An ai-based low-altitude airspace grid dynamic division and management method and system

Abstract

The application relates to an AI-based low-altitude airspace grid dynamic division and management method and system, which comprises the following steps: collecting a multi-source data set corresponding to a target low-altitude airspace in real time, and generating an airspace situation image; performing demand analysis on the airspace situation image to obtain operation demand information, and calculating and generating grid division parameters; generating a dynamic airspace grid through a grid division algorithm, and allocating an airspace resource identifier; obtaining aircraft information and generating planning path information; issuing the planning path information to a management end and performing operation state monitoring; evaluating an airspace situation change amount based on the operation state monitoring result, and triggering a dynamic adjustment mechanism if the change amount exceeds a change threshold; and in summary, the application realizes dynamic management of a low-altitude airspace by collecting multi-source data in real time, generating an airspace situation image, performing demand analysis on the airspace situation image, dynamically dividing a grid, performing collaborative path planning and monitoring adjustment based on the dynamic airspace grid, and the like.

Description

Technical Field

[0001] This application relates to the technical field of low-altitude airspace management, and in particular to an AI-based method and system for dynamic division and management of low-altitude airspace grids. Background Technology

[0002] With the rapid popularization of low-altitude aircraft such as drones and electric vertical take-off and landing aircraft in logistics distribution, urban inspection, emergency rescue and other fields, the intensity of low-altitude airspace resource utilization is increasing significantly. Under this situation, the flexibility and safety of airspace management have become a major demand.

[0003] Current mainstream management technologies rely on pre-set fixed routes and static isolation zones. Such technologies can maintain basic operational order in scenarios where the number of low-altitude aircraft is limited and the mission types are simple. However, when faced with the realities of high-density, multi-type aircraft operating together and rapidly changing operating environments, the limitations of static management models are becoming increasingly apparent. Specifically, the rigid architecture of existing static management models is no longer able to adapt to the dynamic fluctuations of complex flight demands. For example, sudden surges in traffic caused by emergency missions, path interruptions due to temporary weather changes, or airspace function restrictions caused by infrastructure failures directly lead to an imbalance in the spatiotemporal distribution of airspace resources: some areas experience frequent flight conflicts due to over-concentration, requiring inefficient coordination through manual intervention; other areas remain idle for extended periods, resulting in severely underutilized resources. Furthermore, existing technologies lack the ability to continuously perceive and analyze the airspace situation. When aircraft status changes, environmental parameters fluctuate, or external constraints are updated, the existing technological system struggles to achieve real-time reconstruction of the airspace structure. In summary, the response delays caused by these shortcomings not only exacerbate the complexity of conflict resolution but may also trigger a chain of safety risks, failing to meet the dual demands of efficient resource allocation and operational safety assurance for future large-scale low-altitude commercial operations. Summary of the Invention

[0004] To address the aforementioned shortcomings, this application provides an AI-based method and system for dynamic partitioning and management of low-altitude airspace grids.

[0005] The above-mentioned objective of this application is achieved through the following technical solution: A method for dynamic partitioning and management of low-altitude airspace grids based on AI, comprising the following steps: Real-time acquisition of multi-source datasets corresponding to the target low-altitude airspace, and generation of airspace situation profiles based on the multi-source datasets, wherein the multi-source datasets include aircraft status data, flight plan data, airspace environment data, and infrastructure status data; A demand analysis is performed on the airspace situation profile to obtain operational demand information, and grid partitioning parameters are calculated and generated based on the operational demand information. The grid partitioning parameters include grid granularity, grid level, and grid shape. A dynamic airspace grid matching the operational requirements is generated using a preset grid partitioning algorithm, and airspace resource identifiers are assigned to the dynamic airspace grid. Acquire the target aircraft's aircraft information, and perform collaborative path planning based on the dynamic airspace grid and aircraft information to generate planned path information; The planned path information is sent to the management terminal associated with the corresponding target aircraft, and the operational status of the target aircraft is monitored. The changes in the airspace situation of the target low-altitude airspace are assessed based on the results of the operational status monitoring. If the changes in the airspace situation exceed the preset change threshold, a preset dynamic adjustment mechanism is triggered to adjust the dynamic airspace grid.

[0006] The second objective of this invention is achieved through the following technical solution: An AI-based dynamic low-altitude airspace grid partitioning and management system includes: The image generation module is used to collect multi-source datasets corresponding to the target low-altitude airspace in real time, and generate an airspace situation image based on the multi-source datasets. The multi-source datasets include aircraft status data, flight plan data, airspace environment data, and infrastructure status data. The parameter generation module is used to perform demand analysis on the airspace situation profile, obtain operational demand information, and calculate and generate grid division parameters based on the operational demand information. The grid division parameters include grid granularity, grid level, and grid shape. The grid generation module is used to generate a dynamic airspace grid that matches the operational requirements information through a preset grid division algorithm, and to assign airspace resource identifiers to the dynamic airspace grid. The path generation module is used to acquire the target aircraft's aircraft information and perform collaborative path planning based on the dynamic airspace grid and aircraft information to generate planned path information. The status monitoring module is used to send the planned path information to the management terminal associated with the corresponding target aircraft and to monitor the operational status of the target aircraft. The dynamic adjustment module is used to evaluate the change in airspace situation of the target low-altitude airspace based on the monitoring results of the operation status. If the change in airspace situation exceeds the preset change threshold, the preset dynamic adjustment mechanism is triggered to adjust the dynamic airspace grid.

[0007] In summary, this application provides an AI-based method and system for dynamic partitioning and management of low-altitude airspace grids. By collecting multi-source data in real time and generating an airspace situation profile, performing demand analysis on the airspace situation profile and dynamically partitioning the grid, and conducting collaborative path planning and monitoring adjustments based on the dynamic airspace grid, dynamic management of low-altitude airspace can be achieved. This improves the flexibility and real-time performance of low-altitude airspace management, optimizes resource allocation, reduces flight conflicts, and enhances operational safety and efficiency. Attached Figure Description

[0008] Figure 1 This is a flowchart of an embodiment of an AI-based method for dynamic partitioning and management of low-altitude airspace grids according to this application; Figure 2 This is a flowchart of step S10 in an embodiment of an AI-based method for dynamic partitioning and management of low-altitude airspace grids in this application. Figure 3 This is a flowchart of step S13 in an embodiment of an AI-based method for dynamic partitioning and management of low-altitude airspace grids in this application. Figure 4 This is a schematic diagram of the low-altitude airspace grid cell mapping and structure in an embodiment of an AI-based method for dynamic partitioning and management of low-altitude airspace grids according to this application. Detailed Implementation

[0009] The following is in conjunction with the appendix Figures 1-4 This application will be described in further detail.

[0010] In the field of low-altitude airspace management, existing technologies often rely on pre-defined fixed routes or statically isolated airspaces. Their inherent architecture is ill-suited to the high-density, multi-type, and highly dynamic characteristics of low-altitude operations. Specifically, static partitioning mechanisms cannot flexibly adapt to real-time changes in flight demands and operating environments, resulting in uneven spatial and temporal distribution of airspace resources. Some areas may be overcrowded while others may be largely idle, leading to overall low utilization efficiency. At the same time, due to a lack of dynamic perception and rapid response capabilities regarding the airspace situation, existing technologies suffer from delays in conflict resolution and anomaly handling, making it difficult to meet the safety and efficiency requirements of future large-scale low-altitude commercial operations. In particular, the imbalance in the allocation of airspace resources directly restricts the safety and economy of low-altitude operations.

[0011] For example, during peak hours of urban logistics drone delivery, multiple drones simultaneously perform emergency medical supply delivery missions. At this time, a sudden thunderstorm causes some areas to be temporarily designated as no-fly zones. However, because the existing management technology uses static route planning, it is impossible to dynamically adjust the airspace division in real time to avoid areas with severe weather. This causes a large number of drones to gather in the limited safety area, the aircraft density increases sharply, and the probability of flight plan conflicts increases significantly, thus triggering potential collision risks. Furthermore, in this scenario, the static airspace division mechanism is forcibly applied to a dynamically changing environment, which makes it impossible to update airspace resource identification in real time. The aircraft path planning is restricted to a fixed grid, resulting in a continuous decrease in operational efficiency and a reduction in safety margin.

[0012] If the above problems are not addressed, the uneven spatial and temporal distribution of airspace resources may be further exacerbated, leading to some key areas remaining under high load while other areas remain idle for extended periods, resulting in overall resource waste. Simultaneously, the normalization of delayed conflict resolution will accumulate safety risks, potentially triggering flight conflict incidents in high-density operating environments and threatening low-altitude operational safety. Furthermore, the lack of dynamic adjustment capabilities will hinder the adaptability of low-altitude airspace management, failing to support the future large-scale and diversified needs of low-altitude aircraft applications and limiting the sustainable development of the low-altitude economy. In particular, the inability to respond promptly to changes in the airspace situation will lead to delayed management decisions.

[0013] Based on this, in one embodiment, such as Figure 1 As shown, this application discloses an AI-based method for dynamic partitioning and management of low-altitude airspace grids, specifically including the following steps: S10: Real-time acquisition of multi-source datasets corresponding to the target low-altitude airspace, and generation of airspace situation profiles based on the multi-source datasets. The multi-source datasets include aircraft status data, flight plan data, airspace environment data, and infrastructure status data. In this embodiment, low-altitude airspace refers to the airspace above the ground up to a certain altitude, typically used for the operation of low-altitude aircraft such as UAVs and electric vertical takeoff and landing (eVTOL) aircraft. The target low-altitude airspace is the low-altitude airspace range corresponding to the method in this embodiment. Multi-source datasets refer to data sets containing multiple sources and types, including aircraft status data, flight plan data, airspace environment data, and infrastructure status data. Aircraft status data refers to the real-time dynamic parameters of an aircraft operating in the airspace, typically including the aircraft's real-time three-dimensional position, speed, heading, attitude, identification code, and other airborne sensor readings. Flight plan data refers to the flight plan submitted by the aircraft to the airspace-related management system before flight, typically including predetermined takeoff and landing points, planned routes, and planned... Flight time, flight mission type, and aircraft performance parameters, among which flight mission types include logistics delivery, aerial photography and mapping, and police patrol; airspace environment data refers to external natural environment information that affects flight safety and feasibility, usually including real-time meteorological data, geographic information data, and related ecological information, such as bird migration corridors and wildlife habitats; infrastructure status data refers to the status information of ground or airborne fixed facilities that support low-altitude flight operations, usually including take-off and landing field status, signal coverage and quality of communication base stations, availability of navigation facilities, and working status of monitoring equipment; airspace situation profile refers to a digital model or feature representation that comprehensively describes the current and future operational status, environmental conditions, and potential risks of low-altitude airspace based on multi-source datasets.

