A Low-Altitude Airspace Capacity Assessment Method Based on Spatial Parameter Simulation-Driven Relationship Coupling

Abstract

This invention discloses a low-altitude airspace capacity assessment method based on spatial parameter simulation-driven relationship coupling. The method includes: creating a low-altitude airspace capacity simulation model using a simulated spatial morphology dataset to represent urban buildings and low-altitude airspace; conducting simulated flight within the low-altitude airspace capacity simulation model using a UAV parameter set, with a cooperative obstacle avoidance algorithm module used to calculate and obtain the simulated low-altitude airspace capacity; constructing a low-altitude airspace capacity assessment model including a multi-dimensional relationship coupling module, which uses the simulated spatial morphology dataset and the low-altitude airspace capacity dataset for corresponding relationship coupling identification and low-altitude airspace capacity assessment training; and obtaining data from the study area and inputting it into the low-altitude airspace capacity assessment model to obtain the low-altitude airspace capacity. This invention achieves accurate assessment of low-altitude airspace capacity in complex urban building clusters through simulation data production and spatial parameter simulation-driven relationship coupling.

Description

Technical Field

[0001] This invention relates to the field of urban low-altitude airspace capacity assessment and management, and in particular to a low-altitude airspace capacity assessment method based on spatial parameter simulation-driven relationship coupling. Background Technology

[0002] With the rise of the Urban Air Mobility (UAM) concept and the rapid development of drone technology, low-altitude flight activities such as logistics delivery, emergency response, and urban inspection are becoming increasingly frequent. Urban low-altitude airspace, especially the area 120 to 300 meters above the building canopy, is becoming a valuable and limited strategic resource. Accurately assessing the maximum drone traffic that a specific urban spatial structure can support—that is, the low-altitude airspace capacity—is a crucial prerequisite for ensuring flight safety, improving operational efficiency, and scientifically planning airspace. Existing assessment methods are mostly based on open airspace or idealized single-obstacle scenarios, severely neglecting the essential impact of the complexity and diversity of urban building distribution (such as concentrated, dispersed, or strip-shaped) on the operational capacity of drone swarms. The density, layout, and shape of urban buildings, among other spatial morphological indicators, collectively constitute the physical constraints of low-altitude drone operations, and their impact on capacity is non-linear and strongly coupled. Traditional methods, such as geometric calculations based on safety intervals or airspace sector division, cannot dynamically simulate the actual operation of a large number of UAVs coordinating obstacle avoidance in complex building clusters in real time. Furthermore, they struggle to reflect the interaction flows and dynamic bottlenecks generated by obstacle avoidance among multiple UAVs, leading to significant discrepancies between assessment results and actual operational capacity. Current technology lacks consideration for the coordinated operation of multi-configuration mixed fleets. In actual low-altitude operation scenarios, UAVs often include various configurations such as multi-rotor, fixed-wing, and compound-wing aircraft, with vastly different kinematic characteristics (e.g., minimum turning radius, hovering ability, speed range), airspace requirements, and obstacle avoidance behaviors. Existing methods mostly target only a single configuration and cannot assess the complex impact of mixed operations of heterogeneous fleets on overall capacity. Current technologies cannot theoretically reveal which spatial morphology indicators are key influencing factors, nor can they quickly calculate airspace capacity for a new urban area based solely on its morphological indicators, making it difficult to support rapid decision-making in urban planning and airspace management. Summary of the Invention

[0003] The purpose of this invention is to provide a method for assessing low-altitude airspace capacity based on spatial parameter simulation-driven relationship coupling. This method achieves accurate assessment of low-altitude airspace capacity in complex urban building clusters by coupling simulation data production with spatial parameter simulation-driven relationship coupling.

[0004] The objective of this invention is achieved through the following technical solution: A low-altitude airspace capacity assessment method based on spatial parameter simulation-driven relationship coupling, the method comprising: S1. Construct a simulated spatial morphology dataset divided according to city area, city building type, and spatial morphology data hierarchy; construct a low-altitude airspace capacity simulation model including a collaborative obstacle avoidance algorithm module, and use the simulated spatial morphology dataset to create city buildings and low-altitude airspace. S2. Construct a drone parameter set containing drone type combinations. Use the drone parameter set to simulate flight within the low-altitude airspace capacity simulation model. The cooperative obstacle avoidance algorithm module is used to calculate and obtain the maximum number of drones that do not collide during simulated flight as the low-altitude airspace capacity, and obtain the low-altitude airspace capacity dataset divided according to drone type combinations. S3. Construct a low-altitude airspace capacity assessment model that includes a multi-relationship coupling module. The low-altitude airspace capacity assessment model uses the multi-relationship coupling module to perform corresponding relationship coupling identification and low-altitude airspace capacity assessment learning and training by using the simulated spatial morphology dataset and the low-altitude airspace capacity dataset. S4. Obtain spatial morphological data of the study area and input the configuration parameters of UAV type combination into the low-altitude airspace capacity assessment model, and output the assessment results of low-altitude airspace capacity.

