A method for dynamic matching of multi-resolution meteorological grid and wind resistance performance of unmanned aerial vehicle

Abstract

This invention discloses a dynamic matching method for multi-resolution meteorological grids and the wind resistance performance of unmanned aerial vehicles (UAVs). The method includes: establishing a wind resistance parameter database for each UAV model; correcting the nominal wind resistance threshold to an effective wind resistance threshold based on payload, battery level, altitude, gust factor, and mission safety level; performing a full-domain screening using a coarse-resolution grid, activating a fine-resolution grid when wind speed approaches the effective threshold, wind speed gradient is too large, or turbulence intensity exceeds the limit; calculating multi-dimensional adaptability scores for wind speed, crosswind, and turbulence for candidate UAV models and the grid, using the minimum score on the path as the safety bottleneck; selecting the UAV model and flight path combination that meets the threshold and has the optimal overall cost, and updating the parameter database through operational feedback. This invention achieves a balance between computational efficiency and safety accuracy, and significantly improves the accuracy and adaptability of multi-UAV scheduling.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) operational safety assessment technology, and in particular to a dynamic matching method between a multi-resolution meteorological grid and the wind resistance performance of a UAV. Background Technology

[0002] In recent years, with the rapid development of low-altitude applications such as drone logistics delivery, urban air traffic, and emergency rescue, the operational safety of drones in complex urban environments has received increasing attention. Low-altitude wind fields, characterized by uneven spatial distribution, rapid temporal changes, and significant influence from buildings and terrain, are among the most critical meteorological factors affecting drone flight safety. Different drone models exhibit significant differences in wind resistance due to variations in weight, wingspan, power configuration, and control systems. Even for the same model, the actual wind speed threshold it can withstand dynamically changes depending on payload weight, battery level, flight altitude, and control mode when performing different tasks. Therefore, matching appropriate flight plans to drones under different conditions based on real-time wind field conditions is a core issue in ensuring the operational safety of low-altitude drones.

[0003] Currently, a uniform meteorological threshold is commonly used as the judgment standard in drone route planning and flight safety assessment. This involves setting a fixed wind speed limit; all drones are considered flyable below this threshold and deemed flightless above it. This one-size-fits-all approach has significant shortcomings: on the one hand, for drones with strong wind resistance, a uniform conservative threshold may lead to unreasonable idleness of their capacity, reducing scheduling efficiency; on the other hand, for drones with heavy payloads or low battery levels, a uniform aggressive threshold may expose them to meteorological risks beyond their actual capacity. Some studies have attempted to incorporate drone model differences, such as configuring different wind speed limits for different models, but these studies mostly remain at the level of static parameters, failing to incorporate the real-time impact of dynamic factors such as payload changes, battery depletion, and altitude changes on effective wind resistance into the assessment system.

[0004] Meanwhile, the spatial resolution of meteorological data has a decisive impact on the accuracy of flight safety assessments. Currently, low-air meteorological grid data is mainly divided into two types: coarse-resolution grids (such as 50m×50m×30m or even coarser) have a wide coverage and low computational cost, suitable for judging large-scale meteorological trends, but cannot analyze local strong wind characteristics such as narrow wind channels between buildings, crosswind zones around bridges, and turbulent areas around takeoff and landing points; fine-resolution grids (such as 5m×5m×5m or even finer) can accurately depict local wind field changes, but deploying fine grids across the entire area will generate huge computational costs and data storage pressure, making real-time applications difficult in engineering. Existing flight route planning systems often use a fixed meteorological grid resolution; coarse-grid schemes are prone to missing local risk points, while fine-grid schemes are difficult to promote due to excessive computational resource consumption.

[0005] Furthermore, existing methods typically use average or maximum wind speed along a flight path to assess overall safety. This approach overlooks safety bottlenecks along the entire path—even if wind speeds are low for most of the path, the presence of strong winds or turbulence exceeding the aircraft's tolerance in any segment can pose unacceptable safety risks to the entire route. Currently available technologies lack effective solutions for identifying and quantifying these bottlenecks, which can be described as a "weakest link" in flight path safety.

[0006] In summary, existing UAV route planning and meteorological matching methods have the following shortcomings: (1) They do not establish a dynamic mapping relationship between the aircraft's wind resistance capability and real-time mission status; (2) They do not solve the balance problem between low-altitude meteorological grid resolution and computing resources, and lack a triggering mechanism for on-demand activation of fine grids; (3) They do not integrate multi-dimensional meteorological factors such as crosswinds and turbulence with wind speed into the aircraft-meteorological adaptability quantitative evaluation system; (4) The path safety assessment lacks the ability to identify the minimum safety margin bottleneck along the entire route. Therefore, there is an urgent need for an intelligent matching method that can dynamically select the meteorological grid resolution based on wind field changes and output a safe flight plan based on the aircraft's effective wind resistance threshold, in order to overcome the above-mentioned shortcomings. Summary of the Invention

[0007] The purpose of this invention is to provide a dynamic matching method between multi-resolution meteorological grids and the wind resistance performance of UAVs. By using a fine grid trigger function, an effective wind resistance threshold correction function, and a path adaptability score, safe matching of different UAV models under different flight path meteorological conditions can be achieved.

[0008] This invention is achieved using the following technical solution: a dynamic matching method between a multi-resolution meteorological grid and the wind resistance performance of a UAV, comprising the following steps: Establish a wind resistance parameter library for aircraft models to record the nominal wind resistance parameters of at least one UAV model, including the maximum allowable wind speed, the maximum allowable crosswind, and the maximum allowable turbulence intensity; The effective wind resistance threshold is corrected by adjusting the nominal wind resistance parameter to the effective wind resistance threshold under the current mission based on the current mission's load status, power status, flight altitude, gust factor, and mission safety level. A multi-resolution meteorological grid is acquired, and a coarse-resolution meteorological grid and a fine-resolution meteorological grid are maintained. The coarse-resolution meteorological grid covers the entire mission airspace, and the fine-resolution meteorological grid covers a local area in the mission airspace. Fine-grid activation judgment: Based on the effective wind resistance threshold and the wind speed information of the coarse-resolution meteorological grid, it is determined whether to activate the fine-resolution meteorological grid. Adaptability score calculation: For each candidate UAV model and each weather grid combination, an adaptability score is calculated. The adaptability score is used to characterize the flight safety margin of the model under the weather grid conditions. The route and aircraft type recommendation is based on the adaptability scores of each meteorological grid on the candidate flight path, and selects the combination of UAV type and flight path that meets the safety threshold and has the best overall cost.