[0014] Specifically, multi-source datasets corresponding to the target's low-altitude airspace are collected in real time, and an airspace situation profile is generated based on these datasets. The multi-source datasets include aircraft status data, flight plan data, airspace environmental data, and infrastructure status data. One implementation method is to configure multiple independent sensors and data interfaces to obtain information from different data sources. For example, the aircraft position can be obtained through radar, environmental data can be obtained through weather stations, and flight plans can be obtained through a manual data entry system. Collecting and overlaying the above data can serve as the initial construction of the airspace situation profile.

[0015] S20: Perform a demand analysis on the airspace situation profile to obtain operational demand information, and calculate and generate grid partitioning parameters based on the operational demand information. The grid partitioning parameters include grid granularity, grid level, and grid shape. In this embodiment, demand analysis refers to the process of parsing and calculating the airspace situation profile to extract operational demand information for airspace grid division. Operational demand information refers to information reflecting the specific demand for airspace resources by flight activities in low-altitude airspace, such as the number of aircraft, aircraft density, mission type, and distribution of take-off and landing points. Grid division parameters refer to parameters used to guide the generation of airspace grids, including grid granularity, grid level, and grid shape. Grid granularity refers to the size of the grid cell, grid level refers to the hierarchical structure of the grid in the vertical direction, and grid shape refers to the geometric shape of the grid cell.

[0016] Specifically, a demand analysis is performed on the airspace situation profile to obtain operational demand information. Based on this information, grid partitioning parameters are calculated and generated. These parameters include grid granularity, grid level, and grid shape. One approach is to pre-set a fixed set of rules. For example, if the number of aircraft in the airspace exceeds a certain fixed value, the grid granularity is set to a smaller value; otherwise, it is set to a larger value. The grid level and shape can also be preset using similar rules. Another approach is to compare the total number of aircraft displayed in the airspace situation profile with a preset airspace capacity. If the total number of aircraft is close to or exceeds the capacity, the grid granularity is set to a smaller value; otherwise, it is set to a larger value. The grid level can be evenly divided according to the overall altitude range of the airspace, and the grid shape can uniformly adopt a standard rectangle or hexagon.

[0017] S30: Generate a dynamic airspace grid that matches the operational requirements information through a preset grid partitioning algorithm, and assign airspace resource identifiers to the dynamic airspace grid; In this embodiment, the preset grid partitioning algorithm refers to a preset calculation method or calculation model used to cut the three-dimensional airspace into regular or irregular grid units; the dynamic airspace grid refers to an airspace partitioning structure that can be adjusted and optimized according to the real-time airspace situation and operational requirements; the airspace resource identifier refers to a unique identification code assigned to each dynamic airspace grid unit, used to characterize the spatial location, availability, capacity and other attributes of the dynamic airspace grid unit.

[0018] Specifically, a dynamic airspace grid matching the operational requirements is generated using a preset grid partitioning algorithm, and airspace resource identifiers are assigned to the dynamic airspace grid. One implementation is to use a basic geometric partitioning algorithm, for example, uniformly dividing the entire low-altitude airspace into several cubic grids of the same size, where the granularity, hierarchy, and shape of these cubic grids are directly determined by the grid partitioning parameters calculated above. Subsequently, airspace resource identifiers are assigned to each generated cubic grid cell sequentially. Another implementation is to select a preset grid template based on the operational requirements information, for example, using a smaller granularity grid in areas with high demand and a larger granularity grid in areas with low demand. In non-complex airspace scenarios, the selection of the grid template can be based on simple matching rules without involving complex calculations, and the airspace resource identifier can also only contain the coordinate information of the grid cells without including other detailed availability or capacity attributes.

[0019] S40: Acquire the target aircraft's aircraft information and perform collaborative path planning based on the dynamic airspace grid and aircraft information to generate planned path information; In this embodiment, aircraft information refers to the relevant data of the target aircraft, including its registration information, performance parameters, current position, flight intention, etc.; collaborative path planning refers to the process of generating a flight path for the target aircraft that ensures flight safety and improves operational efficiency by combining the aircraft information of multiple aircraft, airspace grid status, airspace rules, and other factors; planned path information refers to the output of the collaborative path planning process, which includes detailed flight path data such as the specific trajectory, altitude, and time of the aircraft from the starting point to the destination.

[0020] Specifically, the process involves acquiring the target aircraft's information and performing collaborative path planning based on the dynamic airspace grid and the aircraft information to generate planned path information. One approach is to receive the flight plan submitted by the target aircraft and use it as aircraft information. Then, within the generated dynamic airspace grid, a shortest path algorithm, such as Dijkstra's algorithm, is used to find a path from the starting point to the destination. This path planning process primarily considers the connectivity of grid cells. Another approach, in non-complex airspace scenarios, involves acquiring the target aircraft's current position and destination information, planning a straight path for the target aircraft within the dynamic airspace grid, and checking whether the path passes through any grid cells marked as impassable. If so, an attempt is made to bypass these impassable grid cells. In this case, the planned path information only contains a series of waypoint coordinates.

[0021] S50: Send the planned path information to the management terminal associated with the corresponding target aircraft and monitor the operational status of the target aircraft; In this embodiment, the management terminal refers to the system or device terminal used to receive, process, and issue instructions, and monitor the operational status of the aircraft. The management terminal typically interacts with the corresponding aircraft for data exchange. Operational status monitoring refers to the real-time tracking and monitoring of the target aircraft's position, speed, altitude, attitude, system health status, etc. during flight.

[0022] Specifically, the planned path information is sent to the management terminal associated with the corresponding target aircraft to guide the generation of flight control commands or heading adjustments. One implementation method is that the planned path information can be broadcast to all management terminals, which can then determine whether it is related to their own associated aircraft. The management terminal periodically receives position reports sent by the aircraft to track its position and monitor its operational status. Another implementation method is that the planned path information is sent to a specific management terminal through point-to-point communication. The management terminal then observes the aircraft's flight trajectory through manual inspection or radar display to determine whether it is flying according to the planned path.

[0023] S60: Based on the monitoring results of the operation status, evaluate the change in the airspace situation of the target low-altitude airspace. If the change in the airspace situation exceeds the preset change threshold, trigger the preset dynamic adjustment mechanism to adjust the dynamic airspace grid.

[0024] In this embodiment, the change in airspace situation refers to an indicator that quantifies the difference between the current situation of low-altitude airspace and the previous moment or the baseline situation, and is used to reflect the degree of change in factors such as airspace environment and flight activities; the preset change threshold refers to a pre-set critical value used to determine whether the change in airspace situation has reached the threshold that needs to be triggered for adjustment; the preset dynamic adjustment mechanism refers to a pre-set strategy and process for re-dividing or optimizing the dynamic airspace grid when the change in airspace situation exceeds the preset threshold.

[0025] Specifically, the change in airspace situation in the target low-altitude airspace is assessed based on the operational status monitoring results. If the change in airspace situation exceeds a preset change threshold, a preset dynamic adjustment mechanism is triggered to adjust the dynamic airspace grid. One implementation method is to count the number of aircraft that deviate from the planned path during the monitoring period. If this number exceeds a fixed change threshold, the airspace situation is considered to have changed, and a preset grid adjustment scheme is triggered, such as uniformly reducing the granularity of all grids. Another implementation method is to periodically compare the current aircraft density with the historical average density. If the difference exceeds a fixed change percentage threshold, the dynamic adjustment mechanism is triggered. This dynamic adjustment mechanism can be limited to merging or splitting grids in local areas.

[0026] Furthermore, in a preferred embodiment of the present invention, the change in airspace situation can be calculated based on the aircraft density increment, collision frequency, and meteorological environment abrupt change index within the dynamic airspace grid. That is, operational status monitoring may include the aircraft density increment and collision frequency within the dynamic airspace grid, combined with the meteorological environment abrupt change index. The specific calculation method is as follows: ,in, The density of aircraft in the target area at the current moment. This is the maximum allowable capacity baseline value for the target area; For time window The number of potential new route conflicts (such as the number of conflicts predicted based on heading angle intersections). This is a penalty coefficient for weather warnings. If severe weather such as thunderstorms and strong winds occurs, [the penalty will be applied]. Take the highest-order value; otherwise, return 0. , , These are preset normalized weighting coefficients, whose sum is 1. If the calculated comprehensive change... If the change exceeds the preset threshold, a preset dynamic adjustment mechanism will be triggered. The preset normalized weight coefficient can be dynamically adjusted and set by technical personnel in the corresponding technical field based on the actual airspace congestion level and historical experience, which will not be elaborated here.

[0027] For example, suppose that in the low-altitude airspace of a certain urban area, a large number of drones need to be managed for logistics delivery and inspection tasks, and the airspace situation in the target area is changeable, with both fixed no-fly zones and dynamic restrictions caused by temporary activities or unsuitable weather conditions.

[0028] First, multi-source datasets corresponding to the target's low-altitude airspace are collected in real time. For example, real-time data on the location, speed, and altitude of logistics drones can be continuously received, along with flight plan data such as delivery and inspection plans pre-submitted by the management terminals of these logistics drones. Simultaneously, real-time meteorological and terrain data of the area are acquired as airspace environmental data, and operational status data of infrastructure such as communication base stations and charging piles are collected. After aggregating these multi-source data, an airspace situation profile is generated. This airspace situation profile can intuitively display the distribution of aircraft, aircraft density, potential conflict points, and the real-time status of the environment and infrastructure within the current airspace.