[0005] To better implement this invention, in method S1, urban building types are classified according to the distribution type of urban buildings. Urban building types include centrally concentrated type, cluster-concentrated type, strip type, centrally radial type, centrally clustered constellation type, and scattered distribution type. The spatial morphology data are obtained by corresponding to the spatial morphology index system. The spatial morphology index system includes building density, average building height, standard deviation of building height, standard deviation of building footprint, average building spacing, standard deviation of building spacing, building compactness, building fractal dimension, building directional concentration, building center of gravity offset, building distribution entropy, and building clustering index.

[0006] Preferably, in method S1, the low-altitude airspace capacity simulation model simulates and generates urban buildings and their layout according to the data parameters of the simulation spatial morphology dataset, and creates a low-altitude airspace located from the top surface of the urban buildings to a height range of N meters below.

[0007] Preferably, the UAV type combination includes multi-rotor UAVs, fixed-wing UAVs, and compound-wing UAVs. The low-altitude airspace capacity simulation model contains a UAV kinematic model. The UAV kinematic model simulates flight and motion constraints in the low-altitude airspace of the low-altitude airspace capacity simulation model according to the UAV type in the UAV parameter set. The UAV kinematic model includes a multi-rotor UAV kinematic module, a fixed-wing UAV kinematic module, and a compound-wing UAV kinematic module. The multi-rotor UAV kinematic module uses a dual integrator model to constrain the simulated flight of the multi-rotor UAV, and the fixed-wing UAV kinematic module uses a Dubins model to constrain the simulated flight of the fixed-wing UAV.

[0008] Preferably, the cooperative obstacle avoidance algorithm module treats the simulated UAV as a cube with a side length of d, where d is the minimum circumscribed circle diameter of the UAV. The cooperative obstacle avoidance algorithm module creates N UAVs to simulate flight in the low-altitude airspace of the low-altitude airspace capacity simulation model using the cooperative obstacle avoidance algorithm. The cooperative obstacle avoidance algorithm includes cooperative obstacle avoidance algorithms based on artificial potential fields, cooperative obstacle avoidance algorithms based on velocity obstacle methods, cooperative obstacle avoidance algorithms based on model predictive control, cooperative obstacle avoidance algorithms based on fast exploration random trees, or cooperative obstacle avoidance algorithms based on reinforcement learning.

[0009] Preferably, the low-altitude airspace capacity acquisition method includes: continuously adding a mixed fleet of aircraft to the low-altitude airspace from the low-altitude airspace side of the low-altitude airspace capacity simulation model. The mixed fleet includes multi-rotor UAVs, fixed-wing UAVs, and / or compound-wing UAVs. The dynamic trajectory of the UAVs in the current low-altitude airspace is displayed in real time. UAVs and buildings, as well as UAVs among themselves, are processed in real time through a cooperative obstacle avoidance algorithm module. The process continues until the last addition of the mixed fleet to the current low-altitude airspace results in a collision. The maximum number of UAVs that did not collide in the last collision is taken as the low-altitude airspace capacity.

[0010] Preferably, in method S3, the multivariate relationship coupling module includes a generalized additive model and a multi-scale geographic weighted regression coupling model. The generalized additive model constructs a nonlinear relationship between spatial morphology indicators and low-altitude airspace capacity. The multi-scale geographic weighted regression model analyzes spatial heterogeneity to obtain the spatial relationship variation coefficient of spatial morphology indicators. The multivariate relationship coupling module uses the nonlinear relationship and the spatial relationship variation coefficient to construct the relationship coupling.

[0011] Preferably, a study area is selected and its spatial morphology data is obtained. A homogeneous fleet of UAVs of the same type is configured in the study area. The low-altitude airspace capacity assessment model outputs the low-altitude airspace capacity assessment results of the homogeneous fleet. The homogeneous fleet is one of the following three cases: a full multi-rotor fleet in which all UAVs are multi-rotor UAVs, a full compound wing fleet in which all UAVs are fixed-wing UAVs, or a full compound wing fleet in which all UAVs are compound wing UAVs. The low-altitude airspace capacity comparison data of the study area under different homogeneous fleet conditions is analyzed. Alternatively, the city building types and spatial morphology data in the simulation spatial morphology dataset can be adjusted and used as an experimental area to analyze the low-altitude airspace capacity comparison data of different city building types or different spatial morphology data under different isomorphic aircraft fleet conditions.