[0009] Furthermore, in the effective wind resistance threshold correction step, the effective wind resistance threshold is calculated according to the following formula: Veff = V0 · κload · κsoc · κalt · κctrl - Δgust - Δsafe Where V0 is the nominal wind resistance threshold, κload is the load correction factor, κsoc is the power correction factor, κalt is the height correction factor, κctrl is the control margin factor, Δgust is the gust margin, and Δsafe is the safety margin.

[0010] Furthermore, the load correction coefficient κload is calculated according to κload = 1 - α·(mload / mmax), where mload is the current load mass, mmax is the maximum allowable load mass, and α is the load influence factor; The power correction coefficient κsoc is calculated according to κsoc = β + γ·(SOC / SOCmax), where SOC is the current remaining battery power, SOCmax is the full battery power, and β and γ are parameters affecting the power. The altitude correction coefficient κalt is calculated according to κalt = 1 - δ·h, where h is the flight altitude and δ is the altitude influence factor; The gust margin Δgust and safety margin Δsafe are preset constants or adjustable parameters that are dynamically configured according to the mission safety level.

[0011] Furthermore, in the fine-grid activation determination step, the fine-resolution meteorological grid is activated when at least one of the following conditions is met: Condition 1: The absolute value of the difference between the effective wind resistance threshold of any candidate aircraft model under the current coarse resolution meteorological grid and the coarse grid wind speed is less than the first trigger threshold; Condition 2: The spatial gradient of wind speed in the current coarse-resolution meteorological grid is greater than the second trigger threshold; Condition 3: The turbulence intensity of the current coarse-resolution meteorological grid is greater than the third trigger threshold.

[0012] Furthermore, in the fine mesh activation determination step, the fine mesh activation flag Ag is determined according to the following activation function: Ag = 1, when minm(Veff^m - Vg) < M0 or ||∇Vg|| > G0 or Kg > K0; Where m is the candidate aircraft model number, Veff^m is the effective wind resistance threshold of the m-th aircraft model, Vg is the wind speed of the coarse-resolution meteorological grid, ∇Vg is the wind speed spatial gradient of the coarse-resolution meteorological grid, Kg is the turbulence intensity of the coarse-resolution meteorological grid, M0 is the first trigger threshold, G0 is the second trigger threshold, and K0 is the third trigger threshold; Ag = 1 indicates that the fine-resolution meteorological grid is activated, and Ag = 0 indicates that it is not activated.

[0013] Furthermore, in the adaptability score calculation step, the adaptability score S{m,g} of the m-th candidate aircraft model in the g-th meteorological grid is calculated according to the following formula: S{m,g} = qg · min{(Veff^m - Vg) / Veff^m, (Leff^m - Lg) / Leff^m, (K_eff^m - Kg) / Keff^m}; Wherein, Veff^m is the effective wind resistance threshold of the m-th aircraft type, Vg is the wind speed of the g-th meteorological grid, Leff^m is the effective crosswind resistance threshold of the m-th aircraft type, Lg is the crosswind speed of the g-th meteorological grid, Keff^m is the effective turbulence resistance threshold of the m-th aircraft type, Kg is the turbulence intensity of the g-th meteorological grid, and qg is the data reliability coefficient of the g-th meteorological grid; when the adaptability score S{m,g} is negative, it indicates that the aircraft type does not have safe flight conditions under the meteorological grid conditions.

[0014] Furthermore, in the path and aircraft recommendation step, the path adaptability score of the candidate path π is taken as the minimum value of the adaptability scores of all meteorological grids on that path, that is: Sπ = min{g∈π} S{m,g}; Select the aircraft type and route combination that satisfies Sπ ≥ Smin and has the lowest overall cost (π), where Smin is a preset minimum safety threshold, and the overall cost (π) includes at least one of flight distance cost, flight time cost, and energy consumption cost.

[0015] Furthermore, in the adaptive score calculation step, when the fine-resolution meteorological grid is not activated, the adaptive score of each candidate aircraft model is calculated only based on the coarse-resolution meteorological grid; when the fine-resolution meteorological grid is activated, the adaptive score of the activated area is switched to be calculated based on the fine-resolution meteorological grid, while the inactive area is still calculated based on the coarse-resolution meteorological grid, forming an adaptive score matrix that combines coarse and fine grids.

[0016] Furthermore, it also includes a feedback update step: after each flight mission, the aircraft type information, payload information, flight altitude, weather grid data experienced, flight attitude correction range, and energy consumption anomaly data for that flight are recorded; if the same aircraft type experiences a predetermined number of attitude protection triggers or energy consumption anomalies under the same weather grid conditions, the effective wind resistance margin of the aircraft type under the same weather grid conditions is reduced by a predetermined percentage, and the reduced effective wind resistance margin is stored in the aircraft type wind resistance parameter library for subsequent matching.

[0017] Furthermore, in the effective wind resistance threshold correction step, the control margin coefficient κctrl is determined according to the flight control mode: κctrl takes a first value in manual control mode and a second value in automatic control mode, and the first value is less than the second value; the mission safety level includes at least low risk level, medium risk level and high risk level, and the safety margin Δsafe increases as the mission safety level increases.

[0018] The beneficial effects of this invention are as follows: This invention achieves several beneficial technical effects by establishing an intelligent matching mechanism between multi-resolution meteorological grids and the dynamic wind resistance capabilities of UAVs: First, it abandons the traditional one-size-fits-all approach to wind speed threshold judgment and incorporates real-time factors such as load, battery power, flight altitude, control mode, gusts, and safety margin into the evaluation system using an effective wind resistance threshold correction formula. This allows for precise quantification of the safety boundaries of different aircraft types under different mission states, significantly improving the accuracy and flexibility of multi-aircraft scheduling decisions. Second, by using a fine-grid activation trigger function, the high-resolution meteorological grid is activated only as needed in high-risk areas such as wind speeds approaching the threshold, excessive wind speed gradients, or turbulence intensity exceeding the standard. This avoids the huge computational cost of a full-domain fine-grid approach and overcomes the shortcomings of simple coarse-grid approaches that miss local risks such as strong winds between buildings, crosswinds in bridge areas, and turbulence at take-off and landing points, achieving the best balance between computational efficiency and safety accuracy.