[0029] Furthermore, a demand analysis is conducted on the airspace situation profile. For example, by analyzing the airspace situation profile, it is identified that during the morning peak hours, the density of logistics drones in specific areas increases significantly. At the same time, a certain sub-area is designated as a temporary no-fly zone due to temporary activities. This yields operational demand information, such as areas with higher drone density requiring more detailed control and temporary no-fly zones requiring avoidance. Based on the operational demand information, grid partitioning parameters are calculated and generated, including grid granularity, grid level, and grid shape. For example, in areas with higher drone density, the grid granularity is calculated to be smaller, while above the temporary no-fly zone, the grid level is set to be impassable. At the edge of the no-fly zone, the grid shape is adjusted to an irregular shape to conform to the boundary.

[0030] Furthermore, by using a preset grid partitioning algorithm and combining the grid partitioning parameters calculated above, a dynamic airspace grid matching the current operational requirements is generated. For example, an adaptive grid algorithm can be used to generate smaller-grained grid cells in areas with high aircraft density and larger-grained grid cells in areas with low aircraft density, and irregularly shaped grids can be generated around no-fly zones. At the same time, a unique airspace resource identifier is assigned to each generated grid cell. This airspace resource identifier not only contains the spatial location information of the grid, but also associates with the current availability status of the grid cell, the maximum number of aircraft it can accommodate, and the expected effective duration.

[0031] Furthermore, when a new target aircraft requests to enter the airspace, such as a drone that needs to perform an emergency delivery mission, the system acquires the aircraft's information, including its take-off and landing points, mission type, and performance parameters. Based on the currently generated dynamic airspace grid and the aircraft information, it performs collaborative path planning. For example, considering the drone's destination and mission priority, it searches the dynamic airspace grid for a safe flight path that avoids no-fly zones and other aircraft while meeting the capacity limits indicated by the airspace resource identifier. Finally, it generates the drone's planned path information.

[0032] Furthermore, the planned path information is sent to the management terminal associated with the corresponding target aircraft. After receiving the planned path information, the management terminal guides the UAV to fly along the planned path. At the same time, it continuously monitors the operational status of the target aircraft, tracks its position, speed, altitude and other parameters in real time, and monitors whether it deviates from the planned path.

[0033] Furthermore, during flight, based on the operational status monitoring results, the changes in the airspace situation of the target low-altitude airspace are continuously assessed. For example, if monitoring detects a sudden and significant increase in the density of aircraft in a certain area, or if new sudden weather conditions cause some airspace to become unsuitable for flight, these sudden situations will lead to an increase in the changes in the airspace situation. If the changes in the airspace situation exceed the preset change threshold, a preset dynamic adjustment mechanism is triggered. For example, the grid division parameters of the area are immediately recalculated, and the grid division algorithm is called to make local or global adjustments to the dynamic airspace grid to adapt to the new airspace situation, ensuring flight safety and improving operational efficiency in airspace operations.

[0034] Based on the above examples, the technical solution of this application can address the shortcomings of traditional low-altitude airspace management methods. Specifically, traditional low-altitude airspace management methods often rely on pre-defined fixed routes or statically isolated airspaces. For example, in the urban area mentioned above, only a few fixed drone routes and permanent no-fly zones may be designated. This fixed and unchanging model is difficult to adapt dynamically when faced with dynamic changes such as a surge in logistics drone traffic during peak periods, temporary no-fly zones, or sudden unsuitable weather conditions. For instance, when the density of aircraft in a certain area is too high, the static grid cannot be refined in time to provide more detailed control, which can easily lead to local congestion. When a temporary no-fly zone appears, the static routes may not be adjusted in time, forcing aircraft to make large-scale detours and reducing operational efficiency.

[0035] This application achieves dynamic perception of low-altitude airspace operation by introducing real-time acquisition of multi-source datasets and generation of airspace situation profiles, in contrast to existing technologies that rely on single or lagging data sources. Furthermore, by performing demand analysis on the airspace situation profiles and calculating and generating dynamic grid partitioning parameters, this application can adaptively adjust the granularity, hierarchy, and shape of the grid according to actual operational needs. This overcomes the problem of existing static partitioning schemes being fixed and unchanging when dealing with high aircraft density and highly dynamic low-altitude operations, making the allocation of airspace resources more accurate and reasonable.

[0036] Furthermore, this application generates a dynamic airspace grid that matches the operational requirements information through a preset grid division algorithm, and assigns an airspace resource identifier to each grid cell. This dynamic airspace grid structure can be adjusted in real time according to changes in the airspace situation. For example, in the example, the grid in areas with high aircraft density can be automatically refined, and the grid at the edge of a temporary no-fly zone can adaptively adjust its shape. Compared with the fixed grids or routes in the prior art, this can significantly improve the spatiotemporal utilization efficiency of airspace resources and effectively reduce the occurrence of inconsistent regional utilization.

[0037] Furthermore, collaborative path planning based on a dynamic airspace grid, taking into account various dynamic constraints such as aircraft information, grid availability, and capacity, can generate planned paths that ensure flight safety and improve operational efficiency. Compared with the simple path planning based solely on fixed routes in existing technologies, this can significantly reduce the probability of flight conflicts and improve flight safety. Finally, through continuous operational status monitoring and a dynamic adjustment mechanism based on changes in airspace situation, this application can quickly respond to sudden changes in airspace situation and adjust the airspace grid in a timely manner to ensure the stability of airspace management. For example, when monitoring detects abnormal congestion or sudden weather changes in a certain area, grid adjustment is automatically triggered to quickly alleviate pressure or guide aircraft to avoid dangerous areas, effectively solving the delay problem in conflict resolution and anomaly handling in existing technologies.

[0038] In summary, the technical solution of this application, by constructing a management model from real-time perception, demand analysis, dynamic partitioning, collaborative planning to continuous monitoring and adaptive adjustment, can effectively solve the shortcomings of existing low-altitude airspace management methods in terms of dynamism, efficiency, and safety. It can improve the flexibility and real-time performance of low-altitude airspace management, optimize resource allocation, and reduce flight conflicts, thereby enhancing operational safety and efficiency.

[0039] In this regard, this application further proposes that, in one embodiment, as follows: Figure 2 As shown, step S10 includes: S11: Identify all aircraft within the target low-altitude airspace, set all aircraft associated with the target low-altitude airspace as the first aircraft set, and set all remaining aircraft as the second aircraft set; In this embodiment, step S11 aims to clarify the scope and priority of data collection. Specifically, by distinguishing between aircraft associated with and entering the target low-altitude airspace, different data collection strategies can be adopted for different aircraft, thereby improving the efficiency and targeting of data collection. The aircraft identification process can be carried out by using a geographic information system in conjunction with the aircraft's real-time location data to determine whether the aircraft is within the geographical boundary of the target low-altitude airspace; or, it can be carried out by the corresponding airspace management system, based on the aircraft's registration information, flight plan reporting information, and other information, combined with airspace division rules, to automatically identify and classify the aircraft.

[0040] S12: Collect flight plan data and real-time status data of the first set of aircraft, and real-time status data of the second set of aircraft, and obtain airspace environment data and infrastructure status data of the target low-altitude airspace as a multi-source dataset. In this embodiment, step S12 aims to adopt differentiated data acquisition strategies for different aircraft sets and supplement airspace environment and infrastructure data to construct a comprehensive multi-source dataset. The first aircraft set is the focus and requires more detailed flight plan data. Furthermore, flight plan data can be obtained from flight service stations, air traffic control systems, or drone operator platforms. Real-time status data can be obtained through ADS-B receivers, radar systems, satellite communications, or drone telemetry links. Airspace environment data can be obtained through weather radar, weather stations, environmental monitoring sensors, or third-party weather service platforms. Infrastructure status data can be obtained through airport management systems, communication and navigation equipment monitoring systems, or IoT sensor networks. Alternatively, aggregated data can be obtained by interfaceing with third-party data providers, such as obtaining flight plans from air traffic control systems, meteorological data from meteorological bureaus, and infrastructure operation data from operators via API interfaces.

[0041] S13: Input the multi-source dataset into the pre-built feature mapping model, and construct a spatial situation profile based on the output of the feature mapping model.

[0042] In this embodiment, step S13 aims to transform heterogeneous multi-source datasets into structured standard feature elements to facilitate the construction and analysis of airspace situation profiles. The feature mapping model is used for data preprocessing and feature engineering extraction. Furthermore, the feature mapping model can be a rule-based expert system that predefines mapping rules from various data types to feature elements. For example, it maps aircraft speed and altitude to flight intensity in dynamic activity elements; it maps meteorological data to weather conditions in ecological constraint elements. Alternatively, the feature mapping model can be a machine learning model, such as a neural network or decision tree, which learns through training to extract key features related to airspace situation from raw data. For example, it learns the relationship between aircraft status and airspace congestion through historical data and maps it to airspace capacity in functional ownership elements.

[0043] Specifically, the proposed solution identifies and classifies aircraft within the target low-altitude airspace, clarifying which aircraft are the primary focus of current airspace management (the first aircraft set) and which are background information (the second aircraft set). Subsequently, data is collected separately for each aircraft set: more comprehensive flight plan and real-time status data are collected for the first aircraft set, while only real-time status data is collected for the second aircraft set. This avoids unnecessary detailed data collection on non-core aircraft, thereby improving the efficiency and relevance of data collection. Simultaneously, combining airspace environmental data and infrastructure status data ensures the comprehensiveness of the multi-source dataset. Finally, the classified and differentiated multi-source dataset is input into a pre-built feature mapping model, which transforms the heterogeneous raw data into standardized feature elements, effectively filtering noise and extracting key information.

[0044] Furthermore, the feature mapping model can specifically be a classification and feature extraction network based on a spatiotemporal graph convolutional network or a multilayer perceptron; for example, the input of the feature mapping model is a multidimensional feature matrix. ,in, The three-dimensional coordinates of the target It is a velocity vector. For timestamps, The target attribute identifiers fed back by the sensors, such as ADS-B and radar echo features, are used. The feature mapping model can extract the spatiotemporal latent variables of the multidimensional feature matrix through forward propagation calculation, and map the fused data to the preset feature set through the Softmax classification layer. The feature set includes static geographic obstacle elements (such as buildings), dynamic air traffic elements (such as drone formations), and temporary restricted area elements.

[0045] It should be noted that those skilled in the art can reproduce the element mapping model based on the examples and content described in this application specification. One specific architecture has been shown in the above example. The specific architecture and parameter application can be adjusted based on the actual scenario requirements, which will not be elaborated here.

