Urban low-altitude unmanned aerial vehicle landing site selection and switching method and device

Abstract

The application discloses a city low-altitude unmanned aerial vehicle landing site selection and switching method and equipment, belongs to the field of data processing, solves the problem of disconnection between site selection and operation, and comprises the following steps: acquiring basic information of a planning area and environmental factor data affecting the operation of a landing site; generating candidate landing site positions and accessible routes according to the basic information; constructing a plurality of simulation scenarios including different environmental factor combinations according to the environmental factor data, and respectively evaluating the operation availability of each candidate landing site in each simulation scenario, so as to determine a main landing site scheme from the candidate landing site positions and a backup landing site scheme different from the main landing site scheme; in the operation stage, the real-time operation availability of the current main landing site is evaluated according to the current environmental factor data, and when the real-time operation availability does not meet a preset condition, a substitute main landing site and a route are selected from the backup landing site scheme, and task switching is performed. The risk disturbance recovery capability of the landing network is improved.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method and equipment for site selection and switching of urban low-altitude unmanned aerial vehicle (UAV) take-off and landing sites. Background Technology

[0002] The location of urban low-altitude drone take-off and landing sites determines the accessibility of drone missions, coverage quality, daily operational throughput, and fault recovery capabilities.

[0003] Existing site selection schemes are mostly based on two-dimensional distance, fixed service radius, construction cost and static demand, and cannot directly incorporate operational risks such as airport opening, runway occupation, temporary airspace control, overlapping no-fly zones, terrain clearance and real-time weather into the site selection-capacity model.

[0004] Current airport status risk assessments are typically geared towards single-mission release, using rules or weighted scoring to determine whether a flight is currently possible. However, the assessment results remain at the mission level and fail to influence airport construction location, hardware capacity, mission redundancy, and alternative topologies. Consequently, sites selected by the planning model may experience prolonged capacity reduction or shutdown in high-frequency risk scenarios, forcing the operating system to temporarily reject missions instead of switching to pre-verified alternative nodes. Summary of the Invention

[0005] To address the aforementioned issues, this application proposes a method for selecting and switching urban low-altitude unmanned aerial vehicle (UAV) take-off and landing sites, including: Acquire basic information about the planning area and data on environmental factors affecting the operation of the take-off and landing field; the basic information includes geospatial data of the planning area, UAV performance parameters, and parameters of take-off and landing field support equipment. Based on the aforementioned basic information, candidate take-off and landing site locations and associated reachable routes are generated; For the candidate take-off and landing sites and the reachable routes, multiple simulation scenarios with different combinations of environmental factors are constructed based on the environmental factor data, and the operational availability of each candidate take-off and landing site in each simulation scenario is evaluated. Based on the operational availability, a primary take-off and landing site scheme is determined from the candidate take-off and landing site locations, along with alternative take-off and landing site schemes that differ from the primary scheme. During the operation phase, the real-time operational availability of the currently used primary take-off and landing field is assessed based on current environmental factor data. When the real-time operational availability does not meet preset conditions, an alternative primary take-off and landing field and corresponding alternative routes are selected from the alternative take-off and landing field schemes, and a task switch is performed.

[0006] In one example, based on the aforementioned basic information, candidate takeoff and landing site locations and associated reachable routes are generated, specifically including: The planning area is divided into several hexagonal service grids according to a first resolution, and grouped according to the parent grid to which each service grid belongs; the parent grid is a hexagonal grid of the planning area divided according to a second resolution, the second resolution being lower than the first resolution, and each parent grid includes multiple service grids; The area comprised of all business grids belonging to the same parent grid is defined as the core area; For any core area, the business grids adjacent to the boundary of that core area in the adjacent core areas are determined as the adjacent overlapping boundary areas of that core area; Candidate take-off and landing sites are set up within the core area, and mission requirement points are set up within the core area and adjacent overlapping boundary areas respectively; For the route combinations between the candidate take-off and landing sites and the mission requirement points, based on the UAV's energy limitations and the flight energy consumption corresponding to the route combinations, the route corresponding to the route combination that meets the energy limitations is selected as the reachable route.

[0007] In one example, based on the drone's energy limitations and the flight energy consumption corresponding to the route combinations, routes corresponding to route combinations that meet the energy limitations are selected as reachable routes, specifically including: Calculate the flight energy consumption required for the UAV to complete the outbound and return journeys along the combined route; Determine the safe available energy for the drone; Determine whether the flight energy consumption meets the energy limit; the energy limit includes: the sum of the outbound flight energy consumption and the return flight energy consumption is not greater than the safe available energy, and the maximum value of the outbound flight energy consumption and the return flight energy consumption is not greater than the preset ratio of the safe available energy; The routes corresponding to the route combinations that meet the energy limits are determined as the reachable routes.

[0008] In one example, the operational availability of each candidate takeoff and landing site in each simulated scenario is evaluated, specifically including: For each candidate take-off and landing site in each simulation scenario, a first-level hard rule judgment is performed to obtain a hard judgment result; the first-level hard rule is used to indicate that when any hard rule is triggered, the candidate take-off and landing site is judged to be unusable in the corresponding simulation scenario; When the hard judgment result is that the first-level hard rule has not been triggered, the multiple risk sub-factors are scored separately. The scores of each risk sub-factor are weighted and summed to obtain a comprehensive safety score; Based on the hard judgment results and the comprehensive safety score, the operational availability of each candidate take-off and landing site in each simulated scenario is determined.

[0009] In one example, after weighted summation of the scores for each risk sub-factor to obtain a comprehensive safety score, the method further includes: Calculate the risk loss of each risk sub-factor based on its weight and the degree of insecurity of each sub-factor; Calculate the risk contribution rate of each risk sub-factor based on the proportion of risk loss of each risk sub-factor in the total risk loss; The risk sub-factors are sorted according to their risk contribution rates from high to low, and used as explanatory information for the comprehensive safety score.

[0010] In one example, based on the operational availability, a primary take-off and landing site is determined from the candidate take-off and landing site locations, along with alternative take-off and landing site options that differ from the primary option. Specifically, this includes: Based on the hard determination results in the operational availability, it is determined whether each candidate take-off and landing site is eligible to be selected in each simulation scenario; a candidate take-off and landing site that triggers a level one hard rule is not eligible to be selected in the corresponding simulation scenario; Based on the comprehensive safety score in the operational availability, the basic coverage quality and baseline capacity of the target candidate take-off and landing sites that are eligible for selection are adjusted in each simulation scenario to obtain the target coverage quality and effective capacity of the target candidate take-off and landing sites in each simulation scenario. The primary take-off and landing site option is determined based on the coverage quality and effective capacity of the target candidate take-off and landing sites; Based on the location of the primary take-off and landing site in the primary take-off and landing site plan, generate alternative take-off and landing site plans that differ from the primary take-off and landing site plan in terms of geographical distribution.

[0011] In one example, the primary take-off and landing site selection is determined based on the coverage quality and effective capacity of the target candidate take-off and landing sites, specifically including: Obtain the construction cost of each target candidate take-off and landing site and the mission requirement frequency of each route; To maximize coverage benefits in each simulated scenario, minimize risk exposure in each simulated scenario, and minimize site construction and hardware deployment costs, a site selection optimization model is constructed. Under the constraint of effective capacity, the location optimization model is solved to obtain the primary takeoff and landing site scheme.

[0012] In one example, based on the location of the primary takeoff and landing site in the primary takeoff and landing site plan, alternative takeoff and landing site plans that differ geographically from the primary takeoff and landing site plan are generated, specifically including: Taking each take-off and landing site in the primary take-off and landing site scheme as the center, a K-ring geographical neighborhood is constructed respectively; the K-ring geographical neighborhood is a set of grid cells in a hierarchical hexagonal grid that is no more than K layers away from the central take-off and landing site; For each K-ring geographical neighborhood, the upper limit of the number of take-off and landing sites that can be built in the neighborhood is determined based on the remaining available budget and the lowest construction cost of candidate take-off and landing sites in the neighborhood; The number of alternative take-off and landing sites constructed within each K-ring geographical neighborhood is limited to the specified upper limit. The alternative take-off and landing site scheme is constrained to be different from the primary take-off and landing site scheme in at least one K-ring geographical neighborhood.

[0013] In one example, selecting an alternative primary takeoff and landing field and its corresponding alternative flight route from the alternative takeoff and landing field options specifically includes: The system obtains the real-time hard determination results, real-time comprehensive safety score, remaining effective capacity, and route energy consumption between the alternative take-off and landing sites and the current mission for each alternative take-off and landing site in the alternative take-off and landing site scheme. Candidate take-off and landing sites are selected based on the following criteria: the real-time hard judgment result indicates that no hard rule has been triggered; the real-time comprehensive safety score is not lower than the basic release threshold; the remaining effective capacity is not less than the number of tasks to be migrated; and the route energy consumption meets the energy limit of the UAV. Based on the delay cost, energy consumption cost, and risk cost caused by the switchover, one of the candidate alternative take-off and landing fields is selected as the alternative primary take-off and landing field, and the corresponding alternative route is determined; the delay cost is the mission delay time caused by the switchover; the energy consumption cost is the increased flight energy consumption caused by the switchover; the risk cost is the increased risk increment caused by the switchover, and the risk increment is determined based on the difference between the comprehensive safety score of the candidate alternative take-off and landing field and the comprehensive safety score of the primary take-off and landing field.

[0014] On the other hand, embodiments of this application provide an urban low-altitude unmanned aerial vehicle (UAV) take-off and landing site selection and switching device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute any of the above-described urban low-altitude UAV take-off and landing site selection and switching methods.

[0015] The above-described technical solutions adopted in the embodiments of this application can achieve the following beneficial effects: By shifting operational risks from post-release conditions to planning variables, we can avoid building sites with high coverage but high frequency of unavailability. Pre-generating geographically differentiated alternative topologies shortens task recovery time under temporary risks.

[0016] This allows for the assessment of the availability of candidate airports during the planning phase using multi-scenario risk assessments, and during the operational phase, it enables the rapid selection of alternative airports and routes from geographically differentiated candidate topologies based on real-time risks, with the operational results fed back into subsequent planning.

[0017] This approach links the planning and site selection of drone take-off and landing sites with real-time risks during the operational phase, enabling pre-designed site plans to provide effective backup support for risk scenarios during operation. Ultimately, it forms a closed-loop system covering the entire lifecycle, from physical route generation and risk scenario assessment to robust site selection, real-time release, primary / backup switchover, and feedback-based optimization. Attached Figure Description

[0018] To more clearly illustrate the technical solution of this application, some embodiments of this application will be described in detail below with reference to the accompanying drawings, in which: Figure 1 A flowchart illustrating a method for selecting and switching urban low-altitude unmanned aerial vehicle (UAV) take-off and landing sites, provided in an embodiment of this application; Figure 2 A schematic diagram of a closed-loop planning and operation process provided in this application embodiment; Figure 3 A schematic diagram illustrating the coupling relationship between candidate takeoff and landing site status risk and location variables, provided in an embodiment of this application; Figure 4 A schematic diagram of a three-level decision chain for the status risk of a candidate takeoff and landing site is provided in an embodiment of this application; Figure 5 This application provides a schematic diagram illustrating the dynamic switching between primary and backup takeoff and landing sites and flight routes during operation, as shown in the embodiments of this application. Figure 6 This is a schematic diagram of a device for selecting and switching locations for urban low-altitude unmanned aerial vehicle (UAV) take-off and landing sites, provided as an embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] Some embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0021] Figure 1 This is a flowchart illustrating a method for selecting and switching take-off and landing sites for urban low-altitude unmanned aerial vehicles (UAVs) according to an embodiment of this application. This process can be executed by computing devices in the relevant field, and certain input parameters or intermediate results in the process can be manually adjusted to help improve accuracy.