[0012] Preferably, a study area is selected and spatial morphology data of the study area is obtained. Different configuration parameters of heterogeneous aircraft are configured in the study area. The configuration parameters of the heterogeneous aircraft are the proportion of three types of aircraft: multi-rotor UAVs, fixed-wing UAVs and compound-wing UAVs. The low-altitude airspace capacity assessment model outputs the low-altitude airspace capacity assessment results under the case of heterogeneous aircraft with different configuration parameters. The low-altitude airspace capacity comparison data of the study area under the case of heterogeneous aircraft with different configuration parameters is analyzed. Alternatively, the city building types and spatial morphology data in the simulation spatial morphology dataset can be adjusted and used as an experimental area to analyze the low-altitude airspace capacity comparison data of different city building types or different spatial morphology data under heterogeneous fleet conditions with different ratio parameters.

[0013] Preferably, in method S3, the multivariate relationship coupling module further uses principal component analysis to perform dimensionality reduction processing on the spatial morphology indicators before relationship coupling identification to obtain the spatial morphology indicators of the main influencing components to participate in the relationship coupling identification process.

[0014] Compared with the prior art, the present invention has the following advantages and beneficial effects: (1) This invention forms a simulated spatial morphology dataset according to the hierarchical structure of urban area, urban building type and spatial morphology data. The city buildings and low-altitude airspace are created through the low-altitude airspace capacity simulation model. The drone parameter set is used to simulate flight in the low-altitude airspace of the low-altitude airspace capacity simulation model. The cooperative obstacle avoidance algorithm module performs fine cooperative obstacle avoidance processing in complex building scenarios and calculates the maximum number of drones that do not collide as the low-altitude airspace capacity. Thus, the spatial morphology and low-altitude airspace associated dataset is obtained. The spatial morphology and low-altitude airspace associated dataset is constructed by corresponding association and combination of the simulated spatial morphology dataset and the low-altitude airspace capacity dataset. The low-altitude airspace capacity assessment model of this invention uses the spatial morphology and low-altitude airspace associated dataset to perform relationship coupling identification and low-altitude airspace capacity assessment training through the multi-relationship coupling module. By establishing the nonlinear and spatially heterogeneous coupling relationship between urban spatial morphology and drone cluster operation capacity, the accurate assessment of low-altitude airspace capacity in complex urban building clusters is realized.

[0015] (2) This invention can obtain a rich database of simulated spatial morphology data and UAV parameter combination driven by coupling relationship. By acquiring new urban spatial morphology data and configuring UAV type combination parameters, it can quickly predict and evaluate high-precision airspace capacity, improve evaluation efficiency, and can be widely applied to decision-making scenarios such as urban planning, dynamic airspace delineation, and flight mission pre-evaluation, providing technical support for the scientific planning, safe management and efficient utilization of low-altitude resources.

[0016] (3) This invention can assess the airspace capacity of heterogeneous fleets with different mixed proportions and perform sensitivity analysis, which makes it easy to determine the advantageous or bottleneck configurations that affect capacity, and provides a scientific basis for air traffic control departments to formulate time-sharing and segment-based operation rules for heterogeneous UAVs.

[0017] (4) The multivariate relationship coupling module of this invention performs dimensionality reduction processing on spatial morphology indicators by principal component analysis, uses the GAM model to capture the nonlinear relationship between spatial morphology indicators and low-altitude airspace capacity, and uses the MGWR model to analyze spatial heterogeneity to obtain the spatial relationship variation coefficient of spatial morphology indicators, thereby realizing the deep coupling of global nonlinear features and local spatial variation features, and revealing the key influence mechanism of urban spatial morphology indicators on low-altitude airspace capacity. Attached Figure Description

[0018] Figure 1 This is a flowchart of the low-altitude airspace capacity assessment method of the present invention. Detailed Implementation