[0019] This invention employs an adaptive scoring mechanism based on the minimum margin principle, uniformly quantifying wind speed margin, crosswind margin, and turbulence margin. It uses the lowest score among all grids along the path as the safety bottleneck for the entire route, thereby accurately identifying the weakest link in the route and effectively preventing safety misjudgments caused by local risk points being masked by averages or maximum values. Finally, through a post-operation feedback update mechanism, the system can continuously optimize effective wind resistance margin parameters based on actual flight data. After long-term operation, it can adaptively improve scheduling accuracy and safety, providing reliable technical support for multi-aircraft mixed operations in complex urban low-altitude environments. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0021] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0023] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0024] The following detailed description of some embodiments of the present invention is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0025] Example 1 See Figure 1 A dynamic matching method between multi-resolution meteorological grids and the wind resistance performance of UAVs includes the following steps: Establish a wind resistance parameter library for aircraft models to record the nominal wind resistance parameters of at least one UAV model, including the maximum allowable wind speed, the maximum allowable crosswind, and the maximum allowable turbulence intensity; The effective wind resistance threshold is corrected by adjusting the nominal wind resistance parameter to the effective wind resistance threshold under the current mission based on the current mission's load status, power status, flight altitude, gust factor, and mission safety level. A multi-resolution meteorological grid is acquired, and a coarse-resolution meteorological grid and a fine-resolution meteorological grid are maintained. The coarse-resolution meteorological grid covers the entire mission airspace, and the fine-resolution meteorological grid covers a local area in the mission airspace. Fine-grid activation judgment: Based on the effective wind resistance threshold and the wind speed information of the coarse-resolution meteorological grid, it is determined whether to activate the fine-resolution meteorological grid. Adaptability score calculation: For each candidate UAV model and each weather grid combination, an adaptability score is calculated. The adaptability score is used to characterize the flight safety margin of the model under the weather grid conditions. The route and aircraft type recommendation is based on the adaptability scores of each meteorological grid on the candidate flight path, and selects the combination of UAV type and flight path that meets the safety threshold and has the best overall cost.

[0026] In this embodiment, a wind resistance parameter database for each aircraft model can be established first through an operating platform (system). Taking a city-level drone logistics delivery platform as an example, this platform manages three types of aircraft models simultaneously: light multi-rotor drones (such as the DJI M300RTK), medium-sized multi-rotor drones (such as the DJI M350), and compound-wing drones (such as a certain model of vertical take-off and landing fixed-wing drone). Platform maintenance personnel enter the nominal wind resistance parameters of each model into the parameter database by consulting the flight manuals provided by the manufacturer or by using wind tunnel test data, as shown in Table 1. Table 1 Nominal wind resistance parameters Meanwhile, the platform configures a default mission safety level for each aircraft model in the parameter library. For example, logistics delivery missions are set to a medium-risk level by default, urban air taxi missions to a high-risk level, and agricultural plant protection missions to a low-risk level. The default safety margin Δ_safe for each safety level is: 1.0 m / s for low-risk level, 1.5 m / s for medium-risk level, and 2.0 m / s for high-risk level. These parameters can be dynamically adjusted based on actual flight data after the system is running.

[0027] When the drone operation platform receives a new flight mission, the system first performs an effective wind resistance threshold correction step. Assume the platform currently has a DJI M300 RTK drone performing a city delivery mission. The specific status parameters of this mission are: current payload mass m_load = 2.0 kg (near full load), remaining battery SOC = 85%, planned flight altitude h = 120 m, automatic control mode, and a medium-risk mission safety level. Current weather forecast indicates gusts, with a gust margin Δgust = 2.5 m / s.

[0028] The system reads the nominal wind resistance threshold V0 = 12 m / s for this model from the model's wind resistance parameter database, and then calculates various correction factors based on the current mission status: Load correction factor κload: calculated according to κload = 1 - α·(mload / m_max). For multi-rotor aircraft, the load influence factor α is empirically taken as 0.15, then κload = 1 - 0.15×(2.0 / 2.7) = 1 - 0.15×0.74 = 1 - 0.111 = 0.889.

[0029] Battery level correction factor κsoc: Calculated according to κsoc = β + γ·(SOC / SOCmax). For this model, β is 0.85, γ is 0.15 (i.e., the correction factor is 1.0 when fully charged and a minimum of 0.85 when empty), and SOCmax = 100%, then κsoc = 0.85 + 0.15×(85 / 100) = 0.85 + 0.1275 = 0.9775.

[0030] Altitude correction factor κalt: Calculated according to κalt = 1 - δ·h, the empirical value of the altitude influence factor δ is 0.002 / m (corresponding to a reduction of approximately 24% in wind resistance at a height of 120m), and h is in meters. Therefore, κalt = 1 - 0.002×120 = 1 - 0.24 = 0.76.

[0031] Control margin coefficient κctrl: This UAV adopts automatic control mode, and the control system has a fast response speed, so κctrl is set to 1.0. If it is in manual control mode, κctrl is set to 0.95 to reflect the response delay and error of manual operation.

[0032] Gust margin Δgust: Based on the gust factor in the weather forecast, the system queries the preset gust margin table. When the gust factor (the ratio of gust speed to average wind speed) is in the range of 1.3 to 1.5, Δgust is taken as 2.5 m / s. In this embodiment, the gust factor is 1.4, so Δgust = 2.5 m / s.

[0033] Safety margin Δsafe: The safety level of the task in this embodiment is medium risk level, Δsafe = 1.5 m / s.

[0034] Substitute into the effective wind resistance threshold calculation formula: Veff = V0 · κload · κsoc · κalt · κctrl - Δgust - Δsafe = 12 × 0.889 × 0.9775 × 0.76 × 1.0 - 2.5 - 1.5 = 12 × 0.660 - 4.0 = 7.92 - 4.0 = 3.92 m / s Calculation results show that although the nominal wind resistance threshold of this model is 12 m / s, under the current conditions of near full load, low to medium battery power, 120 meters altitude, and gusts, its effective wind resistance threshold is only 3.92 m / s, a significant reduction. The system will use this dynamically corrected effective wind resistance threshold, Veff = 3.92 m / s, as the basis for subsequent meteorological matching and safety judgments, rather than directly using the static nominal value of 12 m / s.