[0022] The analysis method involved in the embodiments of this application can be implemented by a terminal device or a server, and this application does not impose any special limitations on it. For ease of understanding and description, the following embodiments are all described in detail using a server as an example.

[0023] It should be noted that the server can be a single device or a system composed of multiple devices, i.e., a distributed server. This application does not make any specific limitations on this.

[0024] Figure 1 The process includes the following steps: S101: Obtain basic information about the planning area and data on environmental factors affecting the operation of the take-off and landing field.

[0025] Basic information refers to the static or quasi-static data necessary for selecting take-off and landing sites and planning flight routes.

[0026] Basic information includes geospatial data of the planning area, drone performance parameters, and parameters of take-off and landing site support equipment.

[0027] Geospatial data, including topographic elevation data, 3D building models, road networks, and population distribution, is used to determine feasible locations for candidate take-off and landing sites and the spatial boundaries of flight zones. For example, the planning area is discretized into a hierarchical hexagonal grid, with different levels of grid serving different purposes for site selection decisions and mission requirement representation.

[0028] The performance parameters of a drone include battery capacity, safe state of charge, propulsion efficiency, cruise speed, cruise drag power, drag coefficient, frontal projection area, and total mass of the drone and its payload. These parameters collectively determine the drone's safe available energy and safe round-trip radius.

[0029] The parameters for runway support equipment include the geographical location of the candidate runway, construction cost, the baseline capacity and deployment cost of various hardware equipment (such as charging facilities and landing pads), and the number and size of runways.

[0030] Environmental factor data refers to various dynamic or semi-dynamic external conditions that affect the take-off, landing, and flight safety of unmanned aerial vehicles (UAVs).

[0031] Environmental data includes takeoff and landing site operation status data, airspace control data, meteorological data, terrain and obstacle data, and mission requirement data.

[0032] The operational status data of the take-off and landing field includes the open / closed status of the candidate take-off and landing field, the runway occupancy status, the take-off and landing field level, etc., which directly determines whether the take-off and landing field is available at a specific time.

[0033] Airspace control data includes the polygon of permanent no-fly zones, the spatial and temporal extent of temporary control areas, and the level of airspace congestion.

[0034] Terrain and obstacle data includes the elevation of the tops of surface obstacles, which is used to compare with the planned flight altitude to calculate terrain clearance.

[0035] The mission requirement data includes the baseline mission frequency and maximum positive deviation for each route, used to express the uncertainty of requirements.

[0036] S102: Based on the basic information, generate candidate take-off and landing site locations and associated reachable routes.

[0037] Candidate take-off and landing sites refer to all available take-off and landing site locations within the planned area. Reachable flight paths refer to flight routes that meet the energy constraints of the UAV and allow for safe round trips between take-off and landing sites or between a take-off and landing site and the mission point.

[0038] S103: Based on the candidate take-off and landing site locations and reachable routes, construct multiple simulation scenarios with different combinations of environmental factors according to environmental factor data, and evaluate the operational availability of each candidate take-off and landing site in each simulation scenario.

[0039] In this example, according to a certain combination logic, the various elements in the environmental factor data are combined according to different state values ​​to form multiple representative risk scenarios.

[0040] Each simulation scenario can represent a normal takeoff and landing field, runway occupancy, temporary air traffic control, strong winds, low visibility, or a combination of multiple events occurring simultaneously. Each scenario has a preset probability of occurrence, which is used for subsequent weighted optimization.

[0041] Operational availability refers to the comprehensive availability measure of a candidate take-off and landing site or take-off and landing site in a specific scenario.

[0042] S104: Based on operational availability, determine the primary take-off and landing site from the candidate take-off and landing site locations, as well as alternative take-off and landing site options that differ from the primary take-off and landing site options.

[0043] S105: During the operation phase, based on current environmental factor data, assess the real-time operational availability of the currently used main take-off and landing field. When the real-time operational availability does not meet the preset conditions, select an alternative main take-off and landing field and the corresponding alternative route from the alternative take-off and landing field options, and perform a mission switch.

[0044] Preset conditions refer to a safety threshold system for determining whether the current take-off and landing site can continue to execute or accept the mission.

[0045] The preset conditions include hard rule release conditions, pending approval release threshold conditions, and safe release threshold conditions.

[0046] When a hard rule is triggered in the real-time availability assessment, or when the overall security score in the real-time availability assessment is lower than the basic release threshold, the real-time availability assessment is determined to not meet the preset conditions.

[0047] When the hard judgment result indicates that the hard rule has not been triggered and the comprehensive security score is not lower than the security release threshold, the real-time operational availability is determined to meet the preset conditions.

[0048] When the hard judgment result indicates that the hard rule has not been triggered, and the comprehensive safety score is higher than the basic release threshold but lower than the safety release threshold, if the manual approval is passed, the real-time operational availability is determined to meet the preset conditions; if the manual approval is not passed, the real-time operational availability is determined to not meet the preset conditions.

[0049] Select suitable alternative landing sites from the pre-built alternative landing site options and migrate the mission to be performed to the alternative site.

[0050] In this example, the safe release threshold can be determined based on airport type, severe weather, and operational intensity, as shown in the following expression:

[0051] In the formula, For a moment The actual effective security clearance threshold; Based on the basic release threshold; Airport status threshold; Meteorological condition thresholds; This is the peak operation threshold. The basic threshold can be set to 0.50, the hub airport threshold to 0.80, the general aviation airport threshold to 0.65, and the severe weather threshold to 0.78.

[0052] It should be noted that, although the embodiments in this application are based on... Figure 1 Steps S101 to S105 will be described sequentially, but this does not mean that steps S101 to S105 must be performed in a strict order. The reason this embodiment follows this order is... Figure 1 The order in which steps S101 to S105 are described is provided to facilitate understanding of the technical solutions of the embodiments of this application by those skilled in the art. In other words, in the embodiments of this application, the order of steps S101 to S105 can be appropriately adjusted according to actual needs.

[0053] pass Figure 1 The method involves site selection and alternative planning based on known mission points. During operation, when the status of the takeoff and landing field changes, an available alternative site is selected from the alternative options, as detailed below: This expands the information dimensions used for site selection decisions from the single geographical or cost information of traditional methods to a multi-source information framework that simultaneously covers static basic conditions and dynamic environmental factors, enabling the site selection process to perceive the dynamic operating environment of candidate sites.

[0054] The selection of take-off and landing sites has been upgraded from selecting isolated locations to generating locations in conjunction with flight routes, ensuring that each candidate take-off and landing site location does not exist in isolation, but is associated with the flight routes that the drone can actually fly to.

[0055] By conducting separate evaluations, a ternary mapping relationship is established between candidate takeoff and landing sites, simulated scenarios, and operational availability. This mapping relationship ensures that each candidate site no longer has a static good or bad evaluation, but rather a availability vector that changes with the scenario. This enables the system to identify sites that, while performing poorly in normal times, remain stable under harsh conditions.

[0056] The availability of the solution across all scenarios is used as the basis for determining the primary solution, ensuring that the final primary solution is the optimal solution from a comprehensive risk and benefit perspective, rather than simply the solution that covers the best.

[0057] The primary and alternative solutions are processed within the same framework determined by operational availability. Both are derived from the same multi-scenario evaluation results, differing only slightly in their decision-making objectives. This ensures that the alternative solutions are replacements with information of equal quality to the primary solution in terms of perceived risk.

[0058] The alternative solutions generated during the planning phase are effectively activated during the operation phase, enabling the evaluation results to be upgraded from merely determining whether a flight is possible to triggering a switch from pre-set alternatives. This achieves a functional leap from passive judgment to proactive decision-making.

[0059] The optimization and generation of alternative solutions have already been completed during the planning phase. During the operation phase, only the usability of the solutions needs to be determined based on the current real-time data, and which solution to select is chosen. The latter is a lightweight selection problem rather than a complete optimization problem. This division of labor allows switching decisions to be completed within seconds, avoiding the long hovering and waiting in the traditional temporary alternative search mode.

[0060] In summary, by shifting operational risks from post-release conditions to planning variables, we can avoid building sites with high coverage but high frequency of unavailability. Pre-generating geographically differentiated alternative topologies shortens task recovery time under temporary risks.

[0061] This allows for the assessment of the availability of candidate airports during the planning phase using multi-scenario risk assessments, and during the operational phase, it enables the rapid selection of alternative airports and routes from geographically differentiated candidate topologies based on real-time risks, with the operational results fed back into subsequent planning.

[0062] This approach links the planning and site selection of drone take-off and landing sites with real-time risks during the operational phase, enabling pre-designed site plans to provide effective backup support for risk scenarios during operation. Ultimately, it forms a closed-loop system covering the entire lifecycle, from physical route generation and risk scenario assessment to robust site selection, real-time release, primary / backup switchover, and feedback-based optimization.

[0063] Achieving a unified closed loop between site selection during the planning phase and safe release during the operation phase improves the physical feasibility, risk interpretability, and disturbance recovery capability of the takeoff and landing network.

[0064] based on Figure 1 In addition to the method described herein, this application also provides some specific implementation schemes and extended schemes of the method, which will be further described below.

[0065] In this example, candidate takeoff and landing site locations and associated reachable routes are generated based on basic information, including the following steps: Step 1: Divide the planning area into several hexagonal service grids according to the first resolution, and group them according to the parent grid to which each service grid belongs; the parent grid is a hexagonal grid of the planning area divided according to the second resolution, which is lower than the first resolution, and each parent grid includes multiple service grids.

[0066] The planning area is divided into regularly shaped grid units according to different spatial resolutions (levels), and there is a parent-child relationship between grids of different levels.

[0067] The H3 hexagonal hierarchical grid system is adopted: the upper layer (parent grid) covers a large geographical area, and the lower layer (child grid) further subdivides the parent grid into smaller hexagonal units, with each child grid uniquely belonging to a parent grid.

[0068] For example, the first resolution is any one of the levels 9 to 11 of the H3 grid system, and the second resolution is any one of the levels 6 to 8 of the H3 grid system.

[0069] Based on this, continuous geographic space is discretized into computable grid cells, transforming the site selection problem from continuous spatial optimization to discrete grid selection, significantly reducing computational complexity. Through the parent-child grid hierarchy, data aggregation and partitioning management can be performed at different precision levels.

[0070] Step 2: Define the area consisting of all business grids belonging to the same parent grid as the core area.

[0071] The core area refers to the collection of business grids that share the same parent grid.

[0072] The core area serves two functions: first, it allows for the selection, construction, and capacity configuration of candidate take-off and landing sites; second, it sets up task demand points within the core area, representing the distribution of tasks that need to be served by that core area.

[0073] Step 3: For any core area, the service grids adjacent to the boundary of the core area in the adjacent core areas are determined as the adjacent overlapping boundary areas of the core area.