[0019] The present invention will be further described in detail below with reference to embodiments: Example like Figure 1As shown, a low-altitude airspace capacity assessment method based on spatial parameter simulation-driven relationship coupling is proposed, the method comprising: S1. Construct a simulated spatial morphology dataset hierarchically based on city area, city building type, and spatial form data. This invention constructs a simulated spatial morphology dataset hierarchically based on city area, city building type, and spatial form data. Adjustments to the internal parameters of these three aspects generate numerous simulated spatial morphology data points, forming a hierarchically structured dataset. In this embodiment, city area is disregarded to reduce the hierarchical dimension (i.e., assuming the same city area), focusing primarily on two levels: city building type and spatial form data. In this invention, city building types are categorized according to their distribution. These types include centrally concentrated, cluster-concentrated, strip-shaped, centrally radial, centrally clustered constellation, and scattered distribution types. The centrally concentrated building distribution is characterized by buildings concentrated in the central area of ​​the airspace, with surrounding buildings generally exhibiting a normal distribution in two dimensions. In this embodiment, a Gaussian distribution is used for the subsequent generation of centrally concentrated building simulations (generating buildings concentrated in the central area, with others generally conforming to a normal distribution; a certain deviation parameter for the normal distribution can be set). The cluster-type building distribution is characterized by buildings forming clusters of relatively independent areas. During building simulation, the ground corresponding to the airspace can be divided into m sub-regions. Buildings within each sub-region are generated in a centrally concentrated manner, with a certain distance maintained between clusters. The linear type primarily targets linear urban built-up areas formed along rivers or canyons. The linear building distribution is characterized by buildings distributed in a linear pattern along a certain direction. During building simulation, a principal direction angle θ is set. The coordinates of the building centers perpendicular to the principal direction follow a uniform distribution, while the coordinates parallel to the principal direction follow a Gaussian distribution. The center-radial type building distribution is characterized by buildings radiating outwards from the center. In the polar coordinates (r, φ) of the center of the center of the radial type, r follows an exponential distribution, and φ follows a uniform distribution. The cluster-constellation type building distribution is characterized by buildings clustered around multiple center points. For example, k center points are preset, and each building belongs to the nearest center point, generating buildings around that center point in a Gaussian distribution. The characteristics of a scattered building distribution are: buildings are randomly and uniformly distributed in the airspace; the center coordinates (x, y) of the building clusters all follow a uniform distribution. Of course, the city building type can also be other types or directly generated randomly from simulated buildings, and the city building type can be obtained subsequently through model clustering.

[0020] Spatial morphology data are compiled into a dataset according to a spatial morphology index system. This system includes spatial morphology indicators such as building density, average building height, standard deviation of building height, standard deviation of building footprint, average building spacing, standard deviation of building spacing, building compactness, building fractal dimension, building directional concentration, building center of gravity offset, building distribution entropy, and building clustering index. In this embodiment, the building density (BD) indicator is defined as the ratio of building footprint to total airspace area: BD = Abuilding / Atotal, where Abuilding is the total building area and Atotal is the total airspace area. The average building height (ABH) indicator is defined as the average height of all buildings: ABH = Σhi / n, where hi is the height of the i-th building and n is the number of buildings. The building height standard deviation (BHSD) indicator reflects the dispersion of building heights. The Building Area Standard Deviation (BASD) reflects the dispersion of building footprints. The Average Building Spacing (ABS) is the average distance between the closest adjacent buildings. The Building Spacing Standard Deviation (BSSD) reflects the dispersion of building spacing. Building Compactness (BC) reflects the regularity of building shapes, BC = 4πA / P², where A is the building area and P is the building perimeter. Building Fractal Dimension (BFD) reflects the complexity of building distribution. Building Orientation Concentration (BOC) reflects the concentration of buildings along their long axis. The Building Centroid Offset (BCO) index reflects the degree of deviation of the building's center of gravity from the geometric center of the airspace.The Building Distribution Entropy (BDE) index reflects the uniformity of the spatial distribution of buildings. The Building Clustering Index (BCI) index reflects the degree of building clustering.

[0021] A low-altitude airspace capacity simulation model is constructed, incorporating a collaborative obstacle avoidance algorithm module. This model utilizes a simulated spatial morphology dataset to create urban buildings and low-altitude airspace. The low-altitude airspace capacity simulation model simulates the creation of urban buildings based on spatial morphology data (which is a crucial indicator for simulation creation). Preferably, the low-altitude airspace capacity simulation model simulates and generates urban buildings and their layout according to the data parameters of the simulated spatial morphology dataset, and creates low-altitude airspace located from the top surface of urban buildings to a height of N meters or less. The low-altitude airspace is defined as the airspace from the skyline of urban buildings to a height of N meters or less. For example, if the height N is 300 meters, then the low-altitude airspace is the airspace below 300 meters and above the top surface of urban buildings.