[0046] Through the above technical solution, this embodiment classifies aircraft within the target low-altitude airspace and adopts differentiated data collection strategies, thereby improving the efficiency and relevance of data collection and avoiding unnecessary data redundancy. By inputting multi-source datasets into a pre-built element mapping model, the standardization of heterogeneous data and extraction of key information are achieved, ensuring the accuracy and real-time nature of the airspace situation profile. Through this data processing method, the generated airspace situation profile can more accurately reflect the actual operational status, potential risks, and resource distribution of the target low-altitude airspace, thus improving the refinement and security of low-altitude airspace management.

[0047] In this regard, this application further proposes that, in one embodiment, as follows: Figure 3 As shown, step S13 includes: S131: Map a multi-source dataset to a predefined set of feature elements through an element mapping model. The set of feature elements includes functional ownership elements, ecological constraint elements, public space elements, and dynamic activity elements. In this embodiment, the feature mapping model maps multi-source datasets to a predefined set of feature elements, aiming to transform the original multi-source datasets into standardized feature elements to facilitate rule generation and situational profile construction. The feature mapping model can be a machine learning-based classifier, such as a support vector machine or neural network, which learns the complex relationships between multi-source data and feature elements through training to achieve automatic data mapping. Alternatively, the feature mapping model can be implemented using an expert knowledge-based rule engine, which presets multiple mapping rules and directly associates specific types of data with corresponding feature elements. The feature element set is a feature representation that comprehensively describes the low-altitude airspace situation, including functional ownership elements, ecological constraint elements, public space elements, and dynamic activity elements. Functional ownership elements refer to the airspace's use, ownership, or management. Permissions can be granted by querying a pre-defined airspace management database to obtain the geographical boundaries, altitude restrictions, and corresponding management agency information of the airspace. Ecological constraints refer to areas with ecological protection requirements for flight activities. This can be achieved by overlaying an ecological protection zone layer on a geographic information system to identify the protected geographical area and corresponding flight restrictions. Public space refers to airspace available for public or specific non-commercial activities. This can be achieved by excluding areas with functional ownership and ecological constraints, defining the remaining open airspace as public space, and determining its usage conditions in conjunction with local regulations. Dynamic activity refers to activities that occur in the airspace in real time and have time-sensitive characteristics. This can be achieved by receiving real-time updates to flight plans, temporary airspace control notices, weather warnings, and other information to dynamically identify and update the spatial and temporal attributes of relevant activities.

[0048] S132: Generate an airspace management rule set based on a set of feature elements. The airspace management rule set includes a subset of static no-fly rules generated based on functional ownership elements and public space elements, a subset of conditional open rules generated based on functional ownership elements and preset functional area management standards, a subset of ecologically sensitive area rules generated based on ecological constraint elements, and a subset of activity constraint rules generated based on dynamic activity elements. In this embodiment, an airspace management rule set is generated based on a set of feature elements. This aims to transform abstract feature elements into specific, executable airspace management instructions. For example, the static no-fly rule subset is generated based on functional ownership elements and public space elements. It is generated by querying areas marked as "no-fly" in functional ownership elements and combining them with areas explicitly prohibited from flying in public space elements to form a fixed list of no-fly areas and their corresponding spatial range and altitude restrictions. The conditional opening rule subset is generated based on functional ownership elements and preset functional area management standards. It is generated by analyzing areas marked as "conditionally open" in functional ownership elements and combining them with preset functional area management standards to form a set of rules containing opening conditions and restrictions. The ecologically sensitive area rule subset is generated based on ecological constraint elements. It is generated by identifying sensitive areas defined in ecological constraint elements and generating flight restriction rules for these areas according to relevant ecological protection regulations. The activity constraint rule subset is generated based on dynamic activity elements. It is generated by monitoring changes in dynamic activity elements in real time and generating temporary flight restriction or guidance rules based on information such as activity type, duration, and scope of impact.

[0049] S133: Construct an airspace situation profile based on the airspace management rule set.

[0050] In this embodiment, an airspace situation profile is constructed based on the airspace management rule set, aiming to structurally generate a comprehensive airspace situation representation from the generated airspace management rule set. The construction of the airspace situation profile can adopt a multi-layer geographic information system overlay method, displaying different types of rule subsets as different layers, and each layer includes its corresponding spatial range, height limit, time attribute, and specific rule content.

[0051] Specifically, this application uses an element mapping model to map multi-source datasets to a predefined set of feature elements, thereby characterizing the static attributes and dynamic changes of airspace from different dimensions. Based on these refined feature elements, an airspace management rule set is generated, encompassing subsets of static no-fly rules, conditional open rules, ecologically sensitive area rules, and activity constraint rules. In this way, the constructed airspace situation profile is no longer merely a data aggregation, but a structured management rule system. This system can transform raw data into information with clear management significance, such as identifying which areas are absolutely no-fly zones, which areas can be opened under specific conditions, which areas require special attention for ecological protection, and which areas are affected by temporary activities. Finally, an airspace situation profile is constructed based on the airspace management rule set, providing a structured airspace view with decision support capabilities, thus laying the foundation for subsequent operational requirements analysis, grid division parameter calculation, and the generation of dynamic airspace grids.

[0052] Through the above technical solutions, this application can transform raw multi-source datasets into structured airspace management rules, thereby constructing a more refined and comprehensive airspace situation profile. This profile not only reflects the physical state of the airspace but also incorporates management logic and constraints such as the airspace's functional attributes, ecological sensitivity, public availability, and real-time dynamic activities. This allows subsequent airspace operation demand analysis to be based on more accurate information, improves the accuracy of grid division parameter calculations, and ultimately ensures that the generation of dynamic airspace grids better matches actual operational needs. Furthermore, it can effectively avoid problems such as unreasonable airspace resource allocation and increased flight conflict risks caused by insufficient information or ambiguous rules, thereby further improving the intelligence and safety level of low-altitude airspace management.

[0053] In this regard, this application further proposes that, in one embodiment, step S20 includes: S21: Analyze the airspace situation profile and extract the operational demand information of the target low-altitude airspace. The operational demand information includes aircraft density distribution, flight plan conflict probability, static no-fly zone information, dynamic no-fly zone information, and preset airspace capacity threshold. In this embodiment, step S21 aims to identify and quantify the key factors affecting airspace grid division from the airspace situation profile, providing a data foundation for subsequent grid parameter calculations. This involves extracting multi-dimensional operational demand information to reflect the current airspace operational status and potential risks. The extraction process can utilize data mining and pattern recognition technologies to analyze various raw data contained in the airspace situation profile, identifying and extracting quantitative indicators such as aircraft density distribution and flight plan conflict probability. Static and dynamic no-fly zone information can be obtained from a pre-defined geographic information system database or real-time updated airspace restriction notices. Pre-defined airspace capacity thresholds can be configured based on historical data, airspace type, and management strategies. Alternatively, a rule-based expert system can be used to predefine various parsing rules and logic. When the airspace situation profile is input, these predefined rules are used to match and extract the corresponding operational demand information.

[0054] S22: Determine the grid granularity based on the aircraft density distribution and airspace capacity threshold. Specifically, this includes: calculating the ratio of the aircraft density distribution to the airspace capacity threshold. When the ratio is higher than a preset first threshold, a corresponding fine-grained grid is used. When the ratio is lower than a preset second threshold, a corresponding coarse-grained grid is used. In this embodiment, step S22 aims to dynamically adjust the grid granularity based on the actual busyness and carrying capacity of the airspace, so as to achieve reasonable allocation and efficient utilization of airspace resources. In busy areas, a fine-grained grid can be used to provide more precise control, while in idle areas, a coarse-grained grid can be used to reduce system overhead. Furthermore, the process of determining the grid granularity can be implemented through a decision engine or lookup table. Specifically, the ratio of the current aircraft density distribution in the airspace to a preset airspace capacity threshold is calculated, and then this ratio is compared with a preset first threshold and a second threshold to select the corresponding grid granularity. Alternatively, fuzzy logic control or a machine learning model can be used, inputting the ratio of the aircraft density distribution to the airspace capacity threshold, outputting continuous suggested grid granularity values, and then discretizing them into several predefined grid granularities.

[0055] S23: Determine the grid level based on static no-fly zone information and dynamic no-fly zone information, specifically including: identifying the static altitude attribute associated with static no-fly zone information and the dynamic altitude attribute associated with dynamic no-fly zone information, and setting the grid level through preset hierarchical control rules, wherein the hierarchical control rules define the main aircraft types and control intensity corresponding to different altitude ranges; In this embodiment, step S23 aims to vertically layer the grid according to the vertical structure and constraints of the airspace to adapt to airspace management for different types of aircraft and different control requirements. By considering both static and dynamic no-fly altitudes, the grid division can be ensured to meet airspace safety requirements. Furthermore, the identification process can be executed through a rule engine. Specifically, it obtains the fixed vertical range (static altitude attribute) from static no-fly zone information and the temporary vertical range (dynamic altitude attribute) from dynamic no-fly zone information. Then, it matches the identified altitude attributes with preset layered control rules to adjust or create corresponding grid levels. Alternatively, it can be achieved through the spatial analysis function of a geographic information system (GIS). Specifically, it imports the three-dimensional spatial data of static and dynamic no-fly zones into the GIS and overlays them with preset layered control rule layers to automatically identify no-fly attributes within different altitude ranges and generate grid division schemes with different layered attributes accordingly.

[0056] S24: Determine the grid shape based on the flight plan conflict probability, static no-fly zone information, and dynamic no-fly zone information. Specifically, this includes: identifying the first spatial contour of static no-fly zone information and the second spatial contour of dynamic no-fly zone information; identifying conflict hotspots and non-conflict hotspots in the target low-altitude airspace based on the flight plan conflict probability; isolating conflict hotspots using irregular grid shapes based on the first and second spatial contours; and using standard grid shapes in non-conflict areas.