[0074] In other words, for any core region, determine the other parent grids that are adjacent to the parent grid of the core region.

[0075] For the service grids included in other parent grids, the service grids adjacent to the service grids on the core area boundary are identified as the adjacent overlapping boundary areas of the core area.

[0076] For example, first divide the service grid and parent grid according to resolution, then determine all service grids under parent grid A as core area A, and among the service grids under other parent grids adjacent to parent grid A, include the service grids adjacent to the boundary of core area A into the adjacent overlapping boundary area of ​​core area A.

[0077] Step 4: Set up candidate take-off and landing sites in the core area, and set up mission requirement points in the core area and adjacent overlapping boundary areas respectively.

[0078] Based on this, regarding data retention and decision-making authority, the core area retains both candidate take-off and landing site locations and mission requirements, allowing for take-off and landing site selection, construction, and capacity configuration within the core area. Adjacent overlapping boundary areas only incorporate mission requirements; candidate take-off and landing site locations are not designated, and they do not participate in construction decisions for the core area's planning.

[0079] By introducing adjacent overlapping boundary regions of the core area, the candidate take-off and landing sites within the core area can cover mission requirements near the boundaries of adjacent core areas.

[0080] At the same time, each candidate take-off and landing site is selected only by its respective core area, thereby avoiding duplicate site selection or construction in the same location by adjacent planned war zones.

[0081] Step 5: For the route combinations between candidate take-off and landing sites and mission requirement points, select the routes corresponding to the route combinations that meet the energy limits as reachable routes, based on the UAV's energy constraints and the flight energy consumption corresponding to the route combinations.

[0082] In this example, the filtering process includes the following steps: Step 51: Calculate the flight energy consumption required for the UAV to complete the outbound and return journeys along the route.

[0083] The calculation along the route involves the specific pairing between each generated candidate takeoff and landing site location and the mission requirement point. For example, if there are M candidate takeoff and landing site locations and N mission requirement points, then the flight energy consumption for each M×N combination needs to be calculated separately.

[0084] The outbound flight refers to the direction from which the UAV flies from the candidate takeoff and landing site to the mission requirement point, while the return flight refers to the direction from the mission requirement point back to the candidate takeoff and landing site. The flight paths in the two directions overlap spatially (the three-dimensional distance is the same), but the energy consumption may differ. This is because the wind direction may change relative to the flight direction (headwind on the outbound flight, tailwind on the return flight), and the direction of altitude change is opposite (climb on the outbound flight, descent on the return flight).

[0085] The required flight energy consumption refers to the actual battery energy consumed by the drone to complete the flight segment, usually measured in watt-hours (Wh) or joules (J).

[0086] Based on this, the three-dimensional flight distance corresponding to the route combination, the equivalent wind speed of the UAV along the flight path, and the altitude change of the UAV on the route combination are obtained.

[0087] Based on the three-dimensional flight distance, equivalent wind speed, and altitude change, calculate the flight energy consumption for the outbound and return journeys, respectively.

[0088] For example, the expression is as follows:

[0089]

[0090] in, Flight energy consumption refers to the battery energy consumed by a drone to complete one flight segment; air density; Cruise speed; This is the equivalent windward speed along the flight path. It is positive when facing the wind and negative when with the wind. Let be the air drag coefficient of the UAV, which is a dimensionless parameter; The area projected onto the windward side of the drone is in square meters. The total mass of the drone and its payload; Let be the acceleration due to gravity, and take . ; This means that only the gravitational potential energy generated during ascent is calculated, and not all potential energy is considered as recoverable energy when the drone descends. To improve efficiency; h is the vertical altitude change of the flight path, i.e., the altitude difference between the mission point and the take-off and landing site; It represents the projected distance of the flight path on a two-dimensional plane, that is, the horizontal straight-line distance between the take-off and landing field and the mission point (the distance on the map). This represents the three-dimensional true length of the flight path, the actual spatial distance the drone traveled.

[0091] Step 52: Determine the safe available energy for the drone.

[0092] Safely available energy can be calculated from the drone's performance parameters during system initialization and then stored. Safely available energy refers to the maximum amount of electricity that the drone is allowed to use for flight missions. It is not equal to the total battery capacity, but rather the total battery capacity minus the safe amount of electricity that must be reserved.

[0093] For example, the expression is as follows:

[0094] in, The safe and available energy that can be used by drones when performing flight missions; For the drone's battery capacity; For a safe state of charge, the value range is: This indicates the minimum percentage of battery charge that must be retained after the flight ends.

[0095] Step 53: Determine whether the flight energy consumption meets the energy limit; the energy limit includes: the sum of the outbound flight energy consumption and the return flight energy consumption is not greater than the safe available energy, and the maximum value of the outbound flight energy consumption and the return flight energy consumption is not greater than the preset ratio of the safe available energy.

[0096] After a drone departs from its launch site to complete a mission and returns, the total energy consumed throughout the entire process must be within the safe and available energy range. This ensures that the drone will not run out of power during its return journey. The expression is as follows:

[0097] in, This indicates the amount of battery energy required for the drone to fly from the take-off and landing site to the mission target point (outbound journey). This indicates the battery energy consumed by the drone to return from the mission requirement point to the take-off and landing site (return trip).

[0098] The greater of the outbound and return energy consumption shall not exceed a preset proportion of safe available energy. The expression is as follows:

[0099] For the one-way safety ratio, even if the total energy consumption for the round trip does not exceed the standard, if a single flight consumes too much power, resulting in a serious lack of safety margin on the return trip, then it is also unqualified.

[0100] For example, let =100Wh A certain flight route consumes 65Wh of energy on the outbound journey and 30Wh on the return journey, with a total round-trip energy consumption of 95Wh ≤ 100Wh (condition one is met). However, the outbound energy consumption is 65Wh >> 70Wh (condition two is not met), so this route is still eliminated. This is because too much electricity has been consumed on the outbound journey, leaving only 35Wh available for the return journey, far below the 30% of safe available energy. If headwinds are encountered or a detour is required, the return journey may not be safe.

[0101] Step 54: Determine the routes corresponding to the route combinations that meet the energy limits as reachable routes.

[0102] The routes corresponding to the route combinations refer to the routes that have been confirmed as actually reachable from the candidate state after the retained route combinations. The routes that have been filtered in this step constitute the set of reachable routes.

[0103] In this example, before calculating flight energy consumption, the process may further include: calculating the maximum one-way flight radius based on the UAV's safe available energy; and filtering route combinations between candidate take-off and landing sites and mission requirement points based on the maximum one-way flight radius, so that the planar distance corresponding to the retained route combinations is not greater than the maximum one-way flight radius.

[0104] The expression is as follows:

[0105] This refers to the maximum one-way flight radius that a drone is allowed to reach when it departs from and safely returns from the takeoff and landing site. This is the theoretical value under ideal, windless conditions. This represents the power of cruise drag.

[0106] Therefore, when generating route combinations between candidate take-off and landing sites and mission requirement points, if the distance between a mission requirement point and a candidate take-off and landing site exceeds [a certain threshold], [further action may be taken]. If so, the combination can be directly excluded from the candidate set without performing any detailed energy consumption calculations.

[0107] The expected technical effect of this initial screening is to avoid performing detailed energy consumption calculations on a large number of invalid combinations that are too far apart. Detailed calculations themselves rely on more detailed data such as wind speed, air density, and climb profile, which are more costly to perform, thus reducing computational overhead.

[0108] In one example, the operational availability of each candidate takeoff and landing site in each simulation scenario is evaluated, including the following steps: Step 1: For each candidate take-off and landing field in each simulation scenario, execute the first-level hard rule judgment to obtain the hard judgment result; the first-level hard rule is used to indicate that when any hard rule is triggered, the candidate take-off and landing field is judged as unusable in the corresponding simulation scenario.

[0109] All candidate take-off and landing sites set up within the core area must be evaluated independently, and the evaluation results cannot be borrowed or transferred between sites.

[0110] Each simulation scenario represents a possible operational environment state (such as a strong wind scenario, runway occupancy scenario, temporary traffic control scenario, etc.). Each candidate takeoff and landing site must be determined independently in each simulation scenario.

[0111] Level 1 hard rule determination refers to the process of checking each hard rule of a candidate take-off and landing field under a specific simulation scenario.

[0112] The hard decision result refers to the conclusion output after the decision, which is used to characterize whether the candidate take-off and landing field is rejected by a vote due to triggering the hard rule in the simulation scenario.

[0113] If any hard rule is triggered, the candidate take-off and landing field will be deemed unavailable in the simulation scenario. This determination is final and cannot be changed, offset, or compensated for by subsequent scoring steps. Being deemed unavailable means that the candidate take-off and landing field completely loses its eligibility to be selected and activated in the simulation scenario.

[0114] For example, hard rules include the closure of take-off and landing fields, unairworthiness of drones, excessive overlap of no-fly zones, conflicts in control time and space, extreme weather, no alternatives for occupied take-off and landing areas, and insufficient terrain clearance.

[0115] A landing pad closure refers to a candidate landing pad being in a non-operational state under a specific simulated scenario, not accepting any drone takeoffs or landings. The trigger condition is that the landing pad is marked as closed due to maintenance, malfunction, management decisions, or its operational time window has expired. The determination method is to read the landing pad's operational status data; the closure is triggered when the status is "closed."

[0116] Unairworthiness of a drone refers to the drone model or specific serial number corresponding to a candidate takeoff and landing site failing to meet airworthiness requirements under a specific scenario. Triggering conditions include the drone failing pre-flight checks, having known unresolved faults, or the drone model being restricted from flight under specific weather conditions (such as strong winds or icing). The determination method combines drone maintenance records, the drone's operational envelope, and the current / simulated scenario conditions.

[0117] Exceeding the no-fly overlap rate means that the degree of spatial overlap between the flight area and the no-fly zone exceeds a preset rejection threshold. The expression is as follows:

[0118] in, Number the candidate take-off and landing field or candidate take-off and landing point. Number the risk scenarios. The flight area is a polygon. The area is a polygon representing a no-fly zone. This represents the function for calculating the area of ​​a polygon.

[0119] The trigger condition for exceeding the no-fly overlap rate is , To preset the rejection threshold, for example, 30%.

[0120] Even partial overlap between flight zones and no-fly zones means that drones may intrude into no-fly airspace during missions. This could involve military zones, takeoff and landing field protection zones, or airspace above important facilities. Setting the overlap rate threshold to a low value (e.g., 30%) reflects a conservative approach to airspace safety.

[0121] Control time-space conflict refers to a prohibitive overlap between the time window of a flight plan and the time window of temporary control. The expression for a control time-space conflict is as follows:

[0122]

[0123] in, The time overlap length, and These are the start and end times of the flight plan. and These are the start and end times of the control measures. This represents the time overlap rate.

[0124] The triggering condition for a control time-space conflict is that the time overlap length is greater than 0, and there is a conflict between the flight plan and the control restrictions during the overlap period (such as the control prohibiting all flight activities).

[0125] Temporary airspace control is typically established for major events, military exercises, or emergencies, during which all flights are prohibited in the relevant airspace. If a flight plan and the control window overlap, the drone cannot perform its mission during that period, and the flight route will be unavailable during that time.