[0022] S2. Construct a UAV parameter set containing combinations of UAV types. Preferably, the UAV types in the UAV type combinations include multi-rotor UAVs, fixed-wing UAVs, and compound-wing UAVs. The low-altitude airspace capacity simulation model contains a UAV kinematic model. The UAV kinematic model simulates flight and motion constraints in the low-altitude airspace of the low-altitude airspace capacity simulation model according to the UAV types in the UAV parameter set. The UAV kinematic model includes a multi-rotor UAV kinematic module, a fixed-wing UAV kinematic module, and a compound-wing UAV kinematic module. The multi-rotor UAV kinematic module uses a dual integrator model to constrain the simulated flight of the multi-rotor UAV. The constraint conditions include: |, Minimum turning radius ,in For acceleration, For maximum acceleration, For flight speed, To achieve the maximum flight speed, the kinematics module for the fixed-wing UAV uses the Dubins model to constrain the simulated flight of the fixed-wing UAV. The constraints include: the rate of change of the heading angle ψ is... ,in L is the roll angle, and L is the wingspan. For flight speed, , The maximum roll angle corresponds to the minimum turning radius. The kinematics module for compound-wing UAVs employs a dual integrator model in the low-speed phase (below a threshold) and a Dubins model in the low-speed phase (above a threshold). The UAV kinematics model sets constraints for each UAV type, including minimum turning radius Rmin, maximum speed Vmax, acceleration limit Amax, and hovering capability. For multi-rotor UAVs, Rmin is smaller, allowing hovering; for fixed-wing UAVs, Rmin is larger, preventing hovering; and for compound-wing UAVs, it combines the characteristics of both multi-rotor and fixed-wing aircraft.

[0023] Within the low-altitude airspace capacity simulation model, simulated flight is conducted using a set of UAV parameters. The cooperative obstacle avoidance algorithm module is used to calculate the maximum number of UAVs that do not collide during simulated flight as the low-altitude airspace capacity, obtaining a low-altitude airspace capacity dataset divided according to UAV type combinations. In some embodiments, the cooperative obstacle avoidance algorithm module treats the simulated UAVs as cubes with side length d, where d is the minimum circumscribed circle diameter of the UAV. During building simulation construction, buildings are composed of a random number of adjacent squares, and buildings are convex polygons; buildings are either adjacent or separated, with no intersections or overlaps, and each building has no voids; for non-adjacent cases, the gap between buildings is greater than or equal to the minimum circumscribed circle diameter d of the UAV, ensuring that UAVs can safely pass between buildings; the generated buildings maintain a safe distance from the upper, lower, left, and right boundaries of the low-altitude airspace, and the safe distance is not less than d. The cooperative obstacle avoidance algorithm module creates N UAVs and simulates their flight in the low-altitude airspace using a cooperative obstacle avoidance algorithm within a low-altitude airspace capacity simulation model. These algorithms include those based on artificial potential fields, velocity-based obstacle avoidance, model predictive control, fast-exploration random trees, or reinforcement learning. The artificial potential field-based cooperative obstacle avoidance algorithm (APF) is explained as follows: The UAV is considered as a particle moving in a potential field. The target point generates an attractive force, and obstacles generate a repulsive force. The UAV moves under the combined force, and obstacle avoidance is achieved by adjusting the potential field parameters. This embodiment uses the artificial potential field-based cooperative obstacle avoidance algorithm as an example for detailed technical description. The resultant force on UAV i is... for: ,in The attractiveness of the target point to the drone i, Let j be the repulsive force exerted by building j on drone i. Let $k$ be the repulsive force exerted by other drones $k$ on drone $i$. The attractive force expression is as follows: Where katt is the attraction coefficient. The position of drone i The target location is shown below. The expression for the repulsive force is as follows: ,when < It takes effect at that time, where krep is the exclusivity coefficient. Let i be the distance between drone i and obstacle j. To influence distance, the following cooperative obstacle avoidance algorithms are used: Velocity Obstacle (VO) based on speed obstacle method: It determines the speed obstacle zone by calculating relative speed and relative position, and selects the speed outside the speed obstacle zone as the UAV's movement speed. Model Predictive Control (MPC) based cooperative obstacle avoidance algorithm: It optimizes the control input in the prediction time domain by establishing a UAV kinematic model to achieve obstacle avoidance and trajectory tracking. Rapidly-exploring Random Tree (RRT) based cooperative obstacle avoidance algorithm: It constructs a search tree through random sampling to quickly explore the feasible path space, suitable for path planning in complex environments. Reinforcement Learning (RL) based cooperative obstacle avoidance algorithm: It learns the optimal obstacle avoidance strategy through interaction with the environment, enabling it to adapt to dynamically changing environmental conditions.

[0024] In some embodiments, the low-altitude airspace capacity acquisition method includes: continuously adding a mixed fleet of drones from the low-altitude airspace side of the low-altitude airspace capacity simulation model to the low-altitude airspace. The mixed fleet includes multi-rotor UAVs, fixed-wing UAVs, and / or compound-wing UAVs. The dynamic trajectories of the UAVs in the current low-altitude airspace are displayed in real time. Real-time cooperative obstacle avoidance processing is performed between UAVs and buildings, and between UAVs themselves, through a cooperative obstacle avoidance algorithm module. The low-altitude airspace capacity is determined by the maximum number of UAVs that did not collide with the last addition of the mixed fleet to the current low-altitude airspace until a collision occurs. This maximum number of UAVs that did not collide in the last addition is taken as the low-altitude airspace capacity. This embodiment uses a mixed fleet containing three different types of UAVs as an example, as shown below: (1) Initialization: Set the airspace boundary, building location, UAV initial position (leftmost side of the airspace), and target position (rightmost side of the airspace).