[0057] In this embodiment, step S24 aims to dynamically adjust the geometry of the grid based on the horizontal structure of the airspace and potential conflict risks to optimize airspace utilization efficiency and flight safety. Specifically, using irregular grids for isolation in conflict hotspot areas can effectively avoid conflicts, while using standard grids in non-conflict areas facilitates management. Furthermore, the grid shape determination process can be achieved collaboratively using spatial geometry algorithms and conflict detection algorithms. Specifically, the two-dimensional plane boundaries, namely the first spatial contour and the second spatial contour, are extracted from static and dynamic no-fly zone information, and the flight plan conflict probability is analyzed. The system identifies potential conflict hotspots in the airspace. Further, within these hotspots, an irregular grid partitioning algorithm is used to generate grid shapes that conform to the no-fly zone outline and conflict zone boundaries to achieve isolation. In non-conflict areas, standard rectangular or hexagonal grids are used. Alternatively, image processing or machine learning methods can be employed. Specifically, the airspace situation profile is converted into a two-dimensional raster image, where different colors or grayscale values ​​represent flight plan conflict probabilities, no-fly zones, and other information. Image segmentation algorithms are then used to identify conflict hotspots and no-fly zone outlines, and the corresponding grid generation module is called for processing.

[0058] Furthermore, in step S24, for no-fly zones or conflict hotspots, to ensure that the mesh boundary does not intersect with the restricted area, a constraint-based Delaunay triangulation algorithm can be used to generate an irregular mesh. The specific execution steps include: S241: Obtain the geographic topological information of no-fly zones or conflict areas, extract the discrete vertex set of their polygon boundaries, and connect adjacent vertices to form a set of constraint edges; S242: Within the effective airspace outside the no-fly zone boundary, based on the local UAV traffic density, a set of internal seed points with uneven density is generated by Poisson disk sampling, where the higher the traffic, the denser the internal seed points. S243: Merge the boundary vertices with the internal seed points and perform Delaunay triangulation with the constraint edges as hard boundary conditions. During the Delaunay triangulation process, it is forcibly guaranteed that no generated triangle mesh edge intersects with the constraint edge. S244: Treat the generated irregular triangles or polygons that conform to the outline of the no-fly zone as independent grid cells, assign them airspace resource identification codes, and set the availability label of the grid cells that fall inside the no-fly zone to unavailable in order to align the physical topology with the airspace rules.

[0059] Specifically, the solution in this application analyzes the airspace situation profile to obtain multi-dimensional operational requirements information, and dynamically adjusts the grid granularity, grid level, and grid shape based on the operational requirements information, so that the generated dynamic airspace grid can more accurately match the actual operational status and management needs of the current airspace.

[0060] Through the above technical solution, this application can achieve refined determination of low-altitude airspace grid division parameters. Specifically, by comprehensively analyzing the airspace situation profile and extracting multi-dimensional operational demand information such as aircraft density distribution, flight plan conflict probability, static no-fly zone information, dynamic no-fly zone information, and preset airspace capacity thresholds, this application can adjust the grid granularity, grid level, and grid shape according to the actual airspace congestion and constraints. This allows the generated dynamic airspace grid to more accurately reflect the current airspace operation status and management needs, effectively avoiding the waste of airspace resources and potential safety hazards caused by traditional fixed grid division or simple rule division. Especially in high-density areas and high-conflict-risk areas, using fine-grained, irregularly shaped grids for isolation can improve airspace utilization efficiency, reduce the probability of flight conflicts, and ensure flight safety. At the same time, using coarse-grained, standard-shaped grids in low-density areas can effectively reduce computational and management load and improve overall operational efficiency.

[0061] In this regard, this application further proposes that, in one embodiment, step S30 includes: S31: Based on the grid division parameters, a preset corresponding grid division algorithm is matched, and the grid division parameters are used as the basic parameters. Combined with the airspace situation profile, the optimization target is set, and the matched grid division algorithm is called to perform iterative optimization calculation based on the target low-altitude airspace to generate a dynamic airspace grid. In this embodiment, the grid partitioning parameters refer to the parameters used to guide the generation of the airspace grid, which define the basic characteristics of the grid, such as granularity, hierarchy, and shape. The preset corresponding grid partitioning algorithm refers to a variety of pre-stored grid generation algorithms, such as those based on quadtrees, octrees, Voronoi diagrams, or adaptive grids. These algorithms can generate grids with different structures and characteristics based on different input parameters and optimization objectives. The matching process of the grid partitioning algorithm aims to select the most suitable algorithm based on the current grid partitioning parameters. The airspace situation profile refers to a comprehensive picture of the current state of the target low-altitude airspace formed after real-time acquisition and processing of multi-source datasets. The description includes information such as aircraft status, flight plan, airspace environment, and infrastructure status, providing basic data for setting optimization goals and conducting evaluations. The optimization goal is a criterion that guides the grid generation algorithm in iterative calculations. For example, it can be set to maximize airspace resource utilization, minimize the probability of flight conflicts, or ensure the avoidance rate of specific areas. Iterative optimization calculation refers to the process by which the grid generation algorithm, based on the initial grid structure, repeatedly adjusts the boundaries, size, or shape of grid cells and evaluates and corrects them according to the optimization goal until the preset convergence condition or number of iterations is reached, thereby generating a dynamic airspace grid that adapts to the current airspace situation.

[0062] S32: Based on the airspace situation profile, perform multi-dimensional state assessment on each grid cell in the generated dynamic airspace grid; In this embodiment, multidimensional state assessment refers to the process of checking the performance and suitability of each grid cell in the generated dynamic airspace grid. Further, multidimensional state assessment includes rule compliance assessment, environmental suitability assessment, and facility availability assessment. Rule compliance assessment aims to check whether the grid cell complies with all relevant airspace management rules, regulations, and restrictions, such as whether it is located in a no-fly zone, restricted flight zone, or temporary flight restriction area, or whether it meets the flight altitude and speed restrictions for a specific aircraft type. Environmental suitability assessment refers to assessing the suitability of the physical environment conditions of the grid cell, including weather conditions, topography, and electromagnetic environment, to determine whether the grid cell area is suitable for flight activities. Facility availability assessment refers to checking the availability of critical infrastructure within or near the grid cell, such as the coverage and performance of communication, navigation, and surveillance equipment, charging facilities, and landing sites, to ensure that aircraft can obtain necessary service support within the grid cell area.

[0063] S33: Assign an airspace resource identifier code to each grid cell based on the multidimensional state assessment results. The airspace resource identifier code includes the spatial location and hierarchical ID of the corresponding grid cell, and is associated with an availability status label, the maximum number of aircraft that can be accommodated, and the expected validity period of the status.

[0064] In this embodiment, the airspace resource identifier code is a unique identifier assigned to each grid cell. It includes not only the precise location information of the grid cell in three-dimensional space and its hierarchical ID, but also several key attributes, including availability status label, maximum number of aircraft that can be accommodated, and expected validity period. The availability status label clarifies whether the grid cell is currently available, restricted, or unavailable, and the specific reason for the restriction. The maximum number of aircraft that can be accommodated clarifies the maximum number of aircraft that the grid cell can safely accommodate in the current state, and is used for airspace capacity management. The expected validity period clarifies the expected time period during which the current availability status and the maximum number of aircraft that can be accommodated are expected to remain valid, providing a time reference for subsequent dynamic adjustments.

[0065] Specifically, the solution in this application generates a dynamic airspace grid that is highly adapted to the current airspace situation by matching the grid division parameters with a preset grid division algorithm and combining the airspace situation profile to set optimization targets for iterative optimization calculation. On this basis, a multi-dimensional state assessment is performed on each grid unit, taking into account its rule compliance, environmental suitability and facility availability. Finally, based on the assessment results, each grid unit is assigned an airspace resource identification code that includes spatial location, hierarchical ID, availability status label, maximum number of aircraft that can be accommodated, and expected validity period of the status.

[0066] For example, to more clearly demonstrate the mapping mechanism from physical spatial domain to digital grid, combined with Figure 4 Please provide a detailed explanation: Based on the grid parameter definitions extracted from situational analysis, namely grid granularity, grid level, and grid shape, a three-dimensional dynamic grid structure was constructed in digital space, including: Multi-layer vertical grid: The target airspace is divided into multiple grids such as H1, H2, and H3 in the vertical height to achieve three-dimensional isolation of flight routes at different flight altitudes; Dynamic grid granularity: To address the uneven distribution of airspace traffic, an adaptive granularity partitioning is adopted. In low-traffic areas far from obstacles, a large-size coarse-grained grid is generated to reduce computational power consumption; while in high-density areas with dense traffic or frequent potential collisions, it is subdivided into a fine-grained grid, thereby providing a more accurate micro-collision avoidance planning space. Irregular mesh: This is the core of the solution to ensure safety. For complex no-fly zone boundaries or terrain undulations, such as the concavity and convexity of building tops, spatial topology algorithms (such as the constraint-based Delaunay triangulation described in the above embodiment) are called to generate irregular meshes that fit the real physical contours of obstacles, such as the smooth concave and wrapping structure on the right side of H1-H3 layers in the figure, which forces the digital boundary to align with the physical restriction zone boundary.

[0067] Meanwhile, to support efficient graph search algorithms, each generated 3D grid cell, such as the labeled numbers ②, ③, and ④, encapsulates structured grid resource attributes, as shown in the information box led out by the line in the figure. Each grid cell includes: an airspace resource identifier (ID): such as ID: 004 or ID: 012, serving as a unique node index in the path search graph model; an availability tag (Avail): indicating the current airspace status of the grid (e.g., Avail: 1 means passable, Avail: 0 means currently restricted or within a no-fly zone); and a maximum capacity (Cap): a threshold value for the number of aircraft that a specific grid can safely accommodate in real time. For example, grid ID: 004 currently has a Cap of 5, indicating sufficient capacity; while grid ID: 012, although available (Avail: 1), currently has a Cap of 0, meaning that the instantaneous flow of this grid cell has reached its limit or the capacity has dropped to a very low level due to weather conditions. When the subsequent path planning algorithm reads this tag, it will automatically increase the cost of passing through this grid cell, thereby guiding aircraft to detour and achieving early congestion relief.

[0068] Through the above technical solution, this application can generate a highly optimized and information-rich dynamic airspace grid, which can effectively solve the shortcomings of traditional grid division methods in dealing with complex low-altitude airspace dynamic changes and multi-dimensional constraints. Through this detailed grid generation and resource identification and allocation mechanism, the airspace management process can more accurately perceive the real-time status and carrying capacity of each grid unit, thereby providing safer and more efficient airspace resources for aircraft, and improving the overall management level and operational efficiency of low-altitude airspace.