[0035] Similarly, the system can use the same correction coefficient framework to calculate the effective crosswind resistance threshold Leff and the effective turbulence resistance threshold Keff, the difference being that their respective nominal reference values ​​and gust margins contain different specific components related to crosswind or turbulence. For simplicity, the following embodiments use wind speed as the main parameter, and the calculation logic for crosswind and turbulence is completely analogous.

[0036] In this embodiment, the system maintains two sets of data simultaneously: a coarse-resolution meteorological grid and a fine-resolution meteorological grid. The data sources for the meteorological grid management module include: numerical weather prediction products from meteorological forecasting centers (such as the output of the WRF model), measured data from miniature meteorological sensors deployed on the rooftops and lampposts of urban buildings, and real-time meteorological observation data transmitted back by the UAV itself during flight. The system uses data assimilation technology to fuse multi-source meteorological data into a unified grid structure.

[0037] A coarse-resolution meteorological grid covers the entire mission airspace. In this embodiment, a grid resolution of 50m (horizontal) × 50m (horizontal) × 30m (vertical) is used. The airspace for urban low-altitude flight missions typically covers an area of ​​approximately 5km × 5km, corresponding to a coarse grid of approximately 100 × 100 × 5 = 50,000 grid cells (vertically from the ground to a height of 150m, divided into approximately 5 layers). This amount of data can be updated in seconds on existing computing platforms.

[0038] The fine-resolution meteorological grid uses a grid resolution of 5m (horizontal) × 5m (horizontal) × 5m (vertical), which requires approximately 1000 times more computation than the coarse grid. Therefore, it is activated only in high-risk local areas identified by the coarse grid. The system establishes a spatial index relationship between the fine-grid data and the coarse-grid data through a geographic information system, ensuring that the fine-grid data can quickly cover the specified local area when activated.

[0039] In this embodiment, after the effective wind resistance threshold correction is completed, the system executes the fine grid activation judgment step. For ease of understanding, the triggering process of the three activation conditions is explained below through three typical scenarios.

[0040] Scenario A: Wind speed approaches wind resistance threshold, triggering fine mesh. Suppose the candidate aircraft for the flight mission received by the platform is a DJI M300 RTK, whose effective wind resistance threshold has been corrected to Veff = 3.92 m / s. The system first traverses all coarse-resolution grids within the mission area, extracting the wind speed Vg for each grid. When the system scans a coarse grid around a certain takeoff and landing point, it finds that the wind speed in that grid is Vg = 2.5 m / s, and calculates Veff - Vg = 3.92 - 2.5 = 1.42 m / s. This value is less than the system's preset first trigger threshold M0 = 3 m / s, indicating that the current wind speed is already close to the effective wind resistance threshold of this aircraft type. Since the spatial resolution of the coarse grid is insufficient to resolve the true wind speed distribution between buildings around the takeoff and landing point—the 2.5 m / s measured by the coarse grid may mask the actual strong winds caused by the local funneling effect between buildings—the system determines that it is necessary to activate the fine-resolution meteorological grid in this area for a refined evaluation.

[0041] Scenario B: Excessive spatial gradient of wind speed triggers fine mesh Suppose that when the system scans a coarse-grid area above a bridge spanning a city river, although the average wind speed in this area is 3.0 m / s, which is not lower than the effective wind resistance threshold of any candidate aircraft model, the system calculates the spatial gradient of wind speed between the coarse grid and its adjacent grids. It finds that the gradient value ||∇Vg|| = 2.0 m / s / m, exceeding the preset second trigger threshold G0 = 1.5 m / s / m. Based on fluid dynamics principles, the system determines that such a large spatial wind speed gradient often indicates the presence of wind shear or local acceleration effects in the area, and the average wind speed of the coarse grid may significantly deviate from the actual wind speed at a point on the flight path. Therefore, the system activates a fine-resolution meteorological grid for the bridge area to accurately analyze the continuous spatial variation of wind speed.

[0042] Scenario C: Turbulence intensity exceeding the limit triggers fine mesh Suppose that when the system scans a narrow airflow area between a cluster of tall buildings, the coarse grid reports a turbulence intensity of Kg = 1.8 m in that area. 2 / s 3 This exceeds the preset third trigger threshold K0 = 1.5 m. 2 / s 3 Excessive turbulence intensity indicates turbulent airflow in the region, with strong vortices and pulsations. Even with low average wind speeds, the instantaneous loads caused by turbulence could exceed the UAV's attitude control capabilities. The system immediately activates a fine-resolution meteorological grid for the region.

[0043] When any of the above conditions are met, the fine-mesh activation flag Ag = 1. The system replaces the coarse-mesh data of the corresponding area with fine-mesh data, and all subsequent adaptive score calculations are based on the fine-mesh. In a specific scenario of this embodiment, the above three activation conditions can occur simultaneously. For example, in a certain task, the system's coarse screening found that the wind speed in a densely built-up area was 2.8 m / s, the wind speed in the neighboring grid was 4.5 m / s (gradient 1.7 m / s / m), and the turbulence intensity was 1.9 m. 2 / s 3 The effective threshold for light aircraft is only 3.9 m / s. When all three conditions are met, the fine grid activation signal is reliably triggered. After the system evaluates the area at a resolution of 5m, it finds a local jet core area with a wind speed of 5.2 m / s at the actual flight path altitude. If it is only based on the average value of 2.8 m / s of the coarse grid, it will be mistakenly judged as safe.

[0044] In this embodiment, after the fine grid is activated, the system proceeds to the adaptability score calculation step. Taking the bridge area in Embodiment 4 as an example, after the fine grid is activated in this area, it contains multiple fine grid cells. The system needs to calculate the adaptability score for each candidate aircraft type and each meteorological grid combination in this area.