[0025] (2) Fleet addition: Add a mixed fleet from the leftmost side of the airspace at fixed time intervals Δt. Each mixed fleet consists of one multi-rotor, one fixed-wing, and one compound-wing UAV, for a total of 3 aircraft.

[0026] (3) Collaborative obstacle avoidance: All UAVs run a collaborative obstacle avoidance algorithm in real time to update their position and speed.

[0027] (4) Collision detection: Detect the distance between the drone and the building, and between drones. If the distance between any two objects is less than the safe distance dsafe = 1.5d, then a collision is determined to have occurred.

[0028] (5) Capacity determination: When a collision occurs, record the total number of UAVs N in the current airspace, then the maximum low-altitude airspace capacity Cmax = N - 3 (minus the last added fleet).

[0029] (6) Repeated simulation: For each urban spatial form, each obstacle avoidance algorithm is used to repeat the simulation 100 times, and the average value is taken as the low-altitude airspace capacity of the form-algorithm combination.

[0030] S3. Construct a low-altitude airspace capacity assessment model including a multivariate relationship coupling module. This model utilizes the simulated spatial morphology dataset and the low-altitude airspace capacity dataset for corresponding relationship coupling identification and low-altitude airspace capacity assessment training through the multivariate relationship coupling module. Preferably, before relationship coupling identification, the multivariate relationship coupling module uses principal component analysis to reduce the dimensionality of spatial morphology indicators to obtain the spatial morphology indicators of the main influencing components for relationship coupling identification. In principal component analysis, the correlation coefficient matrix of all spatial morphology indicators in the spatial morphology indicator system is calculated, and eigenvalue decomposition is performed. The top p principal components with a cumulative variance contribution rate greater than 85% are selected as independent variables for subsequent modeling. The multivariate relationship coupling module includes a generalized additive model (GAM model) and a multi-scale geographic weighted regression coupling model (MGWR model). The generalized additive model constructs the nonlinear relationship between spatial morphology indicators and low-altitude airspace capacity. To capture the global nonlinear influence of principal components on capacity, the generalized additive model constructs the following expression: A spline smoothing function (fi) is used, and the smoothing parameters are determined through cross-validation. The mgcv package in R is used for model fitting to obtain the nonlinear impact curves of each principal component on low-altitude airspace capacity, revealing the nonlinear impact curves of each principal component (i.e., a combination of spatial morphological features) on capacity. A multi-scale geographic weighted regression model is used to analyze spatial heterogeneity and obtain the spatial relationship variation coefficients of spatial morphological indicators. The multivariate relationship coupling module constructs relationships by coupling nonlinear relationships with spatial relationship variation coefficients. The multivariate relationship coupling module constructs the following expression: , Let be the observed value of the explained variable for spatial unit i; Let i be the geographic coordinates of spatial unit i. For spatial element i, the local intercept term; The local regression coefficient of the explanatory variable k in spatial unit i; Let i be the spatial unit and k be the values ​​of the explanatory variable. The random error term of the model is used; the spatial weights are determined using an adaptive bisquare kernel function. (Weights of spatial units i and j) The optimal bandwidth is determined through cross-validation. The larger the bandwidth, the closer the influence of the variables is to global stationarity. Then, the MGWR package in R language is used to fit the model and obtain the spatial distribution map of each regression coefficient.

[0031] The multivariate relationship coupling module uses the global prediction results of the generalized additive model (GAM model) as a benchmark, and corrects it using the local variation coefficients captured by the multi-scale geographical weighted regression coupling model (MGWR model). For any new spatial form, its principal component scores are first calculated and substituted into the GAM model to obtain the benchmark capacity. Then, based on its position in multidimensional space, the MGWR residuals of its neighboring samples are weighted and interpolated to obtain the local correction term. Final capacity The multivariate relation coupling module was validated using the coefficient of determination R², adjusted R², root mean square error RMSE, and mean absolute error MAE.

[0032] S4. Obtain spatial morphological data of the study area and input the configuration parameters of UAV type combination into the low-altitude airspace capacity assessment model, and output the assessment results of low-altitude airspace capacity.