[0069] In this regard, this application further proposes that, in one embodiment, step S31 includes: S311: Based on the airspace situation profile, construct a multi-objective optimization function with the optimization objectives of maximizing airspace resource utilization, minimizing flight conflict probability, and dynamically constrained area avoidance rate. In this embodiment, step S311 aims to provide quantified optimization objectives and evaluation criteria for the generation of dynamic airspace grids. The airspace situation profile provides comprehensive information about the current airspace, including aircraft distribution, environment, and infrastructure. The multi-objective optimization function aims to simultaneously improve airspace utilization efficiency, reduce flight risks, and ensure effective avoidance of restricted areas. Furthermore, the multi-objective optimization function can combine the three optimization objectives into a single comprehensive objective function using a weighted summation method, where the weights can be dynamically adjusted according to actual operational needs and priorities. Alternatively, a Pareto optimization method can be used to generate a series of Pareto optimal solutions, and then a suitable grid partitioning scheme can be selected from the Pareto optimal solution set based on preset decision rules or expert experience.

[0070] S312: Generate an initial grid structure based on the target low-altitude airspace according to the grid division parameters; In this embodiment, an initial grid structure based on the target low-altitude airspace is generated according to the grid partitioning parameters. This initial grid structure serves as the starting point for dynamic airspace grid iterative optimization. The grid partitioning parameters provide the basic geometric and topological information for generating the initial grid, ensuring that the initial structure conforms to the basic requirements of the current airspace. For example, the target low-altitude airspace can be divided into uniform rectangular or hexagonal grid cells according to the grid granularity parameters. It can also be layered vertically according to the grid hierarchy parameters. Furthermore, it can be pre-divided into specific regions using irregular shapes according to the grid shape parameters. In addition, it can also be adaptively subdivided according to the grid granularity parameters based on the quadtree or octree spatial index structure to form initial grids of different granularities. At the same time, the specific regions are pre-processed by combining the grid hierarchy and shape parameters.

[0071] S313: The grid cell boundaries of the initial grid structure are adjusted by iteratively solving the multi-objective optimization function until the multi-objective optimization function converges or reaches the preset number of iterations, thus obtaining a dynamic spatial grid.

[0072] In this embodiment, step S313 is a key step in realizing dynamic mesh optimization. Through an iterative process, the boundaries of the mesh cells are continuously adjusted, enabling the entire mesh structure to better meet the requirements of the multi-objective optimization function, thereby generating an efficient, secure, and adaptable dynamic spatial mesh. Furthermore, convergence conditions or iteration count limitations ensure the effectiveness and computational efficiency of the optimization process. For example, heuristic optimization algorithms such as genetic algorithms or particle swarm optimization algorithms can be used. In each iteration, the multi-objective optimization function value of the current mesh structure is evaluated, and the boundaries of the mesh cells are fine-tuned or reconstructed according to the algorithm's search strategy until the termination condition is met. Alternatively, numerical optimization methods based on gradient descent or simulated annealing can be used, treating the adjustment of the mesh cell boundaries as optimization variables, and gradually approximating the optimal solution by calculating the gradient of the objective function with respect to the optimization variables or by utilizing random perturbations.

[0073] Specifically, the solution in this application constructs a multi-objective optimization function based on airspace situation profiles, which comprehensively considers maximizing airspace resource utilization, minimizing flight conflict probability, and dynamically constrained area avoidance rate, providing clear quantitative objectives for grid optimization. Specifically, an initial grid structure that meets basic requirements is generated using grid partitioning parameters as the starting point for optimization. On this basis, the grid cell boundaries of the initial grid structure are continuously adjusted by iteratively solving the multi-objective optimization function. This iterative process allows the grid structure to gradually evolve from its initial form until it reaches the optimal or suboptimal state, that is, the multi-objective optimization function converges or reaches the preset number of iterations.

[0074] Through the above technical solution, this application can generate a highly adaptable and dynamically responsive dynamic airspace grid, which significantly improves the operational efficiency and safety of low-altitude airspace, effectively reduces the risk of flight conflicts, and ensures effective avoidance of various dynamic constraint areas, thereby solving the problem that traditional grid generation methods are difficult to achieve optimal performance in complex low-altitude environments.

[0075] In this regard, this application further proposes that, in one embodiment, step S40 includes: S41: Analyze aircraft information, identify its associated flight plan data, and extract take-off and landing points, planned routes, and mission types; In this embodiment, parsing the aircraft information refers to performing structured processing and semantic understanding on the received raw data about the target aircraft. Further, parsing includes extracting key fields from the data stream, such as the aircraft's unique identifier, current position, speed, altitude, and intent. Flight plan data refers to the flight plan submitted by the aircraft before executing a mission, including detailed arrangements of the flight mission. Extracting takeoff and landing points, planned routes, and mission types refers to extracting basic information for path planning. The takeoff and landing points clearly define the start and end points of the flight, the planned route clarifies the initial flight intent, and the mission type determines the specific constraints and optimization objectives to be considered during path planning. Furthermore, mission types include, for example, manned transportation, logistics delivery, and inspection operations.

[0076] S42: Based on the take-off and landing points and the planned route, locate several consecutive grid cells in the dynamic airspace grid to form candidate airspace channels; In this embodiment, locating several consecutive grid cells refers to identifying and selecting a series of spatially adjacent and temporally continuous grid cells in the generated dynamic airspace grid based on the aircraft's take-off and landing points and planned flight paths. The identified grid cells together constitute the preliminary flight path area, i.e., the candidate airspace channel. In other words, the candidate airspace channel is one or more consecutive grid cell sequences delineated in the dynamic airspace grid based on the aircraft's initial intentions. It provides a search range for subsequent path search, thereby improving the efficiency and accuracy of path planning.

[0077] S43: Obtain flight safety distance and real-time airspace traffic, use candidate airspace channels as the search space, and generate one or more candidate paths based on flight safety distance, airspace resource identification code and real-time airspace traffic using a path search algorithm; In this embodiment, the safe flight distance refers to the minimum distance that must be maintained to ensure that there are no collisions between aircraft and between aircraft and obstacles. In path planning, the safe flight distance serves as a hard constraint, guiding the path search algorithm to avoid generating unsafe paths. Real-time airspace traffic refers to the number or density of aircraft passing through a certain airspace area within a specific time period. Obtaining real-time airspace traffic helps assess the congestion level of the airspace, thereby selecting grid cells with lower loads during path planning and avoiding local congestion. The airspace resource identifier is a unique identifier for each grid cell in the dynamic airspace grid, which includes information such as the spatial location, level ID, availability status label, maximum number of aircraft that can be accommodated, and expected effective duration of the status of the dynamic airspace grid cell. The path search algorithm can use the information of the airspace resource identifier to determine the availability, capacity limitations, and dynamic changes of the grid cell. The path search algorithm is used to find the optimal path from the starting point to the ending point in a graph structure or grid structure, such as A. Algorithms, Dijkstra's algorithm, etc.

[0078] S44: Perform spatiotemporal conflict detection on candidate paths, filter out a set of safe paths, and determine the planned path information based on the task type.

[0079] In this embodiment, spatiotemporal conflict detection refers to analyzing the generated candidate paths, specifically to predict whether the aircraft will conflict with other aircraft, static obstacles, or dynamic no-fly zones at the same time and in the same space during its flight along the path. It typically involves predicting the future trajectory of the aircraft and collision detection. The safe path set refers to the set of all candidate paths that are determined to be conflict-free and meet safety requirements after spatiotemporal conflict detection. Determining the planning path information based on the task type means selecting the path that best meets the task requirements from the safe path set according to the specific task type of the target aircraft, such as time-sensitive emergency tasks, cost-sensitive logistics tasks, and environmentally sensitive inspection tasks.

[0080] Specifically, to provide a safe and efficient flight path for a target aircraft in a dynamically changing low-altitude airspace environment, this application first analyzes the aircraft information of the target aircraft to obtain its flight intent, including takeoff and landing points, planned routes, and mission types. Then, using the flight intent information, several consecutive grid cells matching the aircraft's intent are located within the previously generated dynamic airspace grid, forming one or more preliminary candidate airspace channels to limit the scope of the path search and improve the efficiency of subsequent calculations. Based on this, to ensure the safety and feasibility of the path, current flight safety clearance and real-time airspace traffic data are obtained. Combined with the airspace resource identifier code contained in each grid cell of the dynamic airspace grid, a path search algorithm is used to search within the candidate airspace channels. The search process also comprehensively considers flight safety clearance. To avoid collisions, airspace resource identifiers are used to determine grid availability and capacity, and real-time airspace traffic is combined to avoid congested areas, thereby generating multiple candidate paths that meet basic safety and passage conditions. To further improve path reliability, spatiotemporal conflict detection is performed on the candidate paths, specifically predicting whether the aircraft will potentially collide with other aircraft, dynamic no-fly zones, or static obstacles in the airspace while flying along any candidate path. Through the above spatiotemporal conflict detection, all conflict-free paths can be filtered out to form a set of safe paths. Finally, from the set of safe paths, the path that best meets the specific needs of the target aircraft is selected as the final planned path information according to its specific mission type. For example, for emergency missions, the shortest time path may be prioritized; for logistics missions, the path with the lowest energy consumption may be prioritized.

[0081] Through the above technical solution, this application can realize dynamic planning of the flight path of the target aircraft. Specifically, by analyzing the aircraft information and flight plan data, and combining the real-time status of the dynamic airspace grid, a flight path that meets the mission requirements and avoids potential conflicts can be generated for the aircraft. Especially in the dynamically changing low-altitude airspace environment, the solution of this application can effectively utilize the grid status information provided by the airspace resource identification code, and combine the flight safety distance and real-time airspace traffic for path search, which has the effect of improving the safety, reliability and efficiency of path planning. Furthermore, through spatiotemporal conflict detection and mission type-based path selection, the conflict of the final planned path can be reduced, thereby improving the safety and efficiency of the aircraft in complex low-altitude airspace.