[0045] Assume the bridge area has three fine grid cells, g1, g2, and g3, corresponding to three waypoints on the flight path. The meteorological parameters for each fine grid cell are shown in the table below: Table 2 Meteorological Parameters

[0046] The platform currently has two candidate drones: Drone A is a DJI M300 RTK, with an effective wind resistance threshold Veff = 3.92 m / s, an effective crosswind resistance threshold Leff = 2.6 m / s (obtained by subtracting the crosswind margin after correcting the nominal 8 m / s with the same correction factor), and an effective turbulence resistance threshold Keff = 0.85 m. 2 / s 3 Drone B is a DJI M350, with an effective wind resistance threshold Veff = 6.8 m / s, an effective crosswind resistance threshold Leff = 4.5 m / s, and an effective turbulence resistance threshold Keff = 1.2 m. 2 / s 3 .

[0047] Taking drone A in grid g1 as an example, calculate the fitness score: First, calculate the three margin ratios: Wind speed margin ratio = (Veff - Vg) / Veff = (3.92 - 4.2) / 3.92 = -0.071 (a negative value, indicating that the wind speed has exceeded the effective wind resistance threshold); Crosswind margin ratio = (Leff - Lg) / Leff = (2.6 - 3.5) / 2.6 = -0.346; Turbulence margin ratio = (Keff - Kg) / Keff = (0.85 - 1.6) / 0.85 = -0.882; Taking the minimum value among the three, min = -0.882, and multiplying it by the data confidence level qg = 0.95, we get: S{A,g1} = 0.95 × (-0.882) = -0.838.

[0048] A negative score indicates that drone A does not meet the conditions for safe flight under grid g1. The weakest of the three risk factors is turbulence intensity—even if the wind speed and crosswind are acceptable, the turbulence has seriously exceeded the standard.

[0049] Let's take drone B in grid g1 as an example again: Wind speed margin ratio = (6.8 - 4.2) / 6.8 = 0.382 Crosswind margin ratio = (4.5 - 3.5) / 4.5 = 0.222 Turbulence margin ratio = (1.2 - 1.6) / 1.2 = -0.333 min = -0.333, S{B,g1} = 0.95 × (-0.333) = -0.316 The same negative value, but close to zero, indicates that drone B is also on the edge of safety in this grid, mainly due to excessive turbulence.

[0050] For drone B in grid g3: Wind speed margin ratio = (6.8 - 3.5) / 6.8 = 0.485 Crosswind margin ratio = (4.5 - 4.0) / 4.5 = 0.111 Turbulence margin ratio = (1.2 - 1.4) / 1.2 = -0.167 min = -0.167, S{B,g3} = 0.95 × (-0.167) = -0.159, which is still negative.

[0051] The above calculations show that, within this fine-grid area, due to the generally high turbulence intensity, the adaptability scores of both UAVs in this area are negative. The system will issue an alarm that there are no safe aircraft types in this area under the current conditions and suggest that the mission be postponed or rerouted.

[0052] In another scenario, assuming the same aircraft is lightly loaded, fully charged, and at low altitude when performing a mission, and the effective wind resistance threshold is increased to Veff = 7.5 m / s, then its score in the g1 grid is positive and greater than Smin (0.2), and the system can determine that it is flyable.

[0053] In this embodiment, after completing the adaptability scoring of all candidate models and grids on all paths, the system proceeds to the path and model recommendation step.

[0054] Suppose a delivery task has three candidate paths π1, π2, and π3 from origin O to destination D, and the platform has two candidate aircraft types, M300 and M350. The system has discretized each path into several waypoints, with each waypoint corresponding to a weather grid (coarse or fine). For each path, the system calculates a path fitness score—taking the minimum fitness score of all grids along that path.

[0055] For example, path π1 passes through a densely built-up area. Although most segments have high scores, when crossing the wind tunnel between two tall buildings, one grid has insufficient crosswind margin, resulting in a score of S = 0.15 for that grid, which is lower than Smin = 0.2. Therefore, the path fitness score of this path is Sπ1 = 0.15, and it is directly excluded.

[0056] Path π2 has a score of no less than 0.25 for all grids, but its flight distance is longer, resulting in a higher overall cost (Cost(π2)). Path π3 has a score of no less than 0.22 for all grids, a shorter flight distance, and a lower overall cost (Cost(π3)). The system selects paths according to the optimization objective (m^, π^) = argmax{m,π} [min{g∈π} S{m,g} - λ·Cost(π)]. In this embodiment, λ is set to 0.01 (cost weighting coefficient), path π3 has a score of 0.22 and a cost of 10 (based on flight distance in kilometers), while path π2 has a score of 0.25 and a cost of 15. For the same aircraft type: The score for path π2 is 0.25 - 0.01 × 15 = 0.10; The score for path π3 is 0.22 - 0.01 × 10 = 0.12; The system ultimately selected path π3, although its minimum adaptability score was slightly lower than that of path π2, it was the optimal choice after considering flight costs. Simultaneously, the system also included safety bottleneck markers for this path in the recommended results—that is, the grid location of the lowest score and the corresponding meteorological elements (e.g., excessive crosswinds at waypoint 12)—for operations dispatchers or pilots to refer to. If necessary, the system can combine the bottleneck markers on the display interface with manual adjustments to the flight path to avoid the bottleneck.

[0057] If both aircraft types on the same route satisfy Sπ ≥ S_min, the system will prioritize recommending the aircraft type with the higher adaptability score (i.e., the aircraft type with a larger safety margin), and provide reasons for the recommendation, such as the M350 aircraft having a 40% higher anti-turbulence margin than the M300 on flight segment 3.

[0058] Example 2 This embodiment fully demonstrates how the system dynamically adjusts the effective wind resistance threshold and switches the grid resolution based on changes in load and power during a delivery task where the outbound journey is fully loaded and the return journey is empty.

[0059] Outbound phase: A certain drone (DJI M300 RTK) performs a delivery mission from delivery station A to user B. At takeoff, mload = 2.5 kg (close to full load of 2.7 kg), SOC = 95%, flight altitude h = 100 m, automatic control mode, mission safety level medium risk (Δsafe = 1.5 m / s), gust margin Δgust = 2.5 m / s.