[0033] In some embodiments, a study area is selected and its spatial morphology data is acquired. A homogeneous fleet of UAVs of the same type is configured within the study area. The low-altitude airspace capacity assessment model outputs the low-altitude airspace capacity assessment results for the homogeneous fleet. The homogeneous fleet can be one of three types: an all-multirotor fleet (all UAVs are multirotor), an all-compound-wing fleet (all UAVs are fixed-wing), or an all-compound-wing fleet (all UAVs are compound-wing). The low-altitude airspace capacity comparison data for the study area under different homogeneous fleet configurations is analyzed. Alternatively, the following method can be used: adjust the simulated spatial morphology dataset to include urban building types and spatial morphology data as an experimental area, and analyze the low-altitude airspace capacity comparison data for different urban building types or different spatial morphology data under different homogeneous fleet configurations. This allows for comparison of the low-altitude airspace capacity of all-multirotor, all-fixed-wing, and all-compound-wing homogeneous fleets under several urban building types, thereby assessing and identifying the most suitable urban building types for each configuration.

[0034] In some embodiments, a study area is selected and its spatial morphology data is acquired. Different configuration parameters are used to configure heterogeneous aircraft fleets within the study area. These parameters refer to the proportions of multi-rotor UAVs, fixed-wing UAVs, and compound-wing UAVs. The low-altitude airspace capacity assessment model outputs low-altitude airspace capacity assessment results for heterogeneous aircraft fleets with different configuration parameters. The low-altitude airspace capacity comparison data for the study area under different configuration parameters is analyzed. Alternatively, the following method can be used: The simulated spatial morphology dataset can be used to concentrate urban building types and spatial morphology data as an experimental area. The low-altitude airspace capacity comparison data for different urban building types or different spatial morphology data under different configuration parameters of heterogeneous aircraft fleets can be analyzed. For each urban building type, the changes in low-altitude airspace capacity under different multi-rotor / fixed-wing / compound-wing configurations are analyzed. By calculating the partial derivative of capacity with respect to the proportion of each configuration, the sensitivity coefficient is determined, achieving heterogeneous aircraft fleet sensitivity analysis.

[0035] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for assessing low-altitude airspace capacity based on spatial parameter simulation-driven relationship coupling, characterized in that: The methods include: S1. Construct a simulated spatial morphology dataset divided according to city area, city building type, and spatial morphology data hierarchy; construct a low-altitude airspace capacity simulation model including a collaborative obstacle avoidance algorithm module, and use the simulated spatial morphology dataset to create city buildings and low-altitude airspace. S2. Construct a drone parameter set containing drone type combinations. Use the drone parameter set to simulate flight within the low-altitude airspace capacity simulation model. The cooperative obstacle avoidance algorithm module is used to calculate and obtain the maximum number of drones that do not collide during simulated flight as the low-altitude airspace capacity, and obtain the low-altitude airspace capacity dataset divided according to drone type combinations. S3. Construct a low-altitude airspace capacity assessment model that includes a multi-relationship coupling module. The low-altitude airspace capacity assessment model uses the multi-relationship coupling module to perform corresponding relationship coupling identification and low-altitude airspace capacity assessment learning and training by using the simulated spatial morphology dataset and the low-altitude airspace capacity dataset. S4. Obtain spatial morphological data of the study area and input the configuration parameters of UAV type combination into the low-altitude airspace capacity assessment model, and output the assessment results of low-altitude airspace capacity.

2. The low-altitude airspace capacity assessment method based on spatial parameter simulation-driven relationship coupling according to claim 1, characterized in that: In method S1, urban building types are classified according to the distribution type of urban buildings. Urban building types include centrally concentrated type, cluster concentrated type, strip type, centrally radial type, centrally clustered constellation type, and scattered distribution type. The spatial morphology data are obtained by corresponding to the spatial morphology index system. The spatial morphology index system includes building density, average building height, standard deviation of building height, standard deviation of building footprint, average building spacing, standard deviation of building spacing, building compactness, building fractal dimension, building directional concentration, building center of gravity offset, building distribution entropy, and building clustering index.

3. The low-altitude airspace capacity assessment method based on spatial parameter simulation-driven relationship coupling according to claim 1, characterized in that: In method S1, the low-altitude airspace capacity simulation model simulates and generates urban buildings and their layout according to the data parameters of the simulation spatial morphology dataset, and creates a low-altitude airspace located from the top surface of urban buildings to a height of N meters below.

4. The low-altitude airspace capacity assessment method based on spatial parameter simulation-driven relationship coupling according to claim 1, characterized in that: The UAV type combination includes multi-rotor UAVs, fixed-wing UAVs, and compound-wing UAVs. The low-altitude airspace capacity simulation model contains a UAV kinematic model. The UAV kinematic model simulates flight and motion constraints in the low-altitude airspace of the low-altitude airspace capacity simulation model according to the UAV type in the UAV parameter set. The UAV kinematic model includes a multi-rotor UAV kinematic module, a fixed-wing UAV kinematic module, and a compound-wing UAV kinematic module. The multi-rotor UAV kinematic module uses a dual integrator model to constrain the simulated flight of the multi-rotor UAV, and the fixed-wing UAV kinematic module uses a Dubins model to constrain the simulated flight of the fixed-wing UAV.