[0082] In this regard, this application further proposes that, in one embodiment, step S43 includes: S431: Set the minimum safe interval constraint for path search based on flight safety distance, extract the passage status, maximum capacity and constraint conditions of each grid cell based on airspace resource identification code, and calculate the current load of each grid cell based on real-time airspace traffic. The passage status includes passable and impassable. In this embodiment, the flight safety distance refers to the minimum distance that must be maintained between different aircraft or between an aircraft and an obstacle during flight to ensure flight safety. Setting a minimum safety distance constraint is to force all generated paths to meet basic safety requirements during path search, preventing collisions between aircraft or with no-fly zones. This minimum safety distance constraint can be dynamically adjusted based on factors such as aircraft type, speed, and airspace environment. For example, a larger safety distance can be set for high-speed aircraft or in complex terrain areas. Furthermore, this can be achieved by marking any path node or path segment that violates this minimum distance as unavailable in the path search algorithm, or by imposing a very high penalty value on paths that violate this constraint in the path evaluation function. The airspace resource identifier code is a unique identifier for each grid cell in the dynamic airspace grid, including key attribute information of that dynamic airspace grid cell. Extracting information from the airspace resource identifier code allows for a comprehensive understanding of the availability, carrying capacity, and limitations of each grid cell during path planning. The passage status can clearly identify the grid cell. Whether an airspace is allowed to pass is defined as follows: "Allowed to pass" indicates permission, and "Not allowed to pass" indicates prohibition (e.g., no-fly zone). Maximum capacity specifies the maximum number of aircraft a grid cell can accommodate at a given time, used to avoid local airspace congestion. Constraints may include altitude restrictions, specific aircraft type restrictions, time window restrictions, etc. Furthermore, the extraction of passage status, maximum capacity, and constraints can be achieved by querying a pre-stored airspace resource database or directly parsing the structure of airspace resource identifiers. Real-time airspace traffic refers to the number of aircraft passing through a specific airspace grid cell at a given time, or the expected number of aircraft passing through. Calculating the current load of each grid cell is to assess the real-time congestion level of each grid cell, thereby avoiding guiding too many aircraft to saturated or soon-to-be-saturated areas during path planning. Furthermore, the current load calculation can be based on real-time monitoring data to count the number of aircraft in the current grid cell, or combined with flight plan data to predict traffic over a future period. For example, the load rate can be obtained by comparing the number of aircraft in the current grid cell with the maximum capacity of that grid cell.

[0083] S432: Within the candidate airspace channel, taking the take-off and landing points as path endpoints, a preset graph search algorithm is used to traverse all grid cell sequences that satisfy the minimum safety interval constraint and whose grid cell status is passable, generating an initial set of feasible paths including several feasible paths. In this embodiment, candidate airspace channels refer to potential flight areas, including a series of continuous grid cells, initially determined based on the aircraft's take-off and landing points and planned routes; graph search algorithms refer to general methods for finding paths in graph structures, such as A... Algorithms, such as Dijkstra's algorithm, breadth-first search, and depth-first search, are used. The starting and ending points are used as path endpoints, meaning the search will start from the starting point and end at the ending point. The traversal process checks whether each potential grid cell sequence satisfies the minimum safe interval constraint and ensures that the passage state of each grid cell must be passable. Only grid cell sequences that satisfy both conditions are considered feasible paths and included in the initial feasible path set.

[0084] S433: Calculate the path length of each feasible path, the total load of the grid cells it contains, and the degree to which the constraint conditions are satisfied as the cost value, and sort them from low to high cost value, and select the N paths with the lowest cost value as candidate paths.

[0085] In this embodiment, cost value refers to a comprehensive indicator for evaluating the quality of a path; path length usually refers to the spatial distance or flight time of the path, serving as an important indicator of efficiency; the total load of the grid cells refers to the sum or average load of the current loads of all grid cells on the path, reflecting the congestion level of the path; the degree of constraint satisfaction quantifies the path's compliance with various airspace restrictions, for example, different penalty scores can be given according to the severity of constraint violations; the calculation of these cost value data usually adopts a weighted summation method to balance the importance of different factors; further, after calculating the cost value of all feasible paths, these feasible paths are sorted in ascending order of cost value, with lower cost value indicating a better path, and the N paths with the lowest cost value are selected as the final candidate paths from the sorted paths; where N is a preset integer representing the number of alternative paths to be provided, this screening process ensures that the final candidate paths are not only safe and feasible, but also perform well in terms of efficiency, congestion level, and constraint satisfaction.

[0086] Furthermore, a multi-objective cost function can be used to calculate the cost of each candidate path after passing through the [missing information]. The value of each grid cell The specific evaluation formula is as follows: ,in, This refers to the spatial distance cost, which is the physical three-dimensional Euclidean distance that the aircraft travels through the grid cells; This refers to the cost of congestion delays, which grows exponentially, as shown in the formula: ,in The current predicted capacity percentage of this grid cell is used to guide subsequent drones to detour. As the capacity approaches saturation, the cost increases dramatically. This refers to the conflict risk penalty. If the grid cell is adjacent to an irregular no-fly zone grid cell, a high boundary risk penalty value can be assigned; otherwise, it is 0. This refers to the penalty for meteorological impacts, specifically the additional cost determined based on the wind shear index or rainfall level at the height of the grid cell; weighting coefficients. , , , The altitude can be dynamically adjusted based on the target aircraft's mission priority. For example, for drones performing emergency medical rescue missions, priority is given to ensuring time, and the altitude will be adjusted accordingly. and reduce This allows it to traverse high-density areas; while ordinary logistics drones are required to be raised. and Guide them to choose a safe and open avoidance route.

[0087] It should be noted that those skilled in the art can fully set and reproduce the spatial distance cost, congestion delay cost, conflict risk penalty, weather impact penalty, and weighting coefficient that may be involved in the cost based on actual needs, according to the examples and content described above.

[0088] Specifically, in the aforementioned collaborative path planning process, to efficiently and safely generate planned path information within the dynamic airspace grid, this application further refines the candidate path generation mechanism. Specifically, firstly, by analyzing flight safety distances, a minimum safety interval constraint is set for path search, ensuring that all subsequently generated paths meet basic flight safety requirements and avoiding potential conflicts between aircraft or with no-fly zones. Simultaneously, using the airspace resource identifier code of each grid cell in the dynamic airspace grid, the passage status, maximum capacity, and other possible constraints of that grid cell are extracted, thereby determining the static and dynamic attributes of each airspace cell. Based on this, combined with real-time airspace traffic data, the current load of each grid cell can be calculated, intuitively reflecting its congestion level. Furthermore, after obtaining the aforementioned key information, the take-off and landing points of the target aircraft can be used as the start and end points of the path, within the pre-determined candidate airspace passage... Within the pathway, a pre-defined graph search algorithm is used to explore paths, strictly adhering to the previously set minimum safety interval constraint during the search process, and only considering grid cells with a passable status. In this way, all feasible grid cell sequences that meet the basic safety and passability conditions are traversed and identified, thus constructing an initial set of feasible paths. To select the optimal candidate paths from this set, each feasible path is further evaluated in multiple dimensions. Specifically, the path length, the total load of its contained grid cells, and the degree to which various constraints are met are calculated, and these evaluation results are set as a unified cost value. Finally, all feasible paths are sorted from low to high according to their cost value, and the N paths with the lowest cost value are selected as the final candidate paths. This ensures that the generated candidate paths not only meet safety requirements but also achieve certain optimizations in terms of efficiency, airspace resource utilization, and compliance.

[0089] Through the above technical solutions, this application comprehensively considers multiple dimensions such as flight safety, airspace resource utilization, and operational efficiency when generating candidate routes. By setting minimum safety interval constraints, the safety of the routes is guaranteed from the source. By combining airspace resource identification codes and real-time airspace traffic, the availability and congestion status of each grid unit can be dynamically perceived, thereby effectively avoiding congested areas and restricted airspace during route search. Furthermore, by adopting a multi-dimensional value assessment and ranking mechanism, the selected candidate routes not only meet the basic feasibility requirements but also achieve a balance between safety, efficiency, and airspace resource utilization. This has the effect of improving the intelligence and refinement of low-altitude airspace route planning and effectively reducing the risk of flight conflicts, thereby improving the utilization efficiency of airspace resources and providing more reliable and efficient route selection for the dynamic management of low-altitude airspace.

[0090] 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 this application.

[0091] In one embodiment, an AI-based dynamic low-altitude airspace grid partitioning and management system is provided, which corresponds one-to-one with the AI-based dynamic low-altitude airspace grid partitioning and management method described in the previous embodiment. The AI-based dynamic low-altitude airspace grid partitioning and management system includes: The image generation module is used to collect multi-source datasets corresponding to the target low-altitude airspace in real time, and generate an airspace situation image based on the multi-source datasets. The multi-source datasets include aircraft status data, flight plan data, airspace environment data, and infrastructure status data. The parameter generation module is used to perform demand analysis on the airspace situation profile, obtain operational demand information, and calculate and generate grid division parameters based on the operational demand information. The grid division parameters include grid granularity, grid level, and grid shape. The grid generation module is used to generate a dynamic airspace grid that matches the operational requirements information through a preset grid division algorithm, and to assign airspace resource identifiers to the dynamic airspace grid. The path generation module is used to acquire the target aircraft's aircraft information and perform collaborative path planning based on the dynamic airspace grid and aircraft information to generate planned path information. The status monitoring module is used to send the planned path information to the management terminal associated with the corresponding target aircraft and to monitor the operational status of the target aircraft. The dynamic adjustment module is used to evaluate the change in airspace situation of the target low-altitude airspace based on the monitoring results of the operation status. If the change in airspace situation exceeds the preset change threshold, the preset dynamic adjustment mechanism is triggered to adjust the dynamic airspace grid.

[0092] Optionally, the portrait generation module includes: The target classification submodule is used to identify all aircraft within the target's low-altitude airspace, set all aircraft associated with the target's low-altitude airspace as the first aircraft set, and set all remaining aircraft as the second aircraft set; The data acquisition submodule is used to collect flight plan data and real-time status data of the first set of aircraft, as well as real-time status data of the second set of aircraft, and to obtain airspace environment data and infrastructure status data of the target low-altitude airspace as a multi-source dataset. The feature mapping submodule is used to input multi-source datasets into a pre-built feature mapping model and construct a spatial situation profile based on the output of the feature mapping model.

[0093] Specific limitations regarding the AI-based low-altitude airspace grid dynamic partitioning and management system can be found in the limitations of the AI-based low-altitude airspace grid dynamic partitioning and management method described above, and will not be repeated here. Each module in the aforementioned AI-based low-altitude airspace grid dynamic partitioning and management system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0094] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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 this application, and should all be included within the protection scope of this application.