[0060] Calculate the effective wind resistance threshold: κload = 1 - 0.15×(2.5 / 2.7) = 1 - 0.139 = 0.861 κsoc = 0.85 + 0.15×(95 / 100) = 0.993 κalt = 1 - 0.002 × 100 = 0.80 κctrl = 1.0 Veff = 12 × 0.861 × 0.993 × 0.80 × 1.0 - 2.5 - 1.5 = 12 × 0.684 - 4.0 = 8.21 - 4.0 = 4.21 m / s The system performed a coarse-grid screening with Veff = 4.21 m / s and found that the wind speed around delivery station A was 2.8 m / s. Since Veff - Vg = 1.41 m / s < M0 = 3 m / s, the fine-grid area around the takeoff and landing point was activated. The fine-grid assessment showed a sudden increase in wind speed (locally reaching 4.5 m / s, exceeding Veff = 4.21 m / s) caused by building bypass at a height of 5m above the ground. The system determined that this takeoff and landing point was currently unsuitable for full-load takeoff and recommended a 3-minute delay until the wind speed decreased or the use of a medium-sized aircraft with stronger wind resistance.

[0061] Return trip phase: The drone returns empty from user B to delivery station A, with mload = 0.5 kg (including packaging weight), SOC = 60% (battery consumed during the return trip), and the flight altitude remains 100 m.

[0062] Calculate the effective wind resistance threshold for the return journey: κload = 1 - 0.15×(0.5 / 2.7) = 1 - 0.028 = 0.972 κsoc = 0.85 + 0.15×(60 / 100) = 0.94 κalt = 0.80 κctrl = 1.0 Veff = 12 × 0.972 × 0.94 × 0.80 × 1.0 - 2.5 - 1.5 = 12 × 0.731 - 4.0 = 8.77 - 4.0 = 4.77 m / s Although the reduced power caused κsoc to decrease from 0.993 to 0.94, the load was significantly reduced (κload increased from 0.861 to 0.972), and the overall effective wind resistance threshold still improved from 4.21 m / s to 4.77 m / s. The system re-screened the coarse grid. At this time, the wind speed around the take-off and landing points was still 2.8 m / s, and Veff - Vg = 1.97 m / s was still less than M0 = 3 m / s, so the fine grid remained active. However, the fine grid showed that the local wind speed in the area had decreased from 4.5 m / s to 3.5 m / s, which was lower than Veff = 4.77 m / s, so the system determined that the return trip could be safely executed.

[0063] If the delivery mission uses the same route for both outbound and return trips, the system's recommended aircraft type may be completely different for the outbound and return journeys. For the outbound journey, the system might recommend upgrading to an M350 or adjusting the flight time, while for the return journey, the system would determine that an M300 is flyable. In this embodiment, during the outbound phase, Smin=0.2. The M300's score in the fine grid at the takeoff and landing point is less than 0.2 due to a local wind speed of 4.5 m / s, so the system temporarily does not recommend flight. After waiting 5 minutes, the local wind speed at the takeoff and landing point drops to 4.0 m / s, the score is recalculated and returns to 0.22, at which point the alarm is deactivated and flight is permitted. This demonstrates the system's responsiveness to the dissipation of local turbulence at the takeoff and landing point.

[0064] Example 3 This embodiment details the process of fine-grid activation and flight path adjustment in special geographical environments.

[0065] A cross-river delivery route needs to cross a 500m long urban road bridge. A coarse grid (50m resolution) shows an average wind speed of Vg = 4.0 m / s in the bridge area, an effective wind resistance threshold of Veff = 6.5 m / s for candidate aircraft, and Veff - Vg = 2.5 m / s > M0 (3 m / s), suggesting the wind speed conditions are safe. However, when calculating the wind speed gradient, the system found that the wind speed difference between the two ends of the bridge and the center of the bridge deck is 1.5 m / s, with a grid spacing of 50m. The gradient value ||∇Vg|| = 1.5 / 50 = 0.03 m / s / m, which does not exceed G0 = 1.5 m / s / m, and the turbulence intensity is also within acceptable limits. However, the system, through a pre-stored geographic information layer, identifies the bridge as belonging to an open wind corridor terrain—with open water below the bridge deck, no obstructions on either side, and the bridge structure potentially generating an airflow acceleration effect. Based on the historical accident area triggering rules configured in the system (an alternative extension scheme for fine mesh activation conditions), the system still activated the fine mesh in the bridge area.

[0066] Fine-grid (5m resolution) data reveals wind field characteristics significantly different from those of the coarse-grid data: at a height of 10m above the central bridge deck, due to changes in the wind profile between the bridge deck and the river surface, as well as the compression effect of the bridge structure on the airflow, the actual wind speed reaches 5.8 m / s, and there is a significant vertical wind shear (the wind speed difference between 5m and 15m above the bridge deck reaches 1.5 m / s). More importantly, downstream of bridge railings and light poles, there is turbulence intensity of Kg = 2.2 m. 2 / s 3 The vortex region is much larger than the Kg = 1.2 m reported by the coarse mesh. 2 / s 3 .

[0067] The system recalculated the adaptability score based on the fine-grid data: at the central height level of the bridge deck (10m), the candidate aircraft's wind speed margin ratio = (6.5 - 5.8) / 6.5 = 0.108, crosswind margin ratio = (4.2 - 3.8) / 4.2 = 0.095, and turbulence margin ratio = (1.0 - 2.2) / 1.0 = -1.2, resulting in a negative score and a judgment of unsafety. However, the system also found that at a height level of 20m above the bridge deck, the wind speed decreased to 4.5 m / s, and the turbulence intensity dropped to 1.3 m. 2 / s 3 All margins met the safety thresholds. Based on this, the system recommended that the UAV increase its flight altitude from the planned 10m to 20m when crossing the bridge (altitude adjustment operation), and suggested a slight course deviation to avoid the turbulent wake area downstream of the light pole. The path adaptability score of the path after altitude adjustment is Sπ = 0.35 (greater than Smin = 0.2). The system marked this scheme as recommended and attached an altitude adjustment command, which the flight control system can use to automatically execute the climb before reaching the bridge.

[0068] Example 4 This embodiment details the complete implementation process of the run feedback update step.