[0126] Extreme weather refers to weather conditions in a simulated scenario or at the time of operation that reach or exceed the weather limits for safe flight of drones.

[0127] The trigger condition is exceeding the weather limit threshold. Different types of drones may have different weather limit thresholds (such as the wind resistance of small multi-rotor drones and large vertical take-off and landing fixed-wing drones are different).

[0128] The determination method involves comparing meteorological data (wind speed, visibility, precipitation type and intensity, icing conditions, etc.) with the limits in the drone's flight manual item by item. If any meteorological parameter exceeds the limit, the system will be triggered.

[0129] Extreme weather poses a direct threat to the flight safety of drones. Winds exceeding their resistance capabilities can lead to loss of control, thunderstorms can damage electronic equipment, and low visibility can impair obstacle avoidance capabilities. These risks cannot be mitigated by adjusting flight routes or adding alternate landing points and must be considered as veto conditions.

[0130] "Occupied take-off and landing area with no alternative" means that the take-off and landing area (runway, landing pad) of the candidate take-off and landing site is occupied and there is no available alternative take-off and landing area under the scenario conditions.

[0131] The triggering condition is that the take-off and landing area is occupied by other drones (such as those that are charging, loading and unloading cargo, or undergoing maintenance), and there are no other available take-off and landing pads or alternative take-off and landing areas in the area.

[0132] The determination method refers to making a judgment based on the occupancy status of the take-off and landing area and the configuration parameters of the take-off and landing field in the simulated scenario.

[0133] If the entire take-off and landing area is occupied, the drone will be unable to land (or take off) even if it flies there, and the mission cannot be completed. If the take-off and landing area has multiple landing pads, being occupied does not mean it is unusable; a rejection is only triggered when the occupation covers the entire take-off and landing area and there is no alternative.

[0134] Insufficient terrain clearance refers to a situation where, on the reachable flight path corresponding to a candidate takeoff and landing site, the minimum clearance between the planned flight altitude of the UAV and the top height of any surface obstacle below is lower than a safe threshold. The terrain clearance expression is as follows:

[0135] in, Number the candidate take-off and landing field or candidate take-off and landing point. Number the risk scenarios. The minimum clearance height for terrain or obstacles corresponding to the flight path; Geographic sampling points on the flight area or route; For drones at sampling points The planned flight altitude; Sampling points The elevation of the top of the terrain, buildings, or other surface obstacles. Both elevations should use the same elevation datum.

[0136] The trigger condition is that the terrain clearance is less than the preset minimum safe clearance height threshold.

[0137] Insufficient airspace can cause drones to collide with buildings or terrain when encountering turbulence, altitude hold errors, or GPS signal anomalies. This hard rule ensures that all reachable flight paths maintain sufficient vertical safety margins at any sampling point.

[0138] The same candidate take-off and landing site may have different judgment results in different simulation scenarios (for example, it passes the hard rule check in a normal scenario, but fails in a strong wind scenario because the wind speed reaches the extreme weather threshold).

[0139] The hard rules cover the core dimensions affecting flight safety, including operational support (closure of take-off and landing fields, occupation of take-off and landing areas), aircraft airworthiness (unairworthiness of drones), airspace safety (overlapping no-fly zones, control conflicts), meteorological safety (extreme weather), and physical safety (terrain clearance), forming a non-overlapping and comprehensive safety filter.

[0140] In this example, a first-level veto variable is defined. Where i is the candidate take-off and landing site number, s is the simulation scenario number, and the expression for the first-level veto variable is as follows:

[0141] This is a set of first-level veto hard rules. Let be the binary indicator of the νth hard rule. The mathematical essence of the max operation is a logical OR operation. Whenever any hard rule is triggered, A value of 1 indicates that no hard rules are triggered. Only 0.

[0142] For example, when the take-off and landing field is closed, the drone is not airworthy, the overlap between the flight area and the no-fly zone reaches 30%, the wind speed reaches 15 m / s, a severe thunderstorm occurs, or the flight altitude exceeds the limit without approval. .otherwise, 0.

[0143] In this example, variables are enabled via candidate takeoff and landing fields. and route activation variables Apply constraints.

[0144]

[0145] when When it is 1, and All are forcibly constrained to 0, when When the value is 0, this constraint does not impose any restrictions, and the activation / enabling of takeoff and landing sites and routes is determined by other constraints (cost, coverage, capacity, etc.). This is a binary decision variable indicating whether route r is activated in the simulation scenario s. A value of 1 indicates that route r is activated in scenario s and can be used to perform tasks, while a value of 0 indicates that route r is not activated in scenario s and does not participate in task execution.

[0146] Step 2: When the hard judgment result is that the first-level hard rule has not been triggered, score the multiple risk sub-factors separately.

[0147] Scoring multiple risk sub-factors separately refers to conducting a refined assessment of the risk status of candidate take-off and landing sites from multiple dimensions after the hard rule check has passed.

[0148] Risk sub-factors include: takeoff and landing field openness status, runway occupancy status, takeoff and landing field level suitability, controlled airspace type, temporary control status, airspace congestion level, terrain clearance height, underlying surface type, no-fly zone overlap, control period matching degree, and peak period suitability. The scores of each sub-factor are mapped to the range [0,1], with higher values ​​indicating greater safety (1 indicates complete safety or complete satisfaction, and 0 indicates high risk or non-satisfaction).

[0149] It should be noted that a low sub-factor score will only lower the overall safety score, but will not directly lead to a veto.

[0150] Step 3: Sum the scores of each risk sub-factor with weights to obtain the comprehensive safety score.

[0151] Weighted summation refers to assigning a weight coefficient to each risk sub-factor, and then multiplying the scores of each sub-factor by their corresponding weights and summing them up.

[0152] For example, the global weights for takeoff and landing field openness, runway occupancy, takeoff and landing field level, controlled airspace, temporary control, busy airspace, terrain clearance, underlying surface, no-fly overlap, control period matching, and peak period adaptation are 0.20, 0.10, 0.10, 0.12, 0.10, 0.08, 0.08, 0.06, 0.06, 0.05, and 0.05, respectively.

[0153] For example, the expression is as follows:

[0154]

[0155] Where i is the candidate take-off and landing site number, and s is the simulation scenario number. The total number of risk sub-factors. Let be the global weight of the nth sub-factor, satisfying that the sum of all weights is 1. For comprehensive safety scoring, As a risk factor.

[0156] Step 4: Based on the hard judgment results and comprehensive safety scores, determine the operational availability of each candidate take-off and landing site in each simulation scenario.

[0157] The hard criteria and overall safety score indicate that operational availability is not determined by a single factor, but rather by two sources. The hard criteria provide a qualification for availability, while the overall safety score provides a quantitative measure of how good it is within the scope of availability.

[0158] The primary veto variable is coupled to the eligibility for activation of candidate take-off and landing sites and the eligibility for activation of flight routes, respectively. The scenario safety score is coupled to the flight route coverage quality and the effective hardware capacity, respectively. This clearly defines the hard judgment result corresponding to the eligibility for activation (access level) and the comprehensive safety score corresponding to the coverage quality and capacity (quality level). Operational availability is a combination of the above two dimensions, including both the veto information of whether it is usable or not (hard judgment result) and the refined measurement of whether it is usable or barely usable (comprehensive safety score).

[0159] In one example, a post-hoc attribution analysis of the composition of the comprehensive safety score is performed to make the site selection and release results interpretable, including the following steps: Step 1: Calculate the risk loss of each risk sub-factor based on its weight and the degree of insecurity.

[0160] To illustrate the impact of each risk sub-factor on the reduction of the scenario safety score, the risk loss for each risk sub-factor is calculated. The risk loss is the product of its weight and the degree of insecurity.

[0161] For example, the The risk loss of each risk sub-factor is:

[0162] i represents the candidate takeoff and landing site number, and s represents the simulation scenario number. For the first The weighted risk loss of each risk sub-factor, As a risk factor, This corresponds to the degree of insecurity of the risk sub-factors.

[0163] Step 2: Calculate the risk contribution rate of each risk sub-factor based on the proportion of risk loss of each risk sub-factor in the total risk loss.

[0164] The risk contribution rate of the nth risk sub-factor is:

[0165] In the formula, For the summation index, Indicates candidate take-off and landing sites and simulated scenarios Total risk loss, For the first The weighted risk loss of each risk sub-factor, Given the risk contribution rate of the nth risk sub-factor, the main risk points are output in order of risk loss contribution, so that the site selection and release results can be interpreted.

[0166] The risk contribution rate represents the percentage of the nth risk sub-factor in the total risk loss. It answers the question of how much responsibility each sub-factor bears for a low overall safety score. The sum of the risk contribution rates of all sub-factors is 100%.

[0167] Step 3: Output the risk sub-factors in descending order of risk contribution rate as explanatory information for the comprehensive safety score.

[0168] Sorted by risk contribution rate from highest to lowest. This is a descending order, with the sub-factors having the highest contribution rate listed first, representing the risk factors that have the greatest impact on reducing the overall safety score. The purpose of this sorting output is to allow users (planners, operators, or system maintenance personnel) to prioritize the most important risk sources without having to calculate and sort them themselves.

[0169] In one example, determining a primary takeoff and landing site from candidate takeoff and landing site locations, and alternative takeoff and landing site sites that differ from the primary takeoff and landing site, based on operational availability, includes the following steps: Step 1: Based on the hard criteria in the availability of the operation, determine whether each candidate take-off and landing site is eligible to be selected in each simulation scenario; candidate take-off and landing sites that trigger the first-level hard rule are not eligible to be selected in the corresponding simulation scenario.

[0170] Selection eligibility refers to the basic threshold conditions for a candidate take-off and landing site to be included in the site selection optimization model.

[0171] Variables enabled for candidate take-off and landing sites A value of 1 indicates that it is enabled (eligible). A value of 0 indicates that it is not enabled (but is eligible).

[0172] The hard decision result has priority and exists in the form of a hard constraint in the site selection model. A site that is rejected outright cannot be selected in any simulation scenario, and no other optimization objective (such as maximizing coverage or minimizing cost) can bypass this constraint.

[0173] Step 2: Based on the comprehensive safety score in the operational availability, adjust the basic coverage quality and baseline capacity of the target candidate take-off and landing sites that are eligible for selection in each simulation scenario to obtain the target coverage quality and effective capacity of the target candidate take-off and landing sites in each simulation scenario; the comprehensive safety score is positively correlated with the basic coverage quality and baseline capacity.

[0174] A higher overall safety score indicates a higher overall safety level for the candidate take-off and landing site in the corresponding scenario.

[0175] Coverage quality measures the actual coverage capability of a candidate takeoff and landing site for mission requirements in the corresponding simulated scenario. The expression is as follows:

[0176] in, This refers to the risk-adjusted target coverage quality, i.e., the target route coverage quality. Based on the quality of physical coverage, For route number, For comprehensive safety scoring.

[0177] The basic physical coverage quality of a flight route (determined by the geometric relationship between the takeoff and landing sites and the mission requirement points) is itself a static value. However, the actual usable coverage quality of the route dynamically decreases depending on the safety score of the takeoff and landing sites. When the overall safety score is 1, the actual coverage quality equals the basic coverage quality, and service capability is fully preserved. When the overall safety score is 0.6, the actual coverage quality is reduced to 60% of the basic coverage quality, and the coverage value contribution of the route in the optimization model decreases accordingly.

