equal to a preset significance level, the threshold is reset. Different smoothing scales The capacity boundary inflection point is redefined using the method described in claim 7; Based on the redefined capacity boundary inflection point, the estimated conflict probability values ​​derived under each load level are reclassified into a first group of data and a second group of data, and the p-values ​​of the first group of data and the second group of data after reclassification are calculated. In response to determining that the p-value is less than a preset significance level, the redefined capacity boundary inflection point is determined to be valid and the redefined capacity boundary inflection point is a significant inflection point; In response to determining that the p-value is greater than or equal to a preset significance level, it is determined that there is currently no significant inflection point, and the load level corresponding to the preset safety threshold on the continuously differentiable response curve is determined to be the conservative limit capacity. The confidence index for calculating the limit capacity assessment value at the significant inflection point includes: Calculate the slope of the continuously differentiable response curve at significant inflection points. ; Obtain the first conflict probability estimate after the significant inflection point. ; Confidence index for determining the limit capacity assessment value at a significant inflection point for ;in, This represents the estimated probability of conflict at a significant inflection point. This is a significant inflection point.

9. The digital twin simulation and dynamic capacity assessment method for low-altitude airspace operation scenarios according to claim 1, characterized in that, After obtaining the limit capacity assessment value at a significant inflection point or the conservative limit capacity at no significant inflection point, the method further includes: Determine the set of uncertainty parameters; the uncertainty parameters include at least meteorological state parameters, infrastructure state parameters, and virtual aircraft performance distribution parameters. The Monte Carlo method is used to randomly sample each uncertainty parameter in the set of uncertainty parameters. Group input parameters; At least based on the The group input parameters are executed respectively. The Monte Carlo stochastic extrapolation and capacity boundary inflection point identification and verification as described in claim 1 are obtained as follows: Conservative limit capacity when there is a significant inflection point or no significant inflection point; wherein, the target low-altitude airspace digital twin model is reused in each simulation, and only the uncertainty parameters are updated incrementally; Determine the limit capacity assessment value at the significant inflection point obtained from the first calculation, or the conservative limit capacity at the absence of a significant inflection point, and the aforementioned... The standard deviation of the limit capacity assessment value at the group's significant inflection point or the conservative limit capacity when there is no significant inflection point is the limit capacity assessment value. Calculate the confidence interval of the ultimate capacity assessment value based on a pre-set confidence level. ;in, This is the lower limit of the confidence interval. This represents the upper limit of the confidence interval; Confidence index for calculating the limit capacity assessment value at the significant inflection point ; Calculate the input sensitivity index: ;in For input sensitivity indicators, The limit capacity assessment value is used; the sensitivity ranking of each uncertainty parameter is calculated based on the input sensitivity index. Overall reliability of the calculated limit capacity assessment: ;in, To assess the overall reliability of the ultimate capacity assessment value, These are the weighting coefficients; Packaging Critical conflict probability corresponding to the limit capacity assessment value , , The capacity assessment results are obtained by ranking the sensitivity of each uncertainty parameter.

10. A digital twin simulation and dynamic capacity assessment device for low-altitude airspace operation scenarios, characterized in that, include: The building module is configured to build a digital twin model of the target low-altitude airspace; The target low-altitude airspace digital twin model includes multiple airspace grid agents, multiple meteorological grid agents, multiple infrastructure grid agents, and an initial availability label. Each airspace grid agent includes static attribute data, each meteorological grid agent includes meteorological status data of the meteorological grid, and each infrastructure grid agent includes infrastructure service status data of the infrastructure grid. The simulation module is configured to use the current environmental conditions as the basis for the state of the target low-altitude airspace digital twin model, with the current environmental conditions as the simulation boundary, progressively increasing the virtual aircraft load level, and executing independently at each load level. The Monte Carlo stochastic simulation is repeated until the simulation termination condition is met, resulting in the load level-conflict probability response curve for low-altitude airspace. Under each virtual aircraft load level, the performance parameters of the virtual aircraft are generated based on differential sampling from a preset probability distribution model. The conflict probability under each load level is determined by... The conflict indicator variables obtained from the second simulation are determined by probability statistics and convergence judgment. It is a positive integer. greater than 0 and less than or equal to positive integers, To determine the number of aircraft load levels; The statistics module is configured to use multi-scale curvature analysis to identify capacity boundary inflection points on the load level-conflict probability response curve, and to determine the validity of capacity boundary inflection points through statistical hypothesis testing, thereby obtaining the limit capacity assessment value when there is a significant inflection point or the conservative limit capacity when there is no significant inflection point. The evaluation module is configured to calculate the confidence index of the limit capacity assessment value at a significant inflection point, and determine the significant inflection point and its confidence index as the capacity assessment result, or determine the conservative limit capacity as the capacity assessment result.

Images