[0069] After three months of operation, a logistics and delivery platform has accumulated a large amount of flight data. Following the information format recorded after each flight mission, each record includes: mission ID, date and time, aircraft type, payload mass, average SOC, flight altitude, the sequence of meteorological grid IDs traversed (including wind speed, crosswind, and turbulence data for each grid), the attitude correction magnitude (correction standard deviations for roll, pitch, and yaw angles) recorded by the IMU (Inertial Measurement Unit) during the flight, and the percentage of energy consumption deviating from the expected value.

[0070] The system operation feedback module performs statistical analysis on the above data. A query was performed using the M300 aircraft under the conditions of Kg ∈ [1.5, 1.8] turbulence range, Vg ∈ [4.0, 5.0] wind speed range, and altitude 80-120m. The results show that during three months of operation, the M300 flew 84 sorties under these grid conditions. Of these, 12 sorties triggered the flight control system's attitude protection (i.e., the attitude correction amplitude exceeded the preset safety threshold), a trigger rate of 14.3%; 8 sorties showed energy consumption anomalies exceeding expectations by more than 20%. The system determined that this trigger rate exceeded the preset alarm threshold (10%), indicating that the effective wind resistance margin setting under these weather conditions was overly optimistic.

[0071] According to the feedback update rules, the system lowers the effective wind resistance margin of the M300 by 10% as a conservative correction under this type of grid condition. Specifically, the original effective wind resistance threshold under this condition, based on Veff = 6.0 m / s (after correction), is reduced by 10% to a new effective wind resistance threshold of Veff_new = 6.0 × 0.9 = 5.4 m / s. The system stores this correction value in the operating correction parameter sub-table of the aircraft's wind resistance parameter library, and records the reason for the correction (high proportion of attitude protection triggered in turbulent intervals) and the correction timestamp.

[0072] In subsequent matching, when the system re-evaluates the M300's flight mission under similar grid conditions (Kg ∈ [1.5, 1.8], Vg ∈ [4.0, 5.0], altitude 80-120m), it will automatically use the downgraded Veffnew = 5.4 m / s for adaptive score calculation. If a candidate grid has a wind speed of 4.8 m / s, the original margin of 6.0 - 4.8 = 1.2 m / s seems safe. After correction, the margin becomes 5.4 - 4.8 = 0.6 m / s. If the score is lower than Smin after deducting the preset safety margin, the system will issue an unsafe judgment, effectively avoiding the previous problem of frequent attitude protection triggers in this weather envelope. With the continuous accumulation of operational feedback, the system's effective wind resistance margin estimation for various aircraft types in different weather envelopes becomes increasingly accurate, significantly improving overall scheduling safety.

[0073] Example 5 This embodiment demonstrates how the system integrates all the aforementioned steps to complete a full task allocation when managing multiple machine models simultaneously.

[0074] At 10:00 AM one morning, the platform simultaneously received three delivery tasks, each located in a different area of ​​the city. The platform's currently available fleet includes: 3 M300 aircraft (in varying conditions), 2 M350 aircraft, and 1 compound airfoil aircraft. The system performs the following processing for each task: Task 1 (Densely Built-up Area in Core Business District): Candidates: M300 (current Veff = 3.5 m / s, due to high load), M350 (Veff = 5.2 m / s), and compound wing (Veff = 7.8 m / s). Coarse-grid screening revealed multiple areas with excessive wind speed gradients in the task area, activating the fine-grid. Fine-grid evaluation showed generally high turbulence intensity in densely built-up areas. M300 scored negative in 3 waypoints (not safe), M350 scored negative in 2 waypoints, and the compound wing scored ≥0.25 at all waypoints. The system recommends the compound wing perform this task, specifying that it fly along a specific road when traversing the densely built-up area (the wind corridor formed by this road has lower turbulence intensity).

[0075] Task 2 (Open Suburban Straight Route): The coarse grid shows uniform wind speed (around 3.0 m / s) along the entire route, with no activation conditions required. The fine grid requires no activation. The effective wind resistance thresholds for the three aircraft models are 3.5, 5.2, and 7.8 m / s, respectively, all significantly higher than 3.0 m / s. The system calculates that the path adaptability scores for each aircraft model are significantly higher than Smin. Therefore, based on the principle of optimal overall cost, the system selects the M300—which has the lowest energy consumption and the route is simple—and ultimately recommends the M300 to execute Task 2.

[0076] Task 3 (Cross-river Bridge and Urban Area Mixed Route): The first half of the route is open river surface (fine grid triggered by bridge areas), and the second half is urban area. The system recommends a composite wing or M350 (with strong wind resistance) for the first half, but both types are currently occupied by Task 1; the M300 has a negative waypoint score in the bridge area section of the first half, failing to meet safety requirements. The system adopts multi-task joint optimization: Task 1 is changed to be performed by the M350 (the M350 only fails to meet certain waypoint requirements in Task 1, which can be resolved by adjusting the altitude), the composite wing is released to perform Task 3, and the M300 performs Task 2. After joint optimization, all three tasks obtained optimal solutions that meet the safety thresholds.

[0077] This embodiment demonstrates that the present invention can not only recommend the optimal combination of aircraft type and flight route for a single mission, but also achieve safe and efficient allocation of fleet resources through joint optimization in multi-mission scenarios. The final scheduling scheme output by the system includes the flight path of each UAV, altitude instructions for each segment, heading deflection suggestions, and safety bottleneck point markers along the entire path, which the flight control system executes accordingly. If sudden changes in weather conditions occur during flight, the system can re-execute the entire process in real time to achieve dynamic replanning. After the mission is completed, all flight data is absorbed by the feedback module for parameter correction in the next round.

[0078] For the foregoing embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to this application.

[0079] The above embodiments describe the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Modifications and variations made by those skilled in the art without departing from the spirit and scope of the invention should be within the protection scope of the appended claims.