[0178] Capacity is used to measure the actual processing capability of various hardware devices at a candidate takeoff and landing site in the corresponding scenario, and is expressed as follows:

[0179] in, For risk scenarios The effective capacity of the ordered hardware unit; Base capacity; Assign a hardware type number; This is the runway occupancy reduction factor, with a value range of (0,1]. A value of 1 indicates no impact from occupancy, and a value of 0 indicates that the runway is completely unusable. It directly corresponds to the runway occupancy status sub-factor, that is, it is directly mapped from the score of the runway occupancy status sub-factor. For example, the runway is completely empty: sub-factor score = 1.0, with a value of 1.0; the runway is partially occupied: sub-factor score = 0.6, with a value of 0.6; the runway is fully occupied: sub-factor score = 0, with a value of 0. This is the weather derating factor, which is the proportional coefficient that the effective capacity of hardware decreases due to unfavorable weather conditions (but not reaching the extreme weather threshold). Its value ranges from (0,1], where 1 indicates no weather impact and 0 indicates that it is completely unusable. Different wind speed levels correspond to different derating factors. For example, the value is 1 for wind speed <5m / s, 0.8 for 5m / s ≤ wind speed <10m / s, and 0.5 for 10m / s ≤ wind speed <15m / s.

[0180] The base capacity of the hardware will also be reduced due to changes in the scenario security score; the lower the overall security score, the smaller the effective capacity of the hardware. , Corrections were made for additional capacity reductions caused by takeoff and landing area occupancy and weather conditions.

[0181] Step 3: Determine the primary take-off and landing site selection based on the coverage quality and effective capacity of the target candidate take-off and landing sites that are eligible for selection.

[0182] Step 31: Obtain the construction cost of each target candidate take-off and landing site and the mission requirement frequency of each route.

[0183] Construction cost represents the fixed, one-time investment required to build a landing field at a candidate site. This includes costs for site preparation, foundation construction, and equipment installation and commissioning. Construction cost is a core parameter in the site selection model, representing the objective of minimizing construction costs; a higher cost indicates a lower likelihood of choosing that location for construction.

[0184] Mission demand frequency indicates the expected number of missions to be executed on a flight route per unit of time (e.g., how many flights per day). It is the core basis for capacity allocation; the higher the frequency, the greater the demand for the hardware processing capacity of the takeoff and landing field.

[0185] The frequency of task requirements includes the baseline task frequency and the maximum positive deviation, which correspond to the task volume under normal circumstances and the additional increment under peak circumstances, respectively.

[0186] Step 32: Construct a site selection optimization model with the goal of maximizing coverage benefits in each simulation scenario, minimizing risk exposure in each simulation scenario, and minimizing site construction costs and hardware deployment costs.

[0187] Among the set of candidate take-off and landing sites eligible for selection, the site selection model aims to maximize the risk-adjusted target coverage quality and minimize construction costs and risk exposure to find the optimal configuration.

[0188] The objective function is expressed as follows:

[0189] In the formula, The probability of scenario s occurring is the likelihood of the s-th risk scenario appearing among all possible scenarios. It is a value between 0 and 1 and can be preset based on historical statistical data. For example, the frequency of strong afternoon winds in the summer in this area is 5 times per month, and the frequency of temporary traffic control during the evening rush hour is 8 times per month. The probability of occurrence is obtained after normalizing the frequency of each scenario.

[0190] Candidate take-off and landing sites Fixed construction costs (including one-time investment costs such as site preparation, foundation construction, equipment installation and commissioning). Candidate take-off and landing sites Deployment hardware type The unit cost refers to the specific type of takeoff and landing support equipment, such as charging piles, helipads, and maintenance equipment.

[0191] To determine whether to construct a candidate take-off and landing site The binary variable, when 1, indicates that in the candidate take-off and landing field When constructing a take-off and landing site, a value of 0 indicates that the site is not among the candidate take-off and landing sites. Construct a take-off and landing field. For candidate take-off and landing sites Deployment hardware type The quantity.

[0192] , and These are, in order, the dimensional reconciliation coefficients for coverage benefits, risk exposure, and switching costs. To simulate the scenario The expected switching cost. This represents the total risk-adjusted target coverage quality provided by the activated routes. This represents the risk exposure cost of having a take-off and landing field activated.

[0193] The primary option corresponds to the optimal solution. The set of candidate take-off and landing sites (i.e., which locations need to be built), and The value of , and the number of various hardware devices configured.

[0194] Step 33: Under the constraint of effective capacity, solve the location optimization model to obtain the primary take-off and landing site scheme.

[0195] The constraint of effective hardware capacity on the frequency of baseline tasks is implemented in the following way: For each candidate takeoff and landing site and each hardware type, determine the set of routes served by that combination.

[0196] For each simulation scenario, capacity inequality constraints are constructed based on the sum of the products of the baseline task frequency and the route activation variable for each route in the route set, the product of the demand fluctuation budget parameter and the global deviation compensation variable, and the sum of the local deviation compensation variables for each route.

[0197] The constraint expression is as follows:

[0198] The capacity inequality constraint is used to limit the sum of the above three items from not exceeding the product of the effective capacity of the candidate take-off and landing field in the corresponding simulation scenario and the number of hardware deployments.

[0199] The global deviation compensation variable and the local deviation compensation variable are non-negative variables, and the sum of the local deviation compensation variable and the global deviation compensation variable for each route is not less than the product of the maximum positive deviation of the route and the route activation variable.

[0200] Specifically, the global deviation compensation variable and the local deviation compensation variable satisfy the following constraints:

[0201]

[0202]

[0203] In the formula, For the airport With hardware type A collection of routes jointly undertaken; The baseline mission frequency for route r; Let r be the activation variable for the flight path. Budget parameters for demand fluctuations; This is a global bias compensation variable; For local deviation compensation variables; Hardware type The number of deployments at candidate takeoff and landing site i; For the route The maximum positive deviation relative to the baseline task frequency, Let τ be the effective capacity of candidate takeoff and landing field i in simulated scenario s for hardware type τ.

[0204] This is a configurable parameter, set by the planner before model solving based on actual operational needs. Its value range is: . The total number of routes served by the combination of candidate takeoff and landing field i and hardware type τ. This control sets an upper limit on the number of routes that can experience simultaneous positive fluctuations. It doesn't specify which routes will fluctuate, but rather limits the maximum number of routes that can fluctuate at the same time. When the value is low, the model assumes fewer routes fluctuating simultaneously, resulting in more compact capacity allocation and lower construction costs; when the value is high, the model assumes more routes fluctuating simultaneously, resulting in more flexible capacity allocation and higher construction costs. Planners can weigh construction costs against operational reliability based on their tolerance for operational uncertainty.

[0205] The essence of this capacity inequality constraint is a capacity robustness constraint under the uncertainty of demand frequency, which solves the problem of how to ensure that the hardware capacity is always sufficient when the task volume of each route fluctuates.

[0206] The actual frequency of missions on each route is not fixed and may deviate positively from the baseline frequency. The problem is that positive fluctuations may occur simultaneously on each route (such as when multiple regions experience temporary emergency delivery needs at the same time).

[0207] If capacity is constrained by simply summing the maximum deviations of all routes, it would be overly conservative, as it is virtually impossible for all routes to reach their peak simultaneously. On the other hand, if fluctuations are completely ignored, it would be too aggressive, and capacity would immediately become insufficient once demand fluctuates.

[0208] This capacity inequality constraint resolves this dilemma through a demand fluctuation budget parameter: when the demand fluctuation budget parameter is 0, no fluctuations are considered (economical but high-risk); when... At that time, all routes simultaneously reach their maximum deviation (absolutely conservative but extremely costly). By adjusting the demand fluctuation budget parameters, a flexible balance can be struck between being overly optimistic and overly conservative.

[0209] Global and local compensation are auxiliary variables introduced through mathematical transformations. Their technical role is to transform the original robust constraints (complex expressions involving the worst-case scenario of all fluctuation combinations) into a system of linear inequalities that can be directly solved. This transforms the combinatorial optimization problem, which originally required enumerating all fluctuation combinations, into a linear form that can be directly handled by a general linear programming / mixed-integer programming solver, greatly reducing the difficulty of solving the problem.

[0210] In this example, during model solving, linear relaxation is solved first. The lower limit of pruning is determined based on the protection quantile of the non-zero relaxation solutions of the site variables. Asymmetric fixing and sorting pruning are performed only on the site variables. Then, the dimensionality-reduced integer model is solved. If the dimensionality-reduced model is infeasible, times out and has no feasible solution, or violates the preset quality threshold, the fixing is lifted and the solution of the full integer model is rolled back to avoid losing feasible solutions due to aggressive pruning.

[0211] This solution strategy is designed for a mixed integer programming model. (Whether to build a website) (Whether to enable / disable) (Is the route activated?) The number of hardware deployments are all integer variables, and the scale is too large to solve directly. Therefore, a dimensionality reduction solution method of linear relaxation and pruning is adopted.

[0212] Solving the site selection optimization model to obtain the primary takeoff and landing site scheme includes the following steps: Step 331: Solve the site selection optimization model by linear relaxation to obtain the relaxed solution of the site construction variables of each candidate take-off and landing site under continuous relaxation.

[0213] The difficulty in solving mixed-integer programming lies in the combinatorial explosion of integer variables. When the number of candidate takeoff and landing sites is n, There are 2^n possible combinations, making direct solution of the complete integer model infeasible for large-scale problems. Linear relaxation reduces the integer variables... , The integer constraints are relaxed to continuous variables in the interval [0,1], which transforms the problem into a linear programming problem (which can be solved in polynomial time), thus quickly obtaining a reference solution.

[0214] Step 332: Determine the lower limit of pruning based on the preset protection quantiles of non-zero values ​​in the relaxation solution.

[0215] In the linear relaxation solution, some site-building variables The value can be non-zero (e.g., 0.85, 0.92, etc.). The lower limit of pruning is determined by the protection quantile, and only applies to site-built variables. Perform asymmetric fixing. Fix the site construction variables with higher values ​​in the relaxation solution to 1 (these candidate landing sites are likely to be selected into the optimal solution), and fix the variables with lower values ​​to 0 (these candidate landing sites are likely not to be selected into the optimal solution), then fix... The value of release , Integer constraints are solved in the reduced-dimensional integer model.

[0216] The meaning of the protective quantile is: select the p quantile of the non-zero value in the relaxed solution as a fixed threshold (e.g., p=0.7), fix the variables with values ​​greater than the threshold as 1, fix the variables with values ​​less than the threshold as 0, while the variables near the threshold are not fixed and are left to be determined by the subsequent complete integer model.

[0217] Step 333: Based on the pruning lower bound, perform asymmetric fixing operations on the site construction variables, while maintaining the integer degrees of freedom for solving the hardware deployment variables and route activation variables, to obtain the dimensionality-reduced integer model.

[0218] The meaning of asymmetric fixation is: for and The strategy for fixing 0 can be asymmetric. For example, it can be more conservative to fix 1 (fixing only those above 0.9) or more aggressive to fix 0 (fixing those below 0.3) to avoid incorrectly excluding truly optimal candidate take-off and landing sites.

[0219] Step 334: Solve the reduced-dimensional integer model to obtain the primary take-off and landing field scheme.