5. The low-altitude airspace capacity assessment method based on spatial parameter simulation-driven relationship coupling according to claim 1, characterized in that: The cooperative obstacle avoidance algorithm module treats the simulated UAV as a cube with a side length of d, where d is the diameter of the UAV's smallest circumscribed circle. The cooperative obstacle avoidance algorithm module creates N UAVs to simulate flight in the low-altitude airspace of the low-altitude airspace capacity simulation model using the cooperative obstacle avoidance algorithm. The cooperative obstacle avoidance algorithm includes cooperative obstacle avoidance algorithms based on artificial potential fields, cooperative obstacle avoidance algorithms based on velocity obstacle methods, cooperative obstacle avoidance algorithms based on model predictive control, cooperative obstacle avoidance algorithms based on fast exploration random trees, or cooperative obstacle avoidance algorithms based on reinforcement learning.

6. The low-altitude airspace capacity assessment method based on spatial parameter simulation-driven relationship coupling according to claim 5, characterized in that: The method for obtaining low-altitude airspace capacity includes: continuously adding a mixed fleet of aircraft to the low-altitude airspace from one side of the low-altitude airspace simulation model. The mixed fleet includes multi-rotor UAVs, fixed-wing UAVs, and / or compound-wing UAVs. The dynamic trajectory of the UAVs in the current low-altitude airspace is displayed in real time. UAVs and buildings, as well as UAVs among themselves, are processed in real time through a cooperative obstacle avoidance algorithm module. The process continues until the last addition of the mixed fleet to the current low-altitude airspace results in a collision. The maximum number of UAVs that did not collide in the last collision is taken as the low-altitude airspace capacity.

7. The low-altitude airspace capacity assessment method based on spatial parameter simulation-driven relationship coupling according to claim 1, characterized in that: In method S3, the multivariate relationship coupling module includes a generalized additive model and a multi-scale geographic weighted regression coupling model. The generalized additive model constructs a nonlinear relationship between spatial morphology indicators and low-altitude airspace capacity. The multi-scale geographic weighted regression model analyzes spatial heterogeneity to obtain the spatial relationship variation coefficient of spatial morphology indicators. The multivariate relationship coupling module uses the nonlinear relationship and the spatial relationship variation coefficient to construct the relationship coupling.

8. The low-altitude airspace capacity assessment method based on spatial parameter simulation-driven relationship coupling according to claim 1, characterized in that: A study area was selected and its spatial morphology data was acquired. A homogeneous fleet of UAVs of the same type was configured in the study area. The low-altitude airspace capacity assessment model output the low-altitude airspace capacity assessment results of the homogeneous fleet. The homogeneous fleet was defined as one of the following three cases: a full multi-rotor fleet in which all UAVs are multi-rotor UAVs, a full compound-wing fleet in which all UAVs are fixed-wing UAVs, or a full compound-wing fleet in which all UAVs are compound-wing UAVs. The low-altitude airspace capacity comparison data of the study area under different homogeneous fleet conditions were analyzed. Alternatively, the city building types and spatial morphology data in the simulation spatial morphology dataset can be adjusted and used as an experimental area to analyze the low-altitude airspace capacity comparison data of different city building types or different spatial morphology data under different isomorphic aircraft fleet conditions.

9. The low-altitude airspace capacity assessment method based on spatial parameter simulation-driven relationship coupling according to claim 1, characterized in that: The study area was selected and its spatial morphology data was obtained. Different configuration parameters were used to configure heterogeneous aircraft fleets in the study area. The configuration parameters of the heterogeneous aircraft fleets were the proportions of multi-rotor UAVs, fixed-wing UAVs and compound-wing UAVs. The low-altitude airspace capacity assessment model output the low-altitude airspace capacity assessment results under the heterogeneous aircraft fleet configuration parameters. The low-altitude airspace capacity comparison data of the study area under the heterogeneous aircraft fleet configuration parameters were analyzed. Alternatively, the city building types and spatial morphology data in the simulation spatial morphology dataset can be adjusted and used as an experimental area to analyze the low-altitude airspace capacity comparison data of different city building types or different spatial morphology data under heterogeneous fleet conditions with different ratio parameters.

10. The low-altitude airspace capacity assessment method based on spatial parameter simulation-driven relationship coupling according to claim 1 or 7, characterized in that: In method S3, before the relationship coupling identification, the multivariate relationship coupling module uses principal component analysis to reduce the spatial morphological indicators to obtain the spatial morphological indicators of the main influencing components to participate in the relationship coupling identification process.

Images