Claims

1. A dynamic matching method between a multi-resolution meteorological grid and the wind resistance performance of a UAV, characterized in that, Includes the following steps: Establish a wind resistance parameter library for aircraft models to record the nominal wind resistance parameters of at least one UAV model, including the maximum allowable wind speed, the maximum allowable crosswind, and the maximum allowable turbulence intensity; The effective wind resistance threshold is corrected by adjusting the nominal wind resistance parameter to the effective wind resistance threshold under the current mission based on the current mission's load status, power status, flight altitude, gust factor, and mission safety level. A multi-resolution meteorological grid is acquired, and a coarse-resolution meteorological grid and a fine-resolution meteorological grid are maintained. The coarse-resolution meteorological grid covers the entire mission airspace, and the fine-resolution meteorological grid covers a local area in the mission airspace. Fine-grid activation judgment: Based on the effective wind resistance threshold and the wind speed information of the coarse-resolution meteorological grid, it is determined whether to activate the fine-resolution meteorological grid. Adaptability score calculation: For each candidate UAV model and each weather grid combination, an adaptability score is calculated. The adaptability score is used to characterize the flight safety margin of the model under the weather grid conditions. The route and aircraft type recommendation is based on the adaptability scores of each meteorological grid on the candidate flight path, and selects the combination of UAV type and flight path that meets the safety threshold and has the best overall cost.

2. The dynamic matching method according to claim 1, characterized in that, In the effective wind resistance threshold correction step, the effective wind resistance threshold is calculated according to the following formula: Veff = V0 · κload · κsoc · κalt · κctrl - Δgust - Δsafe Where V0 is the nominal wind resistance threshold, κload is the load correction factor, κsoc is the power correction factor, κalt is the height correction factor, κctrl is the control margin factor, Δgust is the gust margin, and Δsafe is the safety margin.

3. The dynamic matching method according to claim 2, characterized in that, The load correction factor κload is calculated according to κload = 1 - α·(mload / mmax), where mload is the current load mass, mmax is the maximum allowable load mass, and α is the load influence factor. The power correction coefficient κsoc is calculated according to κsoc = β + γ·(SOC / SOCmax), where SOC is the current remaining battery power, SOCmax is the full battery power, and β and γ are parameters affecting the power. The altitude correction coefficient κalt is calculated according to κalt = 1 - δ·h, where h is the flight altitude and δ is the altitude influence factor; The gust margin Δgust and safety margin Δsafe are preset constants or adjustable parameters that are dynamically configured according to the mission safety level.

4. The dynamic matching method according to claim 1, characterized in that, In the fine-grid activation determination step, the fine-resolution meteorological grid is activated when at least one of the following conditions is met: Condition 1: The absolute value of the difference between the effective wind resistance threshold of any candidate aircraft model under the current coarse resolution meteorological grid and the coarse grid wind speed is less than the first trigger threshold; Condition 2: The spatial gradient of wind speed in the current coarse-resolution meteorological grid is greater than the second trigger threshold; Condition 3: The turbulence intensity of the current coarse-resolution meteorological grid is greater than the third trigger threshold.

5. The dynamic matching method according to claim 4, characterized in that, In the fine mesh activation determination step, the fine mesh activation flag Ag is determined according to the following activation function: Ag = 1, when minm(Veff^m - Vg) < M0 or ||∇Vg|| > G0 or Kg > K0; Where m is the candidate aircraft model number, Veff^m is the effective wind resistance threshold of the m-th aircraft model, Vg is the wind speed of the coarse-resolution meteorological grid, ∇Vg is the wind speed spatial gradient of the coarse-resolution meteorological grid, Kg is the turbulence intensity of the coarse-resolution meteorological grid, M0 is the first trigger threshold, G0 is the second trigger threshold, and K0 is the third trigger threshold; Ag = 1 indicates that the fine-resolution meteorological grid is activated, and Ag = 0 indicates that it is not activated.

6. The dynamic matching method according to claim 1, characterized in that, In the adaptive score calculation step, the adaptive score S{m,g} of the m-th candidate aircraft model in the g-th meteorological grid is calculated according to the following formula: S{m,g} = qg · min{(Veff^m - Vg) / Veff^m, (Leff^m - Lg) / Leff^m, (K_eff^m -Kg) / Keff^m}; Wherein, Veff^m is the effective wind resistance threshold of the m-th aircraft type, Vg is the wind speed of the g-th meteorological grid, Leff^m is the effective crosswind resistance threshold of the m-th aircraft type, Lg is the crosswind speed of the g-th meteorological grid, Keff^m is the effective turbulence resistance threshold of the m-th aircraft type, Kg is the turbulence intensity of the g-th meteorological grid, and qg is the data reliability coefficient of the g-th meteorological grid; when the adaptability score S{m,g} is negative, it indicates that the aircraft type does not have safe flight conditions under the meteorological grid conditions.

7. The dynamic matching method according to claim 1, characterized in that, In the route and aircraft type recommendation step, the route adaptability score of the candidate route π is the minimum value of the adaptability scores of all meteorological grids on that route, that is: Sπ = min{g∈π} S{m,g}; Select the aircraft type and route combination that satisfies Sπ ≥ Smin and has the lowest overall cost (π), where Smin is a preset minimum safety threshold, and the overall cost (π) includes at least one of flight distance cost, flight time cost, and energy consumption cost.

8. The dynamic matching method according to claim 1, characterized in that, In the adaptive score calculation step, when the fine-resolution meteorological grid is not activated, the adaptive score of each candidate aircraft model is calculated only based on the coarse-resolution meteorological grid. When the fine-resolution meteorological grid is activated, the adaptive score of the activated area is switched to be calculated based on the fine-resolution meteorological grid, while the inactive area is still calculated based on the coarse-resolution meteorological grid, forming an adaptive score matrix that combines coarse and fine grids.

9. The dynamic matching method according to claim 1, characterized in that, It also includes a feedback update step: after each flight mission, the aircraft type information, payload information, flight altitude, weather grid data experienced, flight attitude correction range and energy consumption anomaly data of that flight are recorded; if the same aircraft type accumulates a predetermined number of attitude protection triggers or energy consumption anomalies under the same weather grid conditions, the effective wind resistance margin of the aircraft type under the same weather grid conditions is reduced by a predetermined percentage, and the reduced effective wind resistance margin is stored in the aircraft type wind resistance parameter library for subsequent matching.

10. The dynamic matching method according to claim 1, characterized in that, In the effective wind resistance threshold correction step, the control margin coefficient κctrl is determined according to the flight control mode: κctrl takes a first value in manual control mode and a second value in automatic control mode, and the first value is less than the second value; the mission safety level includes at least low risk level, medium risk level and high risk level, and the safety margin Δsafe increases as the mission safety level increases.

Images