[0220] Step 335: When the reduced integer model is infeasible, the solution times out and no feasible solution is obtained, or the solution result violates the preset quality threshold, the fixed variables of the site are released, and the solution of the unpruned complete integer model is rolled back to obtain the primary take-off and landing field scheme.

[0221] While pruning reduces computational complexity, it carries risks. If the pruning lower bound is set too aggressively, it may incorrectly fix key site-building variables in the optimal solution at 0, resulting in the reduced-dimensional integer model having no feasible solution or a severely degraded solution quality (violating the preset quality threshold).

[0222] A rollback mechanism was designed: when the dimensionality reduction model is infeasible, the solution times out and there is no feasible solution, or the solution result violates the preset quality threshold, all site-building variables are unfixed, and the unpruned complete integer model is solved directly. This mechanism ensures the robustness of the method, quickly obtaining the optimal solution through pruning in most cases, and still obtaining a feasible solution through rollback when pruning fails, avoiding the defect of losing the optimal solution due to aggressive pruning.

[0223] Based on this, efficiency is prioritized: for most problem instances, feasible solutions to the dimensionality-reduced integer model are quickly obtained through linear relaxation and pruning. Quality assurance: when pruning leads to a decrease in solution quality, a rollback mechanism ensures the correctness of the final output.

[0224] Step 4: Based on the location of the primary take-off and landing site in the primary take-off and landing site plan, generate alternative take-off and landing site plans that differ from the primary take-off and landing site plan in terms of geographical distribution.

[0225] Differentiation refers to ensuring that the newly constructed take-off and landing sites in the alternative options are spatially sufficiently distant from the primary option, in order to avoid the same risk source (such as the same control area, the same weather unit, or the same terrain risk unit) simultaneously affecting both the primary and alternative nodes. This reduces the probability of the same control area, the same weather unit, or the same terrain risk source simultaneously affecting both the primary and alternative nodes.

[0226] In this example, the specific steps include: Step 41: Construct a K-ring geographical neighborhood for each take-off and landing site in the primary take-off and landing site scheme, with each K-ring geographical neighborhood being a set of grid cells in a hierarchical hexagonal grid that are no more than K layers away from the central take-off and landing site.

[0227] Each takeoff and landing site in the primary takeoff and landing site scheme will serve as the center point for constructing the K-ring neighborhood; that is, the number of K-ring neighborhoods will be equal to the number of primary takeoff and landing sites. The K-ring geographical neighborhood is a spatial metric in a hierarchical hexagonal grid (H3 grid), which is a set of all grid cells within K layers (K rings) of a given takeoff and landing site, centered on the grid where the takeoff and landing site is located.

[0228] In the H3 hexagonal mesh system, rings are defined as hierarchical levels based on the distance between the mesh centers. The first ring consists of six meshes sharing an edge with the center mesh, the second ring is the outermost ring of meshes, and so on. A distance not exceeding K layers encompasses all meshes from the 0th ring (the center mesh itself) to the Kth ring.

[0229] Step 42: For each K-ring geographical neighborhood, determine the upper limit of the number of take-off and landing sites that can be built in that neighborhood based on the remaining available budget and the lowest construction cost of the candidate take-off and landing sites in that neighborhood.

[0230] When generating alternative solutions, a portion of the construction budget has already been allocated to the primary solution (including site construction costs and hardware deployment costs), and the remaining available budget is used for the construction investment of the alternative solutions.

[0231] Within a certain K-ring neighborhood, there may be multiple candidate take-off and landing sites that have not yet been selected by the primary option. Each take-off and landing site has a corresponding construction cost. The minimum cost is taken, and the remaining available budget is divided by this minimum construction cost to obtain the maximum number of take-off and landing sites that can be built in this neighborhood when the entire budget is used for the cheapest site.

[0232] The expression is as follows:

[0233] In the formula, The remaining available budget when generating alternative topologies; For take-off and landing field Centered on the H3 grid Geographical neighborhood; This represents the number of neighborhood expansion layers. Assign candidate airport numbers within the neighborhood. This represents the upper limit of the number of take-off and landing fields that can be built within the i-th K-ring neighborhood, and is a tightened upper bound.

[0234] The calculation method for the maximum number of takeoff and landing sites fully considers the actual cost distribution of candidate sites within the neighborhood. If the neighborhood consists entirely of high-cost sites, the maximum number of sites that can be built in that neighborhood is tightened; if low-cost sites exist, the maximum is relatively loose. This ensures that budget constraints are practically achievable during the generation of alternative solutions.

[0235] Step 43: Constrain the number of alternative take-off and landing sites constructed in each K-ring geographical neighborhood to not exceed the upper limit, and constrain the alternative take-off and landing sites to be different from the primary take-off and landing site in at least one K-ring geographical neighborhood of the primary take-off and landing site.

[0236] Specifically, the expressions for neighborhood construction constraints and geographical differentiation constraints in the candidate topology are as follows:

[0237]

[0238] In the formula, Should a landing field be built for the alternative topology? A binary variable; This is an indicator variable for whether to retain the geographic neighborhood corresponding to the primary takeoff and landing site; The set of newly constructed take-off and landing sites for the primary take-off and landing site option.

[0239] Based on this, the alternative topologies are made different in at least one primary selection area, reducing the probability that the same control area, the same weather unit, or the same terrain risk source will simultaneously affect the primary and backup nodes.

[0240] Based on this, the meaning of the first equation is: within the K-ring neighborhood of any primary take-off and landing site, the number of newly constructed take-off and landing sites for alternative schemes shall not exceed [a certain number]. (The upper bound is determined by the remaining budget and the lowest cost within the neighborhood); the second equation means that the alternative must differ from the primary alternative in at least one primary neighborhood. These two constraints work together to differentiate the alternative from the primary alternative in terms of spatial layout, preventing excessive concentration of primary and alternative sites.

[0241] Therefore, the number of alternative options within the K-ring neighborhood of each primary takeoff and landing site is limited, meaning that alternative takeoff and landing sites cannot be concentrated around the primary takeoff and landing sites. If a primary takeoff and landing site is unavailable due to a control area, weather unit, or terrain risk source in its area, the alternative must have at least one alternative takeoff and landing site outside its K-ring neighborhood that is not affected by the same risk source.

[0242] The design of the K-ring plays a crucial role here. When K=1, the neighborhood only includes the ring neighbors of the central grid, and the primary / backup difference only needs to form within a small spatial range. When K=3, the neighborhood covers a larger area, and the primary / backup difference needs to form over a greater distance. The larger the value of K, the greater the spatial separation between the primary and backup, and the stronger the ability to resist regional risks, but it may affect the coverage quality of the alternative. By adjusting the K value, planners can balance risk resistance and coverage quality.

[0243] In one example, selecting an alternative primary takeoff and landing field from the alternative options and performing a mission switch includes the following steps: Step 1: Obtain the real-time hard judgment results, real-time comprehensive safety score, remaining effective capacity, and route energy consumption between the alternative take-off and landing sites and the current mission in the alternative take-off and landing site scheme.

[0244] The alternative options generated during the planning phase are based on the evaluation results of predicted scenarios and prior data. However, when the actual switch occurs during the operational phase, the environmental conditions may have changed. A landing field in the original alternative options may be experiencing real-time risks (temporary traffic control or sudden weather changes), and its availability may differ from the assessment at the time of planning. Therefore, this step requires reacquiring real-time data for each alternative option at the time of switch to ensure that subsequent selection is based on the current actual conditions rather than the prior assessments from the planning phase.

[0245] Real-time hard judgment result refers to the binary judgment result of whether each alternative take-off and landing site has triggered the first-level hard rule at the current moment. Real-time comprehensive safety score refers to the weighted risk comprehensive safety score of each alternative take-off and landing site at the current moment. Remaining effective capacity refers to the hardware processing capacity of each alternative take-off and landing site that is not currently occupied. Route energy consumption refers to the outbound and return energy consumption between the alternative take-off and landing sites and the current mission.

[0246] Step 2: Select alternative take-off and landing sites that have not triggered hard rules according to real-time hard judgment results, have a real-time comprehensive safety score not lower than the basic release threshold, have remaining effective capacity not less than the number of tasks to be migrated, and have route energy consumption that meets the energy limit of the UAV, as candidate alternative take-off and landing sites.

[0247] Specifically, the expression is as follows:

[0248] in, Candidate alternatives consist of alternative airports and alternative routes; This represents the set of feasible alternatives at the current moment. This is the first-order veto variable at the current time t; The overall safety score at the current time t; Based on the basic release threshold, The remaining effective capacity of the site corresponding to the alternative solution; Workload for tasks that need to be migrated; This indicates that the remaining effective capacity of the alternative airport is sufficient to accommodate the relocation tasks; and Alternative solutions Energy consumption for outbound and return journeys included in the route; Energy is available for the safe use of drones; This refers to the safe energy ratio for a single trip.

[0249] In this example, the feasible set is determined by four filtering criteria; if any criterion is not met, the option is excluded: Condition 1: The real-time hard judgment result is that the hard rule has not been triggered.

[0250] This condition requires that the alternative take-off and landing sites are not currently in a veto state. If If a candidate takeoff and landing site is not selected, it will be excluded due to triggering a hard rule (such as temporary air traffic control coverage, extreme weather, etc.). During the operational phase, real-time hard rule verification with the same criteria will be performed to ensure that the selected takeoff and landing sites have basic safety guarantees at the current moment.

[0251] Condition 2: The real-time comprehensive security score is not lower than the basic release threshold.

[0252] The overall safety score of the candidate take-off and landing sites at the current moment is not lower than the basic clearance threshold (e.g., 0.5). Take-off and landing sites with a score lower than this value have an overall risk level that is too high even if no hard rules are triggered, and are not suitable as candidates.

[0253] Condition 3: The remaining available capacity is not less than the number of tasks to be migrated.

[0254] The remaining available capacity of the alternative landing sites must be sufficient to accommodate the workload of the tasks to be migrated. The remaining available capacity is the unused hardware processing capacity of the landing site at the current moment (such as the number of idle landing pads, charging stations, etc.), which equals the current available capacity minus the occupied portion. The workload of tasks to be migrated refers to the number of tasks that need to be transferred to other landing sites due to the unavailability of the primary landing site. This condition ensures that the alternative landing sites have sufficient physical resources to accommodate newly transferred tasks.

[0255] Condition 4: The energy consumption of the flight path meets the energy limit of the UAV.

[0256] Round-trip total energy consumption constraint: The total round-trip energy consumption from the current mission location to the alternative take-off and landing site and back shall not exceed the safe available energy of the UAV.

[0257] Single-trip energy consumption constraint: The energy consumption of any single trip shall not exceed the preset proportion of safe and usable energy (such as 70%), ensuring that there is sufficient safety margin in either the outbound or return trip.

[0258] Even if the risk and capacity of the alternative take-off and landing sites meet the requirements, the alternative remains unavailable if the drone cannot fly there from its current location. This condition aligns with the energy screening logic of the planning phase, ensuring that the flight path upon which the switching decision depends remains physically reachable at the current moment. This dual verification mechanism has been implemented once during the planning phase and again during the operational phase to address changes in conditions such as real-time wind speed and altitude.