Claims

1. A method for dynamic partitioning and management of low-altitude airspace grids based on AI, characterized in that, Including the following steps: Real-time acquisition of multi-source datasets corresponding to the target low-altitude airspace, and generation of airspace situation profiles based on the multi-source datasets, wherein the multi-source datasets include aircraft status data, flight plan data, airspace environment data, and infrastructure status data; A demand analysis is performed on the airspace situation profile to obtain operational demand information, and grid partitioning parameters are calculated and generated based on the operational demand information. The grid partitioning parameters include grid granularity, grid level, and grid shape. A dynamic airspace grid matching the operational requirements is generated using a preset grid partitioning algorithm, and airspace resource identifiers are assigned to the dynamic airspace grid. Acquire the target aircraft's aircraft information, and perform collaborative path planning based on the dynamic airspace grid and aircraft information to generate planned path information; The planned path information is sent to the management terminal associated with the corresponding target aircraft, and the operational status of the target aircraft is monitored. The changes in the airspace situation of the target low-altitude airspace are assessed based on the results of the operational status monitoring. If the changes in the airspace situation exceed the preset change threshold, a preset dynamic adjustment mechanism is triggered to adjust the dynamic airspace grid.

2. The AI-based method for dynamic partitioning and management of low-altitude airspace grids according to claim 1, characterized in that: The steps of real-time acquisition of multi-source datasets corresponding to the target low-altitude airspace and generation of an airspace situation profile based on the multi-source datasets, wherein the multi-source datasets include aircraft status data, flight plan data, airspace environment data, and infrastructure status data, include the following steps: Identify all aircraft within the target low-altitude airspace, set all aircraft associated with the target low-altitude airspace as the first aircraft set, and set all remaining aircraft as the second aircraft set; Flight plan data and real-time status data of the first set of aircraft and real-time status data of the second set of aircraft were collected respectively, and airspace environment data and infrastructure status data of the target low-altitude airspace were obtained as a multi-source dataset. Multi-source datasets are input into a pre-built feature mapping model, and a spatial situation profile is constructed based on the output of the feature mapping model.

3. The AI-based method for dynamic partitioning and management of low-altitude airspace grids according to claim 2, characterized in that: The step of inputting multi-source datasets into a pre-built feature mapping model and constructing a spatial situation profile based on the output of the feature mapping model includes the following steps: The feature mapping model maps multi-source datasets to a predefined set of feature elements, which includes functional ownership elements, ecological constraint elements, public space elements, and dynamic activity elements. An airspace management rule set is generated based on a set of feature elements. The airspace management rule set includes a subset of static no-fly rules generated based on functional ownership elements and public space elements, a subset of conditional open rules generated based on functional ownership elements and preset functional area management standards, a subset of ecologically sensitive area rules generated based on ecological constraint elements, and a subset of activity constraint rules generated based on dynamic activity elements. An airspace situation profile is constructed based on the airspace management rule set.

4. The AI-based method for dynamic partitioning and management of low-altitude airspace grids according to claim 1, characterized in that: The process of performing a demand analysis on the airspace situation profile to obtain operational demand information, and then calculating and generating grid partitioning parameters based on this information, including grid granularity, grid level, and grid shape, includes the following steps: The airspace situation profile is analyzed to extract the operational demand information of the target low-altitude airspace. The operational demand information includes aircraft density distribution, flight plan conflict probability, static no-fly zone information, dynamic no-fly zone information, and preset airspace capacity threshold. The mesh granularity is determined based on the aircraft density distribution and the airspace capacity threshold. Specifically, this includes: calculating the ratio of the aircraft density distribution to the airspace capacity threshold; when the ratio is higher than a preset first threshold, the corresponding fine-grained mesh is used; when the ratio is lower than a preset second threshold, the corresponding coarse-grained mesh is used. The grid level is determined based on static and dynamic no-fly zone information. Specifically, this includes: identifying the static altitude attributes associated with static no-fly zone information and the dynamic altitude attributes associated with dynamic no-fly zone information, and setting the grid level through preset hierarchical control rules. The hierarchical control rules define the main aircraft types and control intensity corresponding to different altitude ranges. The grid shape is determined based on the flight plan conflict probability, static no-fly zone information, and dynamic no-fly zone information. Specifically, this includes: identifying the first spatial contour of static no-fly zone information and the second spatial contour of dynamic no-fly zone information; identifying conflict hotspots and non-conflict hotspots in the target low-altitude airspace based on the flight plan conflict probability; isolating conflict hotspots using irregular grid shapes based on the first and second spatial contours; and using standard grid shapes in non-conflict areas.

5. The AI-based method for dynamic partitioning and management of low-altitude airspace grids according to claim 1, characterized in that: The step of generating a dynamic airspace grid matching the operational requirements information using a preset grid partitioning algorithm and assigning airspace resource identifiers to the dynamic airspace grid includes the following steps: Based on the matching of grid division parameters, a corresponding grid division algorithm is preset, and the grid division parameters are used as the basic parameters. Combined with the airspace situation profile, the optimization target is set, and the matched grid division algorithm is called to perform iterative optimization calculation based on the target low-altitude airspace to generate a dynamic airspace grid. Based on the airspace situation profile, a multi-dimensional state assessment is performed on each grid cell in the generated dynamic airspace grid. Based on the multidimensional state assessment results, an airspace resource identification code is assigned to each grid cell. The airspace resource identification code contains the spatial location and hierarchical ID of its corresponding grid cell, and is associated with an availability status label, the maximum number of aircraft that can be accommodated, and the expected validity period of the status.

6. The AI-based method for dynamic partitioning and management of low-altitude airspace grids according to claim 5, characterized in that: The steps of matching a preset grid partitioning algorithm based on grid partitioning parameters, using the grid partitioning parameters as basic parameters, setting optimization targets in conjunction with airspace situation profiles, and calling the matched grid partitioning algorithm to perform iterative optimization calculations based on the target low-altitude airspace to generate a dynamic airspace grid include the following steps: Based on the airspace situation profile, a multi-objective optimization function is constructed with the optimization objectives of maximizing airspace resource utilization, minimizing the probability of flight conflict, and dynamically constrained area avoidance rate. An initial grid structure based on the target low-altitude airspace is generated according to the grid division parameters; The boundaries of the grid cells in the initial grid structure are adjusted by iteratively solving a multi-objective optimization function until the multi-objective optimization function converges or reaches the preset number of iterations, thus obtaining a dynamic spatial grid.

7. The AI-based method for dynamic partitioning and management of low-altitude airspace grids according to claim 1, characterized in that: The step of acquiring the target aircraft's aircraft information and performing collaborative path planning based on the dynamic airspace grid and aircraft information to generate planned path information includes the following steps: The aircraft information is analyzed to identify its associated flight plan data and extract take-off and landing points, planned routes, and mission types. Based on the take-off and landing points and the planned flight path, several consecutive grid cells are located in the dynamic airspace grid to form candidate airspace channels; Obtain flight safety distance and real-time airspace traffic, use candidate airspace channels as the search space, and generate one or more candidate paths based on flight safety distance, airspace resource identification code and real-time airspace traffic using a path search algorithm; Spatiotemporal conflict detection is performed on candidate paths to select a set of safe paths, and the planned path information is determined based on the task type.

8. The AI-based method for dynamic partitioning and management of low-altitude airspace grids according to claim 7, characterized in that: The steps of obtaining flight safety distance and real-time airspace traffic, using candidate airspace channels as the search space, and generating one or more candidate paths based on flight safety distance, airspace resource identifiers, and real-time airspace traffic using a path search algorithm, include the following steps: The minimum safe interval constraint for path search is set based on flight safety distance. The accessibility status, maximum capacity and constraint conditions of each grid cell are extracted based on airspace resource identification code. The current load of each grid cell is calculated based on real-time airspace traffic. The accessibility status includes passable and impassable. Within the candidate airspace channel, taking the take-off and landing points as path endpoints, a preset graph search algorithm is used to traverse all grid cell sequences that satisfy the minimum safety interval constraint and whose grid cell status is passable, generating an initial set of feasible paths including several feasible paths. Calculate the path length of each feasible path, the total load of the grid cells it contains, and the degree to which the constraint conditions are satisfied as the cost value. Sort the paths from low to high cost value and select the N paths with the lowest cost value as candidate paths.

9. An AI-based dynamic partitioning and management system for low-altitude airspace grids, characterized in that, include: The image generation module is used to collect multi-source datasets corresponding to the target low-altitude airspace in real time, and generate an airspace situation image based on the multi-source datasets. The multi-source datasets include aircraft status data, flight plan data, airspace environment data, and infrastructure status data. The parameter generation module is used to perform demand analysis on the airspace situation profile, obtain operational demand information, and calculate and generate grid division parameters based on the operational demand information. The grid division parameters include grid granularity, grid level, and grid shape. The grid generation module is used to generate a dynamic airspace grid that matches the operational requirements information through a preset grid division algorithm, and to assign airspace resource identifiers to the dynamic airspace grid. The path generation module is used to acquire the target aircraft's aircraft information and perform collaborative path planning based on the dynamic airspace grid and aircraft information to generate planned path information. The status monitoring module is used to send the planned path information to the management terminal associated with the corresponding target aircraft and to monitor the operational status of the target aircraft. The dynamic adjustment module is used to evaluate the change in airspace situation of the target low-altitude airspace based on the monitoring results of the operation status. If the change in airspace situation exceeds the preset change threshold, the preset dynamic adjustment mechanism is triggered to adjust the dynamic airspace grid.

10. A low-altitude airspace grid dynamic partitioning and management system based on AI according to claim 9, characterized in that, The portrait generation module includes: The target classification submodule is used to identify all aircraft within the target's low-altitude airspace, set all aircraft associated with the target's low-altitude airspace as the first aircraft set, and set all remaining aircraft as the second aircraft set; The data acquisition submodule is used to collect flight plan data and real-time status data of the first set of aircraft, as well as real-time status data of the second set of aircraft, and to obtain airspace environment data and infrastructure status data of the target low-altitude airspace as a multi-source dataset. The feature mapping submodule is used to input multi-source datasets into a pre-built feature mapping model and construct a spatial situation profile based on the output of the feature mapping model.

Images