[0259] Based on this, the four criteria are used to screen candidate take-off and landing sites from four dimensions: safety (no hard rules are triggered), acceptability (overall safety score meets the standard), capacity (sufficient remaining space), and physical accessibility (range and energy consumption are met). If any dimension is not met, the candidate is excluded. Only candidate take-off and landing sites that pass all four tests can enter the candidate alternative set and await the next step of comprehensive cost assessment.

[0260] Step 3: Based on the delay cost, energy consumption cost, and risk cost caused by the switchover, select one from the candidate alternative take-off and landing fields as the alternative primary take-off and landing field, and determine the corresponding route.

[0261] Specifically, the expression is as follows:

[0262] Task delays caused by switching; This refers to the increased flight energy consumption compared to the original mission plan; Loss due to task priority; The overall risk and loss of alternative solutions; , , , These are the weighting coefficients for the corresponding evaluation indicators.

[0263] Based on this, the scheme with the lowest overall cost is selected as the final implementation plan. The routes associated with the alternative primary take-off and landing field in the alternative take-off and landing field schemes are determined as alternative routes.

[0264] It should be noted that the delay cost refers to the mission delay caused by the switchover, that is, how much the mission completion time is delayed compared to the original plan due to the change of takeoff and landing sites. For example, the estimated mission completion time after the alternative takeoff and landing site is: the flight time from the current position of the UAV to the alternative takeoff and landing site, which depends on the distance between the current mission position and the alternative takeoff and landing site and the flight speed.

[0265] Energy cost refers to the increased flight energy consumption caused by the switch. In other words, energy cost means how much more the total flight energy consumption required to complete the same mission after switching to the alternative take-off and landing site is increased compared to the original plan.

[0266] The risk cost is the incremental risk resulting from the switchover, determined by the difference between the overall safety score of the candidate alternative takeoff and landing site and the overall safety score of the primary takeoff and landing site. The lower the cost, the higher the priority.

[0267] In one example, the system records the risk factors, reasons for rejection, security score, threshold, actual delay, task completion status, and primary / backup switchover results for each assessment. The frequency of risk scenarios is updated according to a rolling time window. After obtaining sufficient samples, the weights can be updated based on the correlation between risk factors and actual security events, delays, or task failures, with upper limits for weight changes and a manual approval mechanism set.

[0268] The system can be broken down into spatial computing services, risk rule services, scoring services, robust optimization services, operational release services, and feedback training services. Real-time airport and meteorological data are stored in a short-term cache, while planning schemes, risk records, and model versions are stored in a relational database. Each planning zone is solved in a one-time independent subprocess, with hard timeouts set and idempotent persistence performed on the results.

[0269] In one application scenario, the system constructs airport normal, runway occupied, temporary control, strong wind, and low visibility scenarios for multiple candidate takeoff and landing sites. Candidate site A has better physical coverage, but its safety score and effective capacity decrease in the strong wind and runway occupied scenarios; candidate site B has a slightly higher construction cost but stable availability across scenarios. The robust site selection model configures both A and B in the primary topology and generates a candidate topology that includes candidate site C through geographical neighborhood exclusion constraints. During operation, site A is rejected due to temporary control. Without re-executing global site selection, the system selects site B and its corresponding legal flight path from the candidate topology to complete the task migration, and uses the duration of this control and the handover delay to update the parameters for subsequent scenarios.

[0270] In summary, the planning area is discretized into a hierarchical hexagonal grid, and the safe available energy, safe round-trip radius, and directional three-dimensional energy consumption of candidate routes are calculated. Only physically legal routes that meet the red lines of total round-trip energy consumption and one-way energy consumption are retained.

[0271] Perform data integrity verification on candidate airports and risk scenarios, calculate no-fly overlap rate, minimum airspace and control time overlap rate, and form a first-level veto variable with the maximum value of the hard rule indicator. Only when the first-level veto is not triggered, map the observed values ​​of the eleven risk sub-factors to a safety score between 0 and 1, and then calculate the scenario safety score, risk loss and risk contribution rate.

[0272] By mapping risk outcomes to candidate eligibility, risk-adjusted coverage quality, and risk-adjusted effective capacity, a robust model for site selection, capacity determination, and route activation under risk scenarios is established.

[0273] The primary and alternative topologies are obtained by using distributed adaptive pruning, infeasible rollback, and budget-tightening geographical neighborhood exclusion constraints.

[0274] During operation, hard rules and dynamic scoring are re-executed. When the primary node is unavailable or reduced in capacity, alternative nodes that meet the requirements of risk, range, capacity, and mission priority are selected from the alternative topology.

[0275] Update scenario frequency, factor weights, and planning parameters based on actual risk losses, task delays, and switching results.

[0276] This approach shifts operational risks from post-release conditions to planning variables, avoiding the construction of sites with high coverage but high frequency of unavailability; it maintains a safety baseline with a first-level veto variable, ensuring that no high score can compensate for the risk of absolute no-fly zone; it adjusts coverage quality and effective capacity through risk, enabling robust models to simultaneously defend against demand fluctuations and airport capacity reduction; it pre-generates geographically differentiated alternative topologies, shortening mission recovery time under temporary risks; it interprets release and site selection results using risk loss contribution; and it updates scenario frequency and weights through operational feedback, allowing planning to evolve with actual risks.

[0277] Figure 2 This is a schematic diagram of a closed-loop planning and operation process provided in an embodiment of this application.

[0278] Figure 2 This reflects data flow and decision dependencies. The outputs during the planning phase (primary and alternative topologies) serve as inputs during the runtime phase, and the actual execution results during the runtime are fed back to the scenario parameter updates during the planning phase through a feedback mechanism.

[0279] During the planning phase, the first step is to obtain basic information about the planning area and data on environmental factors that affect the operation of the take-off and landing field.

[0280] Then, candidate take-off and landing sites are generated through spatial partitioning, and physically reachable legal routes are selected based on the three-dimensional energy consumption constraints of the UAV.

[0281] For each candidate airport, the first-level veto variable is calculated under each risk scenario. If the first-level veto is not triggered, the safety score of the risk sub-factor, the comprehensive safety score, and the risk loss contribution are then calculated.

[0282] The first-level veto variables are coupled to the candidate airport activation qualification and route activation qualification, respectively, and the comprehensive safety score is coupled to the route coverage quality and effective hardware capacity, respectively, to construct a robust optimization model for site selection-capacity determination-route activation that includes uncertainty in demand frequency.

[0283] Solving the robust optimization model yields the primary take-off and landing field topology, and generating alternative take-off and landing field topologies through geographical neighborhood exclusion constraints.

[0284] During the operation phase, the real-time primary veto variables, real-time comprehensive safety score, and real-time effective capacity are recalculated. When the primary airport or primary route is rejected, falls below the basic release threshold, or has insufficient capacity, alternative take-off and landing airports and alternative routes that meet the risk, range, and capacity constraints are selected from the alternative take-off and landing airport topologies.

[0285] And update the probability of occurrence of risk scenarios, the weight of risk sub-factors, or planning parameters based on the task execution results.

[0286] Based on this, a complete closed loop has been formed, which includes physical route generation, multi-scenario risk assessment, risk-coupled robust site selection, primary and backup dual topology pre-configuration, real-time safe release and switching, and operational data feedback and re-optimization. This has achieved unified coordination between site selection during the planning period and safe release during the operation period, and improved the physical executability, risk interpretability and disturbance recovery capability of the take-off and landing network.

[0287] Figure 3 This is a schematic diagram illustrating the coupling relationship between the state risk of a candidate takeoff and landing site and the location variables, provided as an embodiment of this application.

[0288] Figure 3 It demonstrates how risk assessment results are mapped to three core parameters of the site selection optimization model.

[0289] For each candidate airport and its corresponding reachable routes, the system collects risk information from four dimensions: airport operation dimension (openness status, runway occupancy, airport class), airspace operation dimension (controlled airspace, temporary control, busy level), geographical security dimension (clearance conditions, underlying surface type, no-fly overlap), and time adaptation dimension (control period matching, peak period adaptation). The above information constitutes a set of risk sub-factors.

[0290] The risk assessment process employs a two-tiered structure: First, a primary hard rule judgment is executed. If any hard rule is triggered, the candidate's eligibility is immediately rejected, and the landing field is excluded from the corresponding scenario. Only when no hard rule is triggered does the comprehensive safety score (safety score) calculation proceed. The 11 sub-factors of the four risk dimensions are scored separately and then weighted and summed to obtain the comprehensive safety score (safety score).

[0291] The risk results are coupled to the site selection model through three output channels. Output 1 (Candidate Qualification): The first-level rejection result constrains the activation qualification of candidate take-off and landing sites; rejected sites cannot be selected for construction in the corresponding simulation scenario. Output 2 (Risk-Adjusted Coverage Quality): The risk-adjusted coverage quality is obtained by multiplying the comprehensive safety score by the basic coverage quality, enabling the site selection model to identify the reduction in the actual coverage contribution of sites with low safety scores when evaluating the coverage value of flight routes. Output 3 (Risk-Adjusted Effective Capacity): This is obtained by applying the comprehensive safety score and a specific capacity reduction factor to the baseline capacity, enabling the robust site selection model to anticipate the decline in the actual usable hardware capacity under different risk scenarios when configuring capacity.

[0292] The risk assessment results are not information labels independent of the site selection decision, but are directly embedded into the constraints and parameters of the optimization model through three output channels. This allows the site selection decision to see the actual availability of each candidate site under different risk scenarios, thereby avoiding the construction of high-coverage but frequently unavailable sites from the source.

[0293] Figure 4 This is a schematic diagram of a three-level decision chain for the status risk of a candidate takeoff and landing site, provided in an embodiment of this application.

[0294] Figure 4 It demonstrates the complete risk assessment and decision-making process from data collection to release determination.

[0295] The system first collects and verifies multi-source input data, including airport operational status, airspace control information, geographic environment data, real-time weather conditions, and mission time windows. After data verification, it proceeds to the first-level decision-making stage: hard rule judgment. The system checks each hard rule to see if it has been triggered (e.g., airport closure, excessive no-fly overlap, extreme weather, etc.). Once any hard rule is triggered, a first-level rejection result is output, directly rejecting the candidate take-off / landing site or mission in the corresponding simulated scenario, along with the reason for rejection, and the process terminates.

[0296] If no hard rules are triggered, the system proceeds to the second level of decision-making, with refined scoring. The system scores 11 risk sub-factors (airport openness, runway occupancy, airport class, controlled airspace, temporary control, airspace congestion, terrain clearance, underlying surface, no-fly zone overlap, control period matching, and peak period suitability) individually, then performs a weighted sum to obtain a comprehensive safety score (scenario safety score). Simultaneously, the system calculates the risk loss and risk contribution rate for each sub-factor to identify the main risk sources of a low comprehensive score.

[0297] After obtaining the comprehensive safety score (safety score), the system proceeds to the third-level decision stage, which involves threshold comparison. The system outputs three levels of judgment results: a score below the basic release threshold indicates that the conditions for flight are not met; a score greater than or equal to the basic release threshold but less than the safety release threshold indicates that the conditions are met (requiring additional operational restrictions or manual approval); and a score greater than or equal to the safety release threshold indicates that the conditions for normal flight are met.

[0298] Based on this, risk assessment and clearance decisions follow a strict sequence. First, hard rules are judged (reject / approve), then soft scores are calculated (safety score), and finally, thresholds are compared (whether clearance is possible). The three-level decision chain is layered and decoupled, with hard rule judgments preceding the absolute priority of veto; soft scores follow, only taking effect after a first-level veto, to finely differentiate candidate take-off and landing sites, and finally output a clear clearance conclusion after comparison with dynamic thresholds.

[0299] Figure 5 This is a schematic diagram illustrating the dynamic switching of primary and backup takeoff and landing sites and flight routes during operation, provided as an embodiment of this application.

[0300] Figure 5 It demonstrates the monitoring, evaluation, and switching execution logic of the planning scheme during the operation phase.

[0301] The system initially executes the primary take-off and landing site and primary route plan. After entering the operation monitoring cycle, it continuously re-evaluates the real-time status of the primary take-off and landing site according to the preset update cycle. The evaluation content includes the first-level veto result, real-time safety score and current effective capacity.

[0302] The evaluation results trigger three judgments: (1) whether a first-level veto is triggered; (2) whether the threshold for safe release is lower than the threshold for safe release; (3) whether the capacity is insufficient (the remaining capacity is less than the number of tasks to be migrated). If none of the three items are triggered, the primary solution will continue to operate normally and the system will continue to monitor periodically; if any item is triggered, the primary solution will be determined to be unavailable and the switching process will be executed.

[0303] The handover process first forms a set of feasible alternatives. Based on the set of feasible alternatives, the optimal replacement node is selected according to the comprehensive cost function to complete the handover of airports and routes and issue the task.

[0304] After the task is completed, the system records the actual execution results (including risk events, delays, switch records, and task completion status), and updates the probability of occurrence of risk scenarios, sub-factor weights, and planning parameters through a feedback mechanism for subsequent planning optimization.

[0305] Based on this, the switching decision is not an ad-hoc relocation, but rather a selection of alternatives from pre-validated candidate topologies, reducing switching time from minutes to seconds. Simultaneously, the switching decision involves two stages: feasibility screening and optimization ranking. First, safety, capacity, and range requirements are ensured, then a comprehensive optimal solution is selected through multi-objective weighted selection. The actual operational results after each switching are fed back to the planning parameters via a feedback loop, forming a complete closed loop of planning, operation, feedback, and replanning. As operational data accumulates, the system's location selection and release decisions continuously self-optimize.

[0306] Based on the same idea, some embodiments of this application also provide devices and non-volatile computer storage media corresponding to the above methods.

[0307] Figure 6 A schematic diagram of a site selection and switching device for urban low-altitude unmanned aerial vehicle (UAV) take-off and landing sites provided in this application embodiment includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the method for site selection and switching of urban low-altitude unmanned aerial vehicle take-off and landing sites as described above.

[0308] Some embodiments of this application provide a non-volatile computer storage medium for selecting and switching urban low-altitude UAV take-off and landing sites, which stores computer-executable instructions capable of executing any of the above-described methods for selecting and switching urban low-altitude UAV take-off and landing sites.

[0309] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the device and medium embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the description of the method embodiments.

[0310] The devices and media provided in this application are one-to-one with the methods. Therefore, the devices and media also have similar beneficial technical effects as their corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the devices and media will not be repeated here.

[0311] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0312] The above description is merely an embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the technical principles of this application should fall within the protection scope of this application.

Claims

1. A method for site selection and switching of urban low-altitude unmanned aerial vehicle (UAV) take-off and landing sites, characterized in that, The method includes: Acquire basic information about the planning area and data on environmental factors affecting the operation of the take-off and landing field; the basic information includes geospatial data of the planning area, UAV performance parameters, and parameters of take-off and landing field support equipment. Based on the aforementioned basic information, candidate take-off and landing site locations and associated reachable routes are generated; For the candidate take-off and landing sites and the reachable routes, multiple simulation scenarios with different combinations of environmental factors are constructed based on the environmental factor data, and the operational availability of each candidate take-off and landing site in each simulation scenario is evaluated. Based on the operational availability, a primary take-off and landing site scheme is determined from the candidate take-off and landing site locations, along with alternative take-off and landing site schemes that differ from the primary scheme. During the operation phase, the real-time operational availability of the currently used primary take-off and landing field is assessed based on current environmental factor data. When the real-time operational availability does not meet preset conditions, an alternative primary take-off and landing field and corresponding alternative routes are selected from the alternative take-off and landing field schemes, and a task switch is performed.

2. The method according to claim 1, characterized in that, Based on the aforementioned basic information, candidate takeoff and landing site locations and associated reachable routes are generated, specifically including: The planning area is divided into several hexagonal service grids according to a first resolution, and grouped according to the parent grid to which each service grid belongs; the parent grid is a hexagonal grid of the planning area divided according to a second resolution, the second resolution being lower than the first resolution, and each parent grid includes multiple service grids; The area comprised of all business grids belonging to the same parent grid is defined as the core area; For any core area, the business grids adjacent to the boundary of that core area in the adjacent core areas are determined as the adjacent overlapping boundary areas of that core area; Candidate take-off and landing sites are set up within the core area, and mission requirement points are set up within the core area and adjacent overlapping boundary areas respectively; For the route combinations between the candidate take-off and landing sites and the mission requirement points, based on the UAV's energy limitations and the flight energy consumption corresponding to the route combinations, the route corresponding to the route combination that meets the energy limitations is selected as the reachable route.

3. The method according to claim 2, characterized in that, Based on the energy limitations of the UAV and the flight energy consumption corresponding to the route combinations, routes corresponding to route combinations that meet the energy limitations are selected as reachable routes, specifically including: Calculate the flight energy consumption required for the UAV to complete the outbound and return journeys along the combined route. Determine the safe available energy for the drone; Determine whether the flight energy consumption meets the energy limit; the energy limit includes: the sum of the outbound flight energy consumption and the return flight energy consumption is not greater than the safe available energy, and the maximum value of the outbound flight energy consumption and the return flight energy consumption is not greater than the preset ratio of the safe available energy; The routes corresponding to the combinations of routes that meet the energy limits are determined as the reachable routes.

4. The method according to claim 2, characterized in that, The operational availability of each candidate takeoff and landing site under each simulation scenario was evaluated, specifically including: For each candidate take-off and landing site in each simulation scenario, a first-level hard rule judgment is performed to obtain a hard judgment result; the first-level hard rule is used to indicate that when any hard rule is triggered, the candidate take-off and landing site is judged to be unusable in the corresponding simulation scenario; When the hard judgment result is that the first-level hard rule has not been triggered, the multiple risk sub-factors are scored separately. The scores of each risk sub-factor are weighted and summed to obtain a comprehensive safety score; Based on the hard judgment results and the comprehensive safety score, the operational availability of each candidate take-off and landing site in each simulated scenario is determined.

5. The method according to claim 4, characterized in that, After weighted summation of the scores for each risk sub-factor to obtain a comprehensive safety score, the method further includes: Calculate the risk loss of each risk sub-factor based on its weight and the degree of insecurity of each sub-factor; Calculate the risk contribution rate of each risk sub-factor based on the proportion of risk loss of each risk sub-factor in the total risk loss; The risk sub-factors are sorted according to their risk contribution rates from high to low, and used as explanatory information for the comprehensive safety score.

6. The method according to claim 4, characterized in that, Based on the operational availability, a primary take-off and landing site is determined from the candidate take-off and landing site locations, along with alternative take-off and landing site options that differ from the primary option. Specifically, this includes: Based on the hard determination results in the operational availability, it is determined whether each candidate take-off and landing site is eligible to be selected in each simulation scenario; a candidate take-off and landing site that triggers a first-level hard rule is not eligible to be selected in the corresponding simulation scenario; Based on the comprehensive safety score in the operational availability, the basic coverage quality and baseline capacity of the target candidate take-off and landing sites that are eligible for selection are adjusted in each simulation scenario to obtain the target coverage quality and effective capacity of the target candidate take-off and landing sites in each simulation scenario; the coverage quality is used to characterize the service coverage capability of the candidate take-off and landing sites to the task requirement points, and the capacity is used to characterize the hardware processing capability of the candidate take-off and landing sites in the corresponding simulation scenario. The primary take-off and landing site option is determined based on the coverage quality and effective capacity of the target candidate take-off and landing sites; Based on the location of the primary take-off and landing site in the primary take-off and landing site plan, generate alternative take-off and landing site plans that differ from the primary take-off and landing site plan in terms of geographical distribution.

7. The method according to claim 6, characterized in that, Based on the coverage quality and effective capacity of the target candidate take-off and landing sites, the primary take-off and landing site selection is determined, specifically including: Obtain the construction cost of each target candidate take-off and landing site and the mission requirement frequency of each route; To maximize coverage benefits in each simulated scenario, minimize risk exposure in each simulated scenario, and minimize site construction and hardware deployment costs, a site selection optimization model is constructed. Under the constraint of effective capacity, the location optimization model is solved to obtain the primary takeoff and landing site scheme.

8. The method according to claim 1, characterized in that, Based on the location of the primary take-off and landing site in the primary take-off and landing site plan, alternative take-off and landing site plans that differ geographically from the primary plan are generated, specifically including: Taking each take-off and landing site in the primary take-off and landing site scheme as the center, a K-ring geographical neighborhood is constructed respectively; the K-ring geographical neighborhood is a set of grid cells in a hierarchical hexagonal grid that is no more than K layers away from the central take-off and landing site; For each K-ring geographical neighborhood, the upper limit of the number of take-off and landing sites that can be built in the neighborhood is determined based on the remaining available budget and the lowest construction cost of candidate take-off and landing sites in the neighborhood; The number of alternative take-off and landing sites constructed within each K-ring geographical neighborhood is limited to the specified upper limit. The alternative take-off and landing site scheme is constrained to be different from the primary take-off and landing site scheme in at least one K-ring geographical neighborhood.

9. The method according to claim 1, characterized in that, Selecting alternative primary takeoff and landing fields and corresponding alternative routes from the candidate takeoff and landing field options specifically includes: The system obtains the real-time hard determination results, real-time comprehensive safety score, remaining effective capacity, and route energy consumption between the alternative take-off and landing sites and the current mission for each alternative take-off and landing site in the alternative take-off and landing site scheme. Candidate take-off and landing sites are selected based on the following criteria: the real-time hard judgment result indicates that no hard rule has been triggered; the real-time comprehensive safety score is not lower than the basic release threshold; the remaining effective capacity is not less than the number of tasks to be migrated; and the route energy consumption meets the energy limit of the UAV. Based on the delay cost, energy consumption cost, and risk cost caused by the switchover, one of the candidate alternative take-off and landing fields is selected as the alternative primary take-off and landing field, and the corresponding alternative route is determined; the delay cost is the mission delay time caused by the switchover; the energy consumption cost is the increased flight energy consumption caused by the switchover; the risk cost is the increased risk increment caused by the switchover, and the risk increment is determined based on the difference between the comprehensive safety score of the candidate alternative take-off and landing field and the comprehensive safety score of the primary take-off and landing field.

10. A device for selecting and switching take-off and landing sites for urban low-altitude unmanned aerial vehicles (UAVs), characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the method for site selection and switching of urban low-altitude unmanned aerial vehicle take-off and landing sites as described in any one of claims 1-8